• State Resolved
  • Locked Locked
  • Replies 7 replies
  • Subscribers 35 subscribers
  • Views 1359 views
  • Users 0 members are here
Support feedback
Options
Options
Related

This thread has been locked.

If you have a related question, please click the "Ask a related question" button in the top right corner. The newly created question will be automatically linked to this question.

Can transceiver troubles

jerome chopard
Prodigy 60 points
Other Parts Discussed in Thread: SN65HVD230

Hi,

I'm using SN65HVD230 CAN transceivers to form a network of sensors. I'm using a baud rate of 100kz and a resistor (R2)

of 50kohm is connected to Rs pin to ensure that transmission will suffer low level of noises. The system works

well for a few days but fail randomly after that and I measure a resistance of a few kilo ohms between Vdd and

CANH (the same measurement read infinite at the beginning of the experiment). If I disconnect power for enough

time everything works again for a few days.

 - my network consists of 16 nodes only

 - each SN65HVD230 is powered (between 3.0 and 3.3V) continuously (I do not use standby mode)

 - I use 6 core flat modular telephone cable for the network (CANH, Vdd, Vdd, Vss, Vss, CANL)

(http://www.altronics.com.au/index.asp?area=item&id=W2313)

 - when not transmitting (recessive) the microchip microcontroller drive D pin high ("CANTX pin will drive VDD

when recessive")

Any help will be greatly appreciated,

Thanks,

Jerome

  • 0
    TI__Mastermind 20590 points

    The schematics and idea's seem okay.  Have you checked that there are no transients or shorts on the bus that might be causing the devices to use the current limiters and also would raise the die temperature thus going into thermal shut down?   Do you have scope plots of the data to ensure there isn't anything un-expected actually going on the bus, ie before and after would be helpful.  Have you changed the CAN transceiver out to see if you may have damaged the unit?  How are you making this measurement between Vcc and CANH? You mention 16 nodes, are they all the same?  How is the network terminated since the cabling in use appears to not have a controlled impedance?  Have you checked for issues with the wiring where shorts could occur?    

    These are a few steps to start debugging. 

     

    -- Scott

  • 0
    Prodigy 20 points

    Jerome, the schemtics do not show any termination resistors accross CANH and CANL.

    A CAN network may work on a lab bench without termination, (short cables) but it really needs a resistive connection between CANH and CANL to operate all the time.

  • 0 in reply to Derek Vanditmars
    Prodigy 60 points

    Sorry, the schematics only show one node with two can ports to enable a daisy chain between node. The network in the field has termination resistors.

  • 0 in reply to Scott Monroe
    Prodigy 60 points

    Thanks for your answer,

    I checked the network after installation before powering the system and there was apparently no shorts. When I do the same measurement after trouble, power removed, multimeter between Vcc and CANH, that's when I find the few kilo ohms resistance. Unfortunately I do not have the capacity to monitor the network in the field.

    The network is composed of 16 identical nodes. Each node has two can ports. they are connected in a daisy chain with termination at the end.

    The problem is recurrent. I tried already 4 times with the same components and every time it stops working after half, one or two days.

    In the datasheet page 6, the maximum rating output current is +-11mA, yet the normal operating condition low level output current is 48 mA. What happen if one of the transceiver stays in dominant mode?

    Thanks

  • 0 in reply to jerome chopard
    TI__Mastermind 28265 points

    What happen if one of the transceiver stays in dominant mode?

    The CAN protocol is based on bus arbitration , that is various drivers have different lengths of dominant states. If for example multiple drivers access the bus at the same time, the driver with the longest dominant state wins the buss access while the other drivers must go recessive.

    If  one driver remains dominant all the time, then something might be wrong within the protocol software or with the specific driver. A consistantly dominant driver will block all communication as it won't relase the bus for any bit transfers.

    Regards,

    Thomas Kugelstadt for Scott Monroe, currently on business travel.

     

  • 0 in reply to Thomas Kugelstadt
    Prodigy 60 points

    Thanks for your answer.

    Is it possible that a transceiver staying in dominant mode overheat?

  • 0 in reply to jerome chopard
    Prodigy 20 points

    Most CAN transceivers have a time-out to prevent staying in dominant mode, and will turn off when they overheat.

    Is there any transient protection, (i.e. transzorb) between the CAN transceiver and the network?

    Although most CAN transceivers have some on-chip ESD and overvoltage protection, this is typically not enough in most non-automotive installations.

    Not sure the exact CAN transceiver you are using, but typically they can handle at least -27 to +40V Bus Fault, and +/-200V Bus transient on the CAN data signals, so adding transzorbs to clamp CAN data signals within these usually fixes CAN transceiver failures in the field.