PIC18FXX8 CANbus conflict

dgi · Joined: 07 Jan 2004 Posts: 11 Location: Philadelphia, USA

I have been working with the PIC18F258, and have encountered intermittent "noise" on the CAN bus which I think is caused by the PIC.

I find out about this from my PC utility program "PCANView USB" which goes "BUSHEAVY" and "BUSOFF" occasionally. Simultaneously, the PIC TXERR and RXERR counters go crazy- count up and down indicating that there is a problem with my physical layer- the signals are not clean and messages are received with many errors.

For a while I didn't know what caused this. But now I have demonstrated that writing to other pins on portB at high frequency causes this problem. For example, executing output_low(PIN_B0) at 10Khz causes noise on the CAN bus. This happens even with FAST_IO, so I don't think it has anything to do with the TRIS registers. And it happens even when the pin's output level is not changed- it is only necessary to write the same thing over and over again, and my CANbus is disabled. At 500 Hz, there is only a little noise detected in the RXERR and TXERR counters. But at 10Khz or 20 KHz, my CAN adapter repeatedly goes BUSOFF. This happens when writing to B0, B1, B4, B5, B6, B7, but not portA or portC.

Anybody else experience the same thing, or have any ideas for what I'm doing wrong or workarounds?

Thanks

DGI.

PCM programmer · Joined: 06 Sep 2003 Posts: 21708

We're using 18F458 on two different boards, and we have
not seen this problem.
On one board, the PIC is running at +5v, so we used the SN65HVD251
driver chip.

On our other board, which runs at +3.3v, we used the SN65HVD231D
driver chip.

I would look closely at grounding, power supply decoupling, etc.

kender · Joined: 09 Aug 2004 Posts: 768 Location: Silicon Valley

Folks,

Will +3.3V and +5V nodes interoperate on the same CAN bus? Are there any caveats? The bus is mechanically small and the cables are shielded.

Cheers!

sjbaxter · Joined: 26 Jan 2006 Posts: 141 Location: Cheshire, UK

It doesn't matter what supply rail you use for the processor and its peripherals, the CAN bus voltage levels are determined by the tranceiver used. Which is usually between 2 and 3 volts differential between CAN_H and CAN_L.

As long as the tranceivers are of a 'common type' accross the network you will have no problems. By 'common type' I mean low speed, high speed, fault tollerant, etc. See the different specs for 82C250, 82C251, TJA1050, etc.
_________________
Regards,
Simon.

kender · Joined: 09 Aug 2004 Posts: 768 Location: Silicon Valley

PCM programmer · Joined: 06 Sep 2003 Posts: 21708

T.I. implies that their transceivers can do it.
http://focus.ti.com/docs/prod/folders/print/sn65hvd231.html
If you download the data sheet on the page above, near the end of it
they describe how it's important that +5v and +3.3v drivers be
interoperable. They show waveforms, but they don't ever actually
draw a conclusion and say "Thus, both types of drivers are interoperable".
That annoys me.

They also have a CAN bus demo kit.
http://focus.ti.com/docs/toolsw/folders/print/canbus-demo.html

The kit consists of 3 boards, as shown in this photo.
http://focus.ti.com/graphics/tool/canbus-demo_400.jpg

The block diagram shows two +3.3v drivers and one +5v driver,
all on the same bus.
http://focus.ti.com/graphics/tool/can_bus_blockdiagram_4.jpg

In the list of features, they say that one of the purposes of the
kit is to demonstrate interoperability between the two types of drivers: