Probe E# — E-Sharp AB
Probe is made by E# (E-Sharp AB) in Sweden. Want your own boards reviewed like this? Book a meeting

AFCKU

OHWR0057 · revision R1

AFCKU (OHWR0057 R1) is an open-hardware circuit board from the CERN Open Hardware Repository, designed in Altium Designer: a 14-layer board of 180.6 × 148.5 mm with 1568 components and 1472 nets, published under CERN-OHL-S-2.0.

review 84% walked · 27 findings: 2 blocker · 12 major · 13 minor (70 items)

Open in the interactive viewerRead the full review

AFCKU (OHWR0057 R1) circuit board, top view: copper, silkscreen and components
Top view, rendered by Probe from the design's own fabrication data.

Design review

Probe recorded 27 findings: 2 blocker · 12 major · 13 minor (70 items) on this design. Each is shown here with its reasoning; the full review links every one to the pins, nets and datasheet pages it rests on.

  1. PowerGroundShortblockerOpen

    'P0V95' is shorted to 'GND' through fitted link(s) R519, R542

    'P0V95' is shorted to 'GND' through fitted zero-ohm link(s) R519, R542: P0V95 -> R519 -> POR_override -> R542 -> GND. Every link on that path is populated, so this is a direct rail-to-ground path in the build as specified. No pin role is involved and none would help — a two-terminal link has no datasheet pin function — which is why the role-layer axis of this check cannot see it. If these are a two-position option network, exactly one link per position should be fitted and the other marked DNP.

    At P0V95

  2. CopperShortBetweenNetsblockerOpen

    GND and NetIC16_11 copper touch on BTM, at (-32.12, 11.44): the two nets are shorted in the copper as exported

    GND and NetIC16_11 copper touch on BTM, at (-32.12, 11.44): the two nets are shorted in the copper as exported. Where: BTM at (-32.12, 11.44) (a pad or track lies wholly under the other net's copper). The netlist still puts the two nets apart, so every connectivity check passes; the copper the fab builds joins them. The usual cause is a pour filled before the other net's via or pad was placed (or moved), so it runs through it without a clearance: refill the zones, check the CAD tool's DRC is clean, and export again. copper-near at a site shows both nets' copper at distance 0.

    At short:GND|NetIC16_11

  3. observationOBS-396majorOpen

    IC16 LMZ31520 SS_SEL has both option straps fitted: R195 (0 ohm to PWRGD) and R196 (0 ohm to PGND) tie the shared power-good net PM.PGOOD to ground

    IC16 LMZ31520 SS_SEL has both option straps fitted: R195 (0 ohm to PWRGD) and R196 (0 ohm to PGND) tie the shared power-good net PM.PGOOD to ground.

    The straps

    IC16 pin 3 SS_SEL (NetIC16_3) connects through R195 (0 ohm, R0402_0R_JUMPER) to PWRGD pin 19 = net PM.PGOOD, and through R196 (0 ohm) to GND. The datasheet offers these as alternatives (LMZ31520 pin functions p.7; slow start section p.17):

    • SS_SEL to PGND: Auto-skip Eco-mode.
    • SS_SEL to PWRGD: FCCM.

    ``mermaid flowchart LR G["GND"] -- "R196 0 ohm" --> S["IC16 SS_SEL<br/>pin 3"] S -- "R195 0 ohm" --> P["PM.PGOOD<br/>R268 10k to P3V3"] P --> M["IC1 LPC1768 MMC<br/>pin 26"] P --> R["IC58 STM811 MR<br/>pin 3"] R -- "RST" --> N["RESETN<br/>IC1 pin 36, IC73 pin 22"] ``

    Fit data

    • The BOM fits both: Mounted Yes, in the 89-part R0402_0R_JUMPER line.
    • The ODB++ assembly data (comp_+_bot) marks R195 and R196 PRP Mounted 'Yes'.
    • Schematic sheet 23 draws both jumpers with no fit marking. The only Mounted No parts on the board are mounting holes, fiducials and test pads.

    Why it matters

    PM.PGOOD is the wired-OR power-good of IC15, IC16, IC61, IC67 and IC69. Fitted together, R195 and R196 short it to ground, so it can never go high: the MMC always reads power-not-good and the STM811 holds RESETN asserted. Likely one of them is meant as an unfitted option.

  4. observationOBS-411majorOpen

    Si549 oscillator option jumpers are all fitted: R224 and R564 short translator outputs IC41 B5 and B6 together and onto DAC2's output, and R383/R384 tie DAC1's output to its own SYNC line and to the OSC2 OE driver

    The Si549 oscillator option jumpers are all fitted: R224 and R564 short translator outputs IC41 B5 and B6 together and onto DAC2's output, and R383/R384 tie DAC1's output to its own SYNC line and to the OSC2 OE driver.

    The OSC1/OSC2 (Si549) control area has 0-ohm option jumpers for several oscillator variants.

    OSC1 side

    PLLDAC_OUT2 = OSC1 pin 1 (OE) + IC82 DAC VOUT through R352 240 ohm.

    JumperJoins PLLDAC_OUT2 toThat net also carries
    R224SI57X_PRI_TUNE/SYNCIC41 output B6 (pin 16); IC82 SYNC (pin 5, PLL_DAC2_SYNC_N) via R353
    R564SI57X_PRI_OEIC41 output B5 (pin 17); Q8 gate; R223 5k1 to P3V3; R563 to OSC1.2

    OSC2 side

    PLLDAC_OUT1 = OSC2 pin 1 (OE) + IC53 DAC VOUT via R218.

    JumperJoins PLLDAC_OUT1 toThat net also carries
    R384SI57X_SEC_TUNE/SYNCIC41 output B7 (pin 15); IC53 SYNC (pin 5) via R449
    R383SI57X_SEC_OEQ16 collector; R379 5k1; R474 to OSC2 pin 2

    Supporting facts

    • IC41 SN74AVC8T245 has DIR tied high (P1V8), so B5/B6/B7 are outputs (A data to B bus; datasheet function table, p.16).
    • Si549 pin 1 is OE by ordering option; the part has no voltage-control input (Si549 datasheet, p.2).
    • R224, R353, R383, R384, R449, R474, R563, R564 are all in the 89-part R0402_0R_JUMPER line, Mounted Yes in the BOM and ODB++. Only mounting holes, fiducials and test pads are Mounted No.

    Why it matters

    With every jumper fitted, two FPGA-driven translator outputs (B5, B6) are shorted together, each DAC's analog output is tied to the digital line that frames its own SPI writes, and the oscillator OE pin sees the DAC voltage. The released assembly data cannot work as built.

  5. observationOBS-412majorOpen

    Supply-select ferrite beads for the OSC2 area are fitted in both positions: FB5+FB6 join P3V3 to P1V8, and FB29+FB7 join P3V3_CLEAN to P1V8_CLEAN

    The OSC2 supply-select ferrite beads are fitted in both positions: FB5 + FB6 join P3V3 to P1V8, and FB29 + FB7 join P3V3_CLEAN to P1V8_CLEAN. R244 + R246 also ground OSC2's CLK- output.

    Schematic sheet 5 draws two either/or supply selections for the secondary Si549 (OSC2), with no DNP or not-fitted marking.

    The pairs

    Fed nodeFrom 3.3 VFrom 1.8 VWhat it feeds
    NetC391_2FB5 from P3V3 (IC43.6 VO2)FB6 from P1V8 (IC43.1 VO1)R380/R381 SEC_SCL/SDA 4k7, R379 OE pull-up, R130 to IC39 PCA9306 VREF2
    OSC_suppFB29 from P3V3_CLEANFB7 from P1V8_CLEAN (P1V8 through L7)OSC2 VDD

    ``mermaid flowchart LR A["P3V3<br/>IC43 ch2"] -- "FB5" --> N["NetC391_2<br/>Si549 I2C pull-ups"] B["P1V8<br/>IC43 ch1"] -- "FB6" --> N C["P3V3_CLEAN"] -- "FB29" --> S["OSC_supp<br/>OSC2 VDD"] D["P1V8_CLEAN"] -- "FB7" --> S ``

    The clock output: R244 (0 ohm) joins OSC2 CLK- (SEC_CLKa_N, OSC2.5) to FPGA_CLK.SEC_CLK_N, and R246 (0 ohm) joins FPGA_CLK.SEC_CLK_N to GND.

    Fit data

    • FB1, FB2, FB3, FB5, FB6, FB7, FB29, FB30 are one BLM21PG221SN1D line, Mounted Yes.
    • R244 / R246 are in the 89-part R0402_0R_JUMPER line, Mounted Yes.
    • ODB++ PRP Mounted Yes on all six. Only mounting holes, fiducials and test pads are Mounted No.

    Why it matters

    With both beads of a pair fitted, P3V3 and P1V8 are shorted through two ferrite beads (a few tens of milliohms at DC), and P3V3_CLEAN and P1V8_CLEAN likewise. R246 grounds a clock output. The intent is plainly one bead per pair: a 3.3 V or 1.8 V option for the Si549 and its I2C.

  6. observationOBS-416majorOpen

    FPGA configuration straps are fitted both ways: R519+R542 short VCCINT (P0V95) to GND through POR_OVERRIDE, and the 1k pull-ups plus 100 ohm pull-downs on M0-M2 and PUDC_B hold all four low whatever SW2 is set to

    The FPGA configuration straps on sheet 13 are fitted both ways: R519 + R542 short VCCINT (P0V95) to GND through POR_OVERRIDE, and 1k pull-ups with 100 ohm pull-downs hold M0-M2 and PUDC_B low whatever SW2 is set to.

    Schematic sheet 13 draws the Kintex UltraScale configuration straps as options, with no fit marking. The BOM and the ODB++ assembly data mark all ten resistors Mounted Yes.

    The straps

    NetIC2 ballUpTo GNDSW2
    POR_overridePOR_OVERRIDER519 0 ohm to P0V95R542 0 ohmnone
    FPGA_M0K7R507 1k to P3V3R521 100 ohmSW2.8
    FPGA_M1L7R509 1k to P3V3R523 100 ohmSW2.7
    FPGA_M2M7R511 1k to P3V3R540 100 ohmSW2.6
    PUDCR7R513 1k to P3V3R541 100 ohmSW2.5

    SW2 positions 1-4 go to GND.

    The numbers

    QuantityValueSource
    POR_OVERRIDEmust be tied to VCCINT or GNDUG575 Table 1-7, p.32
    R519 + R5420 ohm path from VCCINT (P0V95, the 20 A LMZ31520 output) to GNDderived
    Each mode / PUDC line1k over 100 ohm = about 0.3 V, about 3 mA wasted eachderived
    Resulting modeM[2:0] = 000, PUDC_B = 0, whatever SW2 saysderived
    PUDC_B = 0weak pre-configuration pull-ups on all SelectIOUG575, p.32

    Why it matters

    Fitting both POR_OVERRIDE straps shorts VCCINT, and IC16 would run into current limit. The mode can never be anything but 000, so SW2 and the 1k pull-ups do nothing, while QSPI boot (IC65 MT25QL256 on the CCLK/QSPI0 pins) needs M0 high. PUDC_B = 0 matters for the gate-driven LEDs and enables on FPGA pins.

  7. observationOBS-420majorOpen

    FPGA JTAG bypass jumpers are fitted alongside the chain they bypass: R210 ties FPGA TDO straight onto PROG_TDO, which IC75 also drives, and R212/R436 tie the IC47 JTAG fan-out buffer's TCK/TMS outputs back to its inputs

    FPGA JTAG bypass jumpers are fitted together with the chain they bypass: R210 puts FPGA TDO onto PROG_TDO, which IC75 also drives, and R212/R436 tie IC47's TCK/TMS outputs back to its own inputs.

    The chain as drawn

    PROG_TDI (J6 / IC46 mux) goes through R209 into FPGA TDI (IC2 V9). FPGA TDO (U9, FJTAG.TDO) goes through R235 to JTAG1.TDI (FMC1), then through the NC7SZ157 bypass muxes IC49/IC72/IC75 (FMC1, FMC2, RTM). IC75's output Z drives PROG_TDO back to the header.

    ``mermaid flowchart LR H["J6 / IC46<br/>PROG_TDI"] -- "R209" --> F["IC2 FPGA<br/>TDI V9 / TDO U9"] F -- "R235" --> M["IC49 / IC72 / IC75<br/>FMC1, FMC2, RTM"] M -- "IC75 Z" --> P["PROG_TDO<br/>J6.8"] F -- "R210 0 ohm" --> P ``

    The conflicts

    • TDO: R210 (0 ohm) connects FJTAG.TDO directly to PROG_TDO. With R210 and R235 both fitted, two push-pull outputs, FPGA TDO and IC75 Z, drive PROG_TDO together.
    • TCK/TMS: IC47 SN74LV541A buffers PROG_TCK/PROG_TMS to the FPGA, FMCs and RTM (Y2 = FJTAG.TCK, Y8 = FJTAG.TMS; OE1/OE2 = GND, always enabled). R212 and R436 (0 ohm) tie FJTAG.TCK and FJTAG.TMS back onto PROG_TCK/PROG_TMS, the buffer's own inputs (IC47.2-5 and IC47.6-9).

    Fit data

    R209, R210, R212, R235 and R436 (R0402_0R_JUMPER line), IC47 and IC75 are all Mounted Yes in the BOM and ODB++. Only mounting holes, fiducials and test pads are Mounted No.

    When the chain includes an FMC or the RTM, the scan data is corrupted and the outputs fight.

  8. observationOBS-421majorOpen

    Stuffing-option footprints overlap each other and are all marked fitted: 0-ohm pairs R27/R460, R29/R461, R295/R443, R296/R299, R154/R156, R155/R292, R119/R120 and AC-coupling pairs C182/C838, C183/C839 share the same land

    Nine either/or clock-selection pairs have overlapping footprints and both parts marked fitted: R27/R460, R29/R461, R295/R443, R296/R299, R154/R156, R155/R292, R119/R120 and AC-coupling pairs C182/C838, C183/C839.

    Each pair is an either/or clock-source selection feeding one net. The placement record gives the two parts overlapping bodies on the same side.

    The pairs

    PairOption AOption BJoined net
    R27 / R460RTM_CLK_PAMC_CLK_PRTM_IN_OUT_CLK_P
    R29 / R461N side of the above
    R295 / R443CLK3_BIDIR_PCLK1_M2C_PFMC1_1_3_CLK_P
    R296 / R299N side of the above
    R154 / R156, R155 / R292FMC2 equivalents
    C182 / C838GBTCLK1_M2C (from FMC)FLEX_GTP227MGT227_CLK1_P
    C183 / C839N side of the above
    R119 / R120overlap each other and IC52 FIN1101

    Overlap and fit data

    QuantityValueSource
    Overlap, 0402 pairs0.36-0.53 mm² eachplacement record
    Overlap, 0201 pairs0.36 mm²placement record
    Overlap, R119/R120 with IC521.4 mm²placement record
    Example: R27 / R460 bodies (bottom)x -24.05..-23.43, y 50.49..52.01 / x -24.97..-23.45, y 51.41..52.03placement record
    Example: C182 / C183 bodies (top)x -93.6..-92.5, y 71.62..72.72 / x -93.1..-92.0, y 72.12..73.22placement record
    Fit markingall in R0402_0R_JUMPER or CC0201_100NF lines, Mounted Yes; ODB++ PRP Mounted YesBOM; ODB++

    Only mounting holes, fiducials and test pads are Mounted No on this board.

    Why it matters

    Both parts of a pair cannot physically be placed, and if forced they would join two clock sources. With the other both-fitted options on this board (OBS-396, OBS-411, OBS-412, OBS-416, OBS-420), the fit data for option parts is unreliable as released.

  9. observationOBS-425majorOpen

    Every GTH transceiver lane (all 64 RX/TX pins, quads 224-227) has two or three alternate AC-coupling capacitors to different connectors (FMC J15A/J15B, RTM J2, AMC J3), and all of them are marked fitted

    Every used GTH lane of IC2 (64 MGTHRX/MGTHTX balls, quads 224-227) has 2 or 3 series AC-coupling capacitors to different connectors (FMC J15A/J15B, RTM J2, AMC J3), and all of them are marked fitted.

    A script over all 64 used MGTH RX/TX balls (package table: quads 224-227) finds 2-3 coupling capacitors on every net. All are 100 nF 0201 series capacitors to different destinations.

    Examples

    Lane netIC2 ballCapacitorDestination
    Rx227_0_PM2 (MGTHRXP0_227)C180J15A c6 (DP1.DP0_M2C_P)
    C791J2 C7 (GTP3_RX_P)
    Tx225_0_PAH6 (MGTHTXP0_225)C198J15B c2
    C795J2 E3
    C834J3 133 (TX15_P)
    Rx224_0_PAP2C336J3 47 (PIPE1.RX4_P)
    C76J15B b12

    ``mermaid flowchart LR T["IC2 AH6<br/>Tx225_0_P"] -- "C198" --> F["J15B c2<br/>FMC"] T -- "C795" --> R["J2 E3<br/>RTM"] T -- "C834" --> A["J3 133<br/>AMC TX15_P"] ``

    Fit data

    • All the capacitors are in the CC0201_100NF BOM line (418 parts) marked Mounted Yes, and ODB++ PRP says Mounted Yes.
    • For 16 of the capacitor pairs the placement bodies overlap: a shared-pad layout for build-time selection.

    Why it matters

    As released, each transmitter drives two or three connectors at once, and each receiver sees two or three sources plus open stubs. The routing is evidently meant to be chosen per build by fitting one capacitor per lane.

  10. ReferencePlaneDiscontinuitymajorOpen

    45 nets cross the split between the P1V5 and P3V3MP planes on P6 at 66 places; 62 with a plane on the track's other side (fastest: ADD.GA1 at 100 ps 10-90 %, ASSUMED)

    On P6 the P1V5 and P3V3MP planes part, and 45 nets cross from one to the other at 66 place(s). The return current cannot follow a track across; it detours to wherever the two planes are joined (a stitching capacitor, or the source), and the loop it opens radiates and couples. One decision covers the boundary: whether it is stitched near these crossings, or the nets crossing it are slow enough not to care. Tracks that run ALONG this split, down the slot rather than across it, are a finding of their own: theirs is an impedance step, not a detour. At 62 of the 66 crossing(s) the track's other neighbour carries a plane, which the return can use there. 62 of the 66 crossing(s) have a plane on the track's other side at most twice as far away as the broken one (to first order each plane beside a stripline carries a share of the return inversely proportional to its distance from the track). The rest leave most or all of the return to the break, so the finding is a warning.

    • 'ADD.GA1' on L5 at (-138.84, -24.69), 0.34 mm; other side: no plane on GND
    • 'EEPROM_WP' on L5 at (-169.22, 97.19), 0.38 mm; other side: no plane on GND
    • 'PLL_CTRL.VADJ2_PLL_DAC2_SCLK' on L5 at (-164.75, -17.1), 0.28 mm; other side: no plane on GND
    • 'PLL_CTRL.VADJ2_SI57X_SEC_TUNE/SYNC' on L5 at (-164.86, -13.04), 0.5 mm; other side: no plane on GND
    • 'AJTAG.TCK' on L6 at (-158.33, -24.78), 0.36 mm; other side: GND on G3, 0.1 mm from the track against 0.104 mm to P6: 49% of the return on P6
    • 'AJTAG.TDI' on L6 at (-153.15, -24.78), 0.36 mm; other side: GND on G3, 0.1 mm from the track against 0.104 mm to P6: 49% of the return on P6
    • 'AJTAG.TMS' on L6 at (-159.2, -24.78), 0.36 mm; other side: GND on G3, 0.1 mm from the track against 0.104 mm to P6: 49% of the return on P6
    • 'AJTAG.TRST#' on L6 at (-160.29, -24.78), 0.36 mm; other side: GND on G3, 0.1 mm from the track against 0.104 mm to P6: 49% of the return on P6
    • 'EEPROM_WP' on L6 at (-169.2, 78.32), 0.36 mm; other side: GND on G3, 0.1 mm from the track against 0.104 mm to P6: 49% of the return on P6
    • 'FMC1_CRITICAL' on L5 at (-143.17, -24.78), 0.36 mm; other side: GND on GND, 0.103 mm from the track against 0.1 mm to P6: 51% of the return on P6
    • 'FMC1_CRITICAL' on L5 at (-140.37, -24.78), 0.36 mm; other side: GND on GND, 0.103 mm from the track against 0.1 mm to P6: 51% of the return on P6
    • 'FMC1_CRITICAL' on L5 at (-137.18, -24.78), 0.36 mm; other side: GND on GND, 0.103 mm from the track against 0.1 mm to P6: 51% of the return on P6
    • and 54 more

    Every site is a row of this finding's evidence, keyed by net, layers and position. GET /api/odb/boards/58/reference-planes?group=split%40P6%3AP1V5%7CP3V3MP serves each with its own layer_image region; all of them lie in x -170.37..-93.58, y -30.05..103.78 mm (board frame, Y up).

    Read from the copper as drawn on the reference layer, along each track's centreline at exact breakpoints. The reference is the copper beside the track, never the layer's type or name. Whether it matters depends on each signal's edge rate, which nothing on file states for these nets: a slow static line crosses a split harmlessly, a fast edge does not.

    At split@P6:P1V5|P3V3MP

  11. DecapOrdermajorOpen

    DecapOrder: 5 findings (3 components, 2 pins)×5

    5 checked by the review by Claude walker W4 (Cern session) — 3 different reasons

    At IC16.66, IC2, IC2.J6

    Each finding, with its own headline
    • IC16.66IC16 supply pin 66 on 'P12V': the smallest capacitor C210 sits 4.5 mm away, behind C148 at 0.1 mm

      IC16 supply pin 66 on 'P12V': the smallest capacitor C210 sits 4.5 mm away, behind C148 at 0.1 mm. IC16.66: the smallest value, C210 (1 uF), is 4.5 mm from the pin and C148 (10 uF) 0.1 mm: the high-frequency part sits behind more inductance than the bulk part, which needs it least. Swap them, or bring the smallest value to the pin: it serves the fastest current and is the one the loop inductance defeats. Distances are pad centre to pin centre in the board plane. The 5 mm radius, the 10x value step and the 1 mm / 1.5x tie tolerance are Probe's starting values.

