• Resolved

ISO1050: ISO1050DUBR: bus fault voltage value and short circuit test.

Intellectual 460 points

Replies: 4

Views: 148

Part Number: ISO1050

Hi Sir,

We now use ISO1050DUBR,  there's a risk of short circuit between CANL and 24Vdc(system power supply).

Base on this, I made a test like that: When system on communicated, I shorted 24Vdc power to CANL pin, keeps 5 seconds, then disconnected keeps 5 seconds; repeat shorted and disconnected; after 5-6 times like that, communication interrupted.  IC CANL impedance GND2 is about 100 ohms.  

then, this is my doubt: ISO1050DUBR bus fault voltage is -27~40V; how can I understand this value? why I add 24Vdc on CANL will damage the IC.

thanks for your attention, and wait for you.

regards.

  • Hi Cyan, Thank you for using our forums . I will share some initial thoughts below, and our team will follow-up next Tue/Wed due to this being a weekend and also Monday being a local holiday (US Labor Day). The CAN side, runs off a 5V rail, usually regulated down from a 12V or 24V battery/supply. If I understand you correctly, are you expecting a direct short between the system supply and the CAN pins, instead of through the 5V regulator ? Is this risk because the system supply is routed along with the CAN pins (layout)? A block diagram of your system will be very helpful here. To answer your question on the Spec, the numbers you are referring to are the Abs Max specs, these are damage specs.Meaning when the chip is exposed to these values , for any length of time, it can see damage. The actual working range of the IC is 5V +/- 10% for the supply and +/-12V for the CAN pins, called the operating common mode range of the chip. Anything beyond this is not good for long term operating conditions. Now the CAN pins normally have a 30k-ohm input impedance , so when you see 100ohms, it is likely because of damage. Now even though the voltage is 24V (inside of +40), we also have to look at the current limit on the input supply and see what the part was doing when the stress was applied. Potential solutions involve current limiting the supply, using a good TVS diode, or upgrading to the ISO1042 with +/- 70V Abs Max. We still need to check the power levels (current limits). More on this next Tue/Wed, but in the meantime we hope this helps ! Thanks, Abhi

  • In reply to Abhi Aarey:

    Hi Cyan, 

    Wanted to add on to what Abhi has said here. As he mentioned, a block diagram would be helpful. 

    It looks like you are able to induce damage to side 2 of the ISO1050DUB. You are correct, that the absolute max for the CANH and CANL pins is -27 to +40V, so at first glance it is an unexpected result that you were able to damage the device with 24V. However, there are a few possibilities. 

    Are you transmitting on the CAN bus during this fault event (24V shorted to CANL)? One thing to note, is that while the abs max rating for CANH, CANL is -27V to +40V, damage could still occur if there is too much current. Note that the short circuit current limit is only defined for the recommended operating conditions,-12V to +12V. See section 6.6 of the ISO1050 datasheet (page 7). Relevant section pasted below: 

    If the ISO1050 is trying to drive a dominant on the bus while CANL is shorted to 24V, this could lead to quite a fair bit of current inside the device. Perhaps if this fault condition was held for a long duration for a higher current we could damage the driver and therefore see 100 ohms between CANL and GND2. 

    Abhi mentioned that the ISO1042 has an extended abs max range. It also has an extended short-circuit current limit on CANL from -5V to +40V which would easily cover your 24V. See section 6.10 of the ISO1042 datasheet (page 10). Relevant section pasted below: 

    For a new design, I would recommend you switch to the ISO1042DWV as I think it would be much better suited to deal with this fault condition. 

    Best regards, 

    Dan

  • In reply to Dan Kisling:

    Hi Cyan, 

    I haven't heard from you in a few days. Were you resolve your issue with ISO1050?

    Best regards, 

    Dan

  • In reply to Dan Kisling:

    Hi Dan,

    Thanks for your detailed explanation. We understood it now.

    And, will test ISO1042-Q1 on next step; a little pity, it's not a p2p part with ISO1050.

    Thanks all.