AC128..... incredible low "real" gain!!!

Started by pupina, May 24, 2006, 03:01:27 AM

Previous topic - Next topic


Custom pedalmaker Jacques has this to say about the AC128 (which he uses in his Mercer Box):


Fret Wire

Quote from: analogmike on May 25, 2006, 10:47:06 AM
maybe I was a little tough on AC128s for 2 reasons:

1) the misunderstanding that they were often used in the original pedals in the 60s (so far zero out of a few hundred germanium fuzzfaces I and my friends have seen have had AC128s)

2) all the AC128s I bought from several ebay and non ebay sources, and all the ones I pulled from over one hundred reissue UK made arbiter england fuzzfaces sound terrible and have poor specs. I lost a lot of money buying these. And I thought IF there were any good ones available in the UK, then the company making the the reissue fuzzfaces in the UK would be using the good ones.

have fun and be fuzzy!!!

Mike, I haven't tested as many as you've gone through, but enough to see a pattern emerge.

What I've found with a lot of AC's is that they are unstable now. They will show low leakage and proper gain, yet are internally (electronically) unstable. They'll bias up odd too. You can get Q2's collector at 4.5v, and Q1's emitter at zero, but the other voltages are slightly off from the ideal proportions. And they sound like ass: flabby, unresponsive, and too grainy. Don't clean up as well with the volume knob either.

Maybe RG can chime in with the proper electronic term for this internal breakdown. It's like the transistor is in a transitional stage where it still shows low leakage and good gain, but it is actually dying in front of you. It's like the stage the transistor goes through just before it opens up and goes high leakage.

You'll recognize this when you test them. Any type, not just AC's. A good stable Ge, when tested, will within seconds, indicate to a low leakage and it's gain of x. It will then hold on those gain and leakage numbers with almost no fluctuation on the DMM for a decent length of time. I've had good ones hold the same number for over a minute. These are the trannys that sound real good, bias up easy to nice proportionate voltages, clean up good with the volume knob, and hold their bias voltages (don't require re-tweaking every other month).

An unstable Ge will start to settle around a leakage number which may be low, and a particular gain (x), but never actually settle to any one reading. Your meter will never sit still on any one reading. It will always be in a state of fluctuation. If you leave the device on the tester long enough, the leakage will then start to climb up. Not necessarily over the limit for a good fuzz, but it will start to climb nonetheless. These are the Ge's that always seem to bias a little off, don't ever sound right, and are super temp sensitive (worse than normal). And real noticeable in a rig that is not only overdriven, but pushes some air also.

I might as well state the obvious about the above testing scenarios. I'm not talking about the initial heat induced flucuations from handling the devices with your hands or testing them in front of a heating duct. Use tweezers, gloves, and a stable temp room to test and an unstable tranny will still show the same symptoms.

Judging from the descriptions and voltages of a lot of fuzz posts I've read on the forum, I get the feeling a lot of people have used these unstable types. You order trannies, build RG's tester, they test around the right gain and leakage numbers you read about, then you build it and bias it, but it just doesn't seem to sound right. Sound familiar?

Getting back to AC's, in my testing I've found more unstable Ge's with the AC and OC types than I have with some of the american types, especially TI's. But I've also found that any stable Ge of the right gain and leakage sounds good, regardless of the brand. I have one of the original Mayer FF's from the early eighties that still sounds good today. It has BEL AC 128's. It's gotten a little tired over the years: Q2's collector is down to 3.75v. This tells me that these were very stable devices to begin with, and have very slowly increased in leakage over the years. Wish they would all hold up that long.

I think that back in the sixties and seventies, the playing field was level with all the Ge's. Now, forty years later, it appears to me that due to differences in manufacturing methods, some brands of Ge's have survived time better than others. Some brands seem to have yeilded more devices that are still stable today compared to other brands.

I think that not recognizing unstable devices has lead to alot of the brand preferences you read about hear. In my experience, good stable Ge's are all so close in freq resp that they sound boringly alike when properly biased up.