    • IC2IC2 supply pins AC2, AD3, AG2, AH3, C2 on 'MGTAVTT': the smallest capacitor C266 sits 2.0 mm away, behind C273 at 0.6 mm

      IC2 supply pins AC2, AD3, AG2, AH3, C2 on 'MGTAVTT': the smallest capacitor C266 sits 2.0 mm away, behind C273 at 0.6 mm. IC2.AC2: the smallest value, C266 (470 nF), is 2.0 mm from the pin and C273 (4.7 uF) 0.6 mm: the high-frequency part sits behind more inductance than the bulk part, which needs it least. IC2.AD3: the smallest value, C266 (470 nF), is 3.1 mm from the pin and C261 (100 uF) 1.4 mm: the high-frequency part sits behind more inductance than the bulk part, which needs it least. IC2.AG2: the smallest value, C265 (470 nF), is 4.0 mm from the pin and C909 (4.7 uF) 0.6 mm: the high-frequency part sits behind more inductance than the bulk part, which needs it least. IC2.AH3: the smallest value, C265 (470 nF), is 3.2 mm from the pin and C262 (100 uF) 1.4 mm: the high-frequency part sits behind more inductance than the bulk part, which needs it least. IC2.C2: the smallest value, C283 (470 nF), is 3.0 mm from the pin and C253 (100 uF) 1.1 mm: the high-frequency part sits behind more inductance than the bulk part, which needs it least. Swap them, or bring the smallest value to the pin: it serves the fastest current and is the one the loop inductance defeats. Distances are pad centre to pin centre in the board plane. The 5 mm radius, the 10x value step and the 1 mm / 1.5x tie tolerance are Probe's starting values.

    • IC2IC2 supply pins P17, R16, R18, T17, U16, U18, V17, W16, W18 on 'P0V95': the smallest capacitor C906 sits 3.3 mm away, behind C232 at 0.7 mm

      IC2 supply pins P17, R16, R18, T17, U16, U18, V17, W16, W18 on 'P0V95': the smallest capacitor C906 sits 3.3 mm away, behind C232 at 0.7 mm. IC2.P17: the smallest value, C906 (4.7 uF), is 3.3 mm from the pin and C232 (47 uF) 0.7 mm: the high-frequency part sits behind more inductance than the bulk part, which needs it least. IC2.R16: the smallest value, C905 (4.7 uF), is 3.3 mm from the pin and C232 (47 uF) 1.7 mm: the high-frequency part sits behind more inductance than the bulk part, which needs it least. IC2.R18: the smallest value, C906 (4.7 uF), is 4.5 mm from the pin and C232 (47 uF) 0.7 mm: the high-frequency part sits behind more inductance than the bulk part, which needs it least. IC2.T17: the smallest value, C905 (4.7 uF), is 3.7 mm from the pin and C232 (47 uF) 1.7 mm: the high-frequency part sits behind more inductance than the bulk part, which needs it least. IC2.U16: the smallest value, C905 (4.7 uF), is 2.6 mm from the pin and C896 (330 uF) 0.5 mm: the high-frequency part sits behind more inductance than the bulk part, which needs it least. IC2.U18: the smallest value, C905 (4.7 uF), is 4.6 mm from the pin and C901 (100 uF) 1.1 mm: the high-frequency part sits behind more inductance than the bulk part, which needs it least. IC2.V17: the smallest value, C905 (4.7 uF), is 3.7 mm from the pin and C896 (330 uF) 1.8 mm: the high-frequency part sits behind more inductance than the bulk part, which needs it least. IC2.W16: the smallest value, C905 (4.7 uF), is 3.2 mm from the pin and C896 (330 uF) 2.1 mm: the high-frequency part sits behind more inductance than the bulk part, which needs it least. And 1 more pin(s) the same way. Swap them, or bring the smallest value to the pin: it serves the fastest current and is the one the loop inductance defeats. Distances are pad centre to pin centre in the board plane. The 5 mm radius, the 10x value step and the 1 mm / 1.5x tie tolerance are Probe's starting values.

    • IC2IC2 supply pins AG18, AP17, E14, F11, G18, U28, Y29 on 'P1V8': the smallest capacitor C920 sits 3.6 mm away, behind C933 at 1.2 mm

      IC2 supply pins AG18, AP17, E14, F11, G18, U28, Y29 on 'P1V8': the smallest capacitor C920 sits 3.6 mm away, behind C933 at 1.2 mm. IC2.AG18: the smallest value, C920 (4.7 uF), is 3.6 mm from the pin and C933 (47 uF) 1.2 mm: the high-frequency part sits behind more inductance than the bulk part, which needs it least. IC2.AP17: the smallest value, C917 (4.7 uF), is 2.1 mm from the pin and C914 (47 uF) 0.5 mm: the high-frequency part sits behind more inductance than the bulk part, which needs it least. IC2.E14: the smallest value, C275 (470 nF), is 4.4 mm from the pin and C943 (10 uF) 2.8 mm: the high-frequency part sits behind more inductance than the bulk part, which needs it least. IC2.F11: the smallest value, C275 (470 nF), is 3.2 mm from the pin and C943 (10 uF) 1.2 mm: the high-frequency part sits behind more inductance than the bulk part, which needs it least. IC2.G18: the smallest value, C277 (470 nF), is 4.4 mm from the pin and C911 (47 uF) 1.2 mm: the high-frequency part sits behind more inductance than the bulk part, which needs it least. IC2.U28: the smallest value, C919 (4.7 uF), is 4.9 mm from the pin and C941 (47 uF) 1.1 mm: the high-frequency part sits behind more inductance than the bulk part, which needs it least. IC2.Y29: the smallest value, C919 (4.7 uF), is 4.4 mm from the pin and C941 (47 uF) 2.6 mm: the high-frequency part sits behind more inductance than the bulk part, which needs it least. Swap them, or bring the smallest value to the pin: it serves the fastest current and is the one the loop inductance defeats. Distances are pad centre to pin centre in the board plane. The 5 mm radius, the 10x value step and the 1 mm / 1.5x tie tolerance are Probe's starting values.

    • IC2.J6IC2 supply pin J6 on 'MGTAVCC': the smallest capacitor C248 sits 3.2 mm away, behind C237 at 1.1 mm

      IC2 supply pin J6 on 'MGTAVCC': the smallest capacitor C248 sits 3.2 mm away, behind C237 at 1.1 mm. IC2.J6: the smallest value, C248 (470 nF), is 3.2 mm from the pin and C237 (100 uF) 1.1 mm: the high-frequency part sits behind more inductance than the bulk part, which needs it least. Swap them, or bring the smallest value to the pin: it serves the fastest current and is the one the loop inductance defeats. Distances are pad centre to pin centre in the board plane. The 5 mm radius, the 10x value step and the 1 mm / 1.5x tie tolerance are Probe's starting values.

  12. EsdProtectorPlacementmajorOpen

    EsdProtectorPlacement: 5 components×5

    5 checked by the review by Claude walker W4 (Cern session) — 5 different reasons

    At D11, D19, D26, IC4, IC5

    Each finding, with its own headline
    • D11D11.1 on 'AJTAG.TMS': 1 via(s) between J3.166 and D11.1

      D11 (ESD8351XV2T1G) protects 'AJTAG.TMS' at J3.166. Along the copper: 6.76 mm of track from the contact, 1 via(s) before it, on a 2.99 mm branch off the line; the line runs 227.84 mm on to IC46.12; D11 sits 4.68 mm farther from the contact than the nearest place outside J3.

      1 via(s) between J3.166 and D11.1: the strike reaches an inner layer before it is clamped. D11.1 hangs 2.99 mm off the line (limit 1 mm, default: no datasheet figure on file): the line carries the strike past it while the clamp is reached through the branch.

      Measured on the copper the ODB states: tracks, the pads they land on, and the netlist's vias. Not a straight line between centroids.

    • D19D19.1 on 'PROG_TCK': 1 via(s) between J6.6 and D19.1

      D19 (ESD8351XV2T1G) protects 'PROG_TCK' at J6.6. Along the copper: 2.49 mm of track from the contact, 1 via(s) before it, on a 1.54 mm branch off the line; the line runs 128.37 mm on to R212.2; D19 sits at J6, within 0.15 mm of the nearest place outside it to the contact: the track to it is mostly the escape out of J6's own pins, which no placement shortens.

      1 via(s) between J6.6 and D19.1: the strike reaches an inner layer before it is clamped. D19.1 hangs 1.54 mm off the line (limit 1 mm, default: no datasheet figure on file): the line carries the strike past it while the clamp is reached through the branch.

      Measured on the copper the ODB states: tracks, the pads they land on, and the netlist's vias. Not a straight line between centroids.

    • D26D26.3 on 'USB_D_R_N': D26 joins the line 14.77 mm from 2.2 on a 27.51 mm run to IC21.7

      D26 (PRTR5V0U2X) protects 'USB_D_R_N' at 2.2. Along the copper: 14.77 mm of track from the contact, 0 via(s) before it, the line runs through its pad (flow-through); the line runs 27.51 mm on to IC21.7; D26 sits 10.1 mm farther from the contact than the nearest place outside 2.

      D26 joins the line 14.77 mm from 2.2 on a 27.51 mm run to IC21.7: nearer the far end than the contact, so the surge travels most of the line before it is clamped.

      Measured on the copper the ODB states: tracks, the pads they land on, and the netlist's vias. Not a straight line between centroids.

    • IC4IC4.4 on 'IPMB.SCL': IC4.4 hangs 2.38 mm off the line (limit 1 mm, default

      IC4 (BZA462A) protects 'IPMB.SCL' at J3.56. Along the copper: 4.69 mm of track from the contact, 0 via(s) before it, on a 2.38 mm branch off the line; the line runs 376.87 mm on to IC1.24; IC4 sits 2.3 mm farther from the contact than the nearest place outside J3.

      IC4.4 hangs 2.38 mm off the line (limit 1 mm, default: no datasheet figure on file): the line carries the strike past it while the clamp is reached through the branch.

      Measured on the copper the ODB states: tracks, the pads they land on, and the netlist's vias. Not a straight line between centroids.

    • IC5IC5.4 on 'ADD.GA0': IC5.4 hangs 4.82 mm off the line (limit 1 mm, default

      IC5 (BZA462A) protects 'ADD.GA0' at J3.5. Along the copper: 15.29 mm of track from the contact, 0 via(s) before it, on a 4.82 mm branch off the line; the line runs 276.77 mm on to R7.1; IC5 sits 8.36 mm farther from the contact than the nearest place outside J3.

      IC5.4 hangs 4.82 mm off the line (limit 1 mm, default: no datasheet figure on file): the line carries the strike past it while the clamp is reached through the branch.

      Measured on the copper the ODB states: tracks, the pads they land on, and the netlist's vias. Not a straight line between centroids.

  13. PartLifecyclemajorOpen

    PartLifecycle: 12 part numbers×12

    11 Obsolete, 1 Not For New Designs · on 32 placements

    At 2, C115, IC30, IC46, IC73, L11 and 6 more

    Each finding, with its own headline
    • 2ZX62RD-AB-5P8 on 2: the distributor catalogue reports "Obsolete" for Hirose Electric Co Ltd's ZX62RD-AB-5P8

      ZX62RD-AB-5P8 on 2: the distributor catalogue reports "Obsolete" for Hirose Electric Co Ltd's ZX62RD-AB-5P8 — a sourcing/longevity risk to confirm against the MANUFACTURER before acting. That is the maker this board's BOM names (HIROSE (HRS)). A distributor delisting is not the same claim as the part being unavailable; check the vendor's own lifecycle/product-change notice for the status and any named successor. Re-read the source: GET /api/odb/parts/ZX62RD-AB-5P8/providers (which provider said it, the maker and part number it matched, its own words, and when we asked). THE LAST BUY DATE HAS PASSED (2016-03-31, 3,881 day(s) ago), so this is a substitution rather than a buy. The distributor reports NONE left. No approved alternate is recorded for this position.

    • C115ECASD40G157M016K00 on 8 placements (C115, C135, C34, C35, C473, C528, C539, C86): the distributor catalogue reports "Obsolete" for Murata Electronics's ECASD40G157M016K00

      ECASD40G157M016K00 on 8 placements (C115, C135, C34, C35, C473, C528, C539, C86): the distributor catalogue reports "Obsolete" for Murata Electronics's ECASD40G157M016K00 — a sourcing/longevity risk to confirm against the MANUFACTURER before acting. That is the maker this board's BOM names (MURATA). A distributor delisting is not the same claim as the part being unavailable; check the vendor's own lifecycle/product-change notice for the status and any named successor. Re-read the source: GET /api/odb/parts/ECASD40G157M016K00/providers (which provider said it, the maker and part number it matched, its own words, and when we asked). No approved alternate is recorded for this position.

    • IC30LM75AIM/NOPB on 4 placements (IC30, IC31, IC32, IC33): the distributor catalogue reports "Obsolete" for Texas Instruments's LM75AIM/NOPB

      LM75AIM/NOPB on 4 placements (IC30, IC31, IC32, IC33): the distributor catalogue reports "Obsolete" for Texas Instruments's LM75AIM/NOPB — a sourcing/longevity risk to confirm against the MANUFACTURER before acting. That is the maker this board's BOM names (TEXAS INSTRUMENTS). A distributor delisting is not the same claim as the part being unavailable; check the vendor's own lifecycle/product-change notice for the status and any named successor. Re-read the source: GET /api/odb/parts/LM75AIM/NOPB/providers (which provider said it, the maker and part number it matched, its own words, and when we asked). THE LAST BUY DATE HAS PASSED (2024-02-09, 1,010 day(s) ago), so this is a substitution rather than a buy. The distributor reports NONE left. No approved alternate is recorded for this position.

    • IC46SN74CB3T3257PW on IC46: the distributor catalogue reports "Obsolete" for Texas Instruments's SN74CB3T3257PW

      SN74CB3T3257PW on IC46: the distributor catalogue reports "Obsolete" for Texas Instruments's SN74CB3T3257PW — a sourcing/longevity risk to confirm against the MANUFACTURER before acting. That is the maker this board's BOM names (TEXAS INSTRUMENTS). A distributor delisting is not the same claim as the part being unavailable; check the vendor's own lifecycle/product-change notice for the status and any named successor. Re-read the source: GET /api/odb/parts/SN74CB3T3257PW/providers (which provider said it, the maker and part number it matched, its own words, and when we asked). THE LAST BUY DATE HAS PASSED (2024-12-12, 703 day(s) ago), so this is a substitution rather than a buy. The distributor reports 81 left. No approved alternate is recorded for this position.

    • IC738V54816ANLG on IC73: the distributor catalogue reports "Obsolete" for Renesas Electronics Corporation's 8V54816ANLG

      8V54816ANLG on IC73: the distributor catalogue reports "Obsolete" for Renesas Electronics Corporation's 8V54816ANLG — a sourcing/longevity risk to confirm against the MANUFACTURER before acting. This board's BOM names INTEGRATED DEVICE TECHNOLOGY (IDT), and that is one line with Renesas Electronics Corporation: Renesas completed its acquisitions of Intersil on 2017-02-24, Integrated Device Technology on 2019-03-29 and Dialog Semiconductor (which had acquired Adesto Technologies in 2020) on 2021-08-31. A distributor delisting is not the same claim as the part being unavailable; check the vendor's own lifecycle/product-change notice for the status and any named successor. Re-read the source: GET /api/odb/parts/8V54816ANLG/providers (which provider said it, the maker and part number it matched, its own words, and when we asked). THE LAST BUY DATE HAS PASSED (2023-02-10, 1,374 day(s) ago), so this is a substitution rather than a buy. The distributor reports NONE left. No approved alternate is recorded for this position.

    • L11ELLVGG1R0N on 3 placements (L11, L13, L7): the distributor catalogue reports "Obsolete" for Panasonic Industry's ELL-VGG1R0N

      ELLVGG1R0N on 3 placements (L11, L13, L7): the distributor catalogue reports "Obsolete" for Panasonic Industry's ELL-VGG1R0N — a sourcing/longevity risk to confirm against the MANUFACTURER before acting. That is the maker this board's BOM names (PANASONIC). A distributor delisting is not the same claim as the part being unavailable; check the vendor's own lifecycle/product-change notice for the status and any named successor. Re-read the source: GET /api/odb/parts/ELLVGG1R0N/providers (which provider said it, the maker and part number it matched, its own words, and when we asked). THE LAST BUY DATE HAS PASSED (2019-09-25, 2,608 day(s) ago), so this is a substitution rather than a buy. The distributor reports NONE left. No approved alternate is recorded for this position.

    • OSC1549BACB001937ABG on 2 placements (OSC1, OSC2): the distributor catalogue reports "Not For New Designs" for Skyworks Solutions Inc.'s 549BACB001937ABG

      549BACB001937ABG on 2 placements (OSC1, OSC2): the distributor catalogue reports "Not For New Designs" for Skyworks Solutions Inc.'s 549BACB001937ABG — a sourcing/longevity risk to confirm against the MANUFACTURER before acting. This board's BOM names SILICON LABS, and that is one line with Skyworks Solutions Inc.: Skyworks completed its acquisition of Silicon Labs' Infrastructure & Automotive business (timing, isolation) on 2021-07-26; Silicon Labs still sells its other lines under its own name. A distributor delisting is not the same claim as the part being unavailable; check the vendor's own lifecycle/product-change notice for the status and any named successor. Re-read the source: GET /api/odb/parts/549BACB001937ABG/providers (which provider said it, the maker and part number it matched, its own words, and when we asked). No approved alternate is recorded for this position.

    • Q1IRF9317TRPBF on 3 placements (Q1, Q13, Q2): the distributor catalogue reports "Obsolete" for Infineon Technologies's IRF9317TRPBF

      IRF9317TRPBF on 3 placements (Q1, Q13, Q2): the distributor catalogue reports "Obsolete" for Infineon Technologies's IRF9317TRPBF — a sourcing/longevity risk to confirm against the MANUFACTURER before acting. This board's BOM names INTERNATIONAL RECTIFIER, and that is one line with Infineon Technologies: Infineon completed its acquisitions of International Rectifier on 2015-01-13 and Cypress Semiconductor on 2020-04-16. A distributor delisting is not the same claim as the part being unavailable; check the vendor's own lifecycle/product-change notice for the status and any named successor. Re-read the source: GET /api/odb/parts/IRF9317TRPBF/providers (which provider said it, the maker and part number it matched, its own words, and when we asked). THE LAST BUY DATE HAS PASSED (2025-12-31, 319 day(s) ago), so this is a substitution rather than a buy. The distributor reports NONE left. No approved alternate is recorded for this position.

    • Q17Si2315BDS-T1-E3 on Q17: the distributor catalogue reports "Obsolete" for Vishay Siliconix's SI2315BDS-T1-E3

      Si2315BDS-T1-E3 on Q17: the distributor catalogue reports "Obsolete" for Vishay Siliconix's SI2315BDS-T1-E3 — a sourcing/longevity risk to confirm against the MANUFACTURER before acting. That is the maker this board's BOM names (VISHAY SILICONIX). A distributor delisting is not the same claim as the part being unavailable; check the vendor's own lifecycle/product-change notice for the status and any named successor. Re-read the source: GET /api/odb/parts/Si2315BDS-T1-E3/providers (which provider said it, the maker and part number it matched, its own words, and when we asked). THE LAST BUY DATE HAS PASSED (2025-01-21, 663 day(s) ago), so this is a substitution rather than a buy. The distributor reports NONE left. No approved alternate is recorded for this position.

    • Q18IRF9310TRPBF on Q18: the distributor catalogue reports "Obsolete" for Infineon Technologies's IRF9310TRPBF

      IRF9310TRPBF on Q18: the distributor catalogue reports "Obsolete" for Infineon Technologies's IRF9310TRPBF — a sourcing/longevity risk to confirm against the MANUFACTURER before acting. This board's BOM names INTERNATIONAL RECTIFIER, and that is one line with Infineon Technologies: Infineon completed its acquisitions of International Rectifier on 2015-01-13 and Cypress Semiconductor on 2020-04-16. A distributor delisting is not the same claim as the part being unavailable; check the vendor's own lifecycle/product-change notice for the status and any named successor. Re-read the source: GET /api/odb/parts/IRF9310TRPBF/providers (which provider said it, the maker and part number it matched, its own words, and when we asked). THE LAST BUY DATE HAS PASSED (2025-12-31, 319 day(s) ago), so this is a substitution rather than a buy. The distributor reports NONE left. No approved alternate is recorded for this position.

    • Q19IRF9910TRPBF on 6 placements (Q19, Q20, Q21, Q22, Q23, Q24): the distributor catalogue reports "Obsolete" for Infineon Technologies's IRF9910TRPBF

      IRF9910TRPBF on 6 placements (Q19, Q20, Q21, Q22, Q23, Q24): the distributor catalogue reports "Obsolete" for Infineon Technologies's IRF9910TRPBF — a sourcing/longevity risk to confirm against the MANUFACTURER before acting. This board's BOM names INTERNATIONAL RECTIFIER, and that is one line with Infineon Technologies: Infineon completed its acquisitions of International Rectifier on 2015-01-13 and Cypress Semiconductor on 2020-04-16. A distributor delisting is not the same claim as the part being unavailable; check the vendor's own lifecycle/product-change notice for the status and any named successor. Re-read the source: GET /api/odb/parts/IRF9910TRPBF/providers (which provider said it, the maker and part number it matched, its own words, and when we asked). THE LAST BUY DATE HAS PASSED (2023-03-31, 1,325 day(s) ago), so this is a substitution rather than a buy. The distributor reports NONE left. No approved alternate is recorded for this position.

