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

DIOT CRATE CPCI-SERIAL BACKPLANE

OHWR0084 · revision R3

DIOT CRATE CPCI-SERIAL BACKPLANE (OHWR0084 R3) is an open-hardware circuit board from its public repository, designed in Altium Designer: a 14-layer board of 304.8 × 128.5 mm with 141 components and 1225 nets, published under CERN-OHL-W-2.0.

review complete · 2 items need the designer's input · 5 findings: 2 major · 3 minor (16 items)

Open in the interactive viewerRead the full review

DIOT CRATE CPCI-SERIAL BACKPLANE (OHWR0084 R3) circuit board, top view: copper, silkscreen and components
Top view, rendered by Probe from the design's own fabrication data.

Design review

Probe recorded 5 findings: 2 major · 3 minor (16 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. ReferencePlaneDiscontinuitymajorOpen

    3 nets run ALONG the split between the GND and P12V planes on TOP_LAYER at 3 places, the longest 2.51 mm; the impedance rises from 65 ohm to 72 ohm there at worst, 4.8 % of each edge reflected (fastest: 7_DIFF_0_P at 100 ps 10-90 %, ASSUMED)

    On TOP_LAYER the GND and P12V 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 65 ohm; along the slot 66 ohm to 72 ohm, reflecting up to 4.8 % 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 80 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.

    • 'I2C_SDA' on L2 at (108.07, 112.02), 1.3 mm, running ALONG the split for 1.31 mm, not across it, 65 ohm over solid plane, 72 ohm along it: 4.8 % of each edge reflected; other side: GND on L3GND, 0.2 mm from the track against 0.18 mm to TOP_LAYER: 53% of the return on TOP_LAYER
    • 'F_RST' on L2 at (290.08, 133.9), 1.61 mm, running ALONG the split for 1.19 mm, not across it, 65 ohm over solid plane, 66 ohm along it: 0.9 % of each edge reflected; other side: GND on L3GND, 0.2 mm from the track against 0.18 mm to TOP_LAYER: 53% of the return on TOP_LAYER
    • '7_DIFF_0_P' on L2 at (199.46, 111.94), 2.51 mm, running ALONG the split for 2.51 mm, not across it, 65 ohm over solid plane, 66 ohm along it: 0.8 % of each edge reflected; other side: GND on L3GND, 0.2 mm from the track against 0.18 mm to TOP_LAYER: 53% of the return on TOP_LAYER

    Every site is a row of this finding's evidence, keyed by net, layers and position. GET /api/odb/boards/88/reference-planes?group=along%40TOP_LAYER%3AGND%7CP12V serves each with its own layer_image region; all of them lie in x 106.42..291.92, y 110.94..135.04 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 along@TOP_LAYER:GND|P12V

  2. LumpedLengthmajorOpen

    LumpedLength: 2 nets×2

    Checked by the review, stands · Claude walker D (Cern session) — Checked: I2C_SCL/I2C_SDA run from OHWR0083 R5 DIOT SB ZU IC6 (PCA9548A I2C switch) through system-slot J1.B2/C2 to every peripheral slot P1 (J11_Slot1-8.B2/C2), about 3.2 ns one way. The model drives the line through an ASSUMED 25 ohm push-pull source, but I2C is open-drain: the rising edge is set by the pull-ups and bus capacitance (an RC edge far slower than 10 ns), and only the falling edge is driven; the I2C specification (NXP UM10204) also gives receivers 50 ns spike suppression in Fast-mode. No pull-up value per slot or IBIS model is on file, so the reflection cannot be judged here either way. Left open for the owner; the cross-seam level and pull-up check belongs to the crate interconnect review.

    At I2C_SCL, I2C_SDA

    Each finding, with its own headline
    • I2C_SCLI2C_SCL: a transmission line at its edge (3168 ps against Tr/4 = 2500 ps), with no termination in range

