CAN Common-Mode Port

Hardware version

CANBench Duo v1.2 — Schematic-stage refresh of the V1.1 fabricated prototype. V1.2 is electrically identical to V1.1 and carries the InvenTree-canonical component metadata; no V1.2 boards exist yet — testing and bring-up reference the V1.1 hardware.

Other versions: v1.1 — fabricated prototype (current)

The CAN common-mode measurement port is a high-impedance, non-terminating tap on the CAN bus that extracts the common-mode voltage of the CAN-H / CAN-L pair and presents it as a 50 Ω signal at SMA J6 for spectrum-analyser measurement. Common-mode disturbance often dominates the upper-band conducted-emissions signature of a CAN-bus system, and this port makes it directly measurable.

This port does not terminate the CAN bus

The CAN CM port is a high-impedance summing tap. The standard 120 Ω CAN-bus termination must be provided by the DUT and/or the upstream backbone. Connecting the CAN CM port to an unterminated bus will degrade signalling and produce bus errors. The bus-impedance perturbation when the bus is properly terminated (60 Ω each end with the standard 120 Ω terminators) is less than 3 % — negligible.

Overview

This page covers a single sub-circuit — the CAN Common-Mode Measurement Port — drawn on the can_cm_measurement_port KiCad sheet. A matched 1 kΩ summing pair extracts the bus common-mode voltage, an AC-couple pair blocks the bus DC bias, a two-stage π attenuator drops the signal to a level safe for a spectrum-analyser front end, and a multi-stage clamp / TVS cascade protects the analyser at the SMA output. The CAN-H / CAN-L lines arrive on pins 4 and 5 of the M12 N2K connector J10, drawn on the connectors-and-mechanical sheet.

Functional specification and design objectives

The CAN common-mode measurement port must:

• tap the CAN-H / CAN-L pair at the DUT-side M12 N2K connector and present the bus common-mode voltage (V_H + V_L) / 2 as a 50 Ω signal at SMA J6;
• load the bus negligibly — keep the bus-impedance perturbation below 3 % so it does not replace the bus's external 120 Ω termination;
• reject the differential-mode CAN signalling, with a common-mode rejection ratio bounded by the matched summing pair (≈ 60 dB at DC, degrading at HF);
• support the CISPR 25 measurement band, 150 kHz – 108 MHz;
• block the CAN bus DC operating point (nominally 2.5 V common-mode, with margin to 100 V) from reaching the analyser; and
• protect the spectrum-analyser front end from ESD and transients at the SMA via a layered clamp + TVS cascade.

CAN Common-Mode Measurement Port

How it works

The CAN bus enters via pins 4 (NET-H) and 5 (NET-L) of the M12 N2K connector J10 on the connectors-and-mechanical sheet. From there, the chain runs left-to-right along the board's central Y ≈ 90 mm axis through to the SMA J6 at the leftmost edge.

Stage 1 — Common-mode summing

Two equal-weight 1 kΩ thin-film resistors form the summing front-end:

• R41 — 1 kΩ, 0603 thin-film ±0.1 %, between NET-H and RF_CAN1
• R38 — 1 kΩ, 0603 thin-film ±0.1 %, between NET-L and RF_CAN1

By Kirchhoff, the voltage at the summing node RF_CAN1 is (V_H + V_L) / 2 — the common-mode voltage of the CAN bus. The differential-mode component (CAN-H and CAN-L moving in opposite directions, which is how the CAN bus signals data) cancels by symmetry, leaving only the common-mode content at RF_CAN1.

The matching tolerance between R38 and R41 bounds the common-mode rejection ratio (CMRR) of the front-end. With 0.1 % matched thin-film, the theoretical CMRR at DC is about 60 dB. CMRR degrades at higher frequencies because the AC-couple capacitors (next stage) are X7R ±10 % — looser matching than the resistor pair — and because PCB trace asymmetry between the NET-H and NET-L paths adds further imbalance.

