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.

ISO1050: Shorted output on ISO1050DW

Part Number: ISO1050
Other Parts Discussed in Thread: ISO1044, ISO1042

Hello,

We use ISO1050DW transceivers on out PCBs for CAN communication. I wanted some insight on what would have to happen internally on the IC for the CANH/CANL pins to get shorted to GND pin on the transceiver IC. Also, we took the transceiver ICs off of the board to inspect which GND pin was getting shorted to the CANH/CANL pins on the IC, we found that pin10 was the one getting shorted with CANH/CANL on the transceiver.

We have 15 nodes on our CAN network and each node uses a ISO1050DW transceiver for CAN communication. Recently, we have seen that the CANH and/or CANL pins on the transceivers on some of the nodes get shorted to the ground pin on the transceiver after continuous operation for a few hours bringing the whole communication network down. On some transceivers, it is not necessarily shorted, but the resistance is very low. On a good transceiver, we see a resistance of around 7 Mega Ohms. On the faulty transceivers we see resistances in Kilo Ohm range or shorts (very low resistance like 10-100 Ohms). We tried to understand reasons as to why this would happen. The working temperature of each node never goes outside the recommended temperature range for ISO1050DW. We don't see this problem on a huge number of transceivers, but we are seeing it more than we would like to.

  • Abhijit Bidwe1,

    I'm sorry to hear you are experiencing some failures in your CAN network. Having the CAN/CANL lines shorted to ground might be because of a transient overvoltage event over the bus. Are ay TVS diodes being used for bus protection? If so, can you share the TVS diode part number so that I may check to make sure it is effective? Could you also share a schematic of the CAN transceiver in this system along with any components (TVS, termination, choke, etc) on the bus?

    Respectfully,

    Lucas Schulte

  • Hi Lucas,

    Thanks for getting back. We do have an ESD protection diode. The part number for that is NUP2105. I have attached a screenshot of the transceiver schematic. 

    Let me know if you have any questions. 

  • Abhijit,

    It looks like the TVS diode being used has a clamping voltage above the absolute max voltage tolerance of the ISO1050 CANH and CANL. This means that voltage could happen on the bus that damages the ISO1050 and the TVS diode wouldn't protect it.

    Below is a link to a document that explains how to properly protect the ISO1050 as well as the ISO1042. Also, the ISO1042 and ISO1044 are our newer CAN transceivers will several improvements made over the ISO1050. 

    Respectfully,

    Lucas 

  • Hi Lucas, 

    Thanks for sending the application note. We changed the TVS diode to the one that is specified in the application document (CPDT-12V) and I have attached a screenshot of the results we got when we scoped the pins of the ISO1050 IC. 

    The blue line is VCC2 (pin 16), the green and the yellow are the CANH and CANL (pin 13 and pin 12) respectively, the red is 12V supply that is converted to 5V (which is applied at VCC2) via a regulator. 

    We still see transients on the CAN lines after the change of diode. Can you help with this?

    Thanks,

    Abhijit

  • Abhijit,

    Unfortunately it doesn't appear that the image came through. Can you try again? Seeing these waveforms will be very crucial for root cause analysis. 

    Respectfully,

    Lucas

  • Hi Lucas, 

    Thanks for sending the application note. We changed the TVS diode to the one that is specified in the application document (CPDT-12V) and I have attached a screenshot of the results we got when we scoped the pins of the ISO1050 IC. 

    The blue line is VCC2 (pin 16), the green and the yellow are the CANH and CANL (pin 13 and pin 12) respectively, the red is 12V supply that is converted to 5V (which is applied at VCC2) via a regulator. 

    We still see transients on the CAN lines after the change of diode. Can you help with this?

    Thanks,

    Abhijit

  • Lucas,

    Here is the screenshot. 

    The blue line is VCC2 (pin 16), the green and the yellow are the CANH and CANL (pin 13 and pin 12) respectively, the red is 12V supply that is converted to 5V (which is applied at VCC2) via a regulator. 

    Thanks,

    Abhijit

  • Abhijit,

    Unfortunately the time scale of the waveforms is much too large to see meaningful switching patterns and transients. Can you provide waveforms in a simillar timescale as the example below? Ideally catching the non ideal transients and behaviors in question.

    Respectfully,

    Lucas

  • Abhijit,

    Getting waveforms zoomed in on the noise will help eliminate the possibility of these transients just simply being artifacts by the measurement system you are using. Also note that the TVS diode won't eliminate transients but reduce them to Vclamp or below. 

    I am marking this thread as resolved for but please feel free to make a new related thread if you need to continue the conversation.

    Respectfully,

    Lucas

  • Hi Lucas,

    I hope this will give you a better representation. 

  • Hi Abhijit:

    Thanks for sharing the additional waveforms.

    As you can see, the waveforms show a single voltage sample that exceeds the expected value.

    12V supply (red): -25V

    Vcc2 rail (blue): 20V

    CAN_H and CAN_L: (green and yellow): 30V

    Do you know under which test conditions you see these large voltage excursions (normal operation, EMC testing, etc.)? Is it possible to capture these voltage excursions with a higher resolution on the scope? You may trigger the scope with a single rising edge on Vcc2 with a threshold of +6V. Make sure that you adjust the time base of the scope to capture the full duration of the overvoltage condition.

    In addition, can you please confirm if you still see ISO1050 devices getting damaged with the updated TVS diodes?

    The above will confirm if this is an artifact of the scope measurements due to the limited voltage samples (too large of the scope time base) or a real voltage transients coupling into the system.

    Once we rule out any possible scope artifacts, let's start debugging the issue from Vcc2. 20V on Vcc2 exceeds the abs max of ISO1050 (6V) and has the potential to damage the device.

    Thanks,

    Dushmantha

  • Hi Abhijit:

    Just checking in to see if you have any updates.

    Please feel free to reply to this thread if you have any questions.

    Thanks,

    Dushmantha