LISN Measurement Ports

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)

Overview

The LISN measurement ports are two mirror-image RF chains that tap the LISN-filtered DUT-side rails and present each rail's conducted-emissions signature as a 50 Ω signal at a board-mounted SMA. The upper-rail port (RF_LISN_P, SMA J2) taps DUT+; the lower-rail port (RF_LISN_N, SMA J4) taps DUT−. The two ports are full mirror images of each other across the board's Y = 90 mm axis — identical topology, identical component values, identical PCB layout.

This page covers a single sub-circuit — the LISN Measurement Ports — drawn across the lisn_positive_measurement_port and lisn_negative_measurement_port KiCad sheets.

The design is derived from Jay_Diddy_B's EEVblog 5 µH LISN reference design, itself based on the HP / Keysight 11947A transient limiter with high-pass filter.

Functional specification and design objectives

The LISN measurement-port chain is designed to deliver:

• A clean view of conducted emissions across the CISPR 25 measurement band (150 kHz – 108 MHz) with margin to 120 MHz
• A 50 Ω source impedance to the spectrum analyser
• Approximately 10 dB of added attenuation to keep analyser front-ends in their useful dynamic range
• DC blocking up to 48 V so the bench supply's quiescent voltage does not reach the analyser
• A multi-stage protection cascade that limits residual transients at the SMA to a safe level for the analyser

LISN Measurement Ports

The lower-rail mirror has identical topology with mirrored refdes (e.g. R20 → R29, C15 → C20, J2 → J4):

How it works

The signal-flow order from the LISN ladder's DUT-side output to the SMA matches the physical layout of the V1.1 board: right-to-left across the upper portion of the PCB.

Stage 1 — AC-couple from the LISN rail

Three parallel capacitors form the AC-couple from DUT+ to the first internal node RF_LP1:

• C15 — 100 nF / 100 V X7R (0805)
• C18 — 1 µF / 100 V X7R (1210)
• C19 — 1 µF / 100 V X7R (1210)

Total parallel capacitance is about 2.1 µF, which gives a low-frequency cut-off well below the CISPR 25 band edge. The 100 V rating leaves comfortable margin for the design's 48 V supply ceiling.

The lower-rail mirror uses C20, C23, C24 from DUT− to RF_LN1.

Stage 2 — Two-stage π attenuator

The attenuator is built from thin-film 0.1 % matched resistors for impedance precision:

• R20 (17.8 Ω series, 1206 thin-film) between RF_LP1 and RF_LP2
• R24 and R25 (294 Ω each, 0805 thin-film) as shunts on RF_LP1 and RF_LP2
• A second π section uses R22 (45.3 Ω series), R26 / R27 / R28 (130 Ω shunts)

The combined transfer function approximates 10 dB of attenuation across the measurement band, with the input impedance at RF_LP1 ≈ 48 Ω — close to the 50 Ω port target. Numerical verification against the actual V1.2 BOM values is in performance_review/lisn-measurement-ports.md; full S21 / return-loss simulation against PCB parasitics is pending.

Stage 3 — C-L-C high-pass filter

A C-L-C network sets the LF cut-off of the chain:

• C16 and C17 (470 nF each, 0805 X7R) as series elements
• L11 + L12 (470 µH each, in series via Net-(L11-Pad1), shunt to GNDREF) — together a 940 µH choke from RF_LP3 to ground

With two 470 nF capacitors effectively in series (≈ 235 nF) and a 940 µH shunt inductor, the ideal LF corner is:

f_0 ≈ 1 / (2π · √(L · C)) = 1 / (2π · √(940 µH · 235 nF)) ≈ 10.7 kHz

About a decade below the CISPR 25 lower band edge of 150 kHz — well clear of the measurement region. The C-L-C also blocks the upstream LISN-rail DC offset from reaching the diode-clamp stage.

Stage 4 — Multi-stage protection cascade

Four protection stages stack between the attenuator network and the SMA, each progressively limiting voltage and absorbing energy:

1. Outer bipolar 2-diode clamp at RF_LP6. D6 + D11 (series-stacked anode-to-cathode) and D7 + D12 (mirror polarity) form a ±1.2 V clamp using 1N4148W fast-switching diodes. Two-diode series gives roughly half the junction capacitance per leg compared to a single diode — important for keeping HF measurement fidelity (the 1N4148W single-diode capacitance is about 4 pF; the series pair is about 2 pF per leg).

