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

CAJU MODULE

OHWR0001 · revision R1

CAJU MODULE (OHWR0001 R1) is an open-hardware circuit board from the CERN Open Hardware Repository, designed in Altium Designer: a 6-layer board of 36 × 60 mm with 103 components and 142 nets, published under CERN-OHL-W-2.0.

review 90% walked · 8 findings: 4 major · 4 minor (13 items)

Open in the interactive viewerRead the full review

CAJU MODULE (OHWR0001 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 8 findings: 4 major · 4 minor (13 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. observationOBS-72majorOpen

    USB power through D2 back-drives the 5 V LDO U1 output with its input unpowered (VOUT > VIN + 0.3 V abs max)

    When the module is powered from the USB-C connector alone, the 5V rail sits at about 4.7 V (VBUS through Schottky D2) directly on the output of the LDO U1 (LDL1117 5.0 V), while U1's input VIN has no source on this board. That breaks the LDL1117 absolute-maximum rating VOUT ≤ VIN + 0.3 V, and whatever reverse path the pass device has will lift VIN, and with it the carrier's VIN pin on CN1 pin 4.

    What sits between what

    ``mermaid flowchart LR VBUS["CON1 VBUS (A4/B9, A9/B4)"] -->|"D2 CUS08F30 (A→K)"| V5["5V rail"] VIN["VIN (CN1 pin 4, C1, C2 only)"] -->|"U1 pin 3 IN → pins 2/4 OUT"| V5 V5 --> U2["U2 LDL1117 3.3 V"] V5 --> U3["U3 FT230XS VCC"] V5 --> CN2["CN2 pin 3"] ``

    VIN connects only to C1, C2, CN1 pin 4 and U1 pin 3 (net dossier). Nothing on the module drives it when the carrier is absent or unpowered.

    Numbers

    QuantityLimit (datasheet)Actual, USB onlyMargin
    U1 VOUT≤ VIN + 0.3 V≈ 5.0 V − 0.28 V (D2 VF typ at 100 mA) ≈ 4.7 V with VIN undriven (≈ 0 V before any reverse leakage charges it)about −4.4 V (violated)

    Why it matters

    • The LDL1117 datasheet documents no reverse-current or back-feed behaviour (no "reverse" appears in it), so the part is outside its specified conditions in this mode.
    • Any reverse conduction through U1 also puts a few volts on CN1 pin 4, feeding the carrier's VIN net from USB.

    Possible remedies (for the designer)

    • Put a series Schottky or an ideal-diode between U1's output and the 5V rail, so the rail can be OR-ed from D2 and U1 without back-driving U1.
    • Or document that VIN must always be present whenever USB is connected.
  2. DecapViaFanoutmajorOpen

    DecapViaFanout: 2 components×2

    2 checked by the review by Claude walker W4 (Cern session) — 2 different reasons

    At C5, C6

    Each finding, with its own headline
    • C5C5 on '3V3': its rail pad reaches its via over 2.63 mm of track, adding about 0.85 nH to its loop, past Probe's 0.5 nH starting value; put the vias at the pads

      C5 on '3V3': its rail pad reaches its via over 2.63 mm of track, adding about 0.85 nH to its loop, past Probe's 0.5 nH starting value; put the vias at the pads. C5.1 (rail) runs 2.63 mm along its copper to its nearest via, 1.63 mm past the 1 mm beside the pad: 0.85 nH; C5.2 (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 1.73 mm apart (the pad pitch), 0.2 l ln(D/R) nH with R 0.127 mm (the track's width is not read): 0.85 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 1.06 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 0.76 mm depth, ASSUMED). 0.5 nH is Probe's starting value, about the whole mounted inductance of a well-mounted 0402.

