---
title: "How does one coax carry a car camera’s video, control and power?"
topic: "Car camera coax"
answer: "Power rides along as DC, and the data rides on top of it in both directions at once. Four coils keep the data out of the 12 V supply, and each chip subtracts its own signal to hear the other end. On Antmicro’s GMSL2 boards the coils hold 1 kΩ or more over 86% of the band ADI recommends, and over 15 m of coax the video loses 15.9 dB at 3 GHz, where ADI allows 21.1 dB."
url: "https://nearfieldnotes.com/notes/car-camera-coax/"
site: "Near Field Notes"
author: "Michael Walker"
note: 27
family: "Signal integrity"
published: "2026-11-08"
revised: "2026-11-08"
revision: 1
video: "https://nearfieldnotes.com/videos/car-camera-coax/"
---

Note 27 · [Signal integrity](https://nearfieldnotes.com/families/si/) · Rev 1 · [Watch the video · 0:10](https://nearfieldnotes.com/videos/car-camera-coax/)

Car camera coax

# How does one coax carry a car camera’s video, control and power?

**Short answer:** Power rides along as DC, and the data rides on top of it in both directions at once. Four coils keep the data out of the 12 V supply, and each chip subtracts its own signal to hear the other end. On Antmicro’s GMSL2 boards the coils hold 1 kΩ or more over 86% of the band ADI recommends, and over 15 m of coax the video loses 15.9 dB at 3 GHz, where ADI allows 21.1 dB.

The camera sends 6 Gb/s of video and its status replies. The computer sends back control data at 187.5 Mb/s, and 12 V to run the camera. The shield is the ground for all of it.

The figure is 20 ns of the voltage at the camera’s end of the coax.

**Note 27’s opening figure: Four jobs, one coax.** 20 ns at the camera’s connector

![Note 27’s opening figure: Four jobs, one coax](https://nearfieldnotes.com/notes/car-camera-coax/fig-opening.png)

*What the figure shows:* A camera and a computer joined by one coax that carries four jobs: video and status data from the camera, control data and 12 V to it. Below, 20 ns of the voltage on the camera's connector: 12 V with the video and the arriving control data on top of it. Once the camera subtracts its own signal and low-passes what is left, only the control data remains.

The gray trace is everything on the connector: 12 V, the video going out and the control data coming in. Take away the camera’s own signal, low-pass the rest, and the bold trace is what’s left: the control data.

**The numbers**

*Every 1.7 ns of the 20 ns shown. The control data alone is what arrives from the computer, before the camera’s own signal is added.*

| Time | Video alone | Everything | Canceled and low-passed | Control data alone |
|---|---|---|---|---|
| 0.0 ns | 12.200 V | 12.105 V | 11.901 V | 11.890 V |
| 1.7 ns | 11.800 V | 11.702 V | 11.896 V | 11.888 V |
| 3.3 ns | 11.800 V | 11.684 V | 11.898 V | 11.898 V |
| 5.0 ns | 11.800 V | 11.740 V | 11.913 V | 11.958 V |
| 6.7 ns | 12.056 V | 12.092 V | 11.969 V | 12.051 V |
| 8.3 ns | 12.056 V | 12.138 V | 12.048 V | 12.096 V |
| 10.0 ns | 11.944 V | 12.069 V | 12.093 V | 12.107 V |
| 11.7 ns | 12.056 V | 12.179 V | 12.106 V | 12.110 V |
| 13.3 ns | 11.944 V | 12.039 V | 12.108 V | 12.108 V |
| 15.0 ns | 11.944 V | 12.032 V | 12.106 V | 12.073 V |
| 16.7 ns | 12.056 V | 12.056 V | 12.064 V | 11.982 V |
| 18.3 ns | 11.800 V | 11.731 V | 11.984 V | 11.916 V |

## Two boards and a coax

*Antmicro’s open GMSL2 pair*

Antmicro publishes both ends of a GMSL2 link as open hardware. The camera end is a 43 mm square board with a MAX96717 serializer. The camera’s MIPI CSI-2 lanes go in, and one coax comes out.

The computer end is 75 by 60 mm, with a MAX96724 deserializer that takes four cameras. Both boards use FAKRA connectors and [50 Ω](https://nearfieldnotes.com/glossary/#g-characteristic-impedance) coax.

On each board the coax’s center pin meets two paths. One goes through a 100 nF capacitor to the chip. The other turns off at a tap into a ladder of four coils that ends at 12 V.

**Fig. 27: Camera end and computer end.**

![Fig. 27: Camera end and computer end](https://nearfieldnotes.com/notes/car-camera-coax/fig-27.png)

*What the figure shows:* Figure 27: Antmicro's two GMSL2 boards from above, to one scale, joined by a coax. The camera end is 43 mm square with one coax connector; the computer end is 75 by 60 mm with four. On each, the copper from the coax's center pin is lit: the data goes on through a 100 nF capacitor to the chip, and the 12 V turns off at a tap into a ladder of four coils.