    • X17A-8.000MAAE-T on X1: the distributor catalogue reports "Obsolete" for TXC CORPORATION's 7A-8.000MAAE-T

      7A-8.000MAAE-T on X1: the distributor catalogue reports "Obsolete" for TXC CORPORATION's 7A-8.000MAAE-T — a sourcing/longevity risk to confirm against the MANUFACTURER before acting. That is the maker this board's BOM names (TXC CORPORATION). A distributor delisting is not the same claim as the part being unavailable; check the vendor's own lifecycle/product-change notice for the status and any named successor. Re-read the source: GET /api/odb/parts/7A-8.000MAAE-T/providers (which provider said it, the maker and part number it matched, its own words, and when we asked). THE LAST BUY DATE HAS PASSED (2021-09-30, 1,872 day(s) ago), so this is a substitution rather than a buy. The distributor reports NONE left. No approved alternate is recorded for this position.

  14. ReferencePlaneDiscontinuitymajorOpen

    ReferencePlaneDiscontinuity: 20 boards×20

    e.g. “'MLVDS_F.MLVDS_I_8' runs ALONG the split between the MGTAVCC and P0V95 planes on P3 at 1 place, the longest 5.06 mm; the impedance rises from 56 ohm to 62 ohm there at worst, 5.3 % of each edge reflected (fastest: MLVDS_F.MLVDS_I_8 at 100 ps 10-90 %, ASSUMED)” · 20 different statements

    At along@P3:MGTAVCC|P0V95, along@P3:P0V95|PVADJ1, along@P3:P1V5|P1V8, along@P3:P1V8|P3V3, along@P3:P1V8|PVADJ1, along@P3:P1V8|PVADJ2 and 14 more

    Each finding, with its own headline
    • along@P3:MGTAVCC|P0V95'MLVDS_F.MLVDS_I_8' runs ALONG the split between the MGTAVCC and P0V95 planes on P3 at 1 place, the longest 5.06 mm; the impedance rises from 56 ohm to 62 ohm there at worst, 5.3 % of each edge reflected (fastest: MLVDS_F.MLVDS_I_8 at 100 ps 10-90 %, ASSUMED)

      On P3 the MGTAVCC and P0V95 planes part, and 'MLVDS_F.MLVDS_I_8' runs down the slot between them at 1 place(s), rather than across it. This is not a crossing: the return still follows the track, in the plane on its other side and in the slot's edges, so it does not detour. What changes is the track's own impedance. With the plane gone from under it on this side for that length, its field reaches further, to the slot's edges and to the next plane beyond, so its capacitance falls and its impedance rises; every edge travelling down it is partly reflected where it enters the stretch and again where it leaves, by (Z2 - Z1)/(Z2 + Z1). The longer the stretch against the edge's rise, the more of the edge sees the full step. A crossing is generally the worse fault; a mismatch is not free. Over solid plane each track is 56 ohm; along the slot 62 ohm, reflecting up to 5.3 % of each edge (a method-of-moments field solve of the cross-section, one dielectric; were the slot's edges to carry none of the return, up to 64 ohm). No place stays within 10 %, the tolerance controlled impedance is commonly built to (the stackup states none), so the finding is a warning. One decision covers the split line's slot: whether these tracks may run down it, or should be moved over solid plane.

      • 'MLVDS_F.MLVDS_I_8' on L2 at (-60.45, 43.6), 5.57 mm, running ALONG the split for 5.06 mm, not across it, 56 ohm over solid plane, 62 ohm along it: 5.3 % of each edge reflected; other side: GND on G2, 0.103 mm from the track against 0.1 mm to P3: 51% of the return on P3

      Every site is a row of this finding's evidence, keyed by net, layers and position. GET /api/odb/boards/58/reference-planes?group=along%40P3%3AMGTAVCC%7CP0V95 serves each with its own layer_image region; all of them lie in x -61.45..-59.15, y 40.65..48.1 mm (board frame, Y up).

      Read from the copper as drawn on the reference layer, along each track's centreline at exact breakpoints. The reference is the copper beside the track, never the layer's type or name. Whether it matters depends on each signal's edge rate, which nothing on file states for these nets: a slow static line crosses a split harmlessly, a fast edge does not.

    • along@P3:P0V95|PVADJ13 nets run ALONG the split between the P0V95 and PVADJ1 planes on P3 at 3 places, the longest 1.4 mm; the impedance rises from 61 ohm to 68 ohm there at worst, 5.4 % of each edge reflected (fastest: HA2.HA08_N at 100 ps 10-90 %, ASSUMED)

      On P3 the P0V95 and PVADJ1 planes part, and 3 nets run down the slot between them at 3 place(s), rather than across it. This is not a crossing: the return still follows the track, in the plane on its other side and in the slot's edges, so it does not detour. What changes is the track's own impedance. With the plane gone from under it on this side for that length, its field reaches further, to the slot's edges and to the next plane beyond, so its capacitance falls and its impedance rises; every edge travelling down it is partly reflected where it enters the stretch and again where it leaves, by (Z2 - Z1)/(Z2 + Z1). The longer the stretch against the edge's rise, the more of the edge sees the full step. A crossing is generally the worse fault; a mismatch is not free. Over solid plane each track is 61 ohm; along the slot 66 ohm to 68 ohm, reflecting up to 5.4 % of each edge (a method-of-moments field solve of the cross-section, one dielectric; were the slot's edges to carry none of the return, up to 69 ohm). At 1 of the 3 place(s) the rise is within 10 %, the tolerance controlled impedance is commonly built to (the stackup states none); the rest exceed it or cannot be said, so the finding is a warning. One decision covers the split line's slot: whether these tracks may run down it, or should be moved over solid plane.

      • 'HA2.HA11_N' on L1 at (-75.72, 14.46), 2.02 mm, running ALONG the split for 1.4 mm, not across it, 61 ohm over solid plane, 68 ohm along it: 5.4 % of each edge reflected; other side: GND on G1, 0.1 mm from the track against 0.103 mm to P3: 49% of the return on P3
      • 'HA2.HA08_N' on L2 at (-75.81, 14.54), 2.14 mm, running ALONG the split for 1.38 mm, not across it, 61 ohm over solid plane, 67 ohm along it: 5.1 % of each edge reflected; other side: GND on G2, 0.103 mm from the track against 0.1 mm to P3: 51% of the return on P3
      • 'HA2.HA08_P' on L2 at (-75.67, 14.38), 1.93 mm, running ALONG the split for 1.36 mm, not across it, 61 ohm over solid plane, 66 ohm along it: 4.6 % of each edge reflected; other side: GND on G2, 0.103 mm from the track against 0.1 mm to P3: 51% of the return on P3

      Every site is a row of this finding's evidence, keyed by net, layers and position. GET /api/odb/boards/58/reference-planes?group=along%40P3%3AP0V95%7CPVADJ1 serves each with its own layer_image region; all of them lie in x -77.64..-74.13, y 12.56..16.23 mm (board frame, Y up).

      Read from the copper as drawn on the reference layer, along each track's centreline at exact breakpoints. The reference is the copper beside the track, never the layer's type or name. Whether it matters depends on each signal's edge rate, which nothing on file states for these nets: a slow static line crosses a split harmlessly, a fast edge does not.

    • along@P3:P1V5|P1V87 nets run ALONG the split between the P1V5 and P1V8 planes on P3 at 7 places, the longest 39.71 mm; the impedance rises from 53 ohm to 60 ohm there at worst, 6.6 % of each edge reflected (fastest: DDR3_A13 at 100 ps 10-90 %, ASSUMED)

      On P3 the P1V5 and P1V8 planes part, and 7 nets run down the slot between them at 7 place(s), rather than across it. This is not a crossing: the return still follows the track, in the plane on its other side and in the slot's edges, so it does not detour. What changes is the track's own impedance. With the plane gone from under it on this side for that length, its field reaches further, to the slot's edges and to the next plane beyond, so its capacitance falls and its impedance rises; every edge travelling down it is partly reflected where it enters the stretch and again where it leaves, by (Z2 - Z1)/(Z2 + Z1). The longer the stretch against the edge's rise, the more of the edge sees the full step. A crossing is generally the worse fault; a mismatch is not free. Over solid plane each track is 52 ohm to 61 ohm; along the slot 59 ohm to 68 ohm, reflecting up to 6.6 % of each edge (a method-of-moments field solve of the cross-section, one dielectric; were the slot's edges to carry none of the return, up to 69 ohm). At 2 of the 7 place(s) the rise is within 10 %, the tolerance controlled impedance is commonly built to (the stackup states none); the rest exceed it or cannot be said, so the finding is a warning. One decision covers the split line's slot: whether these tracks may run down it, or should be moved over solid plane.

      • 'DDR3_A13' on L2 at (-81.93, 107.1), 2.62 mm, running ALONG the split for 2.05 mm, not across it, 53 ohm over solid plane, 60 ohm along it: 6.6 % of each edge reflected; other side: GND on G2, 0.103 mm from the track against 0.1 mm to P3: 51% of the return on P3
      • 'PM.EN_RTM_PWR' on L2 at (-56.81, 107.18), 39.58 mm, running ALONG the split for 39.16 mm, not across it, 52 ohm over solid plane, 59 ohm along it: 6.1 % of each edge reflected; other side: GND on G2, 0.103 mm from the track against 0.1 mm to P3: 51% of the return on P3
      • 'HB1.HB15_P' on L2 at (-90.51, 90.26), 4.09 mm, running ALONG the split for 3.53 mm, not across it, 61 ohm over solid plane, 68 ohm along it: 5.7 % of each edge reflected; other side: GND on G2, 0.103 mm from the track against 0.1 mm to P3: 51% of the return on P3
      • 'LA2.LA22_P' on L2 at (-88.79, 53.71), 9.29 mm, running ALONG the split for 8.72 mm, not across it, 61 ohm over solid plane, 68 ohm along it: 5.6 % of each edge reflected; other side: GND on G2, 0.103 mm from the track against 0.1 mm to P3: 51% of the return on P3
      • 'DONE' on L2 at (-57.07, 107), 40.51 mm, running ALONG the split for 39.71 mm, not across it, 59 ohm over solid plane, 66 ohm along it: 5.2 % of each edge reflected; other side: GND on G2, 0.103 mm from the track against 0.1 mm to P3: 51% of the return on P3
      • 'LA2.LA22_N' on L2 at (-88.58, 53.69), 9.12 mm, running ALONG the split for 8.65 mm, not across it, 61 ohm over solid plane, 65 ohm along it: 3.5 % of each edge reflected; other side: GND on G2, 0.103 mm from the track against 0.1 mm to P3: 51% of the return on P3
      • 'HB1.HB15_N' on L2 at (-90.93, 90.39), 4.09 mm, running ALONG the split for 3.53 mm, not across it, 61 ohm over solid plane, 65 ohm along it: 3.3 % of each edge reflected; other side: GND on G2, 0.103 mm from the track against 0.1 mm to P3: 51% of the return on P3

      Every site is a row of this finding's evidence, keyed by net, layers and position. GET /api/odb/boards/58/reference-planes?group=along%40P3%3AP1V5%7CP1V8 serves each with its own layer_image region; all of them lie in x -93.45..-35.92, y 48.02..108.3 mm (board frame, Y up).

      Read from the copper as drawn on the reference layer, along each track's centreline at exact breakpoints. The reference is the copper beside the track, never the layer's type or name. Whether it matters depends on each signal's edge rate, which nothing on file states for these nets: a slow static line crosses a split harmlessly, a fast edge does not.

    • along@P3:P1V8|P3V315 nets run ALONG the split between the P1V8 and P3V3 planes on P3 at 15 places, the longest 28.89 mm; the impedance rises from 52 ohm to 59 ohm there at worst, 6.7 % of each edge reflected (fastest: DONE at 100 ps 10-90 %, ASSUMED)

      On P3 the P1V8 and P3V3 planes part, and 15 nets run down the slot between them at 15 place(s), rather than across it. This is not a crossing: the return still follows the track, in the plane on its other side and in the slot's edges, so it does not detour. What changes is the track's own impedance. With the plane gone from under it on this side for that length, its field reaches further, to the slot's edges and to the next plane beyond, so its capacitance falls and its impedance rises; every edge travelling down it is partly reflected where it enters the stretch and again where it leaves, by (Z2 - Z1)/(Z2 + Z1). The longer the stretch against the edge's rise, the more of the edge sees the full step. A crossing is generally the worse fault; a mismatch is not free. Over solid plane each track is 41 ohm to 61 ohm; along the slot 45 ohm to 68 ohm, reflecting up to 6.7 % of each edge (a method-of-moments field solve of the cross-section, one dielectric; were the slot's edges to carry none of the return, up to 69 ohm). At 4 of the 15 place(s) the rise is within 10 %, the tolerance controlled impedance is commonly built to (the stackup states none); the rest exceed it or cannot be said, so the finding is a warning. One decision covers the split line's slot: whether these tracks may run down it, or should be moved over solid plane.

      • 'EN_USB_JTAG' on L2 at (-110.35, -23.37), 14.16 mm, running ALONG the split for 13.53 mm, not across it, 52 ohm over solid plane, 59 ohm along it: 6.7 % of each edge reflected; other side: GND on G2, 0.103 mm from the track against 0.1 mm to P3: 51% of the return on P3
      • 'SCL0' on L2 at (-142.32, -26.21), 17.53 mm, running ALONG the split for 16.97 mm, not across it, 52 ohm over solid plane, 59 ohm along it: 6.3 % of each edge reflected; other side: GND on G2, 0.103 mm from the track against 0.1 mm to P3: 51% of the return on P3
      • 'TRACEDATA2' on L2 at (-124.59, 56.11), 8.99 mm, running ALONG the split for 8.25 mm, not across it, 52 ohm over solid plane, 59 ohm along it: 6.3 % of each edge reflected; other side: GND on G2, 0.103 mm from the track against 0.1 mm to P3: 51% of the return on P3
      • 'CLK_DIRB' on L1 at (-110.99, -23.44), 10.42 mm, running ALONG the split for 9.63 mm, not across it, 45 ohm over solid plane, 51 ohm along it: 6.3 % of each edge reflected; other side: GND on G1, 0.1 mm from the track against 0.103 mm to P3: 49% of the return on P3
      • 'TRACEDATA3' on L1 at (-119.86, -7.65), 1.04 mm, running ALONG the split for 0.48 mm, not across it, 41 ohm over solid plane, 46 ohm along it: 6.1 % of each edge reflected; other side: GND on G1, 0.1 mm from the track against 0.103 mm to P3: 49% of the return on P3
      • 'MGT.FLEX_GTP224_CLK1_N' on L2 at (-126.9, 21.65), 6.38 mm, running ALONG the split for 5.91 mm, not across it, 61 ohm over solid plane, 68 ohm along it: 5.8 % of each edge reflected; other side: GND on G2, 0.103 mm from the track against 0.1 mm to P3: 51% of the return on P3
      • 'FMC1_1_3_CLK_N' on L1 at (-132.12, 20.95), 3.38 mm, running ALONG the split for 2.73 mm, not across it, 61 ohm over solid plane, 68 ohm along it: 5.5 % of each edge reflected; other side: GND on G1, 0.1 mm from the track against 0.103 mm to P3: 49% of the return on P3
      • 'OP15C_C_N' on L1 at (-119.93, -10.05), 2.66 mm, running ALONG the split for 2.1 mm, not across it, 61 ohm over solid plane, 68 ohm along it: 5.5 % of each edge reflected; other side: GND on G1, 0.1 mm from the track against 0.103 mm to P3: 49% of the return on P3
      • 'TRACEDATA0' on L2 at (-122.91, 44.29), 12.49 mm, running ALONG the split for 12.11 mm, not across it, 52 ohm over solid plane, 58 ohm along it: 5.3 % of each edge reflected; other side: GND on G2, 0.103 mm from the track against 0.1 mm to P3: 51% of the return on P3
      • 'CLKS1.CLK0_M2C_P' on L1 at (-134.01, 21.54), 2.9 mm, running ALONG the split for 2.49 mm, not across it, 61 ohm over solid plane, 67 ohm along it: 5.2 % of each edge reflected; other side: GND on G1, 0.1 mm from the track against 0.103 mm to P3: 49% of the return on P3
      • 'TRACECLK' on L1 at (-119.83, -3.5), 3.67 mm, running ALONG the split for 2.75 mm, not across it, 41 ohm over solid plane, 45 ohm along it: 5.1 % of each edge reflected; other side: GND on G1, 0.1 mm from the track against 0.103 mm to P3: 49% of the return on P3
      • 'DONE' on L2 at (-107.98, 107.67), 4.95 mm, running ALONG the split for 4.42 mm, not across it, 59 ohm over solid plane, 65 ohm along it: 4.6 % of each edge reflected; other side: GND on G2, 0.103 mm from the track against 0.1 mm to P3: 51% of the return on P3
      • and 3 more

      Every site is a row of this finding's evidence, keyed by net, layers and position. GET /api/odb/boards/58/reference-planes?group=along%40P3%3AP1V8%7CP3V3 serves each with its own layer_image region; all of them lie in x -152.06..-101.14, y -27.35..108.75 mm (board frame, Y up).

      Read from the copper as drawn on the reference layer, along each track's centreline at exact breakpoints. The reference is the copper beside the track, never the layer's type or name. Whether it matters depends on each signal's edge rate, which nothing on file states for these nets: a slow static line crosses a split harmlessly, a fast edge does not.

    • along@P3:P1V8|PVADJ15 nets run ALONG the split between the P1V8 and PVADJ1 planes on P3 at 5 places, the longest 2.63 mm; the impedance rises from 61 ohm to 68 ohm there at worst, 5.8 % of each edge reflected (fastest: AMC_RTM_PV at 100 ps 10-90 %, ASSUMED)

      On P3 the P1V8 and PVADJ1 planes part, and 5 nets run down the slot between them at 5 place(s), rather than across it. This is not a crossing: the return still follows the track, in the plane on its other side and in the slot's edges, so it does not detour. What changes is the track's own impedance. With the plane gone from under it on this side for that length, its field reaches further, to the slot's edges and to the next plane beyond, so its capacitance falls and its impedance rises; every edge travelling down it is partly reflected where it enters the stretch and again where it leaves, by (Z2 - Z1)/(Z2 + Z1). The longer the stretch against the edge's rise, the more of the edge sees the full step. A crossing is generally the worse fault; a mismatch is not free. Over solid plane each track is 29 ohm to 61 ohm; along the slot 31 ohm to 68 ohm, reflecting up to 5.8 % of each edge (a method-of-moments field solve of the cross-section, one dielectric; were the slot's edges to carry none of the return, up to 69 ohm). At 3 of the 5 place(s) the rise is within 10 %, the tolerance controlled impedance is commonly built to (the stackup states none); the rest exceed it or cannot be said, so the finding is a warning. One decision covers the split line's slot: whether these tracks may run down it, or should be moved over solid plane.

      • 'HA1.HA00_CC_N' on L2 at (-105.04, 54.65), 3.19 mm, running ALONG the split for 2.63 mm, not across it, 61 ohm over solid plane, 68 ohm along it: 5.8 % of each edge reflected; other side: GND on G2, 0.103 mm from the track against 0.1 mm to P3: 51% of the return on P3
      • 'AMC_RTM_PV' on L1 at (-110.99, 100.53), 2.05 mm, running ALONG the split for 1.53 mm, not across it, 52 ohm over solid plane, 58 ohm along it: 5.6 % of each edge reflected; other side: GND on G1, 0.1 mm from the track against 0.103 mm to P3: 49% of the return on P3
      • 'LA1.LA13_N' on L1 at (-105.2, 67.03), 1.73 mm, running ALONG the split for 1.17 mm, not across it, 61 ohm over solid plane, 66 ohm along it: 4 % of each edge reflected; other side: GND on G1, 0.1 mm from the track against 0.103 mm to P3: 49% of the return on P3
      • 'NetLD3_2' on L2 at (-114.4, 100.41), 1.53 mm, running ALONG the split for 0.96 mm, not across it, 29 ohm over solid plane, 31 ohm along it: 3.4 % of each edge reflected; other side: GND on G2, 0.103 mm from the track against 0.1 mm to P3: 51% of the return on P3
      • 'HA1.HA00_CC_P' on L2 at (-105.25, 54.74), 3.19 mm, running ALONG the split for 2.63 mm, not across it, 61 ohm over solid plane, 64 ohm along it: 2.6 % of each edge reflected; other side: GND on G2, 0.103 mm from the track against 0.1 mm to P3: 51% of the return on P3

      Every site is a row of this finding's evidence, keyed by net, layers and position. GET /api/odb/boards/58/reference-planes?group=along%40P3%3AP1V8%7CPVADJ1 serves each with its own layer_image region; all of them lie in x -116.44..-103.88, y 52.1..101.77 mm (board frame, Y up).

      Read from the copper as drawn on the reference layer, along each track's centreline at exact breakpoints. The reference is the copper beside the track, never the layer's type or name. Whether it matters depends on each signal's edge rate, which nothing on file states for these nets: a slow static line crosses a split harmlessly, a fast edge does not.

