TVS0500: 1-Wire I/O Protection for IEC 61000-4-5 1000V/25A, 1.2/50uS

Hi James,

Is the max operating voltage 4 V or is the abs max of the device 4 V? If the later, I'm assuming that the normal operating voltage would be around 3.3 V. If that's the case, then you could use ESDS312 instead which clamps a 1 kV surge (25 A) at 6.5 V. The device should be fine with the clamping voltage, but if not, you could also add a series resistor before the diode to limit the surge energy even more. 

Regards,

Matt Smith 

Hi Matt,

Thanks for the feedback

abs max is 4V and normal is 3.3V.  Clamping at 160% of the abs max is a concern.

The design does have a 20 ohm series resistance after the TVS but nothing in the spec. says that it can handle this.

The other concern I have is that this device could be trying to drive the bus low during this surge, essentially trying to pull 6.5v to ground through a 20 ohm resistance.  I'm not sure what kind of current this would sink but it is likely much higher than the rated current on the device.

Is it the 50uS pulse length that reduces you concern in these cases?

Electrical overstress at these voltages and currents for this amount of time usually will not cause issues in these devices?

Thanks,

Jim

Hi Jim,

I have never seen any type of surge diode that has surge clamping voltages only 0.7 V more than it's working voltage. Certainly not for a 1 kV surge. The abs max of most ICs are rated for constant DC voltages, not transients like surge pulses unless its explicitly stated in the datasheet. The IC should be able to handle higher voltage transients for shorter periods of time. ESDS312 has one of the lowest clamping voltages for a 1 kV surge that I've seen so it will be one of your best bets. The device is spec'd for the 50us voltage pulse(20 us current pulse) you're referring to so that is not a concern. 

Regards,

Matt 

Hey Matt,

Aside from availability issues, I agree that the ESDS312 is very capable with a low clamp voltage.

It looks like the answer for 1-wire parts is clamp with a TVS, add some series resistance, and test until you convince yourself that the 1-wire parts can handle it.  Any recommendation on what that testing should look like? 10 parts in 10 lots with 100 positive and 100 negative pulses per part?