    • C6C6 on '3V3': its rail pad reaches its via over 4.05 mm of track, adding about 1.19 nH to its loop, past Probe's 0.5 nH starting value; put the vias at the pads

      C6 on '3V3': its rail pad reaches its via over 4.05 mm of track, adding about 1.19 nH to its loop, past Probe's 0.5 nH starting value; put the vias at the pads. C6.1 (rail) runs 4.05 mm along its copper to its nearest via, 3.05 mm past the 1 mm beside the pad: 1.19 nH; C6.2 (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 0.90 mm apart (the pad pitch), 0.2 l ln(D/R) nH with R 0.127 mm (the track's width is not read): 1.19 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 0.82 nH (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 0.76 mm depth, ASSUMED). 0.5 nH is Probe's starting value, about the whole mounted inductance of a well-mounted 0402.

  3. FiducialMissingmajorOpen

    FiducialMissing: 2 boards×2

    2 checked by the review by evaluation@esharp.se — 2 different reasons

    At bottom, top

    Each finding, with its own headline
    • bottomThe bottom side carries 82 SMT component(s) and no fiducial was found on it

      The bottom side carries 82 SMT component(s) but no fiducial was found on it — a placement machine has no optical reference to align to. Probe found no pad labelled .pad_usage=g_fiducial/l_fiducial (basis pad_usage); no single-pad, netless, undrilled component whose 0.5-3 mm round pad has a concentric solder-mask opening at least 1.5x its diameter with no other copper inside it (basis geometry); and no component whose refdes starts FID or whose part or package name contains FIDUCIAL (basis footprint-name). If your library names fiducials something Probe does not recognise, mark this benign — disposition_finding(rule='FiducialMissing', entity='bottom', status='benign', reason=…) — the name match only ever silences this check, it never asserts anything. If the side really has no fiducial, say so with status='confirmed': that answers the finding and keeps it gating release.

    • topThe top side carries 19 SMT component(s) and no fiducial was found on it

      The top side carries 19 SMT component(s) but no fiducial was found on it — a placement machine has no optical reference to align to. Probe found no pad labelled .pad_usage=g_fiducial/l_fiducial (basis pad_usage); no single-pad, netless, undrilled component whose 0.5-3 mm round pad has a concentric solder-mask opening at least 1.5x its diameter with no other copper inside it (basis geometry); and no component whose refdes starts FID or whose part or package name contains FIDUCIAL (basis footprint-name). If your library names fiducials something Probe does not recognise, mark this benign — disposition_finding(rule='FiducialMissing', entity='top', status='benign', reason=…) — the name match only ever silences this check, it never asserts anything. If the side really has no fiducial, say so with status='confirmed': that answers the finding and keeps it gating release.

  4. ReferencePlaneDiscontinuitymajorOpen

    ReferencePlaneDiscontinuity: 4 boards×4

    4 checked by the review by Claude walker W4 (Cern session) — 4 different reasons

    At merged:19.02,18.98|19.99,18.98, void:3V3@POWER:hole 2.7,7.48|17.79,56.43, void:3V3@POWER:hole 22.58,20.3|27.43,35.03, void:3V3@POWER:hole 9.32,41.48|30.46,56.26

    Each finding, with its own headline
    • merged:19.02,18.98|19.99,18.98'PWM3' crosses between the drills at (19.02, 18.98) and (19.98, 18.98), whose anti-pads run together in the 3V3 plane on POWER (fastest: PWM3 at 100 ps 10-90 %, ASSUMED)

      Along 'PWM3' on LAYER_LS_2, the 3V3 plane on POWER is cleared round the drills at (19.02, 18.98) and (19.98, 18.98), and the two clearances run together: the straight line from one drill to the other is bare all the way, and the track crosses it, over 0.17 mm of bare stretch at (19.5, 18.98). Past one drill's anti-pad the return flows round on either side; here it cannot pass between the drills and detours round the whole merged clearance - the slot a row of pins or vias cuts in a plane. 0 of the 1 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.