Both boards to one scale, from Antmicro’s renders. The copper from each coax pin is drawn from the board files. From the pin to the tap it carries both. The coax is drawn short.

**The numbers**

| Board | Antmicro design | Chip | Coax connectors |
|---|---|---|---|
| Camera end | GMSL Serializer rev 1.1.1, 43 × 43 mm, 4 layers | MAX96717 | J2, one FAKRA |
| Computer end | GMSL Deserializer rev 1.0.3, 75.1 × 60.1 mm, 4 layers | MAX96724 | J2 to J5, four FAKRA |

Drawn as a circuit, each end is a bias tee. The capacitor passes the data and blocks the DC. The coils pass the DC and block the data.

The two ends are built the same, part for part. Their coils are filter 14 in ADI’s list of power filters recommended for new GMSL2 designs, and Coilcraft’s reference filter SMD-POC-006.

The computer end puts 12 V onto its coax. The camera end takes it off and feeds its regulators.

**Fig. 27b: Both bias tees.**

![Fig. 27b: Both bias tees](https://nearfieldnotes.com/notes/car-camera-coax/fig-27b.png)

*What the figure shows:* Figure 27b: the bias tee at each end of the coax. At the camera, the MAX96717's SIOP pin goes through C30, 100 nF, to the coax connector J2; at a tap between them a ladder of coils runs down to the 12 V that feeds the board's regulators: L6 and L5, 470 nH each, then L4, 6.8 µH, with R17, 5.1 kΩ, across it, then L3, 22 µH, with R16, 5.1 kΩ, across it. The computer end is the same with C27, L13, L9, L5 and R24, L1 and R19, from a 12 V supply into the coax.

Standard symbols, with the boards’ part numbers. The ESD parts on each line and the 12 V decoupling are left out. The table lists them.

**The numbers**

| Part | Camera end | Computer end |
|---|---|---|
| Coupling capacitor | C30, 100 nF | C27, 100 nF |
| 470 nH, Coilcraft PFL1609-471 | L6, L5 | L13, L9 |
| 6.8 µH, Coilcraft 1210POC-682 | L4, with R17 across | L5, with R24 across |
| 22 µH, Coilcraft MSS6132T-223 | L3, with R16 across | L1, with R19 across |
| Damping resistors | R17, R16, 5.1 kΩ | R24, R19, 5.1 kΩ |
| 12 V decoupling | C10 100 nF, C11 10 nF | C6 100 nF, C10 10 nF |
| ESD protection | D2, between the tap and C30 | D4, at the chip |

## Power rides as DC

*One coil isn’t enough*

A coil’s impedance grows with frequency, so it passes DC and blocks fast signals:

`|Z| = 2π f L`

where |Z|: coil’s impedance, its size (Ω); f: frequency (Hz); L: inductance (H).

`= 2π × 10 MHz × 22 µH = 1,382 Ω`

That should make a 22 µH coil a wall at 10 MHz, and Coilcraft’s model of the real part gives 1,451 Ω there. But every coil has some capacitance across its winding. At its self-resonant frequency the two cancel:

`f_SRF = 1/(2π √(L C_p))`

where f_SRF: self-resonant frequency (Hz); L: inductance (H); C_p: capacitance across the coil’s winding (F).

`= 1/(2π √(19.7 µH × 1.72 pF)) = 27.3 MHz`

The model’s inductance sags a little as frequency rises, to the value above at resonance. Past that point the coil acts like a capacitor, and its impedance falls. By 3 GHz the 22 µH coil is down to 31 Ω.

**Fig. 27c: One 22 µH coil.**

![Fig. 27c: One 22 µH coil](https://nearfieldnotes.com/notes/car-camera-coax/fig-27c.png)

*What the figure shows:* Figure 27c: impedance against frequency from 1 MHz to 10 GHz for one 22 µH coil. An ideal coil's impedance, 2πfL, climbs in a straight line. The real coil follows it up to its self-resonance at 27.3 MHz, peaks, and falls, down to 31 Ω at 3 GHz. It holds 1 kΩ only from 7.3 MHz to 100 MHz.

The board’s 22 µH coil against a perfect one. The black line is UG-2212’s recommended 1 kΩ, drawn over its band, 2 MHz to 3 GHz. The circle is the self-resonance.

