# TVS diode for RS485

I'm studying the ISO1500 for an RS485 application. In particular, I'm looking for a TVS diode for bus protection. I've found that a common TVS diode used for RS-485 applications is the SM712 (like Bourns or Littlefuse) because it has asymmetrical reverse stand-off voltages to match the common-mode operating range of RS485 (-7V to 12V).

According to IEC 61000-4-5, for Class 2, a 1kV surge voltage will discharge through a resistor of 42$$\\Omega\$$, producing a current of 24A (as specified here).

I have a doubt regarding how the SM712 can provide surge protection. In SM712 datasheet it seems that the max clamping current is lower than 24A. Moreover, It seems that the clamping voltage is higher than the absolute maximum voltage on the bus pins (±18V from ISO1500 datasheet).

In this case the device will be unprotected by the surge pulse?

• IEC 61000-4-5 is a lightning strike. You don't really expect the IEC 61000-4-5 to protect the IC on its own against a lightning strike do you? Also, the clamping voltages at such extreme currents will always be significantly higher. I'm not sure how this translates into real world survival but it's impractical to find a diode large enough can clamp so tightly at such a high current. If you look up the ESD vulnerability for CMOS circuits it's usually listed in the 100-200V range which is much higher than the actual operating voltage so there seems to be different contexts for the two ratings. – DKNguyen Jun 10 '20 at 17:38
• @DKNguyen basically I agree but it's an indirect lightning strike if anyone is interested in the pedantic truth. Nothing protects a semiconductor from a direct lightning strike! – Andy aka Jun 10 '20 at 20:10

Who says it's a class 2 device? Who says class 2 is 2 kV?

EN 61000-4-5 says this about the voltage impulse: -

So a level 1 device would survive a 500 volt surge and this translates to a surge current of about 11.9 amps. A level 2 device would survive a surge current of 23.8 amps.

The Bourns CDSOT23-SM712 suggests that the limiting voltage will be 14 volts with a surge of 17 amps. You might say that this will still destroy the ISO1500 and so it might but, the ISO1500 has internal protection that can survive an ESD of 4 kV so, it is up to the designer to merge the indirect lightning protection offered by the TVS diode to the levels that are compatible with the ISO1500.

Nobody says it's easy to do.

• Ok thanks for your answer, I'll do some tests. Do you think that two series pulse-proof resistors on the bus lines can help in this case to limit the clamping current? – alediben Jun 11 '20 at 15:45
• (1) No, don't do some tests, do some more work on understanding the problem from what the data sheet tells you. (2) Yes, that's more like it - you need to work out the vulnerable point of the ISO500 and add something (like resistors) that can limit the current. Maybe add a proposal to your question (clearly labeling it new information) and I'll take a look. – Andy aka Jun 11 '20 at 15:48
• Bear also in mind that pretty much all Ethernet lines can survive level 3 or 4 (2 kV and 4 kV). I'm mentioning this because it can shape how you look at the problem. There's also the consideration that if the RS485 cable is shielded, then the applied pulse will be from a 2 ohm source to the shield with the shield grounded end used injected with the surge (ground bounced). – Andy aka Jun 11 '20 at 15:55