TVS2201 clamping diode failed at 27 V

Question

In the development of ESD protection and reverse polarity, I encountered a problem. The circuit is supposed to prevent reverse polarity (it does, we tested it).

During testing, I applied a voltage of 27 V (it is designed for 24 V input voltage) to the input (used power supply: MPS-3005L-3MATRIX). Looking at the documentation of the TVS diode itself: TVS2201 Datasheet, the diodes start clamping at 29.6 V, and the standoff voltage is approximately ±22 V.

In the actual design, I made a mistake by using the 2201 diode instead of the 2701, but this led to the problem. The diode failed at 27 V and then began conducting 2 A. The inrush resistor heated up, and the diode as well, indicating a short circuit or damage. The PCB design itself is in accordance with the datasheet. I am attaching the schematic.

In the corrected design, we will choose the 2701 diode, as originally intended in the initial design. However, from the perspective of preventing a short circuit in production, I am a bit concerned about the conduction of the diode at around 27 V input voltage.

Does anyone have any idea why this happened with such a low voltage, and has anyone else experienced this before?

Answer 1

You are reading the datasheet incorrectly. The 2201 doesn't "start clamping" at 29.6V, rather 29.6V is the clamp voltage when carrying the rated clamp current of 30A (something it can only do for a brief period).

If you look at the app note referenced in the datasheet, you can see a graph showing the typical response curve of these devices.

This shows how the current starts rising well before the Vclamp voltage (though significantly above the stand-off voltage).

For the 22V part, the breakdown voltage Vbr (at which point there is 1mA flowing through the part) is specified in the datasheet as typically being 26.6V, and potentially as low as 25.1V.

Assuming a typical device, and a roughly linear dynamic resistance (a huge approximation, but potentially informative), this means the clamp current will increase to 30A over the 3V range from 26.6V to 29.6V, so a dynamic resistance of 0.1Ω. (The datasheet actually quotes 0.04Ω from 0.5Iclamp to Iclamp, as the resistance isn't really linear, but we're in the right ball part.)

As such, 27V (0.4V above a typical Vbr), could easily lead to 4A of current flow. For a static (as opposed to transient) condition, this will lead to almost immediate failure due to excessive power dissipation. This fits with your observations.

The 2201 part is therefore working exactly as speced, it just isn't the correct part for your design.

You need to select a part with a Vstand-off which covers all the range you view as "normal", and allow for the increase of clamp current from the relevant Vbr to Vclamp. If 27V is that limit, then the 2701 should be fine.

Answer 2

For the TVS2701 this extract is an important part of the device's specification: -

In a nutshell, providing that you keep below the 29.6 volt limit, the current into the device will remain no greater than 1 mA. At 27 volts applied you should be fine.

For the TVS2201 with 27 volts applied it's a different story: -

27 volts is potentially above the point where current is guaranteed to be no more than 1 mA and, if you look at the absolute maximum ratings page, the current only has to rise to 30 mA before all bets are off on this device. Given the type of device, the current will rapidly rise towards 30 mA (and above) as soon as you are exceeding the break-down voltage even by a tiny amount: -

Please also check that you have chosen the right device for use on power lines; the TVSxxx range is suitable for EN 61000-4-5 (indirect lightning protection) on signal lines and, I don't think this applies to power lines such as 24 volts. I hope I'm wrong in this respect but, you ought to check that out.