**The numbers**

*Coilcraft’s model of the coil, fitted to its curve for the whole filter. No DC current through it.*

| Frequency | Ideal, 2πfL | 22 µH coil |
|---|---|---|
| 1 MHz | 138 Ω | 132 Ω |
| 2 MHz | 276 Ω | 262 Ω |
| 10 MHz | 1,382 Ω | 1,451 Ω |
| 30 MHz | 4,147 Ω | 13,711 Ω |
| 100 MHz | 13,820 Ω | 999 Ω |
| 300 MHz | 41,471 Ω | 310 Ω |
| 1 GHz | 138,200 Ω | 92 Ω |
| 3 GHz | 414,706 Ω | 31 Ω |
| 10 GHz | 1,382,000 Ω | 9 Ω |

One big coil

### 1 kΩ over 36% of the band

**Field marks.** The 22 µH coil holds 1 kΩ or more from 7.3 MHz to 100 MHz. That is 36% of 2 MHz to 3 GHz, on a log scale. It shows 262 Ω at 2 MHz and 31 Ω at 3 GHz.

**Look-alike.** A bigger coil. More inductance looks like more blocking, but past 27.3 MHz the 22 µH coil is a capacitor. At 3 GHz it shows 31 Ω, where a 470 nH coil still shows 439 Ω.

**Range.** Self-resonance falls as inductance grows: 676 MHz for the 470 nH coil, 119 MHz for the 6.8 µH and 27.3 MHz for the 22 µH. Coilcraft’s data sheet gives the 22 µH part 21 MHz, typical.

**Habitat.** Every power filter on a camera coax. ADI’s guide has dropped ferrite beads for the job: “Analog Devices is no longer recommending ferrite beads for PoC designs due to their performance sensitivity to increases in temperature and current.”

## Four coils, smallest nearest the line

*Each strong where the others are weak*

No one coil covers 2 MHz to 3 GHz, so the board stacks three kinds. Each peaks at its own self-resonance: the 22 µH coil at 27.3 MHz, the 6.8 µH at 119 MHz and the 470 nH at 676 MHz.

Below its peak each acts like a coil. Above it, each acts like a capacitor.

ADI’s data sheet asks for the smallest coil, the one with the highest self-resonance, on the signal trace. The board puts both 470 nH coils there, then the 6.8 µH, then the 22 µH at the 12 V end.

**Fig. 27d: Three coils, each alone.**

![Fig. 27d: Three coils, each alone](https://nearfieldnotes.com/notes/car-camera-coax/fig-27d.png)

*What the figure shows:* Figure 27d: impedance against frequency for each of the filter's three coil types alone. Each peaks at its self-resonance: the 22 µH at 27.3 MHz, the 6.8 µH at 119 MHz and the 470 nH at 676 MHz. Below its peak each climbs like a coil, above it each falls like a capacitor. None holds 1 kΩ across the band.

Coilcraft’s models of the three parts, each alone with nothing across it. The circle on each is its self-resonance.

**The numbers**

*Coilcraft’s models of the three parts on the board, each coil alone, no resistor across it.*

| Frequency | 22 µH | 6.8 µH | 470 nH |
|---|---|---|---|
| 1 MHz | 132 Ω | 48 Ω | 3 Ω |
| 2 MHz | 262 Ω | 93 Ω | 6 Ω |
| 10 MHz | 1,451 Ω | 442 Ω | 26 Ω |
| 30 MHz | 13,711 Ω | 1,307 Ω | 75 Ω |
| 100 MHz | 999 Ω | 6,234 Ω | 234 Ω |
| 300 MHz | 310 Ω | 2,317 Ω | 687 Ω |
| 1 GHz | 92 Ω | 603 Ω | 1,228 Ω |
| 3 GHz | 31 Ω | 201 Ω | 439 Ω |
| 10 GHz | 9 Ω | 67 Ω | 195 Ω |
| Self-resonance | 27.3 MHz | 119 MHz | 676 MHz |
| Share of 2 MHz to 3 GHz at 1 kΩ or more | 36% | 45% | 14% |

In series their impedances add, as complex numbers:

`Z_PoC = Z_L6 + Z_L5 + (Z_L4 ∥ R) + (Z_L3 ∥ R)`

where Z_PoC: power filter’s impedance, seen from the tap (Ω); Z_L6: impedance of L6, 470 nH (Ω); Z_L5: impedance of L5, 470 nH (Ω); Z_L4: impedance of L4, 6.8 µH (Ω); R: damping resistor across a coil, 5.1 kΩ (Ω); Z_L3: impedance of L3, 22 µH (Ω).

`= (0.5 + j5.5) + (0.5 + j5.5) + (10.7 + j92.5) + (26.7 + j259.6) = 38.4 + j363.1 Ω → 365 Ω`

That is 2 MHz, the bottom of ADI’s band. The two 470 nH coils add almost nothing there, and the sum stops at 365 Ω. At 3 GHz the big coils have turned into small capacitors, and the two 470 nH coils supply most of the 1,098 Ω.

