Author Topic: CAN Bus: EMI capacitor values  (Read 5261 times)

0 Members and 1 Guest are viewing this topic.

Offline PJ BainTopic starter

  • Contributor
  • Posts: 15
  • Country: au
CAN Bus: EMI capacitor values
« on: May 24, 2019, 02:39:54 am »
Hi all,
Using 3 of the TJA1042 CAN transceivers for a new design in conjunction with a i.MX8. The project managers have requested that the 3 channels be capable of 1Mbps each, with a default of 250kbps. I have all the circuits designed except for the EMI capacitor values. These are 2 matched capacitors in the pF range between CANH and GND, and CANL and GND.
From what I can find, the value of these will change depending on the bps. Given they want the potential full 1Mbps available for use, I assume I use the value suitable for this (as it's the smallest). I am struggling to find documentation on how to calculate the values though. I think it is around 35pF from a couple examples I have found. The TVS I am using has a Cd of 25-30pF so if this is the case, I may not need the capacitors at all.
Any links to a resource which shows how to calculate these values based on the kbps?
Thanks

Peter
 

Offline free_electron

  • Super Contributor
  • ***
  • Posts: 8550
  • Country: us
    • SiliconValleyGarage
Re: CAN Bus: EMI capacitor values
« Reply #1 on: May 24, 2019, 03:05:58 am »
Much will depend on your cabling.
The cables form a capacitor as well. length dependent , node tap dependent ...
Are you using common mode chokes ?
Professional Electron Wrangler.
Any comments, or points of view expressed, are my own and not endorsed , induced or compensated by my employer(s).
 

Offline PJ BainTopic starter

  • Contributor
  • Posts: 15
  • Country: au
Re: CAN Bus: EMI capacitor values
« Reply #2 on: May 24, 2019, 03:10:20 am »
The cabling s something I have no control over unfortunately :( I don't even know if they have designed that yet. I do know they place the 120R termination in the cabling where required. I would have preferred a split termination on the PCB, but this device may not always be at the end of the bus.
No I am not using any common mode chokes.
 

Offline Rerouter

  • Super Contributor
  • ***
  • Posts: 4705
  • Country: au
  • Question Everything... Except This Statement
Re: CAN Bus: EMI capacitor values
« Reply #3 on: May 24, 2019, 05:10:18 am »
A good starting reference, pretty much the "Maximum" is 35pF for 1Mbit, so you need to make sure your values are less than that. again its only a maximum, you can be much lower than this without issue, as the termination resistors in general will chew up most of it as its a differential bus.

Just If you can, make sure both Can High and Can Lows protection devices use a common ground point, as that capacitive shunting is generally going to end up being between the 2 data lines via the center ground point.

To make clear, you never had to add physical capacitors, just control how much capacitance there is in your protection devices/

https://www.onsemi.com/pub/Collateral/AND8169-D.PDF

 

Offline SparkyFX

  • Frequent Contributor
  • **
  • Posts: 676
  • Country: de
Re: CAN Bus: EMI capacitor values
« Reply #4 on: May 24, 2019, 05:57:06 am »
Twisted pair cabling would help... but as you got no control over this, line length drastically limits baud rate on CAN.
The other thing is: as your bit timing registers are software, you usually define when to sample for all nodes on the bus, given you know the line length and physical properties of the whole net. Ringing should not be that much of a problem ... i mean afterall it would need to ring differentially and across the terminators.
Support your local planet.
 

Offline PJ BainTopic starter

  • Contributor
  • Posts: 15
  • Country: au
Re: CAN Bus: EMI capacitor values
« Reply #5 on: May 24, 2019, 06:43:11 am »
Rerouter thanks for that. Trying to find those rules of thumb was proving difficult! If that's the case, the TVS is typical 25pF and max 30pF so I won't need to add any extra then.

SparkyFX, yes the cable is twisted pair, I know that much.
 

