TVS breakdown voltage

[Ray91]

Hello,

is the breakdown voltage of a TVS (e.g. SMAJ5.0A) a guaranteed value?

Background: I whish to protect a circuit from overvoltages on the 5V supply which can be normally range from 4.4V to 6V. AN316 would then suggest to use a  SMAJ6.0A or better 6.5A as it says, that the maximum operating voltage should be below the Vrm voltage.

But if I read the datasheet right, the Vbr voltage states a minimum (tested?) voltage at which the reverse current reaches / exceeds 1mA.

If I want to have the protection level as low as possible, I would prefer the SMAJ5, as it has a VBr of 6.4V. I know, if my supply goes up to 6V that I get a higher reverse current, which I can to tolerate. But I would expect, that this current is not higher than the test current of 1mA for the Vbr and therefore the dissipated power in the TVS should well below 10mW. Does an elevated voltage near to the breakdown voltage cause problems over time?

[Peter BENSCH]

No, a TVS breakdown voltage VBR is not a hard guaranteed trip point, but a specification at a test current, here 1 mA. VBR varies with tolerance, temperature (see datasheet and temperature coefficient αT), and part-to-part spread. In addition, the characteristic is not vertical, but has a dynamic resistance (see datasheet, fig. 1 Electrical characteristics).

For a supply that can run up to 6V, a SMAJ5.0A is generally too close to the normal operating range. A TVS is intended mainly for short transient pulses, not for continuous clamping near breakdown.

So the key rule is: choose VRWM above the maximum normal supply voltage and check the full system margins, not only VBR.

Hope that helps?

Regards
/Peter

[Ray91]

Hello, 

thank you for the reply. 

I know that the Vbr, VCl varies with temperature, but the  αT is postive. So I would expect a reduction only with falling temperatures. You state - the Vbr, is the breakdown voltage with a test current of 10mA (I remembered it wrong - the 5V is tested with 10mA not 1mA). So I would asume, that each diode is tested with this test current. As the datasheet states, taht this is a minimum value and the typical value is higher, I would have expected, that all parts are tested. Then the part-to-part variation would only give voltages higher than or at least at the minimum specified break down voltage, if I use the TVS at 25°C or above.

Or does ST only test samples and not each specimen? Or do I have to take into acount, that this voltage / current changes that much through aging?

The dynamic characteristic would only apply - in my opinion - if a would see rapid changes in the applied voltage. But if the voltage changes perhaps with 1% or 2% due to changes of the load or line input of the supply would lie below the breakdown voltage, I would asume the condition as static - because that would be in the range of the ‘knee’. Only if the voltage exeeds the break-down voltage I would se a rapid change of current, because then the diode works on the steep part of the curve.

That would be my expectation and would fit my use-case to have a fast reaction if the supply voltage crosses the breakdown voltage. But - Ok - 10mAs of continous current is quite high.

Best regards 

Ray

[Andrew Neil]

I whish to protect a circuit from overvoltages on the 5V supply which can be normally range from 4.4V to 6V

Sounds more like a job for a voltage regulator capable of up to 6V ?