The 5.1 kΩ resistors across the two big coils flatten their peaks. The 6.8 µH coil does more than fill a gap. Take it out of a filter with one 470 nH coil and the two left resonate together, dipping to 251 Ω at 197 MHz.

**Fig. 27e: The board’s filter.**

![Fig. 27e: The board’s filter](https://nearfieldnotes.com/notes/car-camera-coax/fig-27e.png)

*What the figure shows:* Figure 27e: impedance against frequency for the board's whole filter, four coils in series with 5.1 kΩ across the two big ones. It holds 1 kΩ or more from 5.6 MHz to past 3 GHz, 86% of the band, and is 365 Ω at 2 MHz. Coilcraft's own curve for the same filter runs close to it. The 22 µH coil alone is drawn for comparison.

The whole filter as seen from the tap, against Coilcraft’s published curve for the same four parts. The black line is the 1 kΩ band. The circle is 2 MHz.

**The numbers**

*Coilcraft’s curve is read off its SMD-POC-006 sheet. The model is fitted to it, 8% rms.*

| Frequency | Four coils | Coilcraft’s curve | 22 µH alone |
|---|---|---|---|
| 1 MHz | 185 Ω |  | 132 Ω |
| 2 MHz | 365 Ω | 364 Ω | 262 Ω |
| 10 MHz | 1,834 Ω | 1,854 Ω | 1,451 Ω |
| 30 MHz | 4,556 Ω | 4,474 Ω | 13,711 Ω |
| 100 MHz | 3,207 Ω | 3,245 Ω | 999 Ω |
| 300 MHz | 2,184 Ω | 2,294 Ω | 310 Ω |
| 1 GHz | 2,988 Ω | 3,059 Ω | 92 Ω |
| 3 GHz | 1,098 Ω | 1,227 Ω | 31 Ω |
| 10 GHz | 465 Ω |  | 9 Ω |

The four-coil filter

### 1 kΩ over 86% of the band

**Field marks.** From the tap the filter holds 1 kΩ or more from 5.6 MHz to past 3 GHz. That is 86% of the band, against 36% for the 22 µH coil alone. It shows 365 Ω at 2 MHz and 1,098 Ω at 3 GHz.

**Look-alike.** The 1 kΩ rule. ADI’s guide recommends 1 kΩ from 2 MHz to 3 GHz and lists this exact filter for new designs. Coilcraft’s own curve for it shows 364 Ω at 2 MHz and reaches 1 kΩ at 5.6 MHz. ADI’s echo limit is loose down there. At 2 MHz it asks for the echo to be only 7.8 dB down, and the whole link’s echo is 23.1 dB down.

**Range.** Coilcraft’s curve holds 1 kΩ from 5.6 MHz to 3.6 GHz. The model here follows it to 8% rms. Coilcraft rates the filter for 8 MHz to 3 GHz at 0.6 A, and ADI rates it for 1,000 mA at 85 °C and 600 mA at 125 °C.

**Habitat.** Both of Antmicro’s boards, on the camera’s line and on each of the computer end’s four. On both, the tap sits about 12 mm from the connector.

## Control data comes back on the same wire

*Both ways at once*

The video flows out of the camera at 6 Gb/s. The control data flows in at 187.5 Mb/s, on the same center conductor, at the same time.

ADI’s channel spec plans the control data from 2 MHz to 400 MHz and the video from 100 MHz to 3.5 GHz. It says so plainly: “The energy of the forward and reverse channel overlaps in frequency.” So a filter can’t pull them apart. At the camera’s pin, its own signal is about twice the size of the control data arriving.

Both GMSL2 chips carry “an echo cancellation circuit” for this. Each one knows what it is sending. ADI publishes nothing on how its canceler is built. A textbook canceler learns how its own signal looks at the pin, [reflections](https://nearfieldnotes.com/glossary/#g-reflection) included, from the bits it sent. It subtracts that copy, and the other end’s signal is what’s left.

**Fig. 27f: The control data’s eye at the camera.**

![Fig. 27f: The control data’s eye at the camera](https://nearfieldnotes.com/notes/car-camera-coax/fig-27f.png)

*What the figure shows:* Figure 27f: three eye diagrams of the control data at the camera after its low-pass filter, each two control bits wide. Alone, with no video on the wire, the eye is open by 161 mV. With the video's own signal on the wire and only the low-pass to remove it, the eye is shut. After the camera subtracts its learned copy of its own signal, then low-passes, the eye is open by 142 mV.

After the camera’s low-pass filter, two control bits wide. Alone has no video on the wire. Low-pass is everything on the pin, filtered. Canceled is the pin minus the camera’s learned copy of its own signal, then filtered.