      • 'PWM3' on LAYER_LS_2 at (19.5, 18.98), 0.17 mm between the drills at (19.02, 18.98) and (19.98, 18.98) in POWER; other side: GND on GND_2, 0.55 mm from the track against 0.109 mm to POWER: 83% of the return on POWER

      Every site is a row of this finding's evidence, keyed by net, layers and position. GET /api/odb/boards/4/reference-planes?group=merged%3A19.02%2C18.98%7C19.99%2C18.98 serves each with its own layer_image region; all of them lie in x 18.5..20.5, y 17.9..20.07 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:3V3@POWER:hole 2.7,7.48|17.79,56.4314 nets cross one void in the GND plane on LAYER_LS_2 and the 3V3 plane on POWER at 7 places, the longest 12.78 mm; 17 over tracks of other nets routed inside it (fastest: ADC3 at 100 ps 10-90 %, ASSUMED)

      Along 'RESET' on POWER, the GND plane on LAYER_LS_2 has 12.78 mm under the track with copper of BOOT0, I2C4_SDA, I2C4_SCL_QSPI_CS, AGND, not a plane, at (5.2, 32.81), the same plane on both sides. 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. 17 of the crossings pass over tracks of other nets routed inside the hole on the reference layer: the plane is absent there, and those tracks stand in it. 3 of them run ALONG the bare copper, parallel to its edges, rather than across it (the longest 5.24 mm). 7 places on 14 net(s) cross it, listed below, the longest first. What cleared it: the hole at x 2.7..17.79, y 7.48..56.43. One decision covers the void: whether these tracks may cross it, or the plane should be filled under them. 0 of the 17 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.

      • one crossing of two tracks, seen from each: 'BOOT0' on LAYER_LS_2 at (5.02, 35.08), 0.56 mm over copper of RESET, not a plane; other side: GND on GND_2, 0.55 mm from the track against 0.109 mm to POWER: 83% of the return on POWER / 'I2C4_SCL_QSPI_CS' on LAYER_LS_2 at (10.2, 29.56), 2.96 mm, running ALONG the bare copper for 2.31 mm, not across it over copper of RESET, not a plane; other side: GND on GND_2, 0.55 mm from the track against 0.109 mm to POWER: 83% of the return on POWER / 'I2C4_SDA' on LAYER_LS_2 at (6.64, 30), 3.36 mm, running ALONG the bare copper for 1.81 mm, not across it over copper of RESET, not a plane; other side: GND on GND_2, 0.55 mm from the track against 0.109 mm to POWER: 83% of the return on POWER / 'RESET' on POWER at (5.2, 32.81), 12.78 mm, running ALONG the bare copper for 5.24 mm, not across it over copper of BOOT0, I2C4_SDA, I2C4_SCL_QSPI_CS, AGND, not a plane; other side: GND on GND_1, 0.55 mm from the track against 0.109 mm to LAYER_LS_2: 83% of the return on LAYER_LS_2
      • one crossing of two tracks, seen from each: 'ADC4_DAC1' on LAYER_LS_2 at (11.88, 11.62), 0.56 mm over copper of RESET, not a plane; other side: GND on GND_2, 0.55 mm from the track against 0.109 mm to POWER: 83% of the return on POWER / 'ADC6' on LAYER_LS_2 at (10.51, 11.98), 0.69 mm over copper of RESET, not a plane; other side: GND on GND_2, 0.55 mm from the track against 0.109 mm to POWER: 83% of the return on POWER / 'ADC8' on LAYER_LS_2 at (14.13, 11.62), 0.56 mm over copper of RESET, not a plane; other side: GND on GND_2, 0.55 mm from the track against 0.109 mm to POWER: 83% of the return on POWER / 'GPIO6' on LAYER_LS_2 at (12.5, 11.62), 0.56 mm over copper of RESET, not a plane; other side: GND on GND_2, 0.55 mm from the track against 0.109 mm to POWER: 83% of the return on POWER / 'GPIO7' on LAYER_LS_2 at (13, 11.62), 0.56 mm over copper of RESET, not a plane; other side: GND on GND_2, 0.55 mm from the track against 0.109 mm to POWER: 83% of the return on POWER / 'GPIO8' on LAYER_LS_2 at (13.41, 11.62), 0.56 mm over copper of RESET, not a plane; other side: GND on GND_2, 0.55 mm from the track against 0.109 mm to POWER: 83% of the return on POWER / 'RESET' on POWER at (11.12, 11.8), 5.92 mm over copper of ADC6, ADC4_DAC1, GPIO6, GPIO7, GPIO8, ADC8, ADC7, not a plane; other side: GND on GND_1, 0.55 mm from the track against 0.109 mm to LAYER_LS_2: 83% of the return on LAYER_LS_2
      • one crossing of two tracks, seen from each: 'ADC3' on LAYER_LS_2 at (9.88, 14.08), 0.79 mm over copper of RESET, not a plane; other side: GND on GND_2, 0.55 mm from the track against 0.109 mm to POWER: 83% of the return on POWER / 'RESET' on POWER at (9.7, 14.15), 0.79 mm over copper of ADC3, not a plane; other side: GND on GND_1, 0.55 mm from the track against 0.109 mm to LAYER_LS_2: 83% of the return on LAYER_LS_2
      • 'ADC7' on LAYER_LS_2 at (15.4, 11.14), 0.79 mm over copper of RESET, not a plane; other side: GND on GND_2, 0.55 mm from the track against 0.109 mm to POWER: 83% of the return on POWER
      • 'AGND' on LAYER_LS_2 at (12.62, 29.36), 0.74 mm over copper of RESET, not a plane; other side: GND on GND_2, 0.55 mm from the track against 0.109 mm to POWER: 83% of the return on POWER
      • 'BOOT0' on LAYER_LS_2 at (7.82, 35.08), 0.56 mm over copper of VBAT, not a plane; other side: GND on GND_2, 0.55 mm from the track against 0.109 mm to POWER: 83% of the return on POWER
      • 'I2C1_SDA_UART4_TX' on LAYER_LS_2 at (7.82, 33.86), 0.56 mm over copper of VBAT, not a plane; other side: GND on GND_2, 0.55 mm from the track against 0.109 mm to POWER: 83% of the return on POWER