    • along@P3:P1V8|PVADJ23 nets run ALONG the split between the P1V8 and PVADJ2 planes on P3 at 3 places, the longest 8.8 mm; the impedance rises from 41 ohm to 47 ohm there at worst, 7.1 % of each edge reflected (fastest: TRACEDATA0 at 100 ps 10-90 %, ASSUMED)

      On P3 the P1V8 and PVADJ2 planes part, and 3 nets run down the slot between them at 3 place(s), rather than across it. This is not a crossing: the return still follows the track, in the plane on its other side and in the slot's edges, so it does not detour. What changes is the track's own impedance. With the plane gone from under it on this side for that length, its field reaches further, to the slot's edges and to the next plane beyond, so its capacitance falls and its impedance rises; every edge travelling down it is partly reflected where it enters the stretch and again where it leaves, by (Z2 - Z1)/(Z2 + Z1). The longer the stretch against the edge's rise, the more of the edge sees the full step. A crossing is generally the worse fault; a mismatch is not free. Over solid plane each track is 41 ohm; along the slot 45 ohm to 47 ohm, reflecting up to 7.1 % of each edge (a method-of-moments field solve of the cross-section, one dielectric; were the slot's edges to carry none of the return, up to 49 ohm). At 1 of the 3 place(s) the rise is within 10 %, the tolerance controlled impedance is commonly built to (the stackup states none); the rest exceed it or cannot be said, so the finding is a warning. One decision covers the split line's slot: whether these tracks may run down it, or should be moved over solid plane.

      • 'TRACEDATA0' on L1 at (-84.9, -3.85), 1.97 mm, running ALONG the split for 1.27 mm, not across it, 41 ohm over solid plane, 47 ohm along it: 7.1 % of each edge reflected; other side: GND on G1, 0.1 mm from the track against 0.103 mm to P3: 49% of the return on P3
      • 'I_FMC2_P3V3N' on L2 at (-58.85, -4.71), 1.23 mm, running ALONG the split for 0.58 mm, not across it, 41 ohm over solid plane, 45 ohm along it: 4.9 % of each edge reflected; other side: GND on G2, 0.103 mm from the track against 0.1 mm to P3: 51% of the return on P3
      • 'I_FMC1_P12V0-' on L1 at (-69.85, -4.7), 9.45 mm, running ALONG the split for 8.8 mm, not across it, 41 ohm over solid plane, 45 ohm along it: 4.3 % of each edge reflected; other side: GND on G1, 0.1 mm from the track against 0.103 mm to P3: 49% of the return on P3

      Every site is a row of this finding's evidence, keyed by net, layers and position. GET /api/odb/boards/58/reference-planes?group=along%40P3%3AP1V8%7CPVADJ2 serves each with its own layer_image region; all of them lie in x -86.84..-57.49, y -6.1..-2.85 mm (board frame, Y up).

      Read from the copper as drawn on the reference layer, along each track's centreline at exact breakpoints. The reference is the copper beside the track, never the layer's type or name. Whether it matters depends on each signal's edge rate, which nothing on file states for these nets: a slow static line crosses a split harmlessly, a fast edge does not.

    • along@P3:PVADJ1|PVADJ22 nets run ALONG the split between the PVADJ1 and PVADJ2 planes on P3 at 2 places, the longest 0.42 mm; the impedance rises from 61 ohm to 68 ohm there at worst, 5.7 % of each edge reflected (fastest: LA2.LA28_N at 100 ps 10-90 %, ASSUMED)

      On P3 the PVADJ1 and PVADJ2 planes part, and 2 nets run down the slot between them at 2 place(s), rather than across it. This is not a crossing: the return still follows the track, in the plane on its other side and in the slot's edges, so it does not detour. What changes is the track's own impedance. With the plane gone from under it on this side for that length, its field reaches further, to the slot's edges and to the next plane beyond, so its capacitance falls and its impedance rises; every edge travelling down it is partly reflected where it enters the stretch and again where it leaves, by (Z2 - Z1)/(Z2 + Z1). The longer the stretch against the edge's rise, the more of the edge sees the full step. A crossing is generally the worse fault; a mismatch is not free. Over solid plane each track is 61 ohm; along the slot 65 ohm to 68 ohm, reflecting up to 5.7 % of each edge (a method-of-moments field solve of the cross-section, one dielectric; were the slot's edges to carry none of the return, up to 69 ohm). At 1 of the 2 place(s) the rise is within 10 %, the tolerance controlled impedance is commonly built to (the stackup states none); the rest exceed it or cannot be said, so the finding is a warning. One decision covers the split line's slot: whether these tracks may run down it, or should be moved over solid plane.

      • 'LA2.LA28_P' on L2 at (-102.93, 22.37), 0.99 mm, running ALONG the split for 0.42 mm, not across it, 61 ohm over solid plane, 68 ohm along it: 5.7 % of each edge reflected; other side: GND on G2, 0.103 mm from the track against 0.1 mm to P3: 51% of the return on P3
      • 'LA2.LA28_N' on L2 at (-103.29, 22.69), 0.99 mm, running ALONG the split for 0.42 mm, not across it, 61 ohm over solid plane, 65 ohm along it: 3.2 % of each edge reflected; other side: GND on G2, 0.103 mm from the track against 0.1 mm to P3: 51% of the return on P3

      Every site is a row of this finding's evidence, keyed by net, layers and position. GET /api/odb/boards/58/reference-planes?group=along%40P3%3APVADJ1%7CPVADJ2 serves each with its own layer_image region; all of them lie in x -104.29..-101.78, y 20.87..23.94 mm (board frame, Y up).

      Read from the copper as drawn on the reference layer, along each track's centreline at exact breakpoints. The reference is the copper beside the track, never the layer's type or name. Whether it matters depends on each signal's edge rate, which nothing on file states for these nets: a slow static line crosses a split harmlessly, a fast edge does not.

    • along@P4:FMC1_P3V3|P3V3MP3 nets run ALONG the split between the FMC1_P3V3 and P3V3MP planes on P4 at 3 places, the longest 10.26 mm; the impedance rises from 61 ohm to 68 ohm there at worst, 5.6 % of each edge reflected (fastest: P30_IO7_CD_B1_P at 100 ps 10-90 %, ASSUMED)

      On P4 the FMC1_P3V3 and P3V3MP planes part, and 3 nets run down the slot between them at 3 place(s), rather than across it. This is not a crossing: the return still follows the track, in the plane on its other side and in the slot's edges, so it does not detour. What changes is the track's own impedance. With the plane gone from under it on this side for that length, its field reaches further, to the slot's edges and to the next plane beyond, so its capacitance falls and its impedance rises; every edge travelling down it is partly reflected where it enters the stretch and again where it leaves, by (Z2 - Z1)/(Z2 + Z1). The longer the stretch against the edge's rise, the more of the edge sees the full step. A crossing is generally the worse fault; a mismatch is not free. Over solid plane each track is 61 ohm; along the slot 67 ohm to 68 ohm, reflecting up to 5.6 % of each edge (a method-of-moments field solve of the cross-section, one dielectric; were the slot's edges to carry none of the return, up to 68 ohm). No place stays within 10 %, the tolerance controlled impedance is commonly built to (the stackup states none), so the finding is a warning. One decision covers the split line's slot: whether these tracks may run down it, or should be moved over solid plane.

      • 'TCLKC_P' on L3 at (-120.92, -22.27), 6.18 mm, running ALONG the split for 5.62 mm, not across it, 61 ohm over solid plane, 68 ohm along it: 5.6 % of each edge reflected; other side: GND on G2, 0.1 mm from the track against 0.103 mm to P4: 49% of the return on P4
      • 'P30_IO7_CD_B1_P' on L3 at (-106.95, 0.34), 10.64 mm, running ALONG the split for 10.26 mm, not across it, 61 ohm over solid plane, 67 ohm along it: 5.2 % of each edge reflected; other side: GND on G2, 0.1 mm from the track against 0.103 mm to P4: 49% of the return on P4
      • 'TCLKD_N' on L3 at (-100.83, -17.21), 7.78 mm, running ALONG the split for 6.94 mm, not across it, 61 ohm over solid plane, 67 ohm along it: 4.9 % of each edge reflected; other side: GND on G2, 0.1 mm from the track against 0.103 mm to P4: 49% of the return on P4

      Every site is a row of this finding's evidence, keyed by net, layers and position. GET /api/odb/boards/58/reference-planes?group=along%40P4%3AFMC1_P3V3%7CP3V3MP serves each with its own layer_image region; all of them lie in x -125.89..-97.17, y -23.27..1.47 mm (board frame, Y up).

      Read from the copper as drawn on the reference layer, along each track's centreline at exact breakpoints. The reference is the copper beside the track, never the layer's type or name. Whether it matters depends on each signal's edge rate, which nothing on file states for these nets: a slow static line crosses a split harmlessly, a fast edge does not.

    • along@P4:FMC2_P3V3|P12V'P31_IO2_EF_B2_N' runs ALONG the split between the FMC2_P3V3 and P12V planes on P4 at 1 place, the longest 0.8 mm; the impedance rises from 61 ohm to 68 ohm there at worst, 5.3 % of each edge reflected (fastest: P31_IO2_EF_B2_N at 100 ps 10-90 %, ASSUMED)

      On P4 the FMC2_P3V3 and P12V planes part, and 'P31_IO2_EF_B2_N' runs down the slot between them at 1 place(s), rather than across it. This is not a crossing: the return still follows the track, in the plane on its other side and in the slot's edges, so it does not detour. What changes is the track's own impedance. With the plane gone from under it on this side for that length, its field reaches further, to the slot's edges and to the next plane beyond, so its capacitance falls and its impedance rises; every edge travelling down it is partly reflected where it enters the stretch and again where it leaves, by (Z2 - Z1)/(Z2 + Z1). The longer the stretch against the edge's rise, the more of the edge sees the full step. A crossing is generally the worse fault; a mismatch is not free. Over solid plane each track is 61 ohm; along the slot 68 ohm, reflecting up to 5.3 % of each edge (a method-of-moments field solve of the cross-section, one dielectric; were the slot's edges to carry none of the return, up to 68 ohm). No place stays within 10 %, the tolerance controlled impedance is commonly built to (the stackup states none), so the finding is a warning. One decision covers the split line's slot: whether these tracks may run down it, or should be moved over solid plane.

      • 'P31_IO2_EF_B2_N' on L3 at (-101.24, 29.21), 1.14 mm, running ALONG the split for 0.8 mm, not across it, 61 ohm over solid plane, 68 ohm along it: 5.3 % of each edge reflected; other side: GND on G2, 0.1 mm from the track against 0.103 mm to P4: 49% of the return on P4

      Every site is a row of this finding's evidence, keyed by net, layers and position. GET /api/odb/boards/58/reference-planes?group=along%40P4%3AFMC2_P3V3%7CP12V serves each with its own layer_image region; all of them lie in x -102.79..-99.74, y 28.21..30.36 mm (board frame, Y up).

      Read from the copper as drawn on the reference layer, along each track's centreline at exact breakpoints. The reference is the copper beside the track, never the layer's type or name. Whether it matters depends on each signal's edge rate, which nothing on file states for these nets: a slow static line crosses a split harmlessly, a fast edge does not.

    • along@P4:MGTAVTT|P1V84 nets run ALONG the split between the MGTAVTT and P1V8 planes on P4 at 4 places, the longest 5.48 mm; the impedance rises from 61 ohm to 67 ohm there at worst, 5.2 % of each edge reflected (fastest: P30_IO10_CD_B2_N at 100 ps 10-90 %, ASSUMED)

      On P4 the MGTAVTT and P1V8 planes part, and 4 nets run down the slot between them at 4 place(s), rather than across it. This is not a crossing: the return still follows the track, in the plane on its other side and in the slot's edges, so it does not detour. What changes is the track's own impedance. With the plane gone from under it on this side for that length, its field reaches further, to the slot's edges and to the next plane beyond, so its capacitance falls and its impedance rises; every edge travelling down it is partly reflected where it enters the stretch and again where it leaves, by (Z2 - Z1)/(Z2 + Z1). The longer the stretch against the edge's rise, the more of the edge sees the full step. A crossing is generally the worse fault; a mismatch is not free. Over solid plane each track is 61 ohm; along the slot 66 ohm to 67 ohm, reflecting up to 5.2 % of each edge (a method-of-moments field solve of the cross-section, one dielectric; were the slot's edges to carry none of the return, up to 68 ohm). At 2 of the 4 place(s) the rise is within 10 %, the tolerance controlled impedance is commonly built to (the stackup states none); the rest exceed it or cannot be said, so the finding is a warning. One decision covers the split line's slot: whether these tracks may run down it, or should be moved over solid plane.

      • 'P30_IO10_CD_B2_N' on L3 at (-38.95, 69.44), 6.12 mm, running ALONG the split for 5.48 mm, not across it, 61 ohm over solid plane, 67 ohm along it: 5.2 % of each edge reflected; other side: GND on G2, 0.1 mm from the track against 0.103 mm to P4: 49% of the return on P4
      • 'P30_OUT0_P' on L3 at (-39.47, 76.19), 2 mm, running ALONG the split for 1.49 mm, not across it, 61 ohm over solid plane, 67 ohm along it: 4.9 % of each edge reflected; other side: GND on G2, 0.1 mm from the track against 0.103 mm to P4: 49% of the return on P4
      • 'P30_OUT0_N' on L3 at (-39.29, 76.31), 2.17 mm, running ALONG the split for 1.55 mm, not across it, 61 ohm over solid plane, 66 ohm along it: 4.1 % of each edge reflected; other side: GND on G2, 0.1 mm from the track against 0.103 mm to P4: 49% of the return on P4
      • 'P31_IO1_EF_B2_N' on L3 at (-49.9, 81.78), 1.22 mm, running ALONG the split for 1.01 mm, not across it, 61 ohm over solid plane, 66 ohm along it: 4 % of each edge reflected; other side: GND on G2, 0.1 mm from the track against 0.103 mm to P4: 49% of the return on P4

      Every site is a row of this finding's evidence, keyed by net, layers and position. GET /api/odb/boards/58/reference-planes?group=along%40P4%3AMGTAVTT%7CP1V8 serves each with its own layer_image region; all of them lie in x -51.3..-35.08, y 68.44..82.85 mm (board frame, Y up).

      Read from the copper as drawn on the reference layer, along each track's centreline at exact breakpoints. The reference is the copper beside the track, never the layer's type or name. Whether it matters depends on each signal's edge rate, which nothing on file states for these nets: a slow static line crosses a split harmlessly, a fast edge does not.

    • along@P4:P12V|P1V84 nets run ALONG the split between the P12V and P1V8 planes on P4 at 4 places, the longest 23.61 mm; the impedance rises from 52 ohm to 59 ohm there at worst, 6.3 % of each edge reflected (fastest: P30_IO10_AB_B2_CC_P at 100 ps 10-90 %, ASSUMED)

      On P4 the P12V and P1V8 planes part, and 4 nets run down the slot between them at 4 place(s), rather than across it. This is not a crossing: the return still follows the track, in the plane on its other side and in the slot's edges, so it does not detour. What changes is the track's own impedance. With the plane gone from under it on this side for that length, its field reaches further, to the slot's edges and to the next plane beyond, so its capacitance falls and its impedance rises; every edge travelling down it is partly reflected where it enters the stretch and again where it leaves, by (Z2 - Z1)/(Z2 + Z1). The longer the stretch against the edge's rise, the more of the edge sees the full step. A crossing is generally the worse fault; a mismatch is not free. Over solid plane each track is 52 ohm to 61 ohm; along the slot 59 ohm to 67 ohm, reflecting up to 6.3 % of each edge (a method-of-moments field solve of the cross-section, one dielectric; were the slot's edges to carry none of the return, up to 68 ohm). At 2 of the 4 place(s) the rise is within 10 %, the tolerance controlled impedance is commonly built to (the stackup states none); the rest exceed it or cannot be said, so the finding is a warning. One decision covers the split line's slot: whether these tracks may run down it, or should be moved over solid plane.

      • 'RTM_TCK' on L3 at (-112.62, 92.15), 10.89 mm, running ALONG the split for 10.01 mm, not across it, 52 ohm over solid plane, 59 ohm along it: 6.3 % of each edge reflected; other side: GND on G2, 0.1 mm from the track against 0.103 mm to P4: 49% of the return on P4
      • 'P31_IO2_EF_B2_P' on L3 at (-102.03, 63.51), 24.03 mm, running ALONG the split for 23.51 mm, not across it, 61 ohm over solid plane, 67 ohm along it: 5.2 % of each edge reflected; other side: GND on G2, 0.1 mm from the track against 0.103 mm to P4: 49% of the return on P4
      • 'P31_IO2_EF_B2_N' on L3 at (-102.24, 63.58), 24.21 mm, running ALONG the split for 23.61 mm, not across it, 61 ohm over solid plane, 65 ohm along it: 3.2 % of each edge reflected; other side: GND on G2, 0.1 mm from the track against 0.103 mm to P4: 49% of the return on P4
      • 'P30_IO10_AB_B2_CC_P' on L3 at (-100.38, 50.43), 2.55 mm, running ALONG the split for 1.98 mm, not across it, 61 ohm over solid plane, 63 ohm along it: 1.8 % of each edge reflected; other side: GND on G2, 0.1 mm from the track against 0.103 mm to P4: 49% of the return on P4

      Every site is a row of this finding's evidence, keyed by net, layers and position. GET /api/odb/boards/58/reference-planes?group=along%40P4%3AP12V%7CP1V8 serves each with its own layer_image region; all of them lie in x -117.84..-98.96, y 48.45..93.15 mm (board frame, Y up).

      Read from the copper as drawn on the reference layer, along each track's centreline at exact breakpoints. The reference is the copper beside the track, never the layer's type or name. Whether it matters depends on each signal's edge rate, which nothing on file states for these nets: a slow static line crosses a split harmlessly, a fast edge does not.

    • along@P4:P1V8|P3V3MP3 nets run ALONG the split between the P1V8 and P3V3MP planes on P4 at 3 places, the longest 1.17 mm; the impedance rises from 61 ohm to 67 ohm there at worst, 5.2 % of each edge reflected (fastest: P30_OUT2_N at 100 ps 10-90 %, ASSUMED)

      On P4 the P1V8 and P3V3MP planes part, and 3 nets run down the slot between them at 3 place(s), rather than across it. This is not a crossing: the return still follows the track, in the plane on its other side and in the slot's edges, so it does not detour. What changes is the track's own impedance. With the plane gone from under it on this side for that length, its field reaches further, to the slot's edges and to the next plane beyond, so its capacitance falls and its impedance rises; every edge travelling down it is partly reflected where it enters the stretch and again where it leaves, by (Z2 - Z1)/(Z2 + Z1). The longer the stretch against the edge's rise, the more of the edge sees the full step. A crossing is generally the worse fault; a mismatch is not free. Over solid plane each track is 61 ohm; along the slot 65 ohm to 67 ohm, reflecting up to 5.2 % of each edge (a method-of-moments field solve of the cross-section, one dielectric; were the slot's edges to carry none of the return, up to 68 ohm). At 2 of the 3 place(s) the rise is within 10 %, the tolerance controlled impedance is commonly built to (the stackup states none); the rest exceed it or cannot be said, so the finding is a warning. One decision covers the split line's slot: whether these tracks may run down it, or should be moved over solid plane.

      • 'P31_IO2_AB_B2_P' on L3 at (-28.54, 85.66), 1.64 mm, running ALONG the split for 1.13 mm, not across it, 61 ohm over solid plane, 67 ohm along it: 5.2 % of each edge reflected; other side: GND on G2, 0.1 mm from the track against 0.103 mm to P4: 49% of the return on P4
      • 'P30_OUT2_N' on L3 at (-31.22, 73.51), 1.54 mm, running ALONG the split for 0.97 mm, not across it, 61 ohm over solid plane, 66 ohm along it: 4.5 % of each edge reflected; other side: GND on G2, 0.1 mm from the track against 0.103 mm to P4: 49% of the return on P4
      • 'P31_IO2_AB_B2_N' on L3 at (-28.33, 85.64), 1.78 mm, running ALONG the split for 1.17 mm, not across it, 61 ohm over solid plane, 65 ohm along it: 3.4 % of each edge reflected; other side: GND on G2, 0.1 mm from the track against 0.103 mm to P4: 49% of the return on P4

      Every site is a row of this finding's evidence, keyed by net, layers and position. GET /api/odb/boards/58/reference-planes?group=along%40P4%3AP1V8%7CP3V3MP serves each with its own layer_image region; all of them lie in x -32.69..-27.33, y 72.16..87.65 mm (board frame, Y up).

      Read from the copper as drawn on the reference layer, along each track's centreline at exact breakpoints. The reference is the copper beside the track, never the layer's type or name. Whether it matters depends on each signal's edge rate, which nothing on file states for these nets: a slow static line crosses a split harmlessly, a fast edge does not.

    • along@P6:P0V95|P1V84 nets run ALONG the split between the P0V95 and P1V8 planes on P6 at 4 places, the longest 6.49 mm; the impedance rises from 61 ohm to 68 ohm there at worst, 5.7 % of each edge reflected (fastest: HA1.HA18_P at 100 ps 10-90 %, ASSUMED)

      On P6 the P0V95 and P1V8 planes part, and 4 nets run down the slot between them at 4 place(s), rather than across it. This is not a crossing: the return still follows the track, in the plane on its other side and in the slot's edges, so it does not detour. What changes is the track's own impedance. With the plane gone from under it on this side for that length, its field reaches further, to the slot's edges and to the next plane beyond, so its capacitance falls and its impedance rises; every edge travelling down it is partly reflected where it enters the stretch and again where it leaves, by (Z2 - Z1)/(Z2 + Z1). The longer the stretch against the edge's rise, the more of the edge sees the full step. A crossing is generally the worse fault; a mismatch is not free. Over solid plane each track is 61 ohm; along the slot 66 ohm to 68 ohm, reflecting up to 5.7 % of each edge (a method-of-moments field solve of the cross-section, one dielectric; were the slot's edges to carry none of the return, up to 69 ohm). At 1 of the 4 place(s) the rise is within 10 %, the tolerance controlled impedance is commonly built to (the stackup states none); the rest exceed it or cannot be said, so the finding is a warning. One decision covers the split line's slot: whether these tracks may run down it, or should be moved over solid plane.