**The numbers**

*Over the last 1200 control bits of 2400, sampled where the control data alone opens widest.*

| At the camera’s receiver | Eye opening |
|---|---|
| Control data alone | 161 mV |
| Everything, low-passed | Shut, 77 mV overlap |
| Own signal subtracted, then low-passed | 142 mV |

Echo canceling

### 142 mV of 161 mV

**Field marks.** With no video on the wire, the control data’s [eye](https://nearfieldnotes.com/glossary/#g-eye-diagram) opens 161 mV. With the video on and only the low-pass to remove it, the eye shuts. Subtract the learned copy first, and it opens 142 mV, 89% of the clean eye.

**Look-alike.** A low-pass filter. The control data is slow, so filtering out the fast video looks like enough. In the model the low-pass alone leaves the eye shut. ADI’s channel spec gives the reason: the two channels’ energy overlaps in frequency.

**Range.** After 12 µs of learning, the camera’s own signal left in the control band is 26 dB below where it started.

**Habitat.** GMSL2 and GMSL3, at both ends: ADI’s GMSL3 serializer data sheet repeats the same sentence. TI’s training says FPD-Link uses “a proprietary echo cancelation technique” to send both ways over one conductor.

## Three ways to share a wire

*Split, cancel or take turns*

Canceling is one of three ways links send control data back. The oldest splits the wire by frequency. TI’s DS90UB913A and 914A keep their control channel near 1 to 5 MHz and their video near 70 to 700 MHz, “because they occupy different spaces in the frequency domain.”

The price is a slow back channel. ADI’s GMSL1 serializer, the MAX96705, sends control data as “a series of 1μs pulses, with a maximum raw data rate of 1Mbps.” ADI’s guide asks GMSL1 power filters to be checked from 50 kHz, not 2 MHz.

The third way takes turns. ASA Motion Link’s physical layer “is based on Time Division Duplexing,” and “Downstream and Upstream traffic does not overlap on the wire.” Each 27.38 µs cycle sends video for 25.92 µs and listens for 0.86 µs, which leaves 0.60 µs for gaps.

Nothing overlaps, so there is nothing to cancel. The video has to make up the time it loses, so the line runs faster than the video:

`r_line/r_video = T_cycle/T_down`

where r_line: rate the link runs at while it sends video (Gb/s); r_video: video rate (Gb/s); T_cycle: one send-and-listen cycle (µs); T_down: time in each cycle spent sending video (µs).

`= (27.38 µs)/(25.92 µs) = 1.056`

That is at least 5.6% faster. Each end also holds the video that waits while the link listens. For 8 Gb/s of video:

`B = r_video (T_cycle − T_down)`

where B: video each end holds while the link listens (bit); r_video: video rate (Gb/s); T_cycle: one send-and-listen cycle (µs); T_down: time in each cycle spent sending video (µs).

`= 8 Gb/s × 1.46 µs = 11,680 bits = 1.46 kB`

MIPI publishes no method for A-PHY. A Valens patent, which doesn’t name A-PHY, describes coding the downlink so its low frequencies stay quiet, which “enables band separation between downlink and uplink channels.”

TI publishes nothing on how FPD-Link IV shares its wire. For the DS90UB953, an FPD-Link III part, TI gives the video 20 MHz to 2.1 GHz and the back channel 10 to 50 MHz. The two overlap from 20 to 50 MHz.

**Figure: Back channels, link by link.**

Fig. 27g

| Link | Video | Back | Sharing |
|---|---|---|---|
| GMSL2 | 3 or 6 Gb/s | 187.5 Mb/s | Both at once, each end cancels its echo |
| FPD-Link III, 913A | About 70 to 700 MHz | About 1 to 5 MHz | Split by frequency |
| FPD-Link III, 953 | 4.16 Gb/s | 50 Mb/s | Bands overlap, TI names echo canceling |
| FPD-Link IV, 971 | 7.55 Gb/s | 47.1875 Mb/s | Not published |
| MIPI A-PHY 1.1 | Up to 16 Gb/s | 200 Mb/s | Not published by MIPI |
| ASA Motion Link | 2 to 16 Gb/s | 50 or 100 Mb/s | Takes turns |

From each maker’s data sheet, training or slides. FPD-Link III’s 913A row gives TI’s bands, not rates. The 953’s overlap is worked from its two bands.

## The whole 15 m link

*Pin to pin*

ADI’s channel spec limits the link from pin to pin: camera board, connectors, coax and computer board together. At 6 Gb/s its equation lets the signal lose up to 21.1 dB at 3 GHz.

In decibels, the losses add along the link:

`L_link = L_cam + α ℓ + L_comp`

where L_link: loss of the whole link, pin to pin (dB); L_cam: loss of the camera board (dB); α: cable loss per meter (dB/m); ℓ: cable length (m); L_comp: loss of the computer board (dB).

