Ge transistor bias mystery (higher collector-to-emitter voltage => noise)

[mc50]

Hello, I've had this happen with two Ge transistors from different batches / vendors, a Russian SFT307 (low leakage, hFE around 100ish), and a red dot AC128.

The SFT307 case is probably the most telling, I've put it in a Rangemaster built with this schematic. Then I've put in a trimpot instead of the 3.9K, and biased it at 2V on the collector.

With this setting, I get crackle and static, which gets worse as time goes on. It's not there when I plug in, then you hear it when there's no playing 5m later, and in 15-20m it's like rainfall, the noise gets almost to instrument playing level loudness.

But, cranking up the trimpot until I get to ~7V on the collector gets rid of the noise permanently and gives me tone comparable to what I got at 2V.

(Just to note, it's 2V because this is negative ground, R.G. Keen recommends -7V with a positive ground schematic.)

I also have an AC125 that, while sounding very unremarkable as opposed to the Russian transistor (which is awesome-sounding) has no problem with being biased at 2V on its collector in the same circuit (it's socketed).

Same thing happened with the AC128 - it's installed in a circuit that recommended 0.6V on the collector, but I need to keep that transistor above 2V to not get a similar (albeit less loud) kind of noise from coming in (on its Q1 - it's a 3-transistor overdrive).

Is this regarded as normal? Would you replace the noisy transistors? Would you consider this a transistor fault (in the context of the circuits it's used in, of course)? Is there some parameter I could have measured with an ESR meter that could have told me I might expect this?

Thanks!

[iainpunk]

Some Ge's are just noisy. Even from the same batch some might be noisy and others can be clean as a whistle.
the flipped power supply also provides a worse PSRR, but that shouldn't be the problem here since the noise is bias dependent.

also, that's an odd bias technique, generally people replace the 68k with a 100k pot, i'd replace the 10k output volume pot with a 22k variable resistor, and place the output volume control after the capacitor so that there's no crackle when turning the volume control.

cheers

[mc50]

also, that's an odd bias technique

I was just following PRR's advice here.

[mc50]

Some Ge's are just noisy. Even from the same batch some might be noisy and others can be clean as a whistle.

So the conclusion is: bad transistor for the Rangemaster circuit (too noisy for this purpose assuming that even by ear, and not looking at the numbers, I very much prefer the sound of the pedal at around 2V), find another one to put in (biased at 2V) and move on?

[iainpunk]

i suggest getting a whole bunch of Ge transistors so you can select the least noisy ones in the batch for fuzz or treble boosters and use the other (noisy) ones for percussion synth circuits.

cheers

[Rob Strand]

So the conclusion is: bad transistor for the Rangemaster circuit (too noisy for this purpose assuming that even by ear, and not looking at the numbers, I very much prefer the sound of the pedal at around 2V), find another one to put in (biased at 2V) and move on?

The way people measure the collector voltage in those circuits is ambiguous.      The circuit normally has positive ground and the meaning of 7V is from the +9V rail to the collector ie. treating the +9V as ground.    In the schematic you gave the circuit has been flipped for negative ground so to get the classical "7V" biasing your circuit needs to measure 2V (9V-7V = 2V) on the collector relative to the negative ground.   In simple words 2V *across* the 10k pot on the collector.

[mc50]

In simple words 2V *across* the 10k pot on the collector.

Right, I believe I'm doing it correctly: one multimeter lead on the enclosure (ground) and one on the collector. With 2V measured like that, the Russian transistor is unusably noisy, but sound glorious compared to the mundane AC125.

[mc50]

Actually the question was a bit more philosophical: because the signal swing is similar at 2V and at 7V on the collector (regardless of whether the circuit is positive or negative ground), would it make sense to call a good-sounding transistor "right" for the Rangemaster, even though it would be pretty far from R.G.'s gospel (in voltage and resistance)? Would it make sense to call it "just misbiased" if it makes noise at 2V (but not at 7V), or do we just call it "wrong" for the Rangemaster?

From the above replies I assume the latter is the case, but just wanted it confirmed and put in writing.

[Rob Strand]

Actually the question was a bit more philosophical: because the signal swing is similar at 2V and at 7V on the collector (regardless of whether the circuit is positive or negative ground),

The maximum swing is the same but the signal gain and the non-linearity are quite different, as a result the sound is quite different.

would it make sense to call a good-sounding transistor "right" for the Rangemaster, even though it would be pretty far from R.G.'s gospel (in voltage and resistance)? Would it make sense to call it "just misbiased" if it makes noise at 2V (but not at 7V), or do we just call it "wrong" for the Rangemaster?

