ISO1050DW

Hi Giuliano,

Since the ISO 1050 does not have any low power modes, the RXD pin will always pass the state of the bus. If the signal is dominant the RXD pin will be low and if the signal is recessive the signal will be high.

It looks to me like you are zoomed into an area of the frame where arbitration is taking place. Therefore the local controller could be telling the bus to go recessive, but some other node on the bus is driving the bus dominant. Since it only takes one node to make the bus dominant that is the state the bus will be in. I do not think that there is anything wrong with the device's behavior.

Thanks,

John

Hi John,

probably the image is not so clear. I think, as you wrote, when the bus is recessive the signal RXD have to be high (in NO low power modes).

As you can see in the image, the signal light blue (wich is the RXD pin 3 of ISO1050) tries to go to high level when the bus signal go to recessive (violet signal) but, immidiately it fall down (in a wrong state).

It's true, the image is a zoom on a arbitration frame.

Let me know... Giuliano.

Hi Giuliano,

I apologize, I misinterpreted the picture!

I think we are running into an issue with the isolator on this device not having a glitch filter. If you look at the bus VOD signal (purple signal), there is a glitch on the dominant to recessive edge. Right about in the middle of the falling edge there is a quick down-up-down before going all the way to recessive. The CAN transceiver in this device has a typical hysteresis of 150mV and this down-up-down appears to have a magnitude of approximately two times that. Therefore, the CAN transceiver die is sending a quick pulse to the isolator die which is edge triggered with “low” bandwidth. This isolator is seeing the first downward edge (the intended one) and then seeing the reversal back to dominant edge (the glitch), but since the isolator does not have a very high bandwidth it then misses the subsequent edge back to recessive. Since our isolators are edge-triggered, and it missed the rising edge, it will take a DC refresh cycle to correct the signal back to the true state, which can take around 6us. 

I think that there are at least two possible things we can do to fix this issue.  One is try switching to a split termination, if you are not already doing this. This termination style will filter high frequency noise on the bus and might filter enough of this glitch as to not cause the transceiver to see a dominant edge. For more information on CAN standard and split termination please see the short wiki page that I wrote below:

http://e2e.ti.com/support/interface/industrial_interface/w/industrial_interface/2513.can-termination-techniques.aspx

The second option, which depends on timing for you project, is to switch to the next generation isolated CAN transceivers that we are releasing early 1st quarter next year. These devices will have a built in glitch filter that will filter out high frequency pulses.

Please let me know if you have any questions, and I apologize for the mistake.

Thanks,

John