      Every site is a row of this finding's evidence, keyed by net, layers and position. GET /api/odb/boards/4/reference-planes?group=void%3A3V3%40POWER%3Ahole%202.7%2C7.48%7C17.79%2C56.43 serves each with its own layer_image region; all of them lie in x 3.92..16.68, y 9.99..36.36 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.

    • void:3V3@POWER:hole 22.58,20.3|27.43,35.034 nets cross one void in the GND plane on LAYER_LS_2 and the 3V3 plane on POWER at 2 places, the longest 3.69 mm; 5 over tracks of other nets routed inside it (fastest: NetC30_1 at 100 ps 10-90 %, ASSUMED)

      Along 'NetC30_1' on POWER, the GND plane on LAYER_LS_2 has 3.69 mm under the track with copper of CAN_TX, CAN_RX, not a plane, at (26, 32.77), 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. 5 of the crossings pass over tracks of other nets routed inside the hole on the reference layer: the plane is absent there, and those tracks stand in it. 2 of them run ALONG the bare copper, parallel to its edges, rather than across it (the longest 2.42 mm). 2 places on 4 net(s) cross it, listed below, the longest first. What cleared it: the hole at x 22.58..27.43, y 20.3..35.03. One decision covers the void: whether these tracks may cross it, or the plane should be filled under them. 0 of the 5 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.

