Hi.

I got 8 AC188s the other day and went through them looking for leaky ones, etc.  4 were a disaster, leaking 1mA or more (one was a short circuit!).  2 others were good (200uA).  

But 2 were sort of in-between.  These also had a continuously falling leakage rate over the time I measured them.  They started off with about 350uA, which fell over the course of a minute or two to about 250uA.

Has anybody else noticed leakage dropping with time?  

What do people consider an acceptable "cut-off" for leakage?  300uA? 400uA?

Have read the "Technology of the Fuzz Face" article at GEO? He talks about that subject... well, at least measuring the gain, which is tied to the amount of leakage current.

Most people consider over 300ua suspect. That seems to be the "DIY industry norm". My last batch I got in, had one that started around 325 and dropped to mid 240's in a couple of minutes. Put your finger on it and it goes back up.

That's what you saw, the tranny going back to room temp, so to speak. Whether you used a socket or alligator clips, you were holding it with your fingers to set it up for measurement, and raised it's temp. Most that are in the low 100ua area don't do this as drastically.

Let that tranny sit awhile and try it again but with a cotton glove on to handle it. Bet it stays under 300ua.

Yeah, it makes sense that it's temperature-related.  I was straightening the leads and pushing them into 3 parts of an 8-pin DIL socket, so there was lots of time for them to heat up.  

I guess I can socket the ones that are around the 300uA mark and see whether they sound ok.  I'm using them to make a few easyface-type pedals.

thanks

You know I used to listen to the book worms here.  But I finally started listening to Ammscray when he said ignore numbers and just use your ears.  Throw them into sockets and see what you like best!  I was once brainwashed also with all that book talk but after I tried a bunch of trannys that qualified as failures here, they sounded awsome to my ears!  By all measurements they should have been trashed but they sounded oh so good.  So now I follow the wisdom of Ammscray and just socket them and tweak the bias trimpots to my flavor.  I never measure them anymore and my life has been much much less stressful.  I have a few magic OC44 that sound like crap.  All measures come out to the magic numbers and they sound just moo!  So it really does vary from tranny to tranny.  I will also inject that all book stuff aside the old stuff really does sound better.  No NTE will ever sub a OC.

I've had leakage dropping (and rising) in the past. But those were really poor ones. With AC188K I didn't had a big moveing of the FLG (False Leakage Gain) since the heatsink makes them really consistent.

The leakage measurement is more of an indication of the tranny's life expectancy, not it's sound. You can get good sound with a higher than normal leakage pair. You just won't get a good sound for that long. High leakage also makes it harder to bias.

I curious what forum the "book worms" that brain-washed you are on. On this forum, people like RG and others came up with the accepted 70-130 gain range (Q2 higher than Q1), not from labs, or books, but from actually playing alot of ge fuzz's and measuring the ones that sounded good. Nothing "book worm" about that. So now anyone can build a fuzz if they find Ge's in that range, and they will sound good. An NTE 158 can even sound good, if you can find two that fall in the range for a fuzz. Do you realize that most of the so called experts and pro's use RG's work for the basis of their fuzz's and Rangemasters?

As far as just plugging them in, tweaking bias, and settling on one's that sound good.... nothing the matter with that. That's how they originally built them, except they pulled the trannys out of a bin, soldered them in, and didn't play them. So the bad ones went out too. That's why you hear stories about Roger Mayer having to try 20-30 FF's to find 2 or 3 good ones for Hendrix to use.  That's why Dunlop is still having trouble making their fuzz's sound good.

Just remember about the leakage though. You can find a pair with a great sound, but if the leakage is too high, they may not hold that sound for long. The Mullard's are more consistant, the AC's are not. Measuring saves alot of time when you are using the less consistant types.