2. R21 (5.1 Ω) inter-stage current limiter. Between the outer clamp at RF_LP6 and the inner clamp at RF_LISN_P. When the outer clamp engages, R21 limits the current that flows into the inner stage.

3. Inner bipolar 1-diode clamp at RF_LISN_P. D8 and D9 (anti-parallel 1N4148W pair) provide a tighter ±0.6 V clamp at the SMA-output net. Because R21 has already attenuated the residual transient energy, this stage handles much lower peak current than the outer stage and can clamp tighter without sacrificing reliability.

4. Integrated TVS D10 at the SMA. Tech Public TPAZ1023-02F multi-channel TVS in a DFN1210-6 package. Six channels are available; only Channel 1 (I/O1) is populated on the design — pin 1 connects to RF_LISN_P, pin 3 to GNDREF, pins 2/4/5/6 are deliberately left unconnected. Single-channel population minimises parasitic capacitance loading on the RF node (each active channel adds about 0.3 pF; populating all six would slacken HF response).

The cascade — 2-diode clamp → 5.1 Ω → 1-diode clamp → TVS — successively limits voltage and absorbs energy: the outer 2-diode clamp catches gross transients, R21 limits current into the inner stage, the inner 1-diode clamp tightens the residual transient further, and the TVS catches anything else (especially ESD events with sub-ns rise times).

Stage 5 — SMA output

RF_LISN_P connects to J2 (HCTL HC-SMA6565-13H-G SMA Female Vertical, 50 Ω THT). A 1 MΩ bleeder (R23) ties RF_LISN_P to GNDREF — provides a defined DC reference at the SMA when no analyser is connected, with high enough resistance not to load the RF measurement.

The lower-rail chain ends at J4 (RF_LISN_N) on the same top-extrusion column as J2, mirror-placed across Y = 90.

Performance

Design intent — calculated in performance_review/lisn-measurement-ports.md against V1.2 BOM specs. SPICE simulation against PCB parasitics is pending for full S21 / return-loss verification.

Parameter	Target	Status
Measurement band	150 kHz – 108 MHz (CISPR 25), margin to 120 MHz	Topology supports
Source impedance to analyser	50 Ω	Network input impedance ≈ 48 Ω at LF
Added attenuation	≈ 10 dB	Two-stage π attenuator nominal
DC voltage block	up to 48 V	Limited by C15 / C20 (100 V X7R) — actual margin > 100 V
LF corner	≈ 10.7 kHz (C-L-C high-pass)	Below 150 kHz CISPR band edge
Stray Cj at SMA output	≈ 4.3 pF	D8/D9 ≈ 3 pF + D10 ≈ 0.3 pF + SMA pad ≈ 1 pF
Analyser-side residual transient limit	≤ +10 dBm into 50 Ω	Bounded by the protection cascade
ESD tolerance at SMA	IEC 61000-4-2 ±8 kV air / ±6 kV contact (design target)	TPAZ1023 specific rating verification pending

Not for ISO 7637-2 transient testing

The TVS cascade is designed for ESD and bench-handling events only. ISO 7637-2 Pulse 5a/5b (load dump, 100+ V sustained transients) would exceed the chain's capability and risks damage to the LISN's protection FETs upstream and the TVS at the SMA itself.

PCB Layout

The measurement-port components sit in two rows across the upper portion of the board (LISN+ at Y ≈ 117 mm) and the lower portion (LISN− at Y ≈ 64 mm), mirror-placed across Y = 90 mm. The signal flow runs right-to-left along each row, from the AC-couple cap cluster at the DUT-rail tap (X ≈ 156 mm) through the attenuator and clamp cascade to the SMA at the leftmost board edge (X ≈ 110 mm).

