8

I have 2 questions:

  1. Why using a common-mode choke on serial communication lines (like CAN, RS485) if the signal is differential? Wouldn't the common-mode noise simply be cancelled in the receiver?
  2. I reviewed a project where the CAN had a common-mode choke, but the RS485 did not. Is there a technical reason for that?

Thank you

user
  • 83
  • 5
  • 5
    The common-mode choke also helps to stop the bus radiating. Maybe the edge rate on the RS485 bus was slower? – Kartman Sep 22 '21 at 12:27
  • @Kartman How does that work? because an ideal common-mode choke (coupling k=1) would be like a simple wire for differential signals, so how could it atenuate the higher frequency components of the signal? – user Oct 01 '21 at 14:43
  • 1
    The radiating signals are not necessarily differential. There could be ringing etc. the cm choke works both ways - for external noise coming in and for internal noise going out. – Kartman Oct 01 '21 at 19:56
  • @Kartman But isn't the signal itself the main cause of radiation? I mean, due to the fast rising rates of the edges (as you mentioned)? Because otherwise I don't understand why adding the CM choke only on the CAN, as external signals coming in are similar for both interfaces. – user Oct 02 '21 at 19:26

4 Answers4

10

Wouldn't the common-mode noise simply be cancelled in the receiver?

A lot of noise would be heavily cancelled in a differential receiver but, high frequency stuff can cause problems and, of course a CM choke acts like an inductor and hence increases its series impedance with frequency. High frequency noise beyond the bandwidth limitations of the receiver can still wreak havoc.

I reviewed a project where the CAN had a common-mode choke, but the RS485 did not. Is there a technical reason for that?

There may be but, the devil is in the detail (and the performance testing and the grounding of this and that and what spec it's being tested to and the data bandwidth and the chips in question.......)

Andy aka
  • 434,556
  • 28
  • 351
  • 777
  • What's your take on modern transceivers like [MCP2561FD](https://www.microchip.com/en-us/product/MCP2561FD) using a "SPLIT" output pin for the termination (see page 6), supposedly for common mode stabilization? Reading the datasheet they seem mostly concerned about radiated emissions at higher baudrates (CAN FD) and not so much about radiated/conducted susceptibility. – Lundin Sep 22 '21 at 14:07
  • 1
    The link and the data sheet don't really say much about it. The data sheet covers two versions; one with and one without so there's not much to go on really @Lundin - maybe raise it as a new question? – Andy aka Sep 22 '21 at 15:44
  • I posted a question about this pin over at Codidact here: [CAN "split" pin, bus termination and common mode stabilization](https://electrical.codidact.com/posts/284323). – Lundin Sep 24 '21 at 09:22
7

The differential receiver has a limited range of common mode rejection, on the order of volts. It cannot remove common-mode noise peaks whose voltage exceeds that level. The common mode choke transformer increases that range for high frequency pulses, using two techniques:

  1. It converts some of the high frequency noise (whether common mode or differential) to heat (unlike a transformer, the ferrite in a common-mode choke is purposely lossy)
  2. It subtracts the common mode signal (as a coupled inductor) allowing 10's of volts of high frequency noise difference between the two grounds at the transmitter and the receiver

In answer to your question on why you saw a common mode choke in a CAN bus port but not in a RS-485 port: that's just a decision that those particular engineers made. It has more to do with economics and little to do with differences between CAN bus and RS-485.

Davide Andrea
  • 16,164
  • 4
  • 33
  • 62
4

A common-mode filter will block noise induced on the pair (and its reference ground) from making its way into your system, even if the differential receiver rejects it. In other words, the CM filter prevents EMI/ESD related system upsets in the presence of strong noise.

Even Ethernet, which not only uses differential signaling but also transformer isolation, can benefit from suppressing common-mode noise.

As to why it was applied to CAN bus and not RS-485 in that particular system, I suspect that the designer felt that the places the CAN bus connected to were more likely to have strong electromagnetic noise (like from ignition, fuel injection solenoids, and other power train sources) than the places RS-485 went (in-cabin only.)

hacktastical
  • 49,832
  • 2
  • 47
  • 138
2

Wouldn't the common-mode noise simply be cancelled in the receiver?

Yes, the advantage of differential signaling is to eliminate the "common mode noise" appearing at the receiving end, by "differentiating" the signal:

Tx:
Sig+ = Sig + Noise
Sig- = -Sig + Noise

Rx:
Sig = Sig+ - Sig- = (Sig + Noise) - (-Sig + Noise) => Sig'

schematic

simulate this circuit – Schematic created using CircuitLab

CAN had a common-mode choke, but the RS485 did not

It is specific to the system (design dependency) and applications.

jay
  • 3,781
  • 8
  • 24