Stage 2 — AC-couple

Two parallel 100 nF capacitors block the CAN bus's DC offset (nominally 2.5 V common-mode):

• C25 — 100 nF / 100 V X7R (0805)
• C26 — 100 nF / 100 V X7R (0805)

Combined parallel capacitance is 200 nF. The 100 V rating leaves a huge margin over the 2.5 V CAN-bus quiescent CM and the higher transient CM voltages possible during EMC events.

Stage 3 — Two-stage π attenuator

A two-stage π attenuator brings the common-mode amplitude down to a level safe for the spectrum-analyser input:

RF_CAN5 ─── R42 (68.1 Ω) ─── RF_CAN6 ─── R40 (5.1 Ω) ─── RF_CAN_CM ─── J6
                │                            │
              R46 (91 Ω)                  R47 (91 Ω)
                │                            │
              GNDREF                       GNDREF

All four attenuator resistors are 0.1 % thin-film for impedance precision. The combined transfer function from RF_CAN5 to RF_CAN_CM (loaded with 50 Ω at the SMA) is approximately −10 dB, with an input impedance at RF_CAN5 of about 48 Ω — close to the 50 Ω port target.

Note that the full transfer function from the CAN bus's common-mode voltage to the analyser SMA includes the voltage division across the 1 kΩ summing pair: the high-impedance summing front-end gives up about 21 dB on its own to avoid loading the bus, then the π attenuator adds another 10 dB, for a total of about 31 dB from V_CM at the bus to the analyser. Users back-calculating bus CM levels from analyser readings should account for this.

R40 (5.1 Ω) serves two roles: it is both the second-stage π attenuator's series element AND the current limiter between the outer and inner protection clamps in the next stage.

Stage 4 — Multi-stage protection cascade

The protection cascade follows the same pattern as the LISN measurement ports, but with simpler 1-diode clamps at both stages because CAN-bus signal swings are smaller than the LISN measurement-band signals — single forward V_F drops suffice without conducting at normal operating levels.

1. Outer 1-diode bipolar clamp at RF_CAN6. D25 (cathode on RF_CAN6, anode on GNDREF) and D26 (anode on RF_CAN6, cathode on GNDREF) — anti-parallel 1N4148W pair clamping at ±0.6 V.

2. R40 (5.1 Ω) current limiter. Same R40 that is also the π-attenuator's second-stage series resistor. Limits current into the inner clamp during transient events.

3. Inner 1-diode bipolar clamp at RF_CAN_CM. D22 and D23 mirror the outer clamp arrangement at the SMA-output net.

4. Integrated TVS D24. Tech Public TPAZ1023-02F — same part used on the LISN measurement ports. Only Channel 1 populated; pin 1 to RF_CAN_CM, pin 3 to GNDREF, pins 2/4/5/6 are deliberately left unconnected for minimum parasitic capacitance.

Stage 5 — SMA output

RF_CAN_CM connects to J6 (SMA Female Vertical, 50 Ω) at the top extrusion. R43 (1 MΩ) ties RF_CAN_CM to GNDREF as a defined DC reference at the SMA when no analyser is connected.

Performance

Design intent — calculated in performance_review/can-cm-port.md against V1.2 BOM specs.

Parameter	Target	Status
Measurement band	150 kHz – 108 MHz (CISPR 25)	Topology supports
Source impedance to analyser	50 Ω	Network input impedance ≈ 48 Ω at LF
Added attenuation (π stage)	≈ 10 dB	Two-stage π attenuator nominal
Total bus-to-SMA transfer	≈ −31.5 dB at LF	21 dB summing front-end loss + 10 dB attenuator
DC voltage block	up to 100 V	C25 / C26 100 V rating; CAN bus quiescent CM is 2.5 V
CMRR at DC	≈ 60 dB	Bounded by R38 / R41 0.1 % matching
CMRR at 1 MHz	~ 40 dB	Limited by C25 / C26 X7R ±10 % matching
CMRR at 108 MHz	< 30 dB	Limited by PCB trace asymmetry between NET-H and NET-L paths
CAN bus impedance perturbation	< 3 %	2 kΩ inter-line load vs 60 Ω parallel terminators
Stray Cj at SMA output	≈ 4.3 pF	Same as LISN measurement ports

