Author Topic: CAN bus AC termination according to ISO 15765-4  (Read 6133 times)

0 Members and 1 Guest are viewing this topic.

Offline HwAoRrDkTopic starter

  • Super Contributor
  • ***
  • Posts: 1784
  • Country: gb
CAN bus AC termination according to ISO 15765-4
« on: July 08, 2018, 07:47:51 pm »
I am designing a CAN bus device to fit to my own car that will interface with the bus in two ways: in one mode it will passively listen to existing transmissions on the bus, and in the other it will perform some common diagnostic functions (e.g. OBD2). I managed to get my hands on a copy of the ISO 15765-4 standard document, which contains some requirements for external test equipment. Although my device will be permanently connected, that is essentially what it is.

One of those requirements is regarding CAN bus termination. It states that test equipment must not provide the typical 120R parallel termination (as the vehicle's bus will already be terminated appropriately), but also states that test equipment shall provide AC termination on each of the high and low lines, using a resistor and cap in series to ground (see diagram). It recommends a nominal value for the resistors of 100R and 560pF for the caps.



I have so far not come across this implemented on any designs for CAN-bus devices I have been able to inspect (e.g. published open-source stuff, etc.).

The ISO docs don't say why they recommend this AC termination be implemented, so I'm not sure whether I should add it to the design of my device.

Would it be to combat potential disturbances to the bus signals caused by additional cable length used by any test equipment connected? The standard allows for up to 5 metres. My device won't be connected via a spur anywhere near that length - perhaps 50 or 60cm. Also, in attempting to research further information, I came across TI app note AN-903, wherein it is stated that R and C values for AC termination are usually determined according to impedance and round-trip delay - i.e. the cable length. So this also makes me doubt whether it will be needed for my design.

I don't want to blindly add such termination to my design if it's not necessary in my situation, or might in fact inadvertently have a detrimental effect. Should I? I suppose I could always allocate footprints on the board, and then simply not populate them...
 

Offline HwAoRrDkTopic starter

  • Super Contributor
  • ***
  • Posts: 1784
  • Country: gb
Re: CAN bus AC termination according to ISO 15765-4
« Reply #1 on: July 10, 2018, 02:43:17 pm »
Anyone?
 

Offline JPortici

  • Super Contributor
  • ***
  • Posts: 3810
  • Country: it
Re: CAN bus AC termination according to ISO 15765-4
« Reply #2 on: July 10, 2018, 03:50:38 pm »
IIRC in that document it is said that the AC termination is to reduce ringing at the tool's end, because of the lenght of the cable...
so what you said :)

as always, place the components on the board (the 120 ohm termination too, that's useful when testing in a custom test network), don't mount them
« Last Edit: July 10, 2018, 03:52:45 pm by JPortici »
 

Offline HwAoRrDkTopic starter

  • Super Contributor
  • ***
  • Posts: 1784
  • Country: gb
Re: CAN bus AC termination according to ISO 15765-4
« Reply #3 on: July 10, 2018, 06:33:39 pm »
So would you perhaps agree that such termination would be of quite limited utility with a pretty short length cable spur off the main bus?

Also, I have been reading this app note from UltraCAD, and although I don't fully understand it all, it seems to me to imply that if you don't get the C value right for your bus length, it can definitely have a detrimental effect on signals. I wouldn't like to go add termination that's not suitable for my scenario and cause problems.

I'm hesitant to add the components on to the board regardless, as I don't really have enough space near the connector for the CAN bus. By the time I read about this in the ISO document, I'd already done 95% of the layout of the board, and I would have to juggle a few things around to add it now, which would be a pain. :(
 

Offline Zucca

  • Supporter
  • ****
  • Posts: 4867
  • Country: it
  • EE meid in Itali
Re: CAN bus AC termination according to ISO 15765-4
« Reply #4 on: July 23, 2018, 02:43:57 pm »
https://electronics.stackexchange.com/questions/55389/why-does-the-can-bus-use-a-120-ohm-resistor-as-the-terminating-resistor-and-not

Well I know something in the Automotive.

The R is there to discharge the parassitc C in the cable between the differential CAN lines. Without it the signal does not go down quickly enough at the end of the cable causing a drammatic reduction of Bitrate. I watched it myself with a scope. In other words you want nice and sharps edge, and without that R the Voltage falls down like exp(-t).
Yes you could also think about an impedance matching if you want.

Note there must be only TWO termination R on the CAN bus, generally at the beginning and at the end of the cable.

see page 5 here.
http://www.ti.com/lit/an/slla270/slla270.pdf
You can also intergrate the R in the device board if you are sure to connect it in one of the above places.

That's why in automotive development departments you see a lot of those:

used by EE to get the bus running at 500Kbit/s and playing with a lot of devices... connecting and disconnecting cables/devices.

The two C to ground are to filter the common mode or single ended (i.e. ESD) spikes.
BTW in every CAN transceiver the application note is very helpful.

So would you perhaps agree that such termination would be of quite limited utility with a pretty short length cable spur off the main bus?

Grab a scope and look at the edge, if they are sharp and clean don't use the R because the cable is short enough.
« Last Edit: July 23, 2018, 02:51:01 pm by zucca »
Can't know what you don't love. St. Augustine
Can't love what you don't
 


Share me

Digg  Facebook  SlashDot  Delicious  Technorati  Twitter  Google  Yahoo
Smf