Testing ge diodes for leakage?

[Kroars]

Hello all,

Question: How do you test diodes for leakage?

Perhaps I’ve gotten lucky in the past building circuits that require low leakage ge diodes. This time around I’ve built a couple Nugget pedals and they didn’t quite sound right (not like the others I’ve built anyway). Upon reading through the forums I see that some circuits require diodes that are low leakage (Nugget being one). I’ve scoured the inter webs, including the PedalPCB forums and I can’t find much on how exactly to test a diode for leakage and I’m hoping one of you could help explain.

Thank you very much!

 

[fig]

Some DMM have this feature to get in the ballpark, but accuracy is limited.

I use a Peak DCA55.

Alternatively, some vendors sell matched sets, or even singles in ranges

 

[Bret608]

I feel like I remember reading that you have to measure resistance across the diode one way, and then the other. The higher "reverse resistance", the less leaky the diode is. It's jubal81 who frequents these parts who would be able to explain what I'm talking about. I've actually been meaning to ask about this too. Let me see if I can dig up the thread over at Madbean where I first saw something on this topic.

 

[Kroars]

Some DMM have this feature to get in the ballpark, but accuracy is limited.

I use a Peak DCA55.

Alternatively, some vendors sell matched sets, or even singles in ranges

I have both a DCA 75 and a Fluke 115. I often test for forward voltage, what I’m unsure about is the leakage of the diode.

 

[Kroars]

I feel like I remember reading that you have to measure resistance across the diode one way, and then the other. The higher "reverse resistance", the less leaky the diode is. It's jubal81 who frequents these parts who would be able to explain what I'm talking about. I've actually been meaning to ask about this too. Let me see if I can dig up the thread over at Madbean where I first saw something on this topic.

Awesome, thanks! I’ll keep looking myself and let you know what I find as well. I also have read about taking a resistance measurement both ways but I’m unsure of how high is good. For instance last night I measured a handful of ge diodes and got between 9k-10k -but I don’t know if that’s high, mid or low. Diode leakage seems to be something important only for a few applications (some pedal circuits being one) as there doesn’t seem to be much info on it out there.

 

[fig]

I have both a DCA 75 and a Fluke 115. I often test for forward voltage, what I’m unsure about is the leakage of the diode.

Yeah...have to redact that. Just realized I had never tested the leakage with it. As it turns out, it does not identify Ge or Si, nor gives leakage. My apologies.

 

[Kroars]

Yeah...have to redact that. Just realized I had never tested the leakage with it. As it turns out, it does not identify Ge or Si, nor gives leakage. My apologies.

No worries man. I appreciate you trying to help! Hopefully we’ll all get the answer soon

 

[Kroars]

I’m starting to think testing for diode leakage is akin to asking someone for a left handed hammer....

 

[Kroars]

I feel like I remember reading that you have to measure resistance across the diode one way, and then the other. The higher "reverse resistance", the less leaky the diode is. It's jubal81 who frequents these parts who would be able to explain what I'm talking about. I've actually been meaning to ask about this too. Let me see if I can dig up the thread over at Madbean where I first saw something on this topic.

Ive been going through the DIY Stompbox forum and have found a few threads that mention your suggestion of measuring resistance of the diode in reverse. My question now is how much resistance is considered high leakage and how much is low leakage?

Further, I see that some circuits benefit from having diodes with low leakage -are there circuits that would benefit from diodes with high leakage?

 

[Chuck D. Bones]

Depends on the circuit. The Cornish G2 needs diodes that measure more than 1Meg in reverse. There are circuits that don't care about diode leakage, like the Distortion+. A leaky diode might be useful in a Buzzaround.

 

[Kroars]

Depends on the circuit. The Cornish G2 needs diodes that measure more than 1Meg in reverse. There are circuits that don't care about diode leakage, like the Distortion+. A leaky diode might be useful in a Buzzaround.

Alright, just to make certain I beat this into the ground.... I’ve now tested about a third of these diodes and the best I’ve got was 380k resistance. I know you said to not be surprised if most don’t pass the test, does that mean maybe 1-2 out of 100 might be over 1M?

Here’s a photo of the ge diodes that I’m almost half way through and also a photo of the meter. Am I doing this wrong? Red probe to cathode, black to anode.

Thank you for your time good sir.

 

[Chuck D. Bones]

Look right to me. Have you tried reversing the leads?

 

[jubal81]

Yes, it's rather rare to find them over 1M.
All this talk spurred me to get out the breadboard. Made up a G2 with non-inverting opamps for the input and makeup gain stages and replaced the tone control with the Stupidly Wonderful Tone Control. Sounding pretty damn good. I found four 1N695 diodes that measured over 6M and it roars. No lack of distortion with those puppies. You can see them in the bottom center of the picture.

 

[Kroars]

Look right to me. Have you tried reversing the leads?

Ok, just making sure I’m not extra dumb tonight. Yup, tried both ways. I’ll keep rummaging through my stock pile and perhaps order some more. Thanks again!

 

[Chuck D. Bones]

Whenever I get a reading I don't believe, or don't want to believe, I swap the leads or grab a different meter just in case there is dipshittery afoot.

Guess I should bust out some of my Ge diodes and see if any of them have low leakage. So far I have only used them as hard clippers or as compensating diodes in Tone Benders, so leakage was not an issue.

 

[jcrhee]

is measuring diode leakage the same as measuring forward voltage?

 

[Chuck D. Bones]

No, it's the opposite of that.
DMMs measure resistance by applying a constant current and measuring the voltage drop. This method works because R = V / I. Different resistance ranges require different currents, from a few milliamps in the low resistance ranges to less than a microamp in the high resistance ranges. DMMs use a similar method to measure diode forward voltage, the only difference is that the voltage drop is displayed as voltage instead of resistance and the test current is higher, usually in the 1mA to 5mA range. In the reverse direction, diodes act like resistors, so we use a DMM set to measure megohms. When we use a DMM that way to measure diode leakage, we're actually measuring the diode's reverse voltage at a very small current, on the order of a microamp or less.

 

[Bret608]

Yes! Jubal81 started started that thread I talked about back in the day, and now here he comes with a G2 on the breadboard. Thanks to Jason and Chuck for steering this in the right direction...now I've got to try this out myself.

 

[Bret608]

Just wanted to report back that I finally got some time to try this out according to Chuck's methodology. Interesting results! A couple of old 1n279 that I have measure at 1m or just below in reverse. The best I had beyond that was a 1n60 (O.G. ge, not Schottky like the 1n60p) at about 300k. Most of the rest--unmarked ones, 1n273, etc.--were under 100k. I can see why it would be challenging to find enough for the G2 and similar. Hasn't been a barrier to me yet, but this was a fun exercise and it's good information to have.

 

[Kroars]

Just wanted to report back that I finally got some time to try this out according to Chuck's methodology. Interesting results! A couple of old 1n279 that I have measure at 1m or just below in reverse. The best I had beyond that was a 1n60 (O.G. ge, not Schottky like the 1n60p) at about 300k. Most of the rest--unmarked ones, 1n273, etc.--were under 100k. I can see why it would be challenging to find enough for the G2 and similar. Hasn't been a barrier to me yet, but this was a fun exercise and it's good information to have.

Had the exact same issue. Tested 50+ and only got around to 500k... a lot would start at 1M-2M+ then quickly shoot down to 100k-300k. Ended up ordering a bunch more to see what I can find. What’s the worst that can happen? Gotta go, wife’s yelling for me....