Specific layout choices worth calling out:

• 0.2 mm RF trace widths are NOT a mistake. The archive documents this as a "diminishing returns" decision: given the high component density and short trace runs (≤ 10 mm between adjacent RF components), full controlled-impedance CPWG (1.7 mm trace width) or microstrip (2.8 mm) was judged unnecessary. At the 108 MHz band edge in FR-4, λ/10 ≈ 14 cm — orders of magnitude above the actual trace lengths. The impedance mismatch over such short lengths is electrically lumped.
• F.Cu coplanar ground pour surrounds each RF trace, with the B.Cu continuous GNDREF flood as the reference plane below. This is the CPWG-like geometry, just with thin signal traces.
• Outer clamp pairs (D6+D11, D7+D12) sit as compact 4-component diode islands at RF_LP6 with the GNDREF-side via directly into the F.Cu pour. Mirror on the lower rail with D13+D18, D14+D19.
• TPAZ1023 D10 is placed 2.5 mm from J2's centre pin — minimum-inductance path for the SMA-side clamp. Same for D17 and J4.
• Shunt-choke pair (L11 + L12 in series to GND, mirror with L13 + L14) sits compactly at X = 140.75 mm, 6 mm Y-spacing between the two chokes. Mirror on the lower rail.

See pcb_review/lisn-measurement-ports-layout.md in the source repository for the full per-component coordinate table and verification.

Components

The two ports are full mirror images. Each row covers both rails (positive-rail ref / negative-rail ref); footprint, value, and part are identical across the mirror pair.

Refs (pos / neg)	Value	Function	Datasheet
C15 / C20	100 nF / 100 V X7R, 0805	First AC-couple cap from DUT± rail to RF_LP1/RF_LN1 (parallel with C18/C19)	muRata GCM21BR72A104KA37L
C18, C19 / C23, C24	1 µF / 100 V X7R, 1210	Bulk AC-couple, parallel with C15/C20 (total ≈ 2.1 µF)	muRata GCM32CR72A105KA35L
C16, C17 / C21, C22	470 nF / 50 V X7R, 0805	Series elements of the C-L-C high-pass filter	muRata GCM21BR71H474KA55L
L11, L12 / L13, L14	470 µH, IND-5942	Series pair (940 µH) shunt choke from RF_LP3/RF_LN3 to GNDREF	PROD Tech PSWSAA5942-471M (LCSC C46628485)
R20 / R29	17.8 Ω, 0.1 % thin-film, 1206	First-section π-attenuator series element (RF_LP1 → RF_LP2)	YAGEO RT1206BRD0717R8L
R24, R25 / R33, R34	294 Ω, 0.1 % thin-film, 0805	First-section π-attenuator shunts on RF_LP1/RF_LP2	YAGEO RT0805BRD07294RL
R22 / R31	45.3 Ω, 0.1 % thin-film, 0603	Second-section π-attenuator series element (RF_LP4 → RF_LP6)	YAGEO RT0603BRD0745R3L
R26, R27, R28 / R35, R36, R37	130 Ω, 0.1 % thin-film, 0603	Second-section shunts / HF damping at RF_LP3/RF_LP4/RF_LP6	YAGEO RT0603BRD07130RL
D6, D7, D11, D12 / D13, D14, D18, D19	1N4148W, SOD-123	Outer bipolar 2-diode clamp (±1.2 V) at RF_LP6/RF_LN6	DIODES 1N4148W-7-F
R21 / R30	5.1 Ω, 1 % thick-film, 0603	Inter-stage current limiter between outer and inner clamps	YAGEO AC0603FR-075R1L
D8, D9 / D15, D16	1N4148W, SOD-123	Inner bipolar 1-diode clamp (±0.6 V) at RF_LISN_P/RF_LISN_N	DIODES 1N4148W-7-F
D10 / D17	TPAZ1023-02F, DFN1210-6	Integrated multi-channel TVS at the SMA — only Channel 1 populated	Tech Public TPAZ1023-02F (LCSC C41408050)
R23 / R32	1 MΩ, 1 % thick-film, 0603	DC bleeder from RF_LISN_P/RF_LISN_N to GNDREF	YAGEO RC0603FR-071ML
J2 / J4	SMA Female Vertical, 50 Ω THT	RF output to spectrum analyser / disturbance meter	HCTL HC-SMA6565-13H-G