      • 'FPGA_CLK.SEC_CLK_N' on L5 at (-76.08, 22.7), 5.13 mm, running ALONG the split for 4.47 mm, not across it, 61 ohm over solid plane, 68 ohm along it: 5.7 % of each edge reflected; other side: GND on GND, 0.103 mm from the track against 0.1 mm to P6: 51% of the return on P6
      • 'LA2.LA04_N' on L6 at (-76.04, 22.66), 5.13 mm, running ALONG the split for 4.49 mm, not across it, 61 ohm over solid plane, 68 ohm along it: 5.4 % of each edge reflected; other side: GND on G3, 0.1 mm from the track against 0.104 mm to P6: 49% of the return on P6
      • 'HA1.HA18_P' on L5 at (-81.11, 79.62), 7.17 mm, running ALONG the split for 6.49 mm, not across it, 61 ohm over solid plane, 67 ohm along it: 5.1 % of each edge reflected; other side: GND on GND, 0.103 mm from the track against 0.1 mm to P6: 51% of the return on P6
      • 'HA2.HA18_P' on L5 at (-74.55, 38.18), 4.67 mm, running ALONG the split for 4.1 mm, not across it, 61 ohm over solid plane, 66 ohm along it: 4.4 % of each edge reflected; other side: GND on GND, 0.103 mm from the track against 0.1 mm to P6: 51% of the return on P6

      Every site is a row of this finding's evidence, keyed by net, layers and position. GET /api/odb/boards/58/reference-planes?group=along%40P6%3AP0V95%7CP1V8 serves each with its own layer_image region; all of them lie in x -82.4..-73.23, y 19.85..84.1 mm (board frame, Y up).

      Read from the copper as drawn on the reference layer, along each track's centreline at exact breakpoints. The reference is the copper beside the track, never the layer's type or name. Whether it matters depends on each signal's edge rate, which nothing on file states for these nets: a slow static line crosses a split harmlessly, a fast edge does not.

    • along@P6:P0V95|PVADJ12 nets run ALONG the split between the P0V95 and PVADJ1 planes on P6 at 2 places, the longest 4.58 mm; the impedance rises from 56 ohm to 63 ohm there at worst, 6.3 % of each edge reflected (fastest: MLVDS_DE_7_C at 100 ps 10-90 %, ASSUMED)

      On P6 the P0V95 and PVADJ1 planes part, and 2 nets run down the slot between them at 2 place(s), rather than across it. This is not a crossing: the return still follows the track, in the plane on its other side and in the slot's edges, so it does not detour. What changes is the track's own impedance. With the plane gone from under it on this side for that length, its field reaches further, to the slot's edges and to the next plane beyond, so its capacitance falls and its impedance rises; every edge travelling down it is partly reflected where it enters the stretch and again where it leaves, by (Z2 - Z1)/(Z2 + Z1). The longer the stretch against the edge's rise, the more of the edge sees the full step. A crossing is generally the worse fault; a mismatch is not free. Over solid plane each track is 56 ohm; along the slot 62 ohm to 63 ohm, reflecting up to 6.3 % of each edge (a method-of-moments field solve of the cross-section, one dielectric; were the slot's edges to carry none of the return, up to 64 ohm). No place stays within 10 %, the tolerance controlled impedance is commonly built to (the stackup states none), so the finding is a warning. One decision covers the split line's slot: whether these tracks may run down it, or should be moved over solid plane.

      • 'MLVDS_DE_7_C' on L5 at (-16.08, -2.88), 1.89 mm, running ALONG the split for 1.1 mm, not across it, 56 ohm over solid plane, 63 ohm along it: 6.3 % of each edge reflected; other side: GND on GND, 0.103 mm from the track against 0.1 mm to P6: 51% of the return on P6
      • 'MLVDS_F.MLVDS_I_1' on L6 at (-16.12, -4.59), 5.45 mm, running ALONG the split for 4.58 mm, not across it, 56 ohm over solid plane, 62 ohm along it: 5.1 % of each edge reflected; other side: GND on G3, 0.1 mm from the track against 0.104 mm to P6: 49% of the return on P6

      Every site is a row of this finding's evidence, keyed by net, layers and position. GET /api/odb/boards/58/reference-planes?group=along%40P6%3AP0V95%7CPVADJ1 serves each with its own layer_image region; all of them lie in x -17.12..-14.77, y -8.1..-0.79 mm (board frame, Y up).

      Read from the copper as drawn on the reference layer, along each track's centreline at exact breakpoints. The reference is the copper beside the track, never the layer's type or name. Whether it matters depends on each signal's edge rate, which nothing on file states for these nets: a slow static line crosses a split harmlessly, a fast edge does not.

    • along@P6:P1V5|P3V3MP6 nets run ALONG the split between the P1V5 and P3V3MP planes on P6 at 6 places, the longest 13.44 mm; the impedance rises from 52 ohm to 59 ohm there at worst, 6.3 % of each edge reflected (fastest: EEPROM_WP at 100 ps 10-90 %, ASSUMED)

      On P6 the P1V5 and P3V3MP planes part, and 6 nets run down the slot between them at 6 place(s), rather than across it. This is not a crossing: the return still follows the track, in the plane on its other side and in the slot's edges, so it does not detour. What changes is the track's own impedance. With the plane gone from under it on this side for that length, its field reaches further, to the slot's edges and to the next plane beyond, so its capacitance falls and its impedance rises; every edge travelling down it is partly reflected where it enters the stretch and again where it leaves, by (Z2 - Z1)/(Z2 + Z1). The longer the stretch against the edge's rise, the more of the edge sees the full step. A crossing is generally the worse fault; a mismatch is not free. Over solid plane each track is 52 ohm to 56 ohm; along the slot 55 ohm to 61 ohm, reflecting up to 6.3 % of each edge (a method-of-moments field solve of the cross-section, one dielectric; were the slot's edges to carry none of the return, up to 64 ohm). At 4 of the 6 place(s) the rise is within 10 %, the tolerance controlled impedance is commonly built to (the stackup states none); the rest exceed it or cannot be said, so the finding is a warning. One decision covers the split line's slot: whether these tracks may run down it, or should be moved over solid plane.

      • 'FJTAG.TDI' on L5 at (-129.05, 95.71), 8.55 mm, running ALONG the split for 7.87 mm, not across it, 52 ohm over solid plane, 59 ohm along it: 6.3 % of each edge reflected; other side: GND on GND, 0.103 mm from the track against 0.1 mm to P6: 51% of the return on P6
      • 'FJTAG.TDO' on L5 at (-111.9, 95.85), 13.83 mm, running ALONG the split for 13.09 mm, not across it, 53 ohm over solid plane, 60 ohm along it: 6.3 % of each edge reflected; other side: GND on GND, 0.103 mm from the track against 0.1 mm to P6: 51% of the return on P6
      • 'PLL_CTRL.VADJ2_SI57X_PRI_SCL' on L5 at (-164.77, -9.01), 0.88 mm, running ALONG the split for 0.5 mm, not across it, 52 ohm over solid plane, 57 ohm along it: 4.7 % of each edge reflected; other side: GND on GND, 0.103 mm from the track against 0.1 mm to P6: 51% of the return on P6
      • 'QSPI0_IO0' on L5 at (-113.84, 95.66), 7.21 mm, running ALONG the split for 7 mm, not across it, 56 ohm over solid plane, 61 ohm along it: 4.6 % of each edge reflected; other side: GND on GND, 0.103 mm from the track against 0.1 mm to P6: 51% of the return on P6
      • 'EEPROM_WP' on L6 at (-169.11, 90.32), 14.02 mm, running ALONG the split for 13.44 mm, not across it, 52 ohm over solid plane, 56 ohm along it: 3.1 % of each edge reflected; other side: GND on G3, 0.1 mm from the track against 0.104 mm to P6: 49% of the return on P6
      • 'AJTAG.TDO' on L6 at (-154.89, -24.87), 2.29 mm, running ALONG the split for 1.93 mm, not across it, 52 ohm over solid plane, 55 ohm along it: 2.9 % of each edge reflected; other side: GND on G3, 0.1 mm from the track against 0.104 mm to P6: 49% of the return on P6

      Every site is a row of this finding's evidence, keyed by net, layers and position. GET /api/odb/boards/58/reference-planes?group=along%40P6%3AP1V5%7CP3V3MP serves each with its own layer_image region; all of them lie in x -170.37..-104.09, y -25.9..98.26 mm (board frame, Y up).

      Read from the copper as drawn on the reference layer, along each track's centreline at exact breakpoints. The reference is the copper beside the track, never the layer's type or name. Whether it matters depends on each signal's edge rate, which nothing on file states for these nets: a slow static line crosses a split harmlessly, a fast edge does not.

    • along@P6:P3V3MP|P3V3_CLKSW6 nets run ALONG the split between the P3V3MP and P3V3_CLKSW planes on P6 at 7 places, the longest 7.81 mm; the impedance rises from 20 ohm to 24 ohm there at worst, 8.8 % of each edge reflected (fastest: HA2.HA02_P at 100 ps 10-90 %, ASSUMED)

      On P6 the P3V3MP and P3V3_CLKSW planes part, and 6 nets run down the slot between them at 7 place(s), rather than across it. This is not a crossing: the return still follows the track, in the plane on its other side and in the slot's edges, so it does not detour. What changes is the track's own impedance. With the plane gone from under it on this side for that length, its field reaches further, to the slot's edges and to the next plane beyond, so its capacitance falls and its impedance rises; every edge travelling down it is partly reflected where it enters the stretch and again where it leaves, by (Z2 - Z1)/(Z2 + Z1). The longer the stretch against the edge's rise, the more of the edge sees the full step. A crossing is generally the worse fault; a mismatch is not free. Over solid plane each track is 20 ohm to 61 ohm; along the slot 23 ohm to 68 ohm, reflecting up to 8.8 % of each edge (a method-of-moments field solve of the cross-section, one dielectric; were the slot's edges to carry none of the return, up to 69 ohm). At 3 of the 7 place(s) the rise is within 10 %, the tolerance controlled impedance is commonly built to (the stackup states none); the rest exceed it or cannot be said, so the finding is a warning. One decision covers the split line's slot: whether these tracks may run down it, or should be moved over solid plane.

      • 'VREF2.VREF_B_M2C' on L6 at (-118.12, -9.34), 4.48 mm, running ALONG the split for 3.81 mm, not across it, 20 ohm over solid plane, 24 ohm along it: 8.8 % of each edge reflected; other side: GND on G3, 0.1 mm from the track against 0.104 mm to P6: 49% of the return on P6
      • 'VREF2.VREF_B_M2C' on L6 at (-118, -17.75), 3.16 mm, running ALONG the split for 2.3 mm, not across it, 20 ohm over solid plane, 23 ohm along it: 6.6 % of each edge reflected; other side: GND on G3, 0.1 mm from the track against 0.104 mm to P6: 49% of the return on P6
      • 'HA2.HA18_N' on L5 at (-118.21, 4.39), 2.54 mm, running ALONG the split for 1.65 mm, not across it, 61 ohm over solid plane, 68 ohm along it: 5.7 % of each edge reflected; other side: GND on GND, 0.103 mm from the track against 0.1 mm to P6: 51% of the return on P6
      • 'HA2.HA18_P' on L5 at (-118, 4.39), 1.77 mm, running ALONG the split for 1.48 mm, not across it, 61 ohm over solid plane, 67 ohm along it: 5.1 % of each edge reflected; other side: GND on GND, 0.103 mm from the track against 0.1 mm to P6: 51% of the return on P6
      • 'HA2.HA02_P' on L6 at (-117.99, -2.77), 2.83 mm, running ALONG the split for 2.58 mm, not across it, 61 ohm over solid plane, 66 ohm along it: 4.3 % of each edge reflected; other side: GND on G3, 0.1 mm from the track against 0.104 mm to P6: 49% of the return on P6
      • 'PM.EN_RTM_PWR' on L6 at (-126.92, -7.17), 7.97 mm, running ALONG the split for 7.25 mm, not across it, 52 ohm over solid plane, 55 ohm along it: 3 % of each edge reflected; other side: GND on G3, 0.1 mm from the track against 0.104 mm to P6: 49% of the return on P6
      • 'JTAG2.TRSTn' on L5 at (-135.44, 19.1), 8.99 mm, running ALONG the split for 7.81 mm, not across it, 52 ohm over solid plane, 54 ohm along it: 2 % of each edge reflected; other side: GND on GND, 0.103 mm from the track against 0.1 mm to P6: 51% of the return on P6

      Every site is a row of this finding's evidence, keyed by net, layers and position. GET /api/odb/boards/58/reference-planes?group=along%40P6%3AP3V3MP%7CP3V3_CLKSW serves each with its own layer_image region; all of them lie in x -140.8..-116.89, y -20.21..20.56 mm (board frame, Y up).

      Read from the copper as drawn on the reference layer, along each track's centreline at exact breakpoints. The reference is the copper beside the track, never the layer's type or name. Whether it matters depends on each signal's edge rate, which nothing on file states for these nets: a slow static line crosses a split harmlessly, a fast edge does not.

    • split@P3:P0V95|P1V516 nets cross the split between the P0V95 and P1V5 planes on P3 at 19 places; 15 with a plane on the track's other side (fastest: FPGA_M2 at 100 ps 10-90 %, ASSUMED)

      On P3 the P0V95 and P1V5 planes part, and 16 nets cross from one to the other at 19 place(s). The return current cannot follow a track across; it detours to wherever the two planes are joined (a stitching capacitor, or the source), and the loop it opens radiates and couples. One decision covers the boundary: whether it is stitched near these crossings, or the nets crossing it are slow enough not to care. Tracks that run ALONG this split, down the slot rather than across it, are a finding of their own: theirs is an impedance step, not a detour. At 15 of the 19 crossing(s) the track's other neighbour carries a plane, which the return can use there. 15 of the 19 crossing(s) have a plane on the track's other side at most twice as far away as the broken one (to first order each plane beside a stripline carries a share of the return inversely proportional to its distance from the track). The rest leave most or all of the return to the break, so the finding is a warning.

      • 'LA1.LA13_N' on L1 at (-74.34, 66.58), 0.68 mm; other side: no plane on G1
      • 'LA2.LA05_N' on L2 at (-75.44, 48.74), 0.44 mm; other side: no plane on G2
      • 'LA2.LA05_P' on L2 at (-75.44, 49.42), 0.44 mm; other side: no plane on G2
      • 'LA2.LA06_N' on L2 at (-75.33, 51.47), 0.45 mm; other side: no plane on G2
      • 'FPGA_M2' on L2 at (-62.17, 47.88), 0.4 mm; other side: GND on G2, 0.103 mm from the track against 0.1 mm to P3: 51% of the return on P3
      • 'HA2.HA07_N' on L1 at (-65.45, 56.03), 0.4 mm; other side: GND on G1, 0.1 mm from the track against 0.103 mm to P3: 49% of the return on P3
      • 'HA2.HA07_N' on L1 at (-65.03, 47.88), 0.4 mm; other side: GND on G1, 0.1 mm from the track against 0.103 mm to P3: 49% of the return on P3
      • 'HA2.HA07_P' on L1 at (-65.45, 56.24), 0.4 mm; other side: GND on G1, 0.1 mm from the track against 0.103 mm to P3: 49% of the return on P3
      • 'HA2.HA07_P' on L1 at (-64.81, 47.88), 0.4 mm; other side: GND on G1, 0.1 mm from the track against 0.103 mm to P3: 49% of the return on P3
      • 'LA1.LA11_N' on L1 at (-75.33, 63.98), 0.4 mm; other side: GND on G1, 0.1 mm from the track against 0.103 mm to P3: 49% of the return on P3
      • 'LA1.LA11_P' on L1 at (-75.33, 64.19), 0.4 mm; other side: GND on G1, 0.1 mm from the track against 0.103 mm to P3: 49% of the return on P3
      • 'LA1.LA13_N' on L1 at (-75.28, 65.2), 0.48 mm; other side: GND on G1, 0.1 mm from the track against 0.103 mm to P3: 49% of the return on P3
      • and 7 more

      Every site is a row of this finding's evidence, keyed by net, layers and position. GET /api/odb/boards/58/reference-planes?group=split%40P3%3AP0V95%7CP1V5 serves each with its own layer_image region; all of them lie in x -76.53..-61.17, y 46.68..67.58 mm (board frame, Y up).

      Read from the copper as drawn on the reference layer, along each track's centreline at exact breakpoints. The reference is the copper beside the track, never the layer's type or name. Whether it matters depends on each signal's edge rate, which nothing on file states for these nets: a slow static line crosses a split harmlessly, a fast edge does not.

    • split@P3:P1V5|P1V8155 nets cross the split between the P1V5 and P1V8 planes on P3 at 158 places; 157 with a plane on the track's other side (fastest: AMC_RTM_CRITICAL at 100 ps 10-90 %, ASSUMED)

      On P3 the P1V5 and P1V8 planes part, and 155 nets cross from one to the other at 158 place(s). The return current cannot follow a track across; it detours to wherever the two planes are joined (a stitching capacitor, or the source), and the loop it opens radiates and couples. One decision covers the boundary: whether it is stitched near these crossings, or the nets crossing it are slow enough not to care. Tracks that run ALONG this split, down the slot rather than across it, are a finding of their own: theirs is an impedance step, not a detour. At 157 of the 158 crossing(s) the track's other neighbour carries a plane, which the return can use there. 157 of the 158 crossing(s) have a plane on the track's other side at most twice as far away as the broken one (to first order each plane beside a stripline carries a share of the return inversely proportional to its distance from the track). The rest leave most or all of the return to the break, so the finding is a warning.

      • 'DDR3_A8' on L1 at (-80.14, 107.14), 0.49 mm; other side: no plane on G1
      • 'AMC_RTM_CRITICAL' on L1 at (-98.74, 101.2), 0.42 mm; other side: GND on G1, 0.1 mm from the track against 0.103 mm to P3: 49% of the return on P3
      • 'AMC_RTM_CRITICAL' on L1 at (-86.34, 107.1), 0.57 mm; other side: GND on G1, 0.1 mm from the track against 0.103 mm to P3: 49% of the return on P3
      • 'AMC_RTM_PV' on L1 at (-98.73, 102.53), 0.4 mm; other side: GND on G1, 0.1 mm from the track against 0.103 mm to P3: 49% of the return on P3
      • 'AMC_RTM_PV' on L1 at (-88.48, 107.1), 0.57 mm; other side: GND on G1, 0.1 mm from the track against 0.103 mm to P3: 49% of the return on P3
      • 'FJTAG.TCK' on L2 at (-96.75, 96.43), 0.57 mm; other side: GND on G2, 0.103 mm from the track against 0.1 mm to P3: 51% of the return on P3
      • 'FJTAG.TMS' on L2 at (-97.08, 96.75), 0.57 mm; other side: GND on G2, 0.103 mm from the track against 0.1 mm to P3: 51% of the return on P3
      • 'FPGA_CCLK' on L1 at (-98.5, 98.18), 0.57 mm; other side: GND on G1, 0.1 mm from the track against 0.103 mm to P3: 49% of the return on P3
      • 'FPGA_RESET_DN' on L1 at (-88.73, 55.76), 0.4 mm; other side: GND on G1, 0.1 mm from the track against 0.103 mm to P3: 49% of the return on P3
      • 'HA1.HA00_CC_N' on L2 at (-88.82, 70.03), 0.49 mm; other side: GND on G2, 0.103 mm from the track against 0.1 mm to P3: 51% of the return on P3
      • 'HA1.HA00_CC_P' on L2 at (-88.66, 70.35), 0.45 mm; other side: GND on G2, 0.103 mm from the track against 0.1 mm to P3: 51% of the return on P3
      • 'HA1.HA01_CC_N' on L2 at (-88.74, 68.78), 0.55 mm; other side: GND on G2, 0.103 mm from the track against 0.1 mm to P3: 51% of the return on P3
      • and 146 more

      Every site is a row of this finding's evidence, keyed by net, layers and position. GET /api/odb/boards/58/reference-planes?group=split%40P3%3AP1V5%7CP1V8 serves each with its own layer_image region; all of them lie in x -99.93..-74.78, y 46.44..108.3 mm (board frame, Y up).

      Read from the copper as drawn on the reference layer, along each track's centreline at exact breakpoints. The reference is the copper beside the track, never the layer's type or name. Whether it matters depends on each signal's edge rate, which nothing on file states for these nets: a slow static line crosses a split harmlessly, a fast edge does not.

    • split@P4:MGTAVCC|MGTAVTT36 nets cross the split between the MGTAVCC and MGTAVTT planes on P4 at 38 places; 35 with a plane on the track's other side (fastest: P30_IO10_AB_B2_CC_N at 100 ps 10-90 %, ASSUMED)

      On P4 the MGTAVCC and MGTAVTT planes part, and 36 nets cross from one to the other at 38 place(s). The return current cannot follow a track across; it detours to wherever the two planes are joined (a stitching capacitor, or the source), and the loop it opens radiates and couples. One decision covers the boundary: whether it is stitched near these crossings, or the nets crossing it are slow enough not to care. Tracks that run ALONG this split, down the slot rather than across it, are a finding of their own: theirs is an impedance step, not a detour. At 35 of the 38 crossing(s) the track's other neighbour carries a plane, which the return can use there. 35 of the 38 crossing(s) have a plane on the track's other side at most twice as far away as the broken one (to first order each plane beside a stripline carries a share of the return inversely proportional to its distance from the track). The rest leave most or all of the return to the break, so the finding is a warning.