As an example, when Joe Davisson and Pete Snowberg put up the Fuzz Calculator on Joe's site, I decided to test it against my own biasing. Needless to say, it works very well. Anyways, I used Ge's that I tested and knew to be stable. I used pairs ranging from 40/60 hfe to 225/250 hfe and all different brands. After biasing, I put them in my fuzz jig, played them, and they all sounded identical. Even though they ran from a 40/60 pair to a 225/250hfe pair, and were different brands, every pair had two things in common. They were stable, and they had identical biasing voltages. Obviously, the different pairs required different resistances for Q1 & Q2's collectors. They all had the same sound characteristics, clean-up ability, and pick response. You would have failed a blind test trying to pick out one gain range or brand.

When I read a post using all these descriptive timbre type differences between Ge brands, I figure the poster has either an unstable pair and doesn't realize it, or has not biased one pair correctly versus the other. 

As time goes on, we're going to find more Ge pairs that will test proper gain and leakage, but are unstable.

Ge's are like people, they start dying the day they were born. :icon_sad:

Fret Wire
(Keyser Soze)


finding good Ge transistor in Europe isn't easy. the other day at the electronics store the was a 17 year old guy in front of me, starting on his first fuzz box, asking for GE trannies, being laughed at by the guy behind the counter -- "your're asking for components 40 years old". In the end he did return with some AC188s or so which is sold at 1.5 EUR a piece.

This guy laughed at me about a year ago when I asked for the same item. They certainly didn't have it. By now I know better.

I did get lucky in one store and got two usable (as far as gain is concerned) AC128s (gains 85 and 135) which I've used in a PNP negative ground Fuzz breadboard this morning. I got great gain, but apparently much less volume than I got used to when experimenting with silicone NPNs in the runoffgroove multiface setup. I don't know whether it is because of the PNP config, the different schematic, or the nature of the GE transistors.

I bought a whole bunch of different GE trannies, which I am about to test. I bought a PEAK component tester of Ebay, and it was the great deal, at least one electronics store let me select the transistors before purchase.



QuoteBut I've also found that any stable Ge of the right gain and leakage sounds good, regardless of the brand.

That has been my experience as well...

regards, Jack


FretWire, you have resumed my experiences with the ACxxx family very clearly. If I were looking for AC128s I will buy them from a good source.

I just want to add that fortunately there a lot of good 2S jap germaniums and some 2Ns in my country at a very good price, under a dollar.
So I have quite a collection of germaniums, most of them ( 99% ! ) stable and ready to rock. I'm talking of Toshibas, Matsushitas, Sanyos, Hitachi, some euros and americans if I dig deeper, etc., intended for different applications, ie, driver stages, IF, RF, power, etc., and with different maximun Ic & voltage and transistion frequency. Being stable they are easy to bias, and above all they sit at a fixed collector voltage next time you turn it on, in the abscence of a drastic weather change of course. And they all do the things one expect from a FF. But not all of them sound the same. For example, a couple of 2SB172 sound dark as the 2SD352, 2SD72k & 2SB77, a pair of 2SA49 or 2SA53/52, 2SB65 are more trebly, less farty and roll back differently, two 2SA101/102 or 2N388 have a wider freq response, and so on. And more, silicons do sound different. A pair of BD139, MPSA42 or 2N2219 have their own flavour.

Not being a "tone purist" (lucky me!) or whatever that means, I look for a focused fuzz, not for Jimi's fuzz in that particular song. In fact, of the 3 Ge examples I mentioned above I like them all, maybe the 2SA102 or 2N388 is my favourite, but if I were not so lazy and have more time I will build all of them to enjoy their different tonal characteristics.

As this is a very subjective matter my point is: Which is the best??? None. You like it? Cool! Use it! Who am I to judge your tone?

mac@mac-pc:~$ sudo apt-get install ECC83 EL84

Paul Perry (Frostwave)

Quote from: mac on May 26, 2006, 08:02:28 AM
I just want to add that fortunately there a lot of good 2S jap germaniums and some 2Ns in my country at a very good price, under a dollar.

Bet I wasn't the only person to click on Mac's details to check his country :icon_wink:
Glad there are SOME components you can get in South America!