Gaps & next version

Before next production run

• TPAZ1023 ESD survivability — the "3 A @ 8/20 µs" rating in the BOM is a surge spec, not the IEC 61000-4-2 sub-µs ESD profile; the manufacturer's full datasheet (HBM / CDM / IEC 61000-4-2 ratings) is not publicly indexed in English. Fetch the datasheet and confirm the SMA ESD rating before committing to a production run.
• Full S21 / return-loss verification — SPICE simulation against the netlist (and ideally VNA characterisation of the as-built board) is still pending; the ≈ 10 dB attenuation, 50 Ω port match across the band, and ≈ 10.7 kHz high-pass corner all need numerical confirmation. A SPICE-confirmed return-loss measurement at the SMA would also close the 0.2 mm RF-trace-width question analytically.
• R24 / R25 (R33 / R34) dissipation at fault conditions — the 125 mW π-attenuator shunts may exceed their rating when the cascade is driven to its +10 dBm SMA limit. Confirm via SPICE the maximum allowed DUT-side input level keeps these ≤ 125 mW.
• L11–L14 surge current — confirm via SPICE that peak current through the shunt chokes during an IEC 61000-4-2 Level 4 ESD event stays ≤ the 120 mA rating; specify a higher-current variant in V1.3 if exceeded.

Next version (V1.3)

• 0.2 mm RF trace widths — by design (see PCB Layout) and acceptable at ≤ 108 MHz given the short trace runs; V1.3 may revisit if HF performance falls short of the SPICE/VNA results.
• Sparse SMA ground-via cluster — fewer ground stitching vias around each SMA than the routing memory recommends. Recurring across all three SMAs (J2, J4, and J6 on the CAN CM port). Either V1.3 layout cleanup or empirical confirmation that the IEC 61000-4-2 ESD performance is acceptable as built.
• R21 / R30 (5.1 Ω) thick-film in an otherwise thin-film RF chain — only conducts during clamp events, not under normal operation. V1.3 cleanup candidate to restore the RF-path thin-film discipline.
• Cross-sheet HF tap stub — the DUT-side HF shunt caps (C7, C14, on the LISN supply-path sheet) sit at the LISN ladder end (X = 142 mm), ~ 27 mm away from the DUT banana sockets (J1 / J3 at X = 168.5 mm). The 27 mm trace adds parasitic inductance not bypassed by the HF shunts. V1.3 candidate: add HF shunt caps at the J1 / J3 banana ends.

Measurement workflow

For the operational measurement workflow (spectrum-analyser configuration, baseline noise floor validation, LISN+ / LISN− symmetry comparison, environmental coupling considerations) see the User Manual. The key user-facing points:

• Always terminate unused SMA ports with 50 Ω. Unterminated ports cause false CM signatures, resonances, and unstable traces.
• Maintain repeatable geometry between comparative sweeps. Cable routing, DUT position, and proximity to nearby electronics significantly affect low-level common-mode measurements.
• Compare LISN+ vs LISN− to identify CM (similar signatures) vs DM (asymmetric signatures) content. For full CM / DM separation, pair the CANBench Duo with a CANBench TrueZ.

References

• Jay_Diddy_B, 5 µH LISN for spectrum-analyzer EMC/EMI work (EEVblog forum)
• DIODES Incorporated, 1N4148W-7-F SOD-123 fast switching diode
• Tech Public, TPAZ1023-02F multi-channel ESD/TVS (DFN1210-6)
• HCTL, HC-SMA6565-13H-G SMA Female Vertical 50 Ω
• IEC, IEC 61000-4-2 — Electromagnetic compatibility (EMC) — Electrostatic discharge immunity test
• CISPR, CISPR 25 — Vehicles, boats and internal combustion engines — Radio disturbance characteristics — Limits and methods of measurement

Related pages

• LISN Supply Path — the LISN ladder providing the DUT+ / DUT− rails being tapped by these measurement ports
• CAN Common-Mode Port — the sister measurement chain on the CAN data lines, structurally similar but with simpler clamping
• Connectors and Mechanical — SMA Female Vertical placement on the top extrusion (J2 and J4)
• User Manual — operational measurement workflow, spectrum-analyser setup, and results interpretation