      I2C_SCL: a transmission line at its edge (3168 ps against Tr/4 = 2500 ps), with no termination in range. 3168 ps from OHWR0083 R5:IC6.18 to J11_Slot8.B2 against Tr/4 = 2500 ps (10 ns 10-90 %, group); followed across J1.B2 into OHWR0083 R5 J14.B2 (CPCIS_I2C.SCL, mating confirmed); the connector's own delay is not on file and is not counted. Reflection: 41 ohm line (OHWR0083 R5:L8, 0.15 mm, the run that is most of the path), 3168 ps one way as 9 run(s) in cascade (solved as 7 sections, runs shorter than the time step folded into their neighbours) with 4 via(s) as lumped L and C, driven through 25 ohm ASSUMED (no IBIS output resistance on file) into 5 pF (5 pF per receiver ASSUMED, 1 receiver(s)). Rising, the far load peaks at 3.72 V (13 % over), falls back to 3.19 V, settles at 3.30 V; no VIH on file to judge it against. Falling, it dips to -0.42 V (0.42 V below ground), rings back to 0.11 V, settles at -0.00 V; no VIL on file to judge it against. Connector contacts (their inductance and capacitance) are not modelled, nor the line's losses, nor coupling to its neighbours. Termination: unterminated: about 14 ohm in series at the driver (0.9 (Z0 - Rout)), or 45-47 ohm at the far load (1.1-1.15 Z0), would absorb the return. Tr/4, the 10-200 ohm range and the placement within a lumped length are Probe's starting values. Every path, stub and input is in GET lumped?net=I2C_SCL.

    • I2C_SDAI2C_SDA: a transmission line at its edge (3232 ps against Tr/4 = 2500 ps), with no termination in range

      I2C_SDA: a transmission line at its edge (3232 ps against Tr/4 = 2500 ps), with no termination in range. 3232 ps from OHWR0083 R5:IC6.17 to J11_Slot8.C2 against Tr/4 = 2500 ps (10 ns 10-90 %, group); followed across J1.C2 into OHWR0083 R5 J14.C2 (CPCIS_I2C.SDA, mating confirmed); the connector's own delay is not on file and is not counted. Reflection: 41 ohm line (OHWR0083 R5:L8, 0.15 mm, the run that is most of the path), 3232 ps one way as 13 run(s) in cascade (solved as 10 sections, runs shorter than the time step folded into their neighbours) with 4 via(s) as lumped L and C, driven through 25 ohm ASSUMED (no IBIS output resistance on file) into 5 pF (5 pF per receiver ASSUMED, 1 receiver(s)). Rising, the far load peaks at 3.73 V (13 % over), falls back to 3.18 V, settles at 3.30 V; no VIH on file to judge it against. Falling, it dips to -0.43 V (0.43 V below ground), rings back to 0.12 V, settles at -0.00 V; no VIL on file to judge it against. Connector contacts (their inductance and capacitance) are not modelled, nor the line's losses, nor coupling to its neighbours. Termination: unterminated: about 14 ohm in series at the driver (0.9 (Z0 - Rout)), or 45-47 ohm at the far load (1.1-1.15 Z0), would absorb the return. Tr/4, the 10-200 ohm range and the placement within a lumped length are Probe's starting values. Every path, stub and input is in GET lumped?net=I2C_SDA.

  3. observationOBS-626minorOpen

    PSU presence lines renamed P_PRES1/P_PRES2 on the main backplane (crate-hw #82), but the spec, Front Interface and Power Backplane still say P_PRES0/P_PRES1

    The main backplane names the two PSU presence lines P_PRES1 / P_PRES2, as CERN chose in diot-crate-hw #82 ("to follow the convention that we have PSU1 and PSU2 in the crate"). The DI/OT Crate spec, the Front Interface and the Power Backplane were not renamed and still call the same wires P_PRES0 / P_PRES1, so a reader following names across the crate is off by one.

    The same wire under two names

    ContactOHWR0084 R3 CPCI-SERIAL BACKPLANEMates withName thereSource
    J8.D5 (Utility)P_PRES1OHWR0086 R5 FRONT INTERFACE J3.D5P_PRES0netlists, mating confirmed in OHWR-DIOT-CRATE-V3
    J8.D6 (Utility)P_PRES2OHWR0086 R5 FRONT INTERFACE J3.D6P_PRES1netlists
    Spec Table 2, D5 / D6--P_PRES0 / P_PRES1DI/OT Crate specification v3.1, Table 2, p.7
    Power Backplane PSU slot J1.D3 / J2.D3--P_PRES0 / P_PRES1OHWR0085 R4 POWER BACKPLANE netlist

    On the backplane the two nets continue to the system slot (J2.G7 = P_PRES1, J2.H7 = P_PRES2).

    Design intent

    [diot-crate-hw #82](https://gitlab.com/ohwr/project/diot-crate-hw/-/issues/82) "[EDA-04211-V2]: Rename P_PRES<n> signals" (closed, Done): P_PRES0 → P_PRES1, P_PRES1 → P_PRES2. The backplane is the intended naming.