      • 'P30_IO10_EF_B2_N' on L3 at (-55.44, 70.94), 0.4 mm; other side: no plane on G2
      • 'P30_IO8_EF_B1_N' on L3 at (-55.44, 63.25), 0.4 mm; other side: no plane on G2
      • 'P30_IO9_EF_B2_N' on L3 at (-55.44, 66.23), 0.4 mm; other side: no plane on G2
      • 'FP1_Rx5_N' on L3 at (-58.85, 49.07), 0.3 mm; other side: GND on G2, 0.1 mm from the track against 0.103 mm to P4: 49% of the return on P4
      • 'FP1_Rx5_P' on L3 at (-59.06, 49.07), 0.3 mm; other side: GND on G2, 0.1 mm from the track against 0.103 mm to P4: 49% of the return on P4
      • 'FP1_Rx6_N' on L3 at (-55.44, 58.19), 0.4 mm; other side: GND on G2, 0.1 mm from the track against 0.103 mm to P4: 49% of the return on P4
      • 'FP1_Rx6_P' on L3 at (-55.44, 57.98), 0.4 mm; other side: GND on G2, 0.1 mm from the track against 0.103 mm to P4: 49% of the return on P4
      • 'FP1_Tx6_N' on L3 at (-55.78, 50), 0.84 mm; other side: GND on G2, 0.1 mm from the track against 0.103 mm to P4: 49% of the return on P4
      • 'FP1_Tx6_P' on L3 at (-55.99, 49.78), 0.57 mm; other side: GND on G2, 0.1 mm from the track against 0.103 mm to P4: 49% of the return on P4
      • 'MGT224_CLK1_N' on L3 at (-55.4, 58.95), 0.4 mm; other side: GND on G2, 0.1 mm from the track against 0.103 mm to P4: 49% of the return on P4
      • 'MGT224_CLK1_P' on L3 at (-55.4, 59.16), 0.4 mm; other side: GND on G2, 0.1 mm from the track against 0.103 mm to P4: 49% of the return on P4
      • 'P30_IO10_AB_B2_CC_N' on L3 at (-55.4, 71.96), 0.4 mm; other side: GND on G2, 0.1 mm from the track against 0.103 mm to P4: 49% of the return on P4
      • and 26 more

      Every site is a row of this finding's evidence, keyed by net, layers and position. GET /api/odb/boards/58/reference-planes?group=split%40P4%3AMGTAVCC%7CMGTAVTT serves each with its own layer_image region; all of them lie in x -60.06..-54.2, y 47.92..80.14 mm (board frame, Y up).

      Read from the copper as drawn on the reference layer, along each track's centreline at exact breakpoints. The reference is the copper beside the track, never the layer's type or name. Whether it matters depends on each signal's edge rate, which nothing on file states for these nets: a slow static line crosses a split harmlessly, a fast edge does not.

    • void:P3V3MP@P6:hole -165.33,-15.3|-164.83,-14.8'PLL_CTRL.VADJ2_PLL_DAC1_DIN' crosses one void in the P3V3MP plane on P6 at 1 place, the longest 0.48 mm (fastest: PLL_CTRL.VADJ2_PLL_DAC1_DIN at 100 ps 10-90 %, ASSUMED)

      Along 'PLL_CTRL.VADJ2_PLL_DAC1_DIN' on L5, the P3V3MP plane on P6 has 0.48 mm under the track with no copper, at (-164.95, -15.05), running to the end of the track. No single drill's anti-pad and no landing explains it: a clearance is excused only when it holds a drill and is drawn round it, or the track ends in it on the net's own pad or via, or runs through its own via's clearance and no farther. The return current detours round the void. What cleared it: the hole at x -165.33..-164.83, y -15.3..-14.8. One decision covers the void: whether these tracks may cross it, or the plane should be filled under them. No crossing has a plane on the track's other side, so the whole return meets the break.

      • 'PLL_CTRL.VADJ2_PLL_DAC1_DIN' on L5 at (-164.95, -15.05), 0.48 mm over no copper, to the end of the track; other side: no plane on GND

      Every site is a row of this finding's evidence, keyed by net, layers and position. GET /api/odb/boards/58/reference-planes?group=void%3AP3V3MP%40P6%3Ahole%20-165.33%2C-15.3%7C-164.83%2C-14.8 serves each with its own layer_image region; all of them lie in x -166.08..-163.6, y -16.05..-14.05 mm (board frame, Y up).

      Read from the copper as drawn on the reference layer, along each track's centreline at exact breakpoints. The reference is the copper beside the track, never the layer's type or name. Whether it matters depends on each signal's edge rate, which nothing on file states for these nets: a slow static line crosses a split harmlessly, a fast edge does not.

  15. observationOBS-403minorOpen

    IC16 LMZ31520 (VCCINT, P0V95) switches on by itself when P12V appears, because its inhibit transistor Q12 is biased only from P3V3, a rail that is still off

    IC16 LMZ31520 (VCCINT, P0V95) starts by itself when P12V appears, because its inhibit transistor Q12 is biased only from P3V3, a rail that is still off.

    The inhibit network

    • IC16 INH (pin 16, NetIC16_16) is pulled low only by Q12 (BSS138) drain.
    • Q12 gate (NetQ12_1) has R114 10k to P3V3 and is pulled down by Q25.
    • Q25 gate is PM.EN_VCCINT (IC66 MCP23016 GP0.6, no external pull; IC61's EN has an internal 400 kOhm pull-down while disabled).
    • P3V3 comes from IC43 channel 2, whose EN2 (PM.EN_P3V3) is held low by R69 4k7 until the MMC drives it.
    • INH has an internal pull-up, and the module runs when INH is open (LMZ31520 datasheet, p.19 and p.6).

    ``mermaid flowchart LR P["P3V3<br/>off at first"] -- "R114 10k" --> G["Q12 gate"] E["PM.EN_VCCINT<br/>IC66 GP0.6"] --> Q25["Q25"] Q25 -- "pulls down" --> G G --> Q12["Q12 BSS138"] Q12 -- "pulls INH low" --> I["IC16 INH pin 16<br/>internal pull-up"] ``

    What happens at power-up

    1. Payload power P12V arrives. P3V3 is 0 V, Q12 is off, so IC16 starts and brings up P0V95 before the MMC has done anything.
    2. The MMC enables P3V3. Q12 turns on and inhibits IC16 again.
    3. PM.EN_VCCINT goes high: Q25 on, Q12 off, IC16 runs again.

    Why it matters

    VCCINT's first rise is not under the MMC's sequence control, and the rail goes on, off and on again. UltraScale does allow VCCINT to come up first, so this may not damage anything. But it is probably not what the enable network was drawn to do, since every other rail is held off by a pull-down until the MMC acts.

  16. observationOBS-404minorOpen

    IC69 TPS7A7200 (MGTAVTT, 1.2 V) has about 537 uF on its output against a 200 uF recommended maximum

    IC69 TPS7A7200 (MGTAVTT, 1.2 V) has about 537 µF on its output against a recommended maximum of 200 µF.

    Output setting

    IC69 is set to 1.2 V by its ANY-OUT pins: 100mV, 200mV and 400mV (pins 6, 7, 9) are grounded, 50mV, 800mV and 1.6V (pins 5, 10, 11) are open, giving 0.5 V + 0.7 V. The required 220 pF feed-forward capacitor (C961) is fitted.

    The numbers

    QuantityValueSource
    C249, C253, C261, C2624 × 100 µF (GRM21BC80G107ME15)BOM
    C1761 × 100 µF (CC0805_100UF_6.3V_20%_X5R)BOM
    Others2 × 10 µF, 2 × 4.7 µF, 16 × 0.47 µFBOM
    Total on MGTAVTT, nominalabout 536.9 µFderived
    If the five 100 µF parts lost half their value to DC bias and toleranceabout 265 µFderived
    Recommended COUT4.7 µF min, 200 µF maxTPS7A7200 datasheet, p.4, table 5.3

    The FPGA's MGTAVTT decoupling alone accounts for the excess. Above the recommended range the LDO is outside the conditions its stability and transient response are specified for.

  17. observationOBS-406minorOpen

    IC74 LD1117 (standalone 3.3 V management supply) has about 3.4 uF on its output where the datasheet asks for at least 10 uF for stability

    IC74 LD1117 (standalone 3.3 V management supply) has about 3.4 µF on its output, only 1 µF of it at the regulator, where the datasheet asks for at least 10 µF for stability.

    IC74 LD1117DT33 makes P3V3MP_AUX from P12V_AUX (J4 Mini-Fit, through D28 SS16). Its input has C816 1 µF.

    The numbers

    QuantityValueSource
    Capacitance on P3V3MP_AUX (IC74 VOUT)C817, 1 µF (CC0603_1uF_16V_X7R)netlist; BOM
    Capacitance on P3V3MP, past PPTC F3 (MF-MSMF250/16X-2)19 × CC0201_100NF + 5 × CC0402_100NF = 2.4 µFnetlist; BOM
    Total, nominal1 + 2.4 = about 3.4 µFderived
    Required minimum10 µFLD1117 datasheet, p.1

    This supply feeds the MMC's management rail when the card is powered from J4.

  18. observationOBS-413minorOpen

    Unused translator inputs left floating: IC17 SN74AVC2T245 A2 (pin 9) and IC41 SN74AVC8T245 A8 (pin 10) are on no net while their direction makes them inputs

    Two unused translator inputs float: IC17 SN74AVC2T245 A2 (pin 9) and IC41 SN74AVC8T245 A8 (pin 10) are on no net, and their direction settings make them inputs.

    DeviceUnused pinsDirectionOE
    IC17 SN74AVC2T245RSWRA2 pin 9, B2 pin 4: no netDIR2 (pin 1) high via R401 0 ohm to P1V8, so A2 is an input (A to B)pin 2 low via R402: always enabled
    IC41 SN74AVC8T245RHLA8 pin 10, B8 pin 14: no netDIR (pin 2) = P1V8 (A to B)pin 22 = GND

    IC17 uses only channel 1 (I2C_FPGA_RST_C to FPGA_I2C_RST).

    Both datasheets (recommended operating conditions note 3, p.5) require all unused data inputs to be held at VCCI or GND. The consequence is extra supply current and a toggling unused B output: noise, not a functional fault.

  19. observationOBS-415minorOpen

    IC18 MAX6642 remote-diode input DXP (FPGA die diode) has no filter capacitor; the datasheet asks for 2200 pF from DXP to GND

    IC18 MAX6642 DXP (FPGA die diode input) has no filter capacitor; the datasheet asks for 2200 pF from DXP to GND.

    IC18 measures the Kintex die temperature through the FPGA's thermal diode.

    • DXP: IC18 pin 3 connects on net FPGA_THERM.DXP directly to IC2 ball Y12. The net carries only those two pins, no capacitor to GND.
    • DXN: FPGA_THERM.DXN (IC2 ball Y11) goes to GND through R135 (0 ohm).

    The MAX6642 pin description (p.5) asks for a 2200 pF capacitor between DXP and GND for noise filtering (up to 3300 pF). Without it, noise coupled onto the DXP trace reads as a temperature offset, typically too high.

  20. observationOBS-417minorOpen

    RGB LED3 has the same 330 ohm on all three colours from a 3.3 V anode: red gets about 4 mA, but the green and blue dice (Vf 3.0-3.6 V) get between 0 and about 1 mA and will be dim or dark

    RGB LED3 uses 330 ohm on all three colours from a 3.3 V anode: red gets about 4 mA, but green and blue (Vf 3.0-3.6 V) get between 0 and about 1 mA and will be dim or dark.

    LED3 OVSRRGBCC3 has its common anode (pin 4) on P3V3. Each cathode goes through 330 ohm to a BSS138 low-side switch gated from the FPGA (IC2 F20/G20/U29, each gate with a 10k pull-down).

    ColourResistorSwitchVf at 20 mA, min/typ/max
    RedR277Q71.8 / 2.0 / 2.4 V
    GreenR278Q103.0 / 3.3 / 3.6 V
    BlueR279Q113.0 / 3.3 / 3.6 V

    Vf source: OVSRRGBCC3 datasheet, p.3.

    The numbers

    QuantityValueSource
    Red current(3.3 - 2.0) / 330 = 3.9 mAderived
    Green / blue current(3.3 - 3.0 to 3.3) / 330 = about 0.9 mA at best, zero at typical Vfderived

    No colour weighting intent is recorded on the schematic. The result will vary part to part.

  21. observationOBS-422minorOpen

    No test point on P1V5 (DDR3 VDD/VDDQ and the TPS51200 VTT/VREF input) or on DDR3_VREF, while every other regulated rail has one

    P1V5 (DDR3 supply from IC67) and DDR3_VREF (IC15 REFOUT) have no test point, while every other regulated rail has one.

    P1V5 is made by IC67 TPS7A7200 and feeds the four MT41K DDR3 devices, the FPGA DDR bank and IC15 VLDOIN/REFIN (149 pins). DDR3_VREF has 22 pins. Neither has a test pad.

    Rails that do have one

    • TP1 MGTAVCC, TP3 FMC1_P3V3, TP4 VTT, TP5 FMC2_P3V3
    • TP6 MGTAVTT, TP7 P0V95, TP8 PVADJ1, TP9 PVADJ2
    • TP10 P1V8, TP11 P3V3, B21 P3V3MP, B26 MGT_VCCAUX
    • B5/B6/TPDx GND
  22. observationOBS-423minorOpen

    Schematic title blocks are incomplete: the checker field is blank on every sheet, the Thermometers sheet still carries the placeholder title 'Title 2', and sheet numbering does not match the 29-page set

    The schematic title blocks are incomplete: 'Check.by' is blank on every sheet, the Thermometers sheet is titled 'Title 2', and the sheet numbers do not match the 29-page set.

    Read from the schematic PDF on file (29 pages, extracted title-block text):

    • Checker and date: every sheet's 'Check.by' field is '-' and 'Last Mod.' is '-'.
    • Placeholder title: page 29 (Thermometers.SchDoc, sheet 28 of 29) has the document title 'Title 2' instead of a sheet title.
    • Numbering: 'Sheet x of 29' runs 1-6 on pages 1-6. Pages 7, 8, 9, 11, 12 and 14 carry no parsable sheet number. Pages 10-29 are numbered 9-28 (p.10 FPGA.SchDoc is sheet 9 of 29), so the set never reaches sheet 29 of 29.
  23. observationOBS-540minorOpen

    FMC power-good pull-ups R97-R100 go to the management rail P3V3MP, not to the FMC 3P3V that VITA 57.1 Rule 5.78 names

    On both FMC sites the 10k pull-ups on PG_C2M and PG_M2C (R98/R100 on J15A, R97/R99 on J15B) go to P3V3MP, the always-on MMC management 3.3 V. VITA 57.1 Rule 5.78 asks for them on 3P3V, which on this carrier is FMC1_P3V3/FMC2_P3V3.

    What the standard asks

    RequirementSource
    "The Carrier card shall provide a 10K pull-up resister on each of PG_M2C and PG_C2M signals to 3P3V."ANSI/VITA 57.1-2008, Rule 5.78, p.68

    What the board has

    SiteSignalPull-upRailFMC 3P3V rail on that site
    J15AMISC1.PG_C2M (D1)R98 10kP3V3MPFMC1_P3V3 (C39/D36/D38/D40)
    J15AMISC1.PG_M2C (F1)R100 10kP3V3MPFMC1_P3V3
    J15BMISC2.PG_C2M (D1)R97 10kP3V3MPFMC2_P3V3
    J15BMISC2.PG_M2C (F1)R99 10kP3V3MPFMC2_P3V3

    Why it matters

    P3V3MP is up whenever the AMC has management power, and the FMC 3P3V is switched by IC44. With the module's 3P3V off:

    • PG_M2C reads high through the pull-up even though an unpowered module cannot assert power-good.
    • Both pull-ups source current into the unpowered mezzanine's PG pins.

    The pull-up value is right, but the rail is not the one the standard names.

  24. observationOBS-679minorOpen

    BC847B went end-of-life (Not For New Designs)

    Vendor (DigiKey) lifecycle for BC847B (Nexperia USA Inc.'s listing) changed from "Active" to "Not For New Designs" since this board was reviewed. Placed as Q16, Q26, Q27, Q5, Q6, Q9. A sourcing/longevity risk; consider a replacement before production.

  25. concernCON-001minorOpen

    [eol:BC847B:Not For New Designs] BC847B lifecycle changed from "Active" to "Not For New Designs" since this board was reviewed: OHWR0057 R1 AFCKU. A sourcing and longevity risk. Should this be replaced before the next build?

  26. DecapViaFanoutminorOpen

    C570 on 'P1V8_USB': its rail pad reaches its via over 2.00 mm of track, adding about 0.51 nH to its loop, past Probe's 0.5 nH starting value; put the vias at the pads (a 10 uF bulk capacitor: its inductance matters only far above the band it serves)

    C570 on 'P1V8_USB': its rail pad reaches its via over 2.00 mm of track, adding about 0.51 nH to its loop, past Probe's 0.5 nH starting value; put the vias at the pads (a 10 uF bulk capacitor: its inductance matters only far above the band it serves). C570.1 (rail) runs 2.00 mm along its copper to its nearest via, 1.00 mm past the 1 mm beside the pad: 0.51 nH; C570.2 (ground) not measured: the pad sits in a pour of its net on its own layer, which has no track length to measure. Each run past the pad as one conductor of a pair 1.60 mm apart (the pad pitch), 0.2 l ln(D/R) nH with R 0.127 mm (the track's width is not read): 0.51 nH in all. Put the vias at the pads (or in them), one per pad at least, the rail's and the ground's close together; the capacitor's own inductance as mounted is 0.98 nH (pad 0.10 nH + no via of the first net within 1 mm of its pad: the study's 0.6 nH per mm of 0.61 mm depth, ASSUMED). 0.5 nH is Probe's starting value, about the whole mounted inductance of a well-mounted 0402.

    At C570

  27. ReferencePlaneDiscontinuityminorOpen

    ReferencePlaneDiscontinuity: 6 boards×6

    6 checked by the review by Claude walker W4 (Cern session) — 6 different reasons

    At refchange:GND|P0V95, refchange:GND|P1V5, refchange:GND|P3V3, refchange:GND|P3V3MP, split@P6:P0V95|P1V5, split@P6:P1V5|P1V8

    Each finding, with its own headline
    • refchange:GND|P0V95163 nets change reference between GND and P0V95 at 259 vias, each with another reference that continues across (fastest: AMC_RTM_WARNING at 100 ps 10-90 %, ASSUMED)

      At 259 via(s), on 163 net(s), the signal leaves a layer referenced by one of GND and P0V95 for a layer referenced by the other, and no plane references both layers. Different nets: no via can carry the return across; only decoupling between the two planes near the via can. The capacitor carrying the return across was measured at each via, at the knee of the crossing net's edge against 50 ohm assumed: 259 with none. 259 of them at an ASSUMED edge (no edge on file for the net): state one (net_timing_state) and it is re-read. One decision covers the pair: whether GND and P0V95 are decoupled to each other near these vias. A via whose layers have more references than these two is a site of each pair it joins. Every one is partial: another reference continues across through a stitching via of its net, so only part of the return has nowhere to go, and the finding is the lower tier (info). They are listed with the farthest stitching via first.

