Zener diode test currents

[diydidi]

I built a zener tester a while ago. It uses a LM317 as a constant current source. It is fixed at 5mA by using a 220ohm resistor. 1,25V/220= 0,005A
I recently realised that it wasn't so accurate. measuring a 3v3 0,5W zener is ok, but measuring a 1W zener of the same 3v3 value, results in huge error.
This is because of the dynamic internal resistance of zener diodes. According to the data sheet different values of zener have different test currents where each one would measure accurately its voltage as specified. The 3V3 1W zener I mentioned previously needs 76mA to test spot on 3v3. The 0,5W only needs 5mA to measure precisely 3v3.
Now I can't possibly make my CCS switch between all these currents with a toggle switch for instance, so what would be a better way to modify my current tester??

[nickb]

For the very few occasions then I do need to test a zener, I just use a bench PSU. Set the current limit to the datasheet spec, crank up the volts until the current limit is hit and then just read the voltage off the meter. Usually they fail open or short so it's pretty obvious form the circuit under test that the zener is bad so an actual need to test the voltage is rare indeed.

Anyway, back to your question. You could use a pot and just mark off a rough scale on the panel. I really don't think it needs to be that accurate. If that is not to your liking a 12 position switch set to say 5mA, 10mA, 20mA, 40mA, 75mA, 100mA, 200mA, 400mA, 500mA, 750mA, 1A, 1.5A will cover 99% of cases (watch the LM317 dissipation).

[Chuck H]

Never even considered it! But your observations would explain a few things I've dealt with in my Zener travels. A non obvious reality (to some at least) that is good to be aware of.

[Mick Bailey]

I like Nick's pot idea as it gives more flexibility than switched values. The main problem with Zeners is inaccuracies due to under-current, but once this is overcome the voltage remains relatively stable over the designated range for that particular device. So my perception here would be to increase the current until the voltage just about stabilized and did not increase with a further increase in current (within the maximum rating of the device). The caveat is that if you're checking for a fault rather than just to know the voltage then you'd need to replicate the in-circuit current draw.

A more adventurous route would be to use a processor to do this for you as an interesting side project. The only time I think I've needed to check Zeners is when I've bought diode grab-bags with unknown in-house part numbers or have been unmarked. Otherwise they're open (or more usually) shorted in-circuit, though in the world of infinite possibilities with electronics I suppose its possible for a zener to behave up to a point and then exhibit unusual characteristics.

[diydidi]

Mick great idea. The only thing I noticed is that the pot has litlle change thirough a big part of its rotation and the suddenly becomes very sensitive. I have tried Lin, log and reverse audio taper. I dont know why this is. Doesn’t make sense.
Thoughts?

[nickb]

Mick great idea. The only thing I noticed is that the pot has litlle change thirough a big part of its rotation and the suddenly becomes very sensitive. I have tried Lin, log and reverse audio taper. I dont know why this is. Doesn’t make sense.
Thoughts?

The problem there is the the current and the resistance are reciprocals i.e. i = v/r.

So for a 200 ohm pot this means at 0% rotation you get 5mA, 50% => 11mA, 90%=> 55mA, 95%=> 100mA. I think you can see the problem. Pity I didn't see it before, huh?

Makes me think the switch is the better choice.