Oscillator tolerance requirements?

Hi – I seem to be getting conflicting information on the necessary tolerance for a CAN controller’s oscillator frequency. A Silicon Labs datasheet mentions that their 0.5% oscillator is good enough for CAN. I’ve seen some specs say that 0.05% is good enough. It looks like the official CAN spec calls out 0.5%.

 

Which is it?

 

Thanks!

 

-Michael

Hi Michael,

The tolerance requirement is system specific. The better tolerance, the
smaller TSEG2 (or Phase_Seg 1&2), i.e., the longer Prop_Seg can be used.
Remember, the worst oscillator in the system sets the rule for the whole
system. The SJW is also influenced. The better the worst oscillator in
the system is, the shorter SJW. If you have 50 ppm oscillators in the
system, no SJW is needed as you would always stay in sync.
(Unfortunately, SJW=0 means 1 BTQ, so you can never avoid
resynchronization. Thus, in a system with good oscillators, false edges
might lead to false corrections.)

In most HLPs, the bit timing setting and the oscillator tolerance is a
part of the standard, so there is no need to think about their impact on
systems performance. You get what you get and cannot do anything about it.

To use CAN efficiently, the best way is to have a good system designer
and a proprietary HLP. CanKingdom was designed for that purpose and is
used in advanced designs.

Best regards,
Lars-Berno

Michael J. Noone wrote: --
Archives and useful links: http://groups.yahoo.com/group/CANbus
Subscribe and unsubscribe at www.vector-informatik.com/canlist/
Report any problems to <canlist-owner@...>