Testing diodes for leakage?

[Kroars]

Hello All,

I've built a couple NG-2 pedals in the past and must've just got lucky as this time I'm having issues with the sound.  Apparently this circuit is very picky and requires low leakage ge diodes. 

My question is how exactly do you test diodes for leakage?  I have a fluke 115 and I see folks saying that you need to take a measurement of resistance in reverse and based on the reading you have a low or high leakage diode.  Is this correct? If so, what is considered a high leakage reading and a low leakage reading?

I've scoured the internet trying to find an answer to this question and haven't found much at all.  Really hoping someone here can help.

Thank you kindly!

[antonis]

Hi & Welcome..

Why you don't use any 1N34A, 1N60 or Schottky diodes availiable..??

After all, NG-2 is another (Cornish mode) Big Muff + FuzzFace..

[Kroars]

Those diodes would work, but it is my impression that like ge transistors diodes can have various readings of leakage between the same type of diode.  My question is not which diode to use, but how to test for leakage and also what is a reading of high leakage vs low leakage.  Thanks though.

[antonis]

OK..

Place a high value resistor in series with the diode, get a good DMM, apply a reverse voltage of your taste, measure voltage drop across the resistor and implement Ohm's Law..

Anything less than 300μA (arbitrary value..) should be considered "low"..

[anotherjim]

Is it forward or reverse leakage that's the issue?

[Kroars]

In this particular build it's the reverse leakage.  Particularly 1n34a in this case.  I've got the forward voltage down, I just need to know how to test for reverse leakage and what measurement is considered high leakage and what is considered low.  Thanks!

[r080]

I assume you have a pile of 1N34s and you are trying to choose which ones to use.

Based on the way it is used with an AC signal, I would think you would just want whichever ones have much larger forward current compared to reverse.

That said, you can still find datasheets for old components. With a quick google search, I found one speced at 30uA with 10V reverse voltage, and one with 100uA with 10V reverse voltage. For comparison, a 1n914 datasheet had 10nA at 10V reverse voltage. That should give you an idea of what "low" is, at least for 10V reverse bias.

[Kroars]

Thanks for the response Rob.  I've been using the old sort by ear method and while that does work, I'd like to know how one would go about finding the reverse leakage in a particular diode.  I was under the impression that you can get various readings even with diodes of the exact same lot.  Much like one would test for forward voltage when pairing two diodes for an octave.  While data sheets can certainly be useful, I'd feel better having the knowledge to actually test a particular diode for reverse leakage.

[Kroars]

Finally got an answer and figured I'd post in case anyone else has the same question.  Courtesy of Chuck d Bones:

Set the meter to measure resistance and set the range to manual, 6Meg full scale range. Don't let it autorange, you might get confusing readings. A good diode for this circuit will read at least 1Meg. Higher is better. Use clip leads. Do not touch the diode or leads during the meassurement because your skin resistance and the heat from your body will upset the measurement. Don't be surprised if the majority of your Ge diodes do not pass this test. For comparison purposes, check a Si diode like 1N4148.

[antonis]

Finally got an answer and figured I'd post in case anyone else has the same question.

.... Don't be surprised if the majority of your Ge diodes do not pass this test...

What an answer..!!
(I should notice.. )

[ElectricDruid]

Do not touch the diode or leads during the measurement because your skin resistance and the heat from your body will upset the measurement.

I was about to mention that temperature probably has a bearing, since it does when matching transistors.

[amptramp]

If the diode is in a clear glass or epoxy body, make sure you test it in the dark.  Light falling on the diode junction would cause a photoconductive leakage effect that would not exist inside an enclosure.  Try making measurements with lights on and off.  What may look like black paint may be transparent in the infrared range, so it might affect diodes where the junction is not visible.

[r080]

While data sheets can certainly be useful, I'd feel better having the knowledge to actually test a particular diode for reverse leakage.

Understood. I was thinking the technique Antonis gave would be perfect. The idea is the current through two components in series must be the same. You apply a reverse bias across a combination of the diode under test and a large resistor with a known value (chose one that makes the math easy like 1M). Measure the voltage across the resistor, then, using Ohm's law, you get your leakage current by dividing the voltage reading by your resistor value.

[amptramp]

While data sheets can certainly be useful, I'd feel better having the knowledge to actually test a particular diode for reverse leakage.

Understood. I was thinking the technique Antonis gave would be perfect. The idea is the current through two components in series must be the same. You apply a reverse bias across a combination of the diode under test and a large resistor with a known value (chose one that makes the math easy like 1M). Measure the voltage across the resistor, then, using Ohm's law, you get your leakage current by dividing the voltage reading by your resistor value.

Agreed with one caveat:  the series resistor is in parallel with the voltmeter input impedance, so that has to be calculated in parallel.

[anotherjim]

Doesn't a DMM actually do the necessary on resistance and diode test ranges though? You may not know the actual leakage current from this, but it's certainly the first thing I try if I suspect something is leaky.

[antonis]

Agreed with one caveat:  the series resistor is in parallel with the voltmeter input impedance, so that has to be calculated in parallel.

That's exactly why I posted:

get a good DMM

In the mean of DMM's high input impedance relativey to test series resistance..

For 10V reverse voltage and 100μA expected leakage current, 10k series resistor should be fine and you don't have to bother with DMM's impedance..