I believe RG's 7V is the same as your 2V, ie. 2V across the 10k resistor.   That's generally consider "right" for the classic Rangemaster.   A good deal of commercial clones follow this.

Recently Phend posted a topic in the lounge regarding the biasing, and he's not the only one to think this.  (His 5.845V voltage is measured to "ground" like RG's 7V, so about 3.155V across the 10k).

https://www.diystompboxes.com/smfforum/index.php?topic=126608.msg1211220#msg1211220

[mc50]

Recently Phend posted a topic in the lounge regarding the biasing, and he's not the only one to think this.  (His 5.845V voltage is measured to "ground" like RG's 7V, so about 3.155V across the 10k).

https://www.diystompboxes.com/smfforum/index.php?topic=126608.msg1211220#msg1211220

Very interesting thread. I do prefer 2V (might even go lower than 2V really) even by ear.

Here's a very interesting article (with nice pictures even) that backs up R.G.'s (and my ear-picked) preference: https://www.electrosmash.com/dallas-rangemaster

[Rob Strand]

Here's a very interesting article (with nice pictures even) that backs up R.G.'s (and my ear-picked) preference: https://www.electrosmash.com/dallas-rangemaster

Clearly a lot of people like that otherwise the boutique guys wouldn't copy it.   For me I like the region closer Phend.  I probably should go through the whole process of evaluating the bias point again but these days I don't have the amps that do the Rangmaster justice.

[iainpunk]

wait, you have +2V on the collector right? that's damn near the perfect bias.

my personal rangemaster based design/clone i call the firing range, and i like to bias it at the edge of cutoff, almost but not fully gated.
it uses the BD139, a Si power transistor, it has a lower gain than Si signal transistors, and larger miller capacitance, giving it a more Ge-ish character, while still keeping the clarity and stability of stability. it doesn't ahve the ''mojo sound'' of Ge tho, but i use it more for a punk / riff-raff character, instead of the ''classic rock sound'' normally associated with the range master.
the other transistor is a buffer, to ''correct'' the output impedance.


cheers

[mc50]

wait, you have +2V on the collector right? that's damn near the perfect bias.

I do, yes, with both the Russian transistor and the AC125 (it does require some twiddling of the "3.9K" trimpot, but very, very little - I think I measured once and the trimpot was 3.2K). Except that the Russian one sounds fantastic, clear and chimey and just great, and the AC125 sounds rough and muddy compared to it (although I feel that both sound their best at the 2V mark or just below).

Oh well, I guess I'll keep hunting for more low-leakage, 75-100 gain Russian trannies. One of'em is bound to be quiet and keep the clarity and sparkle.

I wish I knew where to source these things reliably, I'd be free to just perfboard a whole bunch of vintage-style things and no longer depend on kits or prebuilt pedals.

[PRR]

I'm utterly lost. Are you measuring the table to the floor or to the ceiling??

Say "COLLECTOR TO EMITTER" if that is what you are measuring. (Yes, it is usually simpler to measure every point to some designated "ground", but in this thread we have lost our bearings.)

[mc50]

Alright. I have updated the topic's title. The schematic has been linked in the original post, but re-adding here as picture for maximum clarity:



No noise:

Collector to emitter: 0.94V

From (the negative) ground:
C: 7.45V
B: 8.21V
E: 8.39V

Noise:

Collector to emitter: 6.04V

From (the negative) ground:
C: 2V
B: 7.89V
E: 8.05V

FWIW, the transistor is one of these: https://www.ebay.com/itm/254849078091
The other item of the pair has the same issue.

[mc50]

To continue on the quest for maximum clarity, here's the transistor tested:



AFAICT, by numbers alone it should have been perfect for the Rangemaster.

[ManicStrat1982]

Have you checked the leakage of said transistors to see how it coincides with noise? In SOME cases I have noticed high leakage trannies being more static-y and hissy in certain circuits to than lower leakage ones.

[mc50]

Well, my cheapo TC1 tester (pictured above) shows a hFE of 107, Iceo = 31μA.

If you're asking about Icbo, I suppose it's 31/(107+1) = .28μA

I'd say, if I'm not horribly wrong, and if that tool is to be believed (I don't have any other Ge transistor tester), that if leakage is the problem then this transistor is too low leakage.

I thought that's the way to be for a Rangemaster though...