The CMRR vs frequency curve is bounded analytically here; SPICE / VNA characterisation on the as-built V1.2 board would close the analytical bound with measurement. The dominant HF limitation is the AC-couple capacitor matching (X7R ±10 %), not the matched-pair resistor front-end.

Not for ISO 7637-2 transient testing

Same scope clarification as the LISN measurement ports — this is a passive measurement instrument, not a transient-injection compliance test fixture.

PCB Layout

The CAN CM port components sit in a single left-to-right cascade along the board's Y ≈ 90 mm axis — between the upper (LISN+) and lower (LISN−) measurement ports.

The dominant layout constraint is the symmetric placement of R38 and R41. The CAN CM port's CMRR depends critically on matched-length trace geometry from CAN-H and CAN-L (at J10's pins 4 and 5) to the summing node RF_CAN1 (at R38 / R41's RF_CAN1 pads). Even a millimetre of length asymmetry between the NET-H and NET-L traces translates to a CMRR penalty at higher frequencies. The V1.1 layout places R38 and R41 at the same X coordinate with a 1.6 mm Y spacing about the Y = 90 mm axis — supporting symmetric routing in principle. NET-H has 4 F.Cu segments and NET-L has 2 in the as-built layout; a future revision could re-route for closer length matching.

The protection cascade follows the same component-island pattern as the LISN measurement ports: clamp diode pairs as compact 2-component islands at their respective nodes, TPAZ1023 placed within 2.5 mm of the SMA centre pin, F.Cu coplanar ground pour around the RF traces.

R40 (5.1 Ω, 0603 thick-film AC0603) is the only thick-film resistor in the RF path. The rest of the chain uses YAGEO RT-series thin-film at 0.1 % tolerance. R40 only conducts during clamp events; the thick-film tolerance (±1 %, ±200 ppm/°C) is acceptable for the current-limiter role.

See pcb_review/can-cm-port-layout.md in the source repository for the per-component coordinate table.

Components

Ref	Value	Function	Datasheet
R41	1 kΩ ±0.1 %	CM summing resistor on NET-H (CAN-H) path; matched pair with R38	YAGEO RT0603 thin-film
R38	1 kΩ ±0.1 %	CM summing resistor on NET-L (CAN-L) path; matched pair with R41	YAGEO RT0603 thin-film
C25, C26	100 nF / 100 V X7R (0805)	AC-couple pair from summing node RF_CAN1 to RF_CAN5 (200 nF total)	muRata GCM21BR72A104KA37L
R42	68.1 Ω ±0.1 %	π attenuator stage 1 series resistor (RF_CAN5 → RF_CAN6)	YAGEO RT0603 thin-film
R46	91 Ω ±0.1 %	π attenuator shunt to GNDREF at RF_CAN5	YAGEO RT0603 thin-film
R47	91 Ω ±0.1 %	π attenuator shunt to GNDREF at RF_CAN6	YAGEO RT0603 thin-film
R40	5.1 Ω ±1 %	π attenuator stage 2 series resistor / current limiter into the inner clamp (thick-film)	YAGEO AC0603 thick-film
D25, D26	1N4148W	Outer bipolar 1-diode clamp at RF_CAN6 (anti-parallel pair to GNDREF)	DIODES 1N4148W-7-F
D22, D23	1N4148W	Inner bipolar 1-diode clamp at RF_CAN_CM (anti-parallel pair to GNDREF)	DIODES 1N4148W-7-F
D24	TPAZ1023-02F	Integrated multi-channel TVS; only Channel 1 populated (pin 1 = RF_CAN_CM, pin 3 = GNDREF, pins 2/4/5/6 n/c)	Tech Public TPAZ1023-02F
R43	1 MΩ ±1 %	DC bleeder from RF_CAN_CM to GNDREF — defined DC reference at the SMA when no instrument is connected	YAGEO RC0603
J6	SMA Female Vertical, 50 Ω	Panel-mount output to spectrum analyser / disturbance meter at the top extrusion	HCTL HC-SMA6565-13H-G