      • 'MGT.125_GTP225_CLK0_P' at (-35.31, 47.69), TOP -> L5: GND on G1 to GND on GND, P0V95 on P6; partial, GND continues through a via of GND 1.02 mm away; stitching none, 144.2 mm of track past the via on L5: At the 5 GHz knee of a 100 ps edge (1 ns 10-90 %, assumed: no edge is on file for the net) no capacitor carries the return; only the planes do. GND and P0V95 are not a thin pair (at most 0.2 mm), so their own capacitance does not carry the return. To serve it, change layers across the same plane, or move the signal to a layer referenced to ground. None.
      • 'MGT.125_GTP225_CLK0_N' at (-34.96, 47.34), TOP -> L5: GND on G1 to GND on GND, P0V95 on P6; partial, GND continues through a via of GND 0.53 mm away; stitching none, 143.7 mm of track past the via on L5: At the 5 GHz knee of a 100 ps edge (1 ns 10-90 %, assumed: no edge is on file for the net) no capacitor carries the return; only the planes do. GND and P0V95 are not a thin pair (at most 0.2 mm), so their own capacitance does not carry the return. To serve it, change layers across the same plane, or move the signal to a layer referenced to ground. None.
      • 'P31_IO1_CD_B2_P' at (-25.45, 78.71), L3 -> L5: GND on G2, P3V3MP on P4 to GND on GND, P0V95 on P6; partial, GND continues through a via of GND 1.26 mm away; stitching none, 138.0 mm of track past the via on L5: At the 5 GHz knee of a 100 ps edge (100 ps 10-90 %, assumed: no edge is on file for the net) no capacitor carries the return; only the planes do. GND and P0V95 are not a thin pair (at most 0.2 mm), so their own capacitance does not carry the return. To serve it, change layers across the same plane, or move the signal to a layer referenced to ground. None.
      • 'P31_IO1_CD_B2_N' at (-25.45, 78.21), L3 -> L5: GND on G2, P3V3MP on P4 to GND on GND, P0V95 on P6; partial, GND continues through a via of GND 0.78 mm away; stitching none, 137.7 mm of track past the via on L5: At the 5 GHz knee of a 100 ps edge (100 ps 10-90 %, assumed: no edge is on file for the net) no capacitor carries the return; only the planes do. GND and P0V95 are not a thin pair (at most 0.2 mm), so their own capacitance does not carry the return. To serve it, change layers across the same plane, or move the signal to a layer referenced to ground. None.
      • 'PM.1V5_VTT_EN' at (-32.65, 74.57), L2 -> L5: P1V8 on P3, GND on G2 to GND on GND, P0V95 on P6; partial, GND continues through a via of GND 0.74 mm away; stitching none, 135.1 mm of track past the via on L5: At the 5 GHz knee of a 100 ps edge (20 ns 10-90 %, assumed: no edge is on file for the net) no capacitor carries the return; only the planes do. GND and P0V95 are not a thin pair (at most 0.2 mm), so their own capacitance does not carry the return. To serve it, change layers across the same plane, or move the signal to a layer referenced to ground. None.
      • 'VREF2.VREF_A_M2C' at (-71.44, 56.58), TOP -> L6: GND on G1 to P0V95 on P6, GND on G3; partial, GND continues through a via of GND 1 mm away; stitching none, 127.9 mm of track past the via on L6: At the 5 GHz knee of a 100 ps edge (100 ps 10-90 %, assumed: no edge is on file for the net) no capacitor carries the return; only the planes do. GND and P0V95 are not a thin pair (at most 0.2 mm), so their own capacitance does not carry the return. To serve it, change layers across the same plane, or move the signal to a layer referenced to ground. Measured for reference: C907 (4.7 uF 0402) 2.7 mm away: 1.11 nH in the 0.912 mm cavity + 0.30 nH ESL + 0.68 nH mounting (pad 0.07 nH + no via of the first net within 1 mm of its pad: the study's 0.6 nH per mm of 1.02 mm depth, ASSUMED): 65.9 ohm at the knee, 132 % of 50 ohm (assumed track impedance). None.
      • 'DDR3_A7' at (-84.5, 105.84), TOP -> L6: GND on G1 to P0V95 on P6, GND on G3; partial, GND continues through a via of GND 2.77 mm away; stitching none, 122.4 mm of track past the via on L6: At the 5 GHz knee of a 100 ps edge (100 ps 10-90 %, assumed: no edge is on file for the net) no capacitor carries the return; only the planes do. GND and P0V95 are not a thin pair (at most 0.2 mm), so their own capacitance does not carry the return. To serve it, change layers across the same plane, or move the signal to a layer referenced to ground. None.
      • 'DDR3_A2' at (-84.5, 104.92), TOP -> L5: GND on G1 to GND on GND, P0V95 on P6; partial, GND continues through a via of GND 2.65 mm away; stitching none, 122.0 mm of track past the via on L5: At the 5 GHz knee of a 100 ps edge (100 ps 10-90 %, assumed: no edge is on file for the net) no capacitor carries the return; only the planes do. GND and P0V95 are not a thin pair (at most 0.2 mm), so their own capacitance does not carry the return. To serve it, change layers across the same plane, or move the signal to a layer referenced to ground. None.
      • 'DDR3_RAS_N' at (-81.25, 100.65), TOP -> L5: GND on G1 to GND on GND, P0V95 on P6; partial, GND continues through a via of GND 1.03 mm away; stitching none, 119.2 mm of track past the via on L5: At the 5 GHz knee of a 100 ps edge (100 ps 10-90 %, assumed: no edge is on file for the net) no capacitor carries the return; only the planes do. GND and P0V95 are not a thin pair (at most 0.2 mm), so their own capacitance does not carry the return. To serve it, change layers across the same plane, or move the signal to a layer referenced to ground. None.
      • 'DDR3_CAS_N' at (-81.25, 101.57), TOP -> L5: GND on G1 to GND on GND, P0V95 on P6; partial, GND continues through a via of GND 1.95 mm away; stitching none, 118.9 mm of track past the via on L5: At the 5 GHz knee of a 100 ps edge (100 ps 10-90 %, assumed: no edge is on file for the net) no capacitor carries the return; only the planes do. GND and P0V95 are not a thin pair (at most 0.2 mm), so their own capacitance does not carry the return. To serve it, change layers across the same plane, or move the signal to a layer referenced to ground. None.
      • 'DDR3_CKE' at (-84.5, 101.25), TOP -> L5: GND on G1 to GND on GND, P0V95 on P6; partial, GND continues through a via of GND 0.68 mm away; stitching none, 112.9 mm of track past the via on L5: At the 5 GHz knee of a 100 ps edge (100 ps 10-90 %, assumed: no edge is on file for the net) no capacitor carries the return; only the planes do. GND and P0V95 are not a thin pair (at most 0.2 mm), so their own capacitance does not carry the return. To serve it, change layers across the same plane, or move the signal to a layer referenced to ground. None.
      • 'FPGA_CLK.SEC_CLK_P' at (-61.44, 56.58), TOP -> L5: GND on G1 to GND on GND, P0V95 on P6; partial, GND continues through a via of GND 1.41 mm away; stitching none, 99.8 mm of track past the via on L5: At the 5 GHz knee of a 100 ps edge (1 ns 10-90 %, assumed: no edge is on file for the net) no capacitor carries the return; only the planes do. GND and P0V95 are not a thin pair (at most 0.2 mm), so their own capacitance does not carry the return. To serve it, change layers across the same plane, or move the signal to a layer referenced to ground. None.
      • and 247 more

      Every site is a row of this finding's evidence, keyed by net, layers and position. GET /api/odb/boards/58/reference-planes?group=refchange%3AGND%7CP0V95 serves each with its own layer_image region; all of them lie in x -86.5..-3.4, y -8.35..108.25 mm (board frame, Y up).

      The reference on each layer is the first plane the track meets leaving the via, anti-pads and the net's own copper skipped. Whether it matters depends on each signal's edge rate, which nothing on file states for these nets: a slow static line crosses a split harmlessly, a fast edge does not.

    • refchange:GND|P1V5223 nets change reference between GND and P1V5 at 341 vias, each with another reference that continues across (fastest: DDR3_A0 at 100 ps 10-90 %, ASSUMED)

      At 341 via(s), on 223 net(s), the signal leaves a layer referenced by one of GND and P1V5 for a layer referenced by the other, and no plane references both layers. Different nets: no via can carry the return across; only decoupling between the two planes near the via can. The capacitor carrying the return across was measured at each via, at the knee of the crossing net's edge against 50 ohm assumed: 341 with none. 341 of them at an ASSUMED edge (no edge on file for the net): state one (net_timing_state) and it is re-read. One decision covers the pair: whether GND and P1V5 are decoupled to each other near these vias. A via whose layers have more references than these two is a site of each pair it joins. Every one is partial: another reference continues across through a stitching via of its net, so only part of the return has nowhere to go, and the finding is the lower tier (info). They are listed with the farthest stitching via first.

      • 'SI57X_SEC_SDA' at (-167.08, -8.08), TOP -> L5: GND on G1 to GND on GND, P1V5 on P6; partial, GND continues through a via of GND 2.74 mm away; stitching none, 176.1 mm of track past the via on L5: At the 5 GHz knee of a 100 ps edge (10 ns 10-90 %, assumed: no edge is on file for the net) no capacitor carries the return; only the planes do. GND and P1V5 are not a thin pair (at most 0.2 mm), so their own capacitance does not carry the return. To serve it, change layers across the same plane, or move the signal to a layer referenced to ground. Measured for reference: C128 (100 nF 0201) 2.7 mm away: 1.11 nH in the 0.912 mm cavity + 0.21 nH ESL + 1.21 nH mounting (pad 0.05 nH (vias beside the pads) + 1 via pair(s) at 1.70 mm pitch, 0.20 mm drill, 1.02 mm to the farther plane): 79.3 ohm at the knee, 159 % of 50 ohm (assumed track impedance). None.
      • 'IPMB.SDA' at (-155.85, 102.46), TOP -> L6: GND on G1 to P1V5 on P6, GND on G3; partial, GND continues through a via of GND 3.53 mm away; stitching none, 174.7 mm of track past the via on L6: At the 5 GHz knee of a 100 ps edge (10 ns 10-90 %, assumed: no edge is on file for the net) no capacitor carries the return; only the planes do. GND and P1V5 are not a thin pair (at most 0.2 mm), so their own capacitance does not carry the return. To serve it, change layers across the same plane, or move the signal to a layer referenced to ground. None.
      • 'P2_BLUE' at (-169.62, -30.05), TOP -> L5: GND on G1 to GND on GND, P1V5 on P6; partial, GND continues through a via of GND 6.5 mm away; stitching none, 173.6 mm of track past the via on L5: At the 5 GHz knee of a 100 ps edge (100 ps 10-90 %, assumed: no edge is on file for the net) no capacitor carries the return; only the planes do. GND and P1V5 are not a thin pair (at most 0.2 mm), so their own capacitance does not carry the return. To serve it, change layers across the same plane, or move the signal to a layer referenced to ground. None.
      • 'INIT_B' at (-122.48, 108.38), L2 -> L5: P3V3 on P3, GND on G2 to GND on GND, P1V5 on P6; partial, GND continues through a via of GND 6.43 mm away; stitching none, 162.5 mm of track past the via on L5: At the 5 GHz knee of a 100 ps edge (100 ps 10-90 %, assumed: no edge is on file for the net) no capacitor carries the return; only the planes do. GND and P1V5 are not a thin pair (at most 0.2 mm), so their own capacitance does not carry the return. To serve it, change layers across the same plane, or move the signal to a layer referenced to ground. None.
      • 'PM.EN_RTM_PWR' at (-128.41, 108.55), L2 -> L6: P3V3 on P3, GND on G2 to P1V5 on P6, GND on G3; partial, GND continues through a via of GND 8.46 mm away; stitching none, 140.6 mm of track past the via on L6: At the 5 GHz knee of a 100 ps edge (100 ps 10-90 %, assumed: no edge is on file for the net) no capacitor carries the return; only the planes do. GND and P1V5 are not a thin pair (at most 0.2 mm), so their own capacitance does not carry the return. To serve it, change layers across the same plane, or move the signal to a layer referenced to ground. None.
      • 'PM.EN_FMC1_P12V' at (-132.37, 96.4), TOP -> L6: GND on G1 to P1V5 on P6, GND on G3; partial, GND continues through a via of GND 0.58 mm away; stitching none, 128.4 mm of track past the via on L6: At the 5 GHz knee of a 100 ps edge (20 ns 10-90 %, assumed: no edge is on file for the net) no capacitor carries the return; only the planes do. GND and P1V5 are not a thin pair (at most 0.2 mm), so their own capacitance does not carry the return. To serve it, change layers across the same plane, or move the signal to a layer referenced to ground. None.
      • 'PLL_CTRL.VADJ2_SI57X_PRI_SDA' at (-171.42, -19.15), TOP -> L5: GND on G1 to GND on GND, P1V5 on P6; partial, GND continues through a via of GND 1.6 mm away; stitching none, 122.6 mm of track past the via on L5: At the 5 GHz knee of a 100 ps edge (10 ns 10-90 %, assumed: no edge is on file for the net) no capacitor carries the return; only the planes do. GND and P1V5 are not a thin pair (at most 0.2 mm), so their own capacitance does not carry the return. To serve it, change layers across the same plane, or move the signal to a layer referenced to ground. None.
      • 'DDR3_A1' at (-81.22, 104.5), TOP -> L2: GND on G1 to P1V5 on P3, GND on G2; partial, GND continues through a via of GND 2.43 mm away; stitching none, 121.8 mm of track past the via on L2: At the 5 GHz knee of a 100 ps edge (100 ps 10-90 %, assumed: no edge is on file for the net) no capacitor carries the return; only the planes do. GND and P1V5 are not a thin pair (at most 0.2 mm), so their own capacitance does not carry the return. To serve it, change layers across the same plane, or move the signal to a layer referenced to ground. Measured for reference: C776 (100 nF 0201) 2.2 mm away: 0.23 nH in the 0.203 mm cavity + 0.21 nH ESL + 0.31 nH mounting (pad 0.05 nH (vias beside the pads) + 1 via pair(s) at 0.80 mm pitch, 0.20 mm drill, 0.31 mm to the farther plane): 23.6 ohm at the knee, 47 % of 50 ohm (assumed track impedance). None.
      • 'DDR3_A11' at (-81.22, 105.75), TOP -> L2: GND on G1 to P1V5 on P3, GND on G2; partial, GND continues through a via of GND 2.48 mm away; stitching none, 121.8 mm of track past the via on L2: At the 5 GHz knee of a 100 ps edge (100 ps 10-90 %, assumed: no edge is on file for the net) no capacitor carries the return; only the planes do. GND and P1V5 are not a thin pair (at most 0.2 mm), so their own capacitance does not carry the return. To serve it, change layers across the same plane, or move the signal to a layer referenced to ground. Measured for reference: C776 (100 nF 0201) 3.0 mm away: 0.26 nH in the 0.203 mm cavity + 0.21 nH ESL + 0.31 nH mounting (pad 0.05 nH (vias beside the pads) + 1 via pair(s) at 0.80 mm pitch, 0.20 mm drill, 0.31 mm to the farther plane): 24.4 ohm at the knee, 49 % of 50 ohm (assumed track impedance). None.
      • 'DDR3_A13' at (-84.52, 106.75), TOP -> L2: GND on G1 to P1V5 on P3, GND on G2; partial, GND continues through a via of GND 2.05 mm away; stitching none, 121.7 mm of track past the via on L2: At the 5 GHz knee of a 100 ps edge (100 ps 10-90 %, assumed: no edge is on file for the net) no capacitor carries the return; only the planes do. GND and P1V5 are not a thin pair (at most 0.2 mm), so their own capacitance does not carry the return. To serve it, change layers across the same plane, or move the signal to a layer referenced to ground. None.
      • 'DDR3_A12' at (-81.25, 103.5), TOP -> L2: GND on G1 to P1V5 on P3, GND on G2; partial, GND continues through a via of GND 2.48 mm away; stitching none, 121.7 mm of track past the via on L2: At the 5 GHz knee of a 100 ps edge (100 ps 10-90 %, assumed: no edge is on file for the net) no capacitor carries the return; only the planes do. GND and P1V5 are not a thin pair (at most 0.2 mm), so their own capacitance does not carry the return. To serve it, change layers across the same plane, or move the signal to a layer referenced to ground. Measured for reference: C776 (100 nF 0201) 2.0 mm away: 0.22 nH in the 0.203 mm cavity + 0.21 nH ESL + 0.31 nH mounting (pad 0.05 nH (vias beside the pads) + 1 via pair(s) at 0.80 mm pitch, 0.20 mm drill, 0.31 mm to the farther plane): 23.4 ohm at the knee, 47 % of 50 ohm (assumed track impedance). None.
      • 'DDR3_A10' at (-81.25, 102.5), TOP -> L2: GND on G1 to P1V5 on P3, GND on G2; partial, GND continues through a via of GND 2.52 mm away; stitching none, 121.0 mm of track past the via on L2: At the 5 GHz knee of a 100 ps edge (100 ps 10-90 %, assumed: no edge is on file for the net) no capacitor carries the return; only the planes do. GND and P1V5 are not a thin pair (at most 0.2 mm), so their own capacitance does not carry the return. To serve it, change layers across the same plane, or move the signal to a layer referenced to ground. Measured for reference: C776 (100 nF 0201) 2.2 mm away: 0.23 nH in the 0.203 mm cavity + 0.21 nH ESL + 0.31 nH mounting (pad 0.05 nH (vias beside the pads) + 1 via pair(s) at 0.80 mm pitch, 0.20 mm drill, 0.31 mm to the farther plane): 23.6 ohm at the knee, 47 % of 50 ohm (assumed track impedance). None.
      • and 329 more

      Every site is a row of this finding's evidence, keyed by net, layers and position. GET /api/odb/boards/58/reference-planes?group=refchange%3AGND%7CP1V5 serves each with its own layer_image region; all of them lie in x -179.03..-38, y -35.81..110.55 mm (board frame, Y up).

      The reference on each layer is the first plane the track meets leaving the via, anti-pads and the net's own copper skipped. Whether it matters depends on each signal's edge rate, which nothing on file states for these nets: a slow static line crosses a split harmlessly, a fast edge does not.

    • refchange:GND|P3V3142 nets change reference between GND and P3V3 at 254 vias, each with another reference that continues across (fastest: AMC_RTM_CRITICAL at 100 ps 10-90 %, ASSUMED)

      At 254 via(s), on 142 net(s), the signal leaves a layer referenced by one of GND and P3V3 for a layer referenced by the other, and no plane references both layers. Different nets: no via can carry the return across; only decoupling between the two planes near the via can. The capacitor carrying the return across was measured at each via, at the knee of the crossing net's edge against 50 ohm assumed: 254 with none. 254 of them at an ASSUMED edge (no edge on file for the net): state one (net_timing_state) and it is re-read. One decision covers the pair: whether GND and P3V3 are decoupled to each other near these vias. A via whose layers have more references than these two is a site of each pair it joins. Every one is partial: another reference continues across through a stitching via of its net, so only part of the return has nowhere to go, and the finding is the lower tier (info). They are listed with the farthest stitching via first.

      • 'OVERTEMPn' at (-144.25, 71.9), TOP -> L2: GND on G1 to P3V3 on P3, GND on G2; partial, GND continues through a via of GND 4.96 mm away; stitching none, 307.5 mm of track past the via on L2: At the 5 GHz knee of a 100 ps edge (100 ps 10-90 %, assumed: no edge is on file for the net) no capacitor carries the return; only the planes do. GND and P3V3 are not a thin pair (at most 0.2 mm), so their own capacitance does not carry the return. To serve it, change layers across the same plane, or move the signal to a layer referenced to ground. None.
      • 'OVERTEMPn' at (-155.35, 71.81), TOP -> L2: GND on G1 to P3V3 on P3, GND on G2; partial, GND continues through a via of GND 3.02 mm away; stitching none, 307.5 mm of track past the via on L2: At the 5 GHz knee of a 100 ps edge (100 ps 10-90 %, assumed: no edge is on file for the net) no capacitor carries the return; only the planes do. GND and P3V3 are not a thin pair (at most 0.2 mm), so their own capacitance does not carry the return. To serve it, change layers across the same plane, or move the signal to a layer referenced to ground. None.
      • 'OVERTEMPn' at (-44.77, -23.97), TOP -> L2: GND on G1 to P3V3 on P3, GND on G2; partial, GND continues through a via of GND 2.85 mm away; stitching none, 307.5 mm of track past the via on L2: At the 5 GHz knee of a 100 ps edge (100 ps 10-90 %, assumed: no edge is on file for the net) no capacitor carries the return; only the planes do. GND and P3V3 are not a thin pair (at most 0.2 mm), so their own capacitance does not carry the return. To serve it, change layers across the same plane, or move the signal to a layer referenced to ground. None.
      • 'OVERTEMPn' at (-155.35, 84.96), TOP -> L2: GND on G1 to P3V3 on P3, GND on G2; partial, GND continues through a via of GND 2.07 mm away; stitching none, 307.5 mm of track past the via on L2: At the 5 GHz knee of a 100 ps edge (100 ps 10-90 %, assumed: no edge is on file for the net) no capacitor carries the return; only the planes do. GND and P3V3 are not a thin pair (at most 0.2 mm), so their own capacitance does not carry the return. To serve it, change layers across the same plane, or move the signal to a layer referenced to ground. None.
      • 'OVERTEMPn' at (-151.73, -9.2), TOP -> L2: GND on G1 to P3V3 on P3, GND on G2; partial, GND continues through a via of GND 1.84 mm away; stitching none, 307.5 mm of track past the via on L2: At the 5 GHz knee of a 100 ps edge (100 ps 10-90 %, assumed: no edge is on file for the net) no capacitor carries the return; only the planes do. GND and P3V3 are not a thin pair (at most 0.2 mm), so their own capacitance does not carry the return. To serve it, change layers across the same plane, or move the signal to a layer referenced to ground. None.
      • 'OVERTEMPn' at (-42.05, -27.05), TOP -> L2: GND on G1 to P3V3 on P3, GND on G2; partial, GND continues through a via of GND 0.97 mm away; stitching none, 307.5 mm of track past the via on L2: At the 5 GHz knee of a 100 ps edge (100 ps 10-90 %, assumed: no edge is on file for the net) no capacitor carries the return; only the planes do. GND and P3V3 are not a thin pair (at most 0.2 mm), so their own capacitance does not carry the return. To serve it, change layers across the same plane, or move the signal to a layer referenced to ground. None.
      • 'DONE' at (-153.35, 82.63), TOP -> L2: GND on G1 to P3V3 on P3, GND on G2; partial, GND continues through a via of GND 1.11 mm away; stitching none, 284.6 mm of track past the via on L2: At the 5 GHz knee of a 100 ps edge (100 ps 10-90 %, assumed: no edge is on file for the net) no capacitor carries the return; only the planes do. GND and P3V3 are not a thin pair (at most 0.2 mm), so their own capacitance does not carry the return. To serve it, change layers across the same plane, or move the signal to a layer referenced to ground. None.
      • 'AMC_RTM_WARNING' at (-156.55, 90.19), TOP -> L2: GND on G1 to P3V3 on P3, GND on G2; partial, GND continues through a via of GND 3.51 mm away; stitching none, 215.2 mm of track past the via on L2: At the 5 GHz knee of a 100 ps edge (100 ps 10-90 %, assumed: no edge is on file for the net) no capacitor carries the return; only the planes do. GND and P3V3 are not a thin pair (at most 0.2 mm), so their own capacitance does not carry the return. To serve it, change layers across the same plane, or move the signal to a layer referenced to ground. None.
      • 'SCL0' at (-137.16, 68.15), TOP -> L2: GND on G1 to P3V3 on P3, GND on G2; partial, GND continues through a via of GND 2.85 mm away; stitching none, 215.1 mm of track past the via on L2: At the 5 GHz knee of a 100 ps edge (10 ns 10-90 %, assumed: no edge is on file for the net) no capacitor carries the return; only the planes do. GND and P3V3 are not a thin pair (at most 0.2 mm), so their own capacitance does not carry the return. To serve it, change layers across the same plane, or move the signal to a layer referenced to ground. None.
      • 'SCL0' at (-150.48, -9.2), TOP -> L2: GND on G1 to P3V3 on P3, GND on G2; partial, GND continues through a via of GND 1.19 mm away; stitching none, 215.1 mm of track past the via on L2: At the 5 GHz knee of a 100 ps edge (10 ns 10-90 %, assumed: no edge is on file for the net) no capacitor carries the return; only the planes do. GND and P3V3 are not a thin pair (at most 0.2 mm), so their own capacitance does not carry the return. To serve it, change layers across the same plane, or move the signal to a layer referenced to ground. None.
      • 'SCL0' at (-48.58, -26.3), TOP -> L2: GND on G1 to P3V3 on P3, GND on G2; partial, GND continues through a via of GND 0.77 mm away; stitching none, 215.1 mm of track past the via on L2: At the 5 GHz knee of a 100 ps edge (10 ns 10-90 %, assumed: no edge is on file for the net) no capacitor carries the return; only the planes do. GND and P3V3 are not a thin pair (at most 0.2 mm), so their own capacitance does not carry the return. To serve it, change layers across the same plane, or move the signal to a layer referenced to ground. None.
      • 'SDA0' at (-136.02, 68.86), TOP -> L2: GND on G1 to P3V3 on P3, GND on G2; partial, GND continues through a via of GND 2.14 mm away; stitching none, 215.0 mm of track past the via on L2: At the 5 GHz knee of a 100 ps edge (10 ns 10-90 %, assumed: no edge is on file for the net) no capacitor carries the return; only the planes do. GND and P3V3 are not a thin pair (at most 0.2 mm), so their own capacitance does not carry the return. To serve it, change layers across the same plane, or move the signal to a layer referenced to ground. None.
      • and 242 more

      Every site is a row of this finding's evidence, keyed by net, layers and position. GET /api/odb/boards/58/reference-planes?group=refchange%3AGND%7CP3V3 serves each with its own layer_image region; all of them lie in x -177.34..-31.93, y -35.12..110.55 mm (board frame, Y up).

