Fwd: Fwd: CAN HW compability

Thank you Daniel for your very detailed response. It makes very much sense.

Looking at the MCP2551 transciever it reads: "The MCP2551 is a high-speed
CAN, fault-tolerant device"

Which is a little confusing sense it seems it can't be both high-speed  and
fault tolerant, but reading a little more its seems (to me) the meaning is
the transciever is hight speed (ISO 11898-2) with fault tolerant
capabilities that are not related directly to the ISO specifications (e.g.
Detection of ground fault)

Anyway, the MCP2551 has a pin to pin replacement called the MCP2561 which
is ISO-11898-5 compatible. I know it is not a fault tolerant transciver
(i.e. ISO 11898-3) but i am wondering if it is a "little more" compatible

Also, if i were to summarise your given options:

   1. Replace the transciever to support FT CAN (ISO 11898-3) and try to
   figure out the correct termination to be used
   2. Use the highest termination and shortest cables as possible and hope
   for the best
      1. Actually it is exactly what i have done. i used the no termination
      (the natural transciever termination is about 47kohm) and
shortest cables.
      As mentioned in previous mail it seems to work but the signal looks very
      bad via the osciloscope

If there are any other notes i will be happy to hear

Best Regards

Job


2015-05-08 12:52 GMT+03:00 Bram Kerkhof <[email protected]>:

> That’s because you are probably working on a fault-tolerant CAN network
> using FTCAN transceivers where the termination is distributed over the
> network. Each node has its own internal switchable termination (depending
> on the active/error state of the transceiver), with the values calculated
> so that the total termination value is around 100 ohm for the bus. Because
> the termination is switched, you cannot measure it when the nodes are
> unpowered (which is why you don’t see it).
>
>
>
> If that is the case, you are using the wrong transceivers. You are
> probably knee-deep in TJA1054 territory.
>
>
>
> Secondly, if you get the correct transceivers you still need to figure out
> the termination. Internal termination for FTCAN networks with a variable
> number of nodes is calculated in such a way that the nodes that are
> required (i.e.: always connected to the bus) have relatively low internal
> termination values so they contribute most of the overall bus resistance,
> while nodes that are optional have higher values so the impact of their
> absence on overall termination is limited. There are no general rules on
> how to distribute the termination values over the network, although they
> are generally lower for nodes on longer cable stubs. You want the bus to be
> as close as possible to 100 ohms, but 120 ohms (due to absent optional
> nodes) will do just fine. It would not surprise me that every OEM has their
> own “secret sauce” for FTCAN transceiver/termination requirements.
>
>
>
> Thirdly, most transceivers have minimal termination values they can
> support, so an FT CAN network with only two nodes (the side of the parking
> assistance module) may require some creative tinkering to get it to work,
> especially if the other node has a relatively high termination due to the
> fact that it is an optional extra in the design.
>
>
>
> It’s most likely going to be relatively easy to switch out the parking
> assist module for something of your own making: either pick a high-enough
> resistance value so your impact on the already balanced bus is negligible
> (I would venture 250 to 300 ohms times the total number of supported nodes
> in the network), or open up the module that you want to replace and figure
> it out using a magnifier and a good cup of coffee.
>
> For the side of the parking assist module, pick the lowest resistance your
> transceiver can support, keep the bus as short as possible and hope for the
> best. Bus resistance will be far above the ideal 100 ohms, but for really
> short runs it may just work. Another solution might be that you add a
> separate (non-switched) terminator in the network, but all bets are off how
> it will respond – I would assume that FT operation is out of the question.
>
>
>
> Disclaimer: there are a lot of assumptions in this message, and FTCAN can
> be a tricky beast to conquer. Don’t take my words for gospel ;-)
>
>
>
> cheers,
>
> Bram
>
>
>
> *From:* [email protected] [mailto:
> [email protected]] *On Behalf Of *Job Race
> *Sent:* vrijdag 8 mei 2015 9:14
> *To:* [email protected]
> *Cc:* [email protected]
> *Subject:* Re: [CANLIST] Fwd: CAN HW compability
>
>
>
> Hi Daniel
>
> Thank you for your answer
>
> BTW by "Sensors" i mean "Parking asistance module". sorry.
>
>
>
> Adding termination was my first response but:
>
> 1. Adding any resistor between CANH & CANL smaller than 200kohm resulted
> in complete bus break down i.e. signal almost straight. as you increse the
> resistance the signal is shorter and shorter until it is gone. not to
> mention two 120 ohm.
>
> 2. The original entire network does not appear to have termination
> resistors. The way i know that is that before i disconnected the parking
> assistance module i disconnected car battery and tested resistance between
> CANH & CANL and it read about 1M ohm (double checked) but as you can see in
> picture 2 the signal looks very good
>
> PS: the MCP2551 has basic (unpowered) 47kohm resistance between CANH & CANL
>
>
>
> Best Regards
>
> Job
>
>
>
>
>
> 2015-05-07 16:12 GMT+03:00 Daniel Perron <[email protected]>:
>
> Hi Job,
>
>
>
> The first picture really show you that you are missing the terminator
> resistor.  High rise and a very slow drop! Also you have the ringing echo
> on the high rise side.
>
>
>
> You need to add a terminator. What is the value you put there?
>
>
>
> By definition  you need terminator on both ends.
>
> Is this sensor at the end of the bus? There is a good chance it isn’t so
> you will need to add a terminator on the sensor also. (2 terminators
> needed!).
>
>
>
>
>
> Daniel
>
>
>
>
>
>
>