Gaps & next version

Before next production run

• R38 / R41 matching. Resistor matching directly bounds the CMRR of the summing pair. Confirm the BOM specifies matched-pair-grade ±0.1 % parts and that any future substitution preserves the matching.
• TPAZ1023 ESD/TVS rating. The integrated-TVS specifications (clamp voltage, per-channel capacitance, peak-pulse current) are nominal datasheet values from an LCSC-hosted redirect; the manufacturer ("Tech Public") has no publicly indexed datasheet. Treat as datasheet-cited and verify against board behaviour where measurable.
• SMA ground via cluster around J6 is sparse — recurring observation across the RF ports; either fix in the next revision or accept and confirm by ESD test.

Next version (V1.3)

• CMRR HF bound is set by the C25 / C26 X7R ±10 % matching (looser than the R38 / R41 0.1 % pair); at 1 MHz CMRR may drop to ~ 40 dB. Switch C25 / C26 to a C0G matched pair (as on the LISN ladder) to tighten the HF CMRR.
• NET-H / NET-L trace asymmetry — the as-built layout has 4 F.Cu segments on NET-H versus 2 on NET-L. Re-route for symmetric matched-length traces to improve CMRR.
• R40 thick-film (AC0603 ±1 %, ±200 ppm/°C) in an otherwise thin-film RF chain — cleanup candidate (same observation as R21 on the LISN ports); acceptable for the current-limiter role today.

References

• DIODES Incorporated, 1N4148W-7-F SOD-123 switching diode (D22, D23, D25, D26)
• Tech Public, TPAZ1023-02F multi-channel TVS (D24)
• muRata, GCM21BR72A104KA37L 100 nF / 100 V X7R MLCC (C25, C26)
• YAGEO, RT0603 thin-film 0.1 % series (R38, R41, R42, R46, R47)
• YAGEO, AC0603 thick-film 1 % series (R40)
• YAGEO, RC0603 1 MΩ 1 % (R43)
• HCTL, HC-SMA6565-13H-G SMA Female Vertical 50 Ω (J6)
• ISO, ISO 11898-2 — Road vehicles — Controller area network (CAN) — Part 2: High-speed medium access unit (the 120 Ω termination requirement this port does not replace)
• NMEA, NMEA 2000 Standard — M12 Micro-C Code A pinout (J10 pin 4 = NET-H, pin 5 = NET-L)
• IEC, CISPR 25 — conducted-emissions measurement band (150 kHz – 108 MHz); IEC 61000-4-2 — electrostatic discharge

Related pages

• LISN Measurement Ports — the sister RF measurement chain; topologically similar but with 2-diode-series outer clamps for the higher LISN measurement-band amplitudes
• LISN Supply Path — the LISN ladder also delivers DC supply to the DUT via the M12 N2K connector (J10 pin 2 / pin 3); the CAN-H / CAN-L lines (pins 4 / 5) tapped here arrive on the same connector
• Connectors and Mechanical — J10 M12 N2K pin map (Shield, NET-S, NET-C, CAN-H, CAN-L) and J6 SMA placement on the top extrusion
• Power Indicator LED — supply-fault indicator; Q1 is confirmed isolated from the CAN measurement nets
• User Manual → Measurement Procedure — operational measurement workflow including the "external 120 Ω termination required" caveat