      The reference on each layer is the first plane the track meets leaving the via, anti-pads and the net's own copper skipped. Whether it matters depends on each signal's edge rate, which nothing on file states for these nets: a slow static line crosses a split harmlessly, a fast edge does not.

    • refchange:GND|P3V3MP254 nets change reference between GND and P3V3MP at 371 vias, each with another reference that continues across (fastest: ADD.GA1 at 100 ps 10-90 %, ASSUMED)

      At 371 via(s), on 254 net(s), the signal leaves a layer referenced by one of GND and P3V3MP for a layer referenced by the other, and no plane references both layers. Different nets: no via can carry the return across; only decoupling between the two planes near the via can. The capacitor carrying the return across was measured at each via, at the knee of the crossing net's edge against 50 ohm assumed: 371 with none. 371 of them at an ASSUMED edge (no edge on file for the net): state one (net_timing_state) and it is re-read. One decision covers the pair: whether GND and P3V3MP are decoupled to each other near these vias. A via whose layers have more references than these two is a site of each pair it joins. Every one is partial: another reference continues across through a stitching via of its net, so only part of the return has nowhere to go, and the finding is the lower tier (info). They are listed with the farthest stitching via first.

      • 'AJTAG.TRST#' at (-161.5, 49.67), TOP -> L6: GND on G1 to P3V3MP on P6, GND on G3; partial, GND continues through a via of GND 1.69 mm away; stitching none, 238.4 mm of track past the via on L6: At the 5 GHz knee of a 100 ps edge (20 ns 10-90 %, assumed: no edge is on file for the net) no capacitor carries the return; only the planes do. GND and P3V3MP are not a thin pair (at most 0.2 mm), so their own capacitance does not carry the return. To serve it, change layers across the same plane, or move the signal to a layer referenced to ground. None.
      • 'AJTAG.TCK' at (-161.63, 38), TOP -> L6: GND on G1 to P3V3MP on P6, GND on G3; partial, GND continues through a via of GND 2.28 mm away; stitching none, 223.4 mm of track past the via on L6: At the 5 GHz knee of a 100 ps edge (2 ns 10-90 %, assumed: no edge is on file for the net) no capacitor carries the return; only the planes do. GND and P3V3MP are not a thin pair (at most 0.2 mm), so their own capacitance does not carry the return. To serve it, change layers across the same plane, or move the signal to a layer referenced to ground. None.
      • 'AJTAG.TMS' at (-161.63, 36.05), TOP -> L6: GND on G1 to P3V3MP on P6, GND on G3; partial, GND continues through a via of GND 3.85 mm away; stitching none, 223.0 mm of track past the via on L6: At the 5 GHz knee of a 100 ps edge (2 ns 10-90 %, assumed: no edge is on file for the net) no capacitor carries the return; only the planes do. GND and P3V3MP are not a thin pair (at most 0.2 mm), so their own capacitance does not carry the return. To serve it, change layers across the same plane, or move the signal to a layer referenced to ground. None.
      • 'AJTAG.TDI' at (-157.79, 35.4), TOP -> L6: GND on G1 to P3V3MP on P6, GND on G3; partial, GND continues through a via of GND 2.63 mm away; stitching none, 222.6 mm of track past the via on L6: At the 5 GHz knee of a 100 ps edge (2 ns 10-90 %, assumed: no edge is on file for the net) no capacitor carries the return; only the planes do. GND and P3V3MP are not a thin pair (at most 0.2 mm), so their own capacitance does not carry the return. To serve it, change layers across the same plane, or move the signal to a layer referenced to ground. None.
      • 'AJTAG.TDO' at (-157.79, 37.35), TOP -> L6: GND on G1 to P3V3MP on P6, GND on G3; partial, GND continues through a via of GND 0.68 mm away; stitching none, 217.4 mm of track past the via on L6: At the 5 GHz knee of a 100 ps edge (2 ns 10-90 %, assumed: no edge is on file for the net) no capacitor carries the return; only the planes do. GND and P3V3MP are not a thin pair (at most 0.2 mm), so their own capacitance does not carry the return. To serve it, change layers across the same plane, or move the signal to a layer referenced to ground. None.
      • 'TCLKC_P' at (-6.3, -6.5), TOP -> L3: GND on G1 to GND on G2, P3V3MP on P4; partial, GND continues through a via of GND 2.29 mm away; stitching none, 206.7 mm of track past the via on L3: At the 5 GHz knee of a 100 ps edge (1 ns 10-90 %, assumed: no edge is on file for the net) no capacitor carries the return; only the planes do. GND and P3V3MP are not a thin pair (at most 0.2 mm), so their own capacitance does not carry the return. To serve it, change layers across the same plane, or move the signal to a layer referenced to ground. None.
      • 'TCLKD_N' at (-7.99, -7.57), TOP -> L3: GND on G1 to GND on G2, P3V3MP on P4; partial, GND continues through a via of GND 2.43 mm away; stitching none, 194.9 mm of track past the via on L3: At the 5 GHz knee of a 100 ps edge (1 ns 10-90 %, assumed: no edge is on file for the net) no capacitor carries the return; only the planes do. GND and P3V3MP are not a thin pair (at most 0.2 mm), so their own capacitance does not carry the return. To serve it, change layers across the same plane, or move the signal to a layer referenced to ground. None.
      • 'TCLKD_P' at (-8, -8.15), TOP -> L3: GND on G1 to GND on G2, P3V3MP on P4; partial, GND continues through a via of GND 2.32 mm away; stitching none, 194.2 mm of track past the via on L3: At the 5 GHz knee of a 100 ps edge (1 ns 10-90 %, assumed: no edge is on file for the net) no capacitor carries the return; only the planes do. GND and P3V3MP are not a thin pair (at most 0.2 mm), so their own capacitance does not carry the return. To serve it, change layers across the same plane, or move the signal to a layer referenced to ground. None.
      • 'PROGRAM_B' at (-142.25, 101.28), TOP -> L5: GND on G1 to GND on GND, P3V3MP on P6; partial, GND continues through a via of GND 3.62 mm away; stitching none, 188.0 mm of track past the via on L5: At the 5 GHz knee of a 100 ps edge (100 ps 10-90 %, assumed: no edge is on file for the net) no capacitor carries the return; only the planes do. GND and P3V3MP are not a thin pair (at most 0.2 mm), so their own capacitance does not carry the return. To serve it, change layers across the same plane, or move the signal to a layer referenced to ground. None.
      • 'IPMB.SCL' at (-155.37, 100.33), TOP -> L6: GND on G1 to P3V3MP on P6, GND on G3; partial, GND continues through a via of GND 3.37 mm away; stitching none, 185.8 mm of track past the via on L6: At the 5 GHz knee of a 100 ps edge (10 ns 10-90 %, assumed: no edge is on file for the net) no capacitor carries the return; only the planes do. GND and P3V3MP are not a thin pair (at most 0.2 mm), so their own capacitance does not carry the return. To serve it, change layers across the same plane, or move the signal to a layer referenced to ground. None.
      • 'TCLKA_N' at (-130.61, -2.92), TOP -> L5: GND on G1 to GND on GND, P3V3MP on P6; partial, GND continues through a via of GND 1.64 mm away; stitching none, 175.7 mm of track past the via on L5: At the 5 GHz knee of a 100 ps edge (1 ns 10-90 %, assumed: no edge is on file for the net) no capacitor carries the return; only the planes do. GND and P3V3MP are not a thin pair (at most 0.2 mm), so their own capacitance does not carry the return. To serve it, change layers across the same plane, or move the signal to a layer referenced to ground. None.
      • 'TCLKA_N' at (-128.48, -15.15), TOP -> L5: GND on G1 to GND on GND, P3V3MP on P6; partial, GND continues through a via of GND 1.14 mm away; stitching none, 175.7 mm of track past the via on L5: At the 5 GHz knee of a 100 ps edge (1 ns 10-90 %, assumed: no edge is on file for the net) no capacitor carries the return; only the planes do. GND and P3V3MP are not a thin pair (at most 0.2 mm), so their own capacitance does not carry the return. To serve it, change layers across the same plane, or move the signal to a layer referenced to ground. None.
      • and 359 more

      Every site is a row of this finding's evidence, keyed by net, layers and position. GET /api/odb/boards/58/reference-planes?group=refchange%3AGND%7CP3V3MP serves each with its own layer_image region; all of them lie in x -170.07..-4.3, y -33.24..104.95 mm (board frame, Y up).

      The reference on each layer is the first plane the track meets leaving the via, anti-pads and the net's own copper skipped. Whether it matters depends on each signal's edge rate, which nothing on file states for these nets: a slow static line crosses a split harmlessly, a fast edge does not.

    • split@P6:P0V95|P1V55 nets cross the split between the P0V95 and P1V5 planes on P6 at 5 places, all with a plane on the track's other side (fastest: INIT_B at 100 ps 10-90 %, ASSUMED)

      On P6 the P0V95 and P1V5 planes part, and 5 nets cross from one to the other at 5 place(s). The return current cannot follow a track across; it detours to wherever the two planes are joined (a stitching capacitor, or the source), and the loop it opens radiates and couples. One decision covers the boundary: whether it is stitched near these crossings, or the nets crossing it are slow enough not to care. Tracks that run ALONG this split, down the slot rather than across it, are a finding of their own: theirs is an impedance step, not a detour. At 5 of the 5 crossing(s) the track's other neighbour carries a plane, which the return can use there. Every crossing has a plane on the track's other side at most twice as far away as the broken one, so at most 51% of the return meets the break and the rest follows the track on that plane (to first order each plane beside a stripline carries a share of the return inversely proportional to its distance from the track). The finding is the lower tier (info).

      • 'GTP0-3_CLK_OUT_N' on L6 at (-93.46, 109.69), 0.41 mm; other side: GND on G3, 0.1 mm from the track against 0.104 mm to P6: 49% of the return on P6
      • 'GTP0-3_CLK_OUT_P' on L6 at (-93.45, 109.48), 0.4 mm; other side: GND on G3, 0.1 mm from the track against 0.104 mm to P6: 49% of the return on P6
      • 'INIT_B' on L5 at (-93.45, 109.3), 0.4 mm; other side: GND on GND, 0.103 mm from the track against 0.1 mm to P6: 51% of the return on P6
      • 'PM.1V5_VTT_EN' on L5 at (-93.45, 103.53), 0.57 mm; other side: GND on GND, 0.103 mm from the track against 0.1 mm to P6: 51% of the return on P6
      • 'PROGRAM_B' on L5 at (-93.45, 108.33), 0.4 mm; other side: GND on GND, 0.103 mm from the track against 0.1 mm to P6: 51% of the return on P6

      Every site is a row of this finding's evidence, keyed by net, layers and position. GET /api/odb/boards/58/reference-planes?group=split%40P6%3AP0V95%7CP1V5 serves each with its own layer_image region; all of them lie in x -94.66..-92.25, y 102.33..110.69 mm (board frame, Y up).

      Read from the copper as drawn on the reference layer, along each track's centreline at exact breakpoints. The reference is the copper beside the track, never the layer's type or name. Whether it matters depends on each signal's edge rate, which nothing on file states for these nets: a slow static line crosses a split harmlessly, a fast edge does not.

    • split@P6:P1V5|P1V87 nets cross the split between the P1V5 and P1V8 planes on P6 at 9 places, all with a plane on the track's other side (fastest: PROGRAM_B at 100 ps 10-90 %, ASSUMED)

      On P6 the P1V5 and P1V8 planes part, and 7 nets cross from one to the other at 9 place(s). The return current cannot follow a track across; it detours to wherever the two planes are joined (a stitching capacitor, or the source), and the loop it opens radiates and couples. One decision covers the boundary: whether it is stitched near these crossings, or the nets crossing it are slow enough not to care. Tracks that run ALONG this split, down the slot rather than across it, are a finding of their own: theirs is an impedance step, not a detour. At 9 of the 9 crossing(s) the track's other neighbour carries a plane, which the return can use there. Every crossing has a plane on the track's other side at most twice as far away as the broken one, so at most 51% of the return meets the break and the rest follows the track on that plane (to first order each plane beside a stripline carries a share of the return inversely proportional to its distance from the track). The finding is the lower tier (info).

      • 'FJTAG.TDI' on L5 at (-104.73, 97.08), 0.4 mm; other side: GND on GND, 0.103 mm from the track against 0.1 mm to P6: 51% of the return on P6
      • 'PM.1V5_VTT_EN' on L5 at (-104.73, 99.33), 0.4 mm; other side: GND on GND, 0.103 mm from the track against 0.1 mm to P6: 51% of the return on P6
      • 'PM.1V5_VTT_EN' on L5 at (-94.1, 102.23), 0.87 mm; other side: GND on GND, 0.103 mm from the track against 0.1 mm to P6: 51% of the return on P6
      • 'PM.EN_P1V5' on L6 at (-104.73, 102.65), 0.4 mm; other side: GND on G3, 0.1 mm from the track against 0.104 mm to P6: 49% of the return on P6
      • 'PM.PGOOD' on L6 at (-97.1, 101.95), 0.57 mm; other side: GND on G3, 0.1 mm from the track against 0.104 mm to P6: 49% of the return on P6
      • 'PROGRAM_B' on L5 at (-104.73, 99.75), 0.4 mm; other side: GND on GND, 0.103 mm from the track against 0.1 mm to P6: 51% of the return on P6
      • 'PROGRAM_B' on L5 at (-99.45, 101.95), 0.4 mm; other side: GND on GND, 0.103 mm from the track against 0.1 mm to P6: 51% of the return on P6
      • 'QSPI0_IO1' on L5 at (-104.73, 99.11), 0.4 mm; other side: GND on GND, 0.103 mm from the track against 0.1 mm to P6: 51% of the return on P6
      • 'QSPI0_IO2' on L5 at (-104.73, 98.3), 0.4 mm; other side: GND on GND, 0.103 mm from the track against 0.1 mm to P6: 51% of the return on P6

      Every site is a row of this finding's evidence, keyed by net, layers and position. GET /api/odb/boards/58/reference-planes?group=split%40P6%3AP1V5%7CP1V8 serves each with its own layer_image region; all of them lie in x -105.93..-93.1, y 96.08..103.65 mm (board frame, Y up).

      Read from the copper as drawn on the reference layer, along each track's centreline at exact breakpoints. The reference is the copper beside the track, never the layer's type or name. Whether it matters depends on each signal's edge rate, which nothing on file states for these nets: a slow static line crosses a split harmlessly, a fast edge does not.

Main parts

Part numberMakerDescriptionWhere
2998KEYSTONEBattery Holder whit 2 PinBH2
39-28-1043MOLEXConnector 4 MaleJ4
511BBA125M000BAGRSILICON LABSCrystal Oscillator 6 pins, Si511 (LVDS/LVPECL/HCSL/Dual CMOS)OSC3
5120913-1TYCO ELECTRONICSSingle Pin FemaleJ12
7A-8.000MAAE-TTXC CORPORATIONQuartz CrystalX1
7M-12.000MAAJ-TTXC CORPORATIONCrystal 4 pins 1=XTAL Control 2=GND3=XTAL 4=GNDX2
87832-1420MOLEXConnector 14 MaleJ6
8V54816ANLGINTEGRATED DEVICE TECHNOLOGY (IDT)16-Port, Bi-directional M-LVDS Clock Cross-Point SwitchIC73
ABS25-32.768KHZ-TABRACONCrystal 4 pins 1=XTAL Control 2 And 3=NC 4=XTALX3
AT24C64D-SSHM-TATMEL64K (8192 x 8), I2C-Compatible (2-Wire) Serial EEPROMIC71
AT24MAC602-XHMATMEL1.7V to 5.5V, I2C-Compatible (2-wire) 2-Kbit (256 x 8) Serial EEPROMIC27
BC847BVNXP SEMICONDUCTORS45V, 100mA NPN General Purpose Double TransistorT1
FT4232H-56QFUTURE TECHNOLOGY DEVICES INTERNATIONAL (FTDI)Quad High Speed USB to Multipurpose UART/MPSSE ICIC21
FTSH-110-01-F-DV-KSAMTECConnector 20 MaleJ9
HDT0001C&KPush-Button Switch SPDTPB1
HSME-C120AVAGO TECHNOLOGIESLED Green Avago C120LED2
HSMR-C120AVAGO TECHNOLOGIESLED Blue Avago C120LED1
IRF9310TRPBFINTERNATIONAL RECTIFIER-30V -20A P-Channel HEXFET® Power MOSFETQ18
LD1117DT33CTRST MICROELECTRONICS3-Terminal Regulator 1=GND 2=IN 3=OUTIC74
LMZ31520RLGTEXAS INSTRUMENTS20A Power Module with 3V-14.5V Input in QFN PackageIC16
LPC1768FBD100NXP SEMICONDUCTORSLPC1768 Series 64kB 100MHz 32-Bit ARM Cortex-M3 MicrocontrollerIC1
M93C46-WMN6TPST MICROELECTRONICS2.5 to 5.5 V, 1Kb (128x8) Serial Microwire Bus EEPROMIC60
MAX6642ATT90+TMAXIM±1°C, SMBus-Compatible Remote/Local Temperature Sensor with Overtemperature AlarmIC18
MCP23016-I/SOMICROCHIP TECHNOLOGY16-bit I/O I2C I/O ExpanderIC66
MCP79410-I/SNMICROCHIP TECHNOLOGYI2C Real-Time Clock/Calendar with EEPROM, SRAM, Unique ID and Battery SwitchoverIC9
MF-MSMF250/16X-2BOURNSFuseF3
MIC5219-3.3YM5MICROCHIP TECHNOLOGY500mA Peak Output LDO RegulatorIC19
MT25QL256ABA8ESF-0AATMICRON TECHNOLOGY256Mbit (32MB), 2.7V-3.6V, Multiple I/O Serial NOR Flash Memory,with Reset# & Hold#IC65
OVSRRGBCC3OPTEK / TT ELECTRONICSRight Angle Tricolor (RGB) Surface Mount LED, Body 3.2x1mmLED3
PRTR5V0U2XNXP SEMICONDUCTORSUltra Low Capacitance Double Rail-to-Rail ESD Protection DiodeD26
Si2315BDS-T1-E3VISHAY SILICONIXP-CHANNEL MOSFET WITH DIODEQ17
SK26ATAIWAN SEMICONDUCTORSchottky DiodeD7
SKHHLMA010ALPSPush-Button Switch SPDTPB2
SKV585811-CUWAKEFIELD-VETTEHeat Sink Without PadHS1
SN65MLVD082DGGTEXAS INSTRUMENTS8-Channel Half-Duplex M-LVDS Line TransceiverIC59
SN74AUP1T87DCKRTexas InstrumentsLow Power, 1.8/2.5/3.3-V In, 3.3-V CMOS Out, 2-Input Positive Exclusive-NOR Gate, DCK0005A, LARGE T&RIC45
SN74CB3T3257PWTEXAS INSTRUMENTS4-BIT 1-OF-2 FET Multiplexer/Demultiplexer 2.5V/3.3V Low-Voltage Bus Switch 5V Tolerant ShifterIC46
SN74LV541APWRTEXAS INSTRUMENTSOctal Buffer/Line Driver with 3-State OutputsIC47
STM811SW16FST MICROELECTRONICSMicroprocessor Reset CircuitIC58
TL431ACDBZTTEXAS INSTRUMENTSPrecision Reference 1=Cathode 2=Adj 3=AnodeIC54

About this design

Design
AFCKU — OHWR0057 R1
Designed in
Altium Designer
Board
14 copper layers, 180.6 × 148.5 mm, 1568 components, 1472 nets
Licence
CERN-OHL-S-2.0
Source
https://gitlab.com/ohwr/project/afcku at commit b44b584723ce
Attribution
CERN Open Hardware Repository, project ohwr/project/afcku (commit b44b584723cedf02f06646c2e3646280a440b17f, 2025-04-09). Licence: CERN-OHL-S-2.0. Source: https://gitlab.com/ohwr/project/afcku. This is a snapshot of the copy published on the OHWR GitLab pages (R1), not CERN's released record. CERN's source of truth is EDMS (https://edms.cern.ch), and the current release may be newer.
Reviewed with
Probe by E# (E-Sharp AB)