      • one crossing of two tracks, seen from each: 'CAN_RX' on LAYER_LS_2 at (26.41, 34.36), 0.56 mm over copper of NetC30_1, not a plane; other side: GND on GND_2, 0.55 mm from the track against 0.109 mm to POWER: 83% of the return on POWER / 'CAN_TX' on LAYER_LS_2 at (26.1, 32.93), 3.03 mm, running ALONG the bare copper for 2.35 mm, not across it over copper of NetC30_1, not a plane; other side: GND on GND_2, 0.55 mm from the track against 0.109 mm to POWER: 83% of the return on POWER / 'NetC30_1' on POWER at (26, 32.77), 3.69 mm, running ALONG the bare copper for 2.42 mm, not across it over copper of CAN_TX, CAN_RX, not a plane, to the end of the track; other side: GND on GND_1, 0.55 mm from the track against 0.109 mm to LAYER_LS_2: 83% of the return on LAYER_LS_2
      • one crossing of two tracks, seen from each: 'GPIO9' on LAYER_LS_2 at (23.18, 22.51), 0.56 mm over copper of NetC30_1, not a plane; other side: GND on GND_2, 0.55 mm from the track against 0.109 mm to POWER: 83% of the return on POWER / 'NetC30_1' on POWER at (23, 21.69), 1.84 mm over copper of GPIO9, not a plane, to the end of the track; other side: GND on GND_1, 0.55 mm from the track against 0.109 mm to LAYER_LS_2: 83% of the return on LAYER_LS_2

      Every site is a row of this finding's evidence, keyed by net, layers and position. GET /api/odb/boards/4/reference-planes?group=void%3A3V3%40POWER%3Ahole%2022.58%2C20.3%7C27.43%2C35.03 serves each with its own layer_image region; all of them lie in x 21.72..28, y 19.72..35.68 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:3V3@POWER:hole 9.32,41.48|30.46,56.264 nets cross one void in the 3V3 plane on POWER at 8 places, the longest 3.94 mm; 8 over tracks of other nets routed inside it (fastest: AGND at 100 ps 10-90 %, ASSUMED)

      Along 'AGND' on LAYER_LS_2, the 3V3 plane on POWER has 3.94 mm under the track with copper of 5V, not a plane, at (19.21, 49.73), the same plane on both sides. 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. 8 of the crossings pass over tracks of other nets routed inside the hole on the reference layer: the plane is absent there, and those tracks stand in it. 3 of them run ALONG the bare copper, parallel to its edges, rather than across it (the longest 2.64 mm). 8 places on 4 net(s) cross it, listed below, the longest first. What cleared it: the hole at x 9.32..30.46, y 41.48..56.26. One decision covers the void: whether these tracks may cross it, or the plane should be filled under them. 0 of the 8 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.

      • 'AGND' on LAYER_LS_2 at (19.21, 49.73), 3.94 mm, running ALONG the bare copper for 2.64 mm, not across it over copper of 5V, not a plane; other side: GND on GND_2, 0.55 mm from the track against 0.109 mm to POWER: 83% of the return on POWER
      • 'GPIO3' on LAYER_LS_2 at (17.29, 43.59), 3.59 mm, running ALONG the bare copper for 2.26 mm, not across it over copper of NetC14_1, not a plane; other side: no plane on GND_2
      • 'VREF+' on LAYER_LS_2 at (22.99, 49.75), 2.58 mm, running ALONG the bare copper for 1.12 mm, not across it over copper of 5V, not a plane; other side: GND on GND_2, 0.55 mm from the track against 0.109 mm to POWER: 83% of the return on POWER
      • 'AVCC' on LAYER_LS_2 at (21.74, 49.74), 1.3 mm over copper of 5V, not a plane; other side: GND on GND_2, 0.55 mm from the track against 0.109 mm to POWER: 83% of the return on POWER
      • 'AVCC' on LAYER_LS_2 at (24.38, 52.2), 1.07 mm over copper of 5V, not a plane; other side: GND on GND_2, 0.55 mm from the track against 0.109 mm to POWER: 83% of the return on POWER
      • 'AGND' on LAYER_LS_2 at (24.3, 52.9), 0.92 mm over copper of 5V, not a plane; other side: GND on GND_2, 0.55 mm from the track against 0.109 mm to POWER: 83% of the return on POWER
      • 'AVCC' on LAYER_LS_2 at (17.94, 48.18), 0.73 mm over copper of NetC14_1, not a plane; other side: GND on GND_2, 0.55 mm from the track against 0.109 mm to POWER: 83% of the return on POWER
      • 'VREF+' on LAYER_LS_2 at (17.44, 47.54), 0.71 mm over copper of NetC14_1, not a plane; other side: GND on GND_2, 0.55 mm from the track against 0.109 mm to POWER: 83% of the return on POWER