    Why it matters

    Nothing is miswired: the copper is continuous and the interconnect composes. The risk is documentation only. Carrying the #82 rename into the spec, the Front Interface and the Power Backplane schematics would remove the off-by-one.

  4. ReferencePlaneDiscontinuityminorOpen

    2 nets cross the split between the GND and P12V planes on TOP_LAYER at 4 places, all with a plane on the track's other side (fastest: F_IO1 at 2 ns 10-90 %, suggested by its name)

    On TOP_LAYER the GND and P12V planes part, and 2 nets cross from one to the other at 4 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 4 of the 4 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 53% 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).

    • 'F_IO1' on L2 at (289.32, 133.3), 0.16 mm; other side: GND on L3GND, 0.2 mm from the track against 0.18 mm to TOP_LAYER: 53% of the return on TOP_LAYER
    • 'F_IO1' on L2 at (292.04, 133.96), 0.23 mm; other side: GND on L3GND, 0.2 mm from the track against 0.18 mm to TOP_LAYER: 53% of the return on TOP_LAYER
    • 'P_IO0' on L2 at (289.32, 130.04), 0.23 mm; other side: GND on L3GND, 0.2 mm from the track against 0.18 mm to TOP_LAYER: 53% of the return on TOP_LAYER
    • 'P_IO0' on L2 at (293.24, 133.96), 0.23 mm; other side: GND on L3GND, 0.2 mm from the track against 0.18 mm to TOP_LAYER: 53% of the return on TOP_LAYER

    Every site is a row of this finding's evidence, keyed by net, layers and position. GET /api/odb/boards/88/reference-planes?group=split%40TOP_LAYER%3AGND%7CP12V serves each with its own layer_image region; all of them lie in x 288.24..294.32, y 128.96..135.04 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@TOP_LAYER:GND|P12V

  5. ConnectorDomainMixedminorOpen

    ConnectorDomainMixed: 11 components×11

    Checked by the review, stands · Claude walker D (Cern session) — Checked: two supplies on one connector is the crate design, not an accident. J8 (Utility) and J7 (Ext PSU) are power-distribution connectors whose +12V / +5V contacts are assigned in DI/OT Crate spec v3.1 Table 2 and sec.3.1 item 6; J1 (system slot P1) and J11_Slot1-8 (peripheral slot P1) are CompactPCI Serial P1, which carries the slot supply (+12 V on A1/D1/E1/G1/H1/J1/K1) and +5 V on B1, declared cpci-serial-system-p1 / cpci-serial-peripheral-p1 (no contact map on file to suppress it). Left for the owner to confirm against PICMG CPCI-S.0.

    At J1, J11_Slot1, J11_Slot2, J11_Slot3, J11_Slot4, J11_Slot5 and 5 more

    Each finding, with its own headline
    • J1J1 takes 2 different supplies off the board on one connector: P12V (12.0 V), P5V (5.0 V)

      J1 carries 2 different supplies out of the board on one connector: P12V (12.0 V) on A1/D1/E1/G1/H1/J1/K1, P5V (5.0 V) on B1. 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.

    • J11_Slot1J11_Slot1 takes 2 different supplies off the board on one connector: P12V (12.0 V), P5V (5.0 V)

      J11_Slot1 carries 2 different supplies out of the board on one connector: P12V (12.0 V) on A1/D1/E1/G1/H1/J1/K1, P5V (5.0 V) on B1. 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.

    • J11_Slot2J11_Slot2 takes 2 different supplies off the board on one connector: P12V (12.0 V), P5V (5.0 V)

      J11_Slot2 carries 2 different supplies out of the board on one connector: P12V (12.0 V) on A1/D1/E1/G1/H1/J1/K1, P5V (5.0 V) on B1. 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.

    • J11_Slot3J11_Slot3 takes 2 different supplies off the board on one connector: P12V (12.0 V), P5V (5.0 V)

      J11_Slot3 carries 2 different supplies out of the board on one connector: P12V (12.0 V) on A1/D1/E1/G1/H1/J1/K1, P5V (5.0 V) on B1. 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.

    • J11_Slot4J11_Slot4 takes 2 different supplies off the board on one connector: P12V (12.0 V), P5V (5.0 V)

      J11_Slot4 carries 2 different supplies out of the board on one connector: P12V (12.0 V) on A1/D1/E1/G1/H1/J1/K1, P5V (5.0 V) on B1. 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.