`= 0.87 dB + 0.9 dB/m × 15 m + 1.57 dB = 15.94 dB`

That is the four coils at each end and new coax, which loses 0.9 dB/m at 3 GHz by ADI’s data sheet. The coils count twice. A weak filter loads the line, which costs loss, and it reflects. ADI limits the echo because the spec uses it to judge “the high-speed transmitter energy reflected into the low-speed receiver.”

**Fig. 27h: The whole link.**

![Fig. 27h: The whole link](https://nearfieldnotes.com/notes/car-camera-coax/fig-27h.png)

*What the figure shows:* Figure 27h: the whole link pin to pin, camera board, 15 m of coax and computer board, from 10 MHz to 3.5 GHz. Above, the signal that arrives: with the board's four coils at each end it is 15.9 dB down at 3 GHz, where ADI's limit is 21.1 dB; with one 22 µH coil at each end, 20.6 dB. Below, the echo sent back: the four coils stay under ADI's limit everywhere, the lone coils break it from 576 MHz.

Camera board, 15 m of new coax and computer board, pin to pin, with the same filter at both ends, against the GMSL2 channel spec’s limits for 6 Gb/s. Echoes between the two ends make both curves ripple every 7 MHz. Loss is drawn through the middle of its ripple, and echo along its peaks.

**The numbers**

*ADI’s loss limit is 21.1 dB at 3 GHz. Aged cable is 1.1 dB/m at 3 GHz against 0.9 new. A plus sign is room to spare.*

| Link | Loss at 3 GHz | Closest to the loss limit | Echo against its limit |
|---|---|---|---|
| Four coils, new cable | 15.9 dB | +3.1 dB | +11.5 dB |
| 22 µH alone, new cable | 20.6 dB | +0.2 dB | Over from 576 MHz |
| Four coils, aged cable | 18.9 dB | +2.0 dB | +12.2 dB |
| 22 µH alone, aged cable | 23.7 dB | −3.0 dB | Over from 735 MHz |
| Cable alone, new | 13.5 dB |  |  |
| Cable alone, aged | 16.5 dB |  |  |

The link

### 15.9 dB of 21.1 dB

**Field marks.** With four coils at each end, the video arrives 15.9 dB down at 3 GHz, 5.2 dB inside ADI’s limit. The camera board takes 0.87 dB, the coax 13.5 dB and the computer board 1.57 dB. The echo stays at least 11.5 dB under its limit.

**Look-alike.** Passing on loss. With one 22 µH coil at each end the loss still passes, 20.6 dB against 21.1 dB. The echo breaks ADI’s limit from 576 MHz up.

**Range.** Aged coax loses 1.1 dB/m at 3 GHz, so 15 m costs 16.5 dB. The four-coil link is then 18.9 dB down and still passes. The lone-coil link is 23.7 dB down, 2.6 dB past the limit.

**Habitat.** Long runs at 6 Gb/s. ADI’s MAX96792A data sheet rates GMSL2’s 6 Gb/s mode for 15 m of foam coax at 105 °C.

## Which way back?

*From the chips’ part numbers*

The serializer and deserializer settle it. Their maker’s papers say how control data gets back, or say nothing.

Key to back channels

### How does my camera’s link send control data back?

1. 1a The chips are ADI’s GMSL parts Go to 2
2. 1b Another family Go to 3
3. 2a GMSL2 or GMSL3, such as the MAX96717 Both ways at once, each end cancels its echo
4. 2b GMSL1, such as the MAX96705 A slow control channel, 1 Mb/s at most
5. 3a TI’s FPD-Link III Go to 4
6. 3b Another family Go to 5
7. 4a DS90UB913A or 914A Split by frequency
8. 4b DS90UB953 Overlapping bands, and TI names echo canceling
9. 5a ASA Motion Link Takes turns
10. 5b TI’s FPD-Link IV or MIPI A-PHY Not published

Antmicro’s boards end at 2a. For A-PHY, a Valens patent describes separate bands, but MIPI doesn’t say A-PHY uses it.

## Stack the coils yourself

*Twelve filters, one 22 µH coil*

Choose which coils go between the line and the 22 µH coil, and whether 5.1 kΩ resistors sit across the big ones. The curve is the filter seen from the tap.

The readings run the whole 15 m link with that filter at both ends, as in Fig. 27h.

**Note 27’s explorer: The filter from the tap.**

![Note 27’s explorer: The filter from the tap](https://nearfieldnotes.com/notes/car-camera-coax/fig-explore.png)

*What the figure shows:* Impedance against frequency of the power filter as built from the chosen coils, with the 22 µH coil always in it, against UG-2212's 1 kΩ from 2 MHz to 3 GHz. The circle marks the filter's lowest point in that band.