      Every site is a row of this finding's evidence, keyed by net, layers and position. GET /api/odb/boards/4/reference-planes?group=void%3A3V3%40POWER%3Ahole%209.32%2C41.48%7C30.46%2C56.26 serves each with its own layer_image region; all of them lie in x 15.7..25.76, y 41.72..53.9 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.

  5. observationOBS-73minorOpen

    BOM part numbers for the LDOs U1 and U2 (LDL117S50R, LDL117S33R) are not ST order codes: the datasheet lists LDL1117S50R and LDL1117S33R

    The designer's BOM and the schematic give U1 as "STMicroelectronics / LDL117S50R" and U2 as "LDL117S33R". The LDL1117 datasheet's order-code table has no such codes. Its 5.0 V and 3.3 V codes are LDL1117S50R and LDL1117S33R (one more "1"). A buyer ordering the BOM text as written gets no match, and could substitute a different part.

    Evidence

    PositionBOM / schematic textDatasheet order code (Table 6)Output
    U1LDL117S50RLDL1117S50R5.0 V
    U2LDL117S33RLDL1117S33R3.3 V
    • Designer BOM: Bill of Materials [ALL] - CAJU MODULE_V1.1.xlsx, rows U1 and U2 ("SMD Integrated Circuit / Linear Voltage Regulator (LDO) 5V 1.2A / SOT-223", "... 3.3V 1.2A ...").
    • The schematic POWER SUPPLY sheet (page 3) prints the same short codes under U1 and U2.
    • The description (1.2 A, SOT-223, 5 V and 3.3 V) matches the LDL1117 family exactly, so the intent is clear and this is a typo, not a wrong part.

    Remedy

    Correct the manufacturer part numbers to LDL1117S50R and LDL1117S33R in the library and the BOM.

  6. observationOBS-75minorOpen

    No fiducials on either side of the board, and the fabrication package has no panel to carry them, although both sides hold 0.5 mm-pitch parts

    The CAJU module has no global or local fiducials on the top or bottom side. The released fabrication data contains no panel drawing that could supply them. Both sides carry fine-pitch parts: U5 (STM32H743, LQFP-100, 0.5 mm pitch) on top, and U4 (LAN8742A, QFN-24, 0.5 mm pitch) with U3 (SSOP-16) on the bottom. Pick-and-place alignment then depends on the assembler adding panel fiducials, or on board-edge or pad-pattern vision.

    What was checked

    SourceResult
    ODB++ export (both sides)no pad with fiducial usage, no FID component; the top and bottom copper/paste/silk renders show vias and pads only
    Pick-and-place file PICK AND PLACE_CAJU MODULE_V1.1.csvno fiducial entries
    Gerber set CAJU_MODULE_V1.1.zipa single 36 × 60 mm outline (GKO), no panel or rail layers
    Fine-pitch partsU5 LQFP-100 0.5 mm (top), U4 QFN-24 0.5 mm (bottom), U3 SSOP-16 0.65 mm (bottom)

    Why it matters

    With no fiducials, a double-sided build with 0.5 mm-pitch leads relies on the assembler's panelisation to supply global fiducials. Local fiducials for the QFP are absent either way. This is a yield and placement-accuracy risk, not a functional defect.

    Remedy

    Add two or three global fiducials per side, if space allows at the board corners, or release a panel drawing with rail fiducials. Consider a local fiducial pair at U5.