    • J11_Slot5J11_Slot5 takes 2 different supplies off the board on one connector: P12V (12.0 V), P5V (5.0 V)

      J11_Slot5 carries 2 different supplies out of the board on one connector: P12V (12.0 V) on A1/D1/E1/G1/H1/J1/K1, P5V (5.0 V) on B1. 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.

    • J11_Slot6J11_Slot6 takes 2 different supplies off the board on one connector: P12V (12.0 V), P5V (5.0 V)

      J11_Slot6 carries 2 different supplies out of the board on one connector: P12V (12.0 V) on A1/D1/E1/G1/H1/J1/K1, P5V (5.0 V) on B1. 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.

    • J11_Slot7J11_Slot7 takes 2 different supplies off the board on one connector: P12V (12.0 V), P5V (5.0 V)

      J11_Slot7 carries 2 different supplies out of the board on one connector: P12V (12.0 V) on A1/D1/E1/G1/H1/J1/K1, P5V (5.0 V) on B1. 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.

    • J11_Slot8J11_Slot8 takes 2 different supplies off the board on one connector: P12V (12.0 V), P5V (5.0 V)

      J11_Slot8 carries 2 different supplies out of the board on one connector: P12V (12.0 V) on A1/D1/E1/G1/H1/J1/K1, P5V (5.0 V) on B1. 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.

    • J7J7 takes 2 different supplies off the board on one connector: P12V (12.0 V), P5V (5.0 V)

      J7 carries 2 different supplies out of the board on one connector: P12V (12.0 V) on P1/P2, P5V (5.0 V) on P5. 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.

    • J8J8 takes 2 different supplies off the board on one connector: P12V (12.0 V), P5V (5.0 V)

      J8 carries 2 different supplies out of the board on one connector: P12V (12.0 V) on P8/P9, P5V (5.0 V) on P5. 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.

Main parts

Part numberMakerDescriptionWhere
51940-650AMPHENOLThru-Hole, Lead 3.43mm, PowerBlade Vertical Power Connector 3ACP 2P 24S 4DCP ContactsJ7, J8
Use only in PCBFiducial TargetFTG1, FTG2, FTG3
10052829-101LFAMPHENOL72 Contacts Receptacle, 6 Columns, 4 Pairs per ColumnJ1, J11_Slot1, J11_Slot2, J11_Slot3, J11_Slot4, J11_Slot5 and 4 more
CC0402_10NF_16V_10%_X7RGENERICNon-polarised capacitorC1, C10, C11, C12, C2, C3 and 6 more
10114760-101LFAMPHENOL96 Contacts Header, 8 Columns, 4 Pairs per ColumnJ23_Slot1, J23_Slot2, J23_Slot3, J23_Slot4, J23_Slot5, J23_Slot6 and 10 more
EDA-04126-V1-0CERNPICMG CompactPCI Serial Backplane Guide ElementGUIDE1, GUIDE10, GUIDE11, GUIDE12, GUIDE13, GUIDE14 and 11 more
R0402_1K_1%_0.0625W_100PPMGENERICResistor - 1%R1, R10, R11, R12, R13, R14 and 14 more
10052842-101LFAMPHENOL96 Contacts Receptacle, 8 Columns, 4 Pairs per ColumnJ13_Slot1, J13_Slot2, J13_Slot3, J13_Slot4, J13_Slot5, J13_Slot6 and 22 more

About this design

Design
DIOT CRATE CPCI-SERIAL BACKPLANE — OHWR0084 R3
Designed in
Altium Designer
Board
14 copper layers, 304.8 × 128.5 mm, 141 components, 1225 nets
Licence
CERN-OHL-W-2.0
Source
https://edms.cern.ch/item/EDA-04211-V3-0/0 at commit EDMS EDA-042
Attribution
CERN EDMS item EDA-04211-V3-0 'DI/OT Crate - CPCI-Serial Backplane': document 3086352 v.1, retrieved 2026-10-18T12:45:59Z. CERN EDA-04211 V3, DI/OT Crate - CPCI-Serial Backplane. Copyright CERN 2019, 2023-2024. Licence: CERN OHL-W v2.0+. Project: https://gitlab.com/ohwr/project/diot-crate-hw
Reviewed with
Probe by E# (E-Sharp AB)