Coilcraft’s models, as in Fig. 27e. The circle is the filter’s lowest point between 2 MHz and 3 GHz. Loss and echo are judged against ADI’s limits, with new coax.

**The numbers**

*Every filter has the 22 µH coil. The whole link uses the same filter at both ends and new cable.*

| 470 nH coils, 6.8 µH, resistors | At 1 kΩ or more | Lowest in the band | Loss at 3 GHz | Echo |
|---|---|---|---|---|
| Two, with 6.8 µH, no 5.1 kΩ | 86% | 366 Ω at 2 MHz | 15.9 dB | Inside |
| Two, with 6.8 µH, with 5.1 kΩ | 86% | 365 Ω at 2 MHz | 15.9 dB | Inside |
| One, with 6.8 µH, no 5.1 kΩ | 79% | 361 Ω at 2 MHz | 16.0 dB | Inside |
| One, with 6.8 µH, with 5.1 kΩ | 79% | 360 Ω at 2 MHz | 16.0 dB | Inside |
| None, with 6.8 µH, no 5.1 kΩ | 66% | 231 Ω at 3 GHz | 16.2 dB | Inside |
| None, with 6.8 µH, with 5.1 kΩ | 65% | 230 Ω at 3 GHz | 16.3 dB | Inside |
| Two, no 6.8 µH, no 5.1 kΩ | 64% | 273 Ω at 2 MHz | 16.0 dB | Inside |
| Two, no 6.8 µH, with 5.1 kΩ | 64% | 272 Ω at 2 MHz | 16.0 dB | Inside |
| One, no 6.8 µH, no 5.1 kΩ | 48% | 251 Ω at 200 MHz | 16.2 dB | Over from 165 MHz |
| One, no 6.8 µH, with 5.1 kΩ | 48% | 267 Ω at 2 MHz | 16.2 dB | Over from 192 MHz |
| None, no 6.8 µH, no 5.1 kΩ | 36% | 31 Ω at 3 GHz | 20.6 dB | Over from 576 MHz |
| None, no 6.8 µH, with 5.1 kΩ | 35% | 31 Ω at 3 GHz | 20.7 dB | Over from 576 MHz |

#### Conditions

- **Boards:** Antmicro’s GMSL Serializer rev 1.1.1 and GMSL Deserializer rev 1.0.3, Apache-2.0, from their KiCad files: the camera end and channel A of the computer end. The renders are Antmicro’s, with makers’ marks painted out.
- **Coils:** Coilcraft’s lumped models of the PFL1609-471, 1210POC-682 and MSS6132-223. The MSS6132T-223 on the boards has no published model, so the MSS6132-223 stands in, with its capacitance, one resistor and one loss term refitted to Coilcraft’s SMD-POC-006 curve. The PFL1609 model stops at 1 GHz, so 0.07 pF across it was fitted above that. The whole filter matches the curve to 8% rms from 1 MHz to 10 GHz. No DC current, ideal resistors, and the 12 V side an AC ground.
- **Link:** A cascade of each board’s trace, as a line with the impedance and delay from Antmicro’s own field-solver results, the 100 nF capacitor, the ESD parts’ capacitance, a stand-in for each FAKRA connector, and 15 m of coax. The coax’s loss has a part that grows with the square root of frequency and a part that grows with frequency, fitted to 0.9 dB/m at 3 GHz new and 1.1 dB/m aged. Limits from Equations 3 and 4 of the GMSL2 channel spec.
- **Camera’s pin:** Random 6 Gb/s bits launched at 400 mV with 50 ps edges, and 187.5 Mb/s control bits launched at 250 mV with 2.3 ns edges, the data sheets’ typical figures. The camera hears its own signal, a −22 dB reflection near its connector and a −30 dB one from the far end. The canceler is a textbook LMS design spanning 48 video bits at eight phases, learning over 2,400 control bits, 12.8 µs. The receiver’s low-pass is second order at 150 MHz.
- **Opening:** 20 ns near the end of that run, once the canceler has learned.
- **Sums:** ASA Motion Link’s line rate and buffer above are worked from Kang and Gorshe’s burst timing. The gaps’ total is what the cycle leaves after the two bursts.

Simulated only. Nothing on this page has been measured on a board yet.

#### Limits

- ADI publishes nothing on its canceler. The one here is a textbook design, so Fig. 27f shows what canceling can do, not what the MAX96717 does.
- Random bits stand in for GMSL2’s 9b10b code, which is DC-balanced and puts less of the video’s power near DC.
- The 22 µH stand-in resonates at 27.3 MHz, against 21 MHz typical on the MSS6132T-223’s data sheet.
- DC current lowers a coil’s inductance and moves its resonance. ADI’s rule of thumb lets saturation current cut it by no more than 10%.
- The FAKRA connectors are stand-ins, not field-solved.
- The video’s eye at 6 Gb/s isn’t shown. Its loss and echo are.