  7. ReferencePlaneDiscontinuityminorOpen

    2 nets cross the slot of 6 drills from (14.5, 36.08) to (16.46, 36.19), whose anti-pads run together in the 3V3 plane on POWER, between 2 pairs of drills (fastest: UART8_RX at 3 ns 10-90 %, suggested by its name)

    The 3V3 plane on POWER is cleared round a field of 6 drills from (14.5, 36.08) to (16.46, 36.19), and the clearances run together: the straight line from one drill to the next is bare all the way. 2 nets cross between drills of it at 2 place(s), each listed below with its two drills. Past one drill's anti-pad the return flows round on either side; here it cannot pass between the drills and detours round the whole merged clearance - the slot a row of pins or vias cuts in a plane. One decision covers the slot: whether these tracks may cross it, or the clearances should leave copper between the drills. 0 of the 2 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.

    • 'I2C4_SDA' on LAYER_LS_2 at (15.42, 35.16), 0.17 mm between the drills at (15, 35.4) and (15.9, 35.02) in POWER; other side: GND on GND_2, 0.55 mm from the track against 0.109 mm to POWER: 83% of the return on POWER
    • 'UART8_RX' on LAYER_LS_2 at (15.98, 36.21), 0.23 mm between the drills at (15.5, 36.27) and (16.45, 36.19) in POWER; other side: GND on GND_2, 0.55 mm from the track against 0.109 mm to POWER: 83% of the return on POWER

    Every site is a row of this finding's evidence, keyed by net, layers and position. GET /api/odb/boards/4/reference-planes?group=merged%3A14.5%2C36.08%7C16.46%2C36.19 serves each with its own layer_image region; all of them lie in x 14.41..16.98, y 34.14..37.35 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 merged:14.5,36.08|16.46,36.19

  8. ConnectorDomainMixedminorOpen

    CN2 takes 2 different supplies off the board on one connector: 5V (5.0 V), 3V3 (3.3 V)

    CN2 carries 2 different supplies out of the board on one connector: 5V (5.0 V) on 3, 3V3 (3.3 V) on 1/2. Declare the intent. Two rails on one connector is ordinary in a docking or power-distribution interface and a hazard in a signal one, and the difference is not visible from the netlist — but note that adjacent contacts at different voltages are exactly what a mirrored, reversed or off-by-one cable shorts together, so the cost of getting the mating wrong is a rail collision rather than a dead signal. If this is a standard connector, declare its family (connector_family_declare) and file the standard's contact map: supplies on the contacts it assigns to power are then not reported.

    At CN2

Main parts

Part numberMakerDescriptionWhere
B3U-1100PCH1
BD12520A03050580LBCN3
CUS08F30,H3FD2
DS04-254-1-01BK-SMTDIP1
ECS-.327-6-34S-TRECSX3
FT230XS-RU3
LAN8742A-CZU4
LDL117S33RU2
LDL117S50RU1
M95P32-EXMNT/EU6
STM32H743VIT6U5
TYPE-C-31-M-12KHECON1
2-826632-0TE ConnectivityCN1, CN2
ECS-250-12-33-AGN-TR3ECSX1, X2
CSL1001BT1CROHM SemiconductorLD1, LD2, LD3, LD4, LD5
BLM18AG601SN1DMurataL1, L2

About this design

Design
CAJU MODULE — OHWR0001 R1
Designed in
Altium Designer
Board
6 copper layers, 36 × 60 mm, 103 components, 142 nets
Licence
CERN-OHL-W-2.0
Source
https://gitlab.com/ohwr/project/caju-module at commit 63756fd00502
Attribution
CERN Open Hardware Repository, project ohwr/project/caju-module (commit 63756fd00502b8b2baaebd5a74230a2e7427d707, 2025-11-14). Licence: CERN-OHL-W-2.0. Source: https://gitlab.com/ohwr/project/caju-module. 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)