Offline MagicSmoker

  • Super Contributor
  • ***
  • Posts: 1408
  • Country: us
Re: CAN Bus: EMI capacitor values
« Reply #6 on: May 24, 2019, 01:01:35 pm »
...The project managers have requested that the 3 channels be capable of 1Mbps each, with a default of 250kbps. ...

I hope these are short cables, because twisted pair for this application typically has a capacitance of 30-35pF per meter.
 

Offline Rerouter

  • Super Contributor
  • ***
  • Posts: 4705
  • Country: au
  • Question Everything... Except This Statement
Re: CAN Bus: EMI capacitor values
« Reply #7 on: May 24, 2019, 01:52:03 pm »
Quote
recommended that the maximum capacitance of the protective network measured from each signal line to ground should be less than 35 pF for 1.0 Mbits/s

The network capacitance can be higher, for a normal 1mbit network it should be less than 40m of wire... with each nodes input capacitance chipping it down a little.

http://www.ti.com/lit/an/slla270/slla270.pdf

 

Offline Niklas

  • Frequent Contributor
  • **
  • Posts: 404
  • Country: se
Re: CAN Bus: EMI capacitor values
« Reply #8 on: May 24, 2019, 10:39:17 pm »
Depending on EMC immunity demands, those capacitors could also help you to pass a BCI test.

With 3 transceivers on the same PCBA, make sure you have enough buffering capacitance close to the supply pin on each of them. A small series resistor for the 5 V supply could also reduce the load step seen by the voltage regulator.

If you cannot control what twisted pair cable that the customer is going to use, at least try to specify the nominal impedance of 120 ohms to reduce common mode emissions.
 

Offline PJ BainTopic starter

  • Contributor
  • Posts: 15
  • Country: au
Re: CAN Bus: EMI capacitor values
« Reply #9 on: June 05, 2019, 11:28:07 pm »
Depending on EMC immunity demands, those capacitors could also help you to pass a BCI test.

With 3 transceivers on the same PCBA, make sure you have enough buffering capacitance close to the supply pin on each of them. A small series resistor for the 5 V supply could also reduce the load step seen by the voltage regulator.

If you cannot control what twisted pair cable that the customer is going to use, at least try to specify the nominal impedance of 120 ohms to reduce common mode emissions.

I have a 0.1uF and a 10uF ceramic on VCC and a 0.1uF on VIO for each transceiver. The resistor isn't a bad idea, I'll look into that one.
 

Offline PJ BainTopic starter

  • Contributor
  • Posts: 15
  • Country: au
Re: CAN Bus: EMI capacitor values
« Reply #10 on: June 05, 2019, 11:32:49 pm »
Much will depend on your cabling.
The cables form a capacitor as well. length dependent , node tap dependent ...
Are you using common mode chokes ?

I wasn't going to use CMF, but now it looks like I am :P Issue is the guys writing the requirements have said the default is 250kbps, but they want the bus capable of the full 1Mbps. I am trying to get them to pick one or the other as from what I can tell, the capacitors and CMF will be different for both.

Any suggestions on a CMF? I'm using a really nice ones for the USB2 and USB3 which have built in ESD protection as well and a flow through design which makes layout super nice.
 

Offline Niklas

  • Frequent Contributor
  • **
  • Posts: 404
  • Country: se
Re: CAN Bus: EMI capacitor values
« Reply #11 on: June 06, 2019, 01:28:41 pm »
A 51 uH CMC from Murata's DLW43SH-series would be a good start. Then you also have several second source alternatives available from Bourns, Epcos/TDK and others.
You don't need to worry too much about different capacitance values for different baudrates, start with 22pF 50-100V NP0/C0G. Also the tolerance on the CMC is quite wide. The edges on the CAN signals from the transceiver are not affected by the baudrate, only the time between them will change. On older transceivers the enable pin also doubled as slew rate control with an external resistor, but nowadays it is only on/off.
 
The following users thanked this post: PJ Bain


Share me

Digg  Facebook  SlashDot  Delicious  Technorati  Twitter  Google  Yahoo
Smf