#### Files

- sim/site.py Every curve and number on this page, from the three sims below
- sim/poc.py The coils’ models, fitted to Coilcraft’s curve for the whole filter
- sim/channel.py The whole link’s loss and echo, and ADI’s limits
- sim/echo.py The camera’s pin and the echo canceler
- post/site_boards.py The board renders in Fig. 27, makers’ marks painted out
- runs/layout/Both boards’ copper, read from their KiCad files

#### Sources

- [Antmicro, GMSL Serializer, the camera end’s design](https://github.com/antmicro/gmsl-serializer)
- [Antmicro, GMSL Deserializer, the computer end’s design](https://github.com/antmicro/gmsl-deserializer)
- [ADI, GMSL2 channel specification, its bands, limits and echo](https://www.analog.com/media/en/technical-documentation/user-guides/gmsl2-channel-specification-user-guide.pdf)
- [ADI, UG-2212 GMSL2 hardware design guide, its 1 kΩ, filter 14 and ferrite beads](https://www.analog.com/media/en/technical-documentation/user-guides/gmsl2-hardware-design-guide.pdf)
- [ADI, MAX96717F data sheet, its echo canceler, cable loss and coil order](https://www.farnell.com/datasheets/3812033.pdf)
- [ADI, MAX96716F data sheet, its control data’s swing and edges](https://www.analog.com/media/en/technical-documentation/data-sheets/max96716f.pdf)
- [ADI, MAX96792A data sheet, its 15 m at 6 Gb/s](https://www.analog.com/media/en/technical-documentation/data-sheets/max96792a.pdf)
- [ADI, MAX96793 data sheet, GMSL3’s echo canceler](https://www.analog.com/media/en/technical-documentation/data-sheets/max96793.pdf)
- [ADI, MAX96705 data sheet, GMSL1’s control channel](https://www.analog.com/media/en/technical-documentation/data-sheets/MAX96705.pdf)
- [Coilcraft, Document 1560, SMD-POC-006 and its impedance curve](https://www.coilcraft.com/getmedia/e758a227-f461-4472-a019-c08100514698/doc1560_SMD-POC-006.pdf)
- [Coilcraft, MSS6132T data sheet, its self-resonance](https://www.coilcraft.com/pdfs/mss6132t.pdf)
- Coilcraft SPICE models: [PFL1609](https://www.coilcraft.com/getmedia/369e42d0-504a-4d6b-af80-cd4f9c929916/spice_pfl1609.pdf), [1210POC](https://www.coilcraft.com/getmedia/af27a66a-21f1-4562-a79e-8f06dbb0779c/spice_1210poc.pdf), [MSS6132](https://www.coilcraft.com/getmedia/4467fc62-6bf5-4ee0-94b3-2a8168442d01/spice_mss6132.pdf)
- [TI, Analog Applications Journal SLYT581, the 913A and 914A bands](https://www.ti.com/lit/an/slyt581/slyt581.pdf)
- [TI, DS90UB953-Q1 data sheet, its bands and back channel](https://www.ti.com/lit/ds/symlink/ds90ub953-q1.pdf)
- [TI, DS90UB971-Q1 data sheet, its rates](https://www.ti.com/lit/ds/symlink/ds90ub971-q1.pdf)
- [TI Precision Labs, What is FPD-Link?, its echo cancelation](https://www.ti.com/video/6094247787001)
- [MIPI, DevCon 2022 PHY roadmap, A-PHY’s rates](https://www.mipi.org/hubfs/DevCon-2022/2022-MIPI-DevCon-PHY-Roadmap.pdf)
- [Valens, US patent 11,038,726 B2, band separation](https://patents.google.com/patent/US11038726B2)
- [ASA, Motion Link tutorial, TDD and its speed grades](https://auto-serdes.org/wp-content/uploads/2022/08/ASA-tutorial-Sep-2021.pdf)
- [Kang and Gorshe, IEEE 802.3dm, ASA Motion Link’s burst timing](https://grouper.ieee.org/groups/802/3/dm/public/0724/kang_3dm_01b_2407.pdf)

What should the next note be? [Ask for one](https://nearfieldnotes.com/#ask-for-a-note)

---

Source: https://nearfieldnotes.com/notes/car-camera-coax/

Cite as: Michael Walker, “Car camera coax”, Near Field Notes, Note 27, Rev 1, 8 Nov 2026. https://nearfieldnotes.com/notes/car-camera-coax/

Figures: free to reuse under CC BY 4.0 (https://nearfieldnotes.com/license/). Each one's image is linked under its name. Credit Michael Walker, Near Field Notes, and link the figure: this address plus #fig- and its number, like https://nearfieldnotes.com/notes/car-camera-coax/#fig-27.