High Gain / High Leakage PNP Germaniums

[italianguy63]

I am getting around to buiding some Fuzz Faces.  I got a nice assortment of germainium PNP's (2N404, 2N404A, 2N1305, and 2N1309's).

In testing them, I found especially the High Gain ones (2N1309) have higher leakage numbers.  The leakage is almost always above the "published acceptance" level of 300, with most all of them around 500.  But when you subtract the leakage out, they still have nice high gains (around 150 hfe).  Are these useable, and is the high leakage just indicitive of the high gain PNP or the 2N1309's in general?

Thanks!

MC

[Govmnt_Lacky]

Only you can decide that! 

Build it.... Rock it..... Decide on it!

Good Luck!

The numbers used for testing Germ transistors is merely a guideline to assist you. You and your ears will decide what is good or not.

[italianguy63]

I am putting in sockets so I can mix and match (that's the plan).  I have gains from about 35 to a 150.  Leakage values all over the place, but all the 1309's seem to test high.  I only had 2 DOA out of about 40 that I bought.  I am just wondering if the high leakage ones are useable in a high gain setup in Q2? 

[Govmnt_Lacky]

Some claim that the "ideal" values for a Fuzz Face is 70 and 100 for Q1 and Q2 respectively but, your ears will guide you 

[italianguy63]

I'd love for someone who has done a bunch to chime in... I had also heard higher leakage in Q1 is wanted too.. It would be great to know what gains and what leakage combinations (ranges) work.  I'm afraid it is mostly just "voodoo.. "   

[kingswayguitar]

i made a fuzz face with crazy high leakage pnp.
sounds great just lacks a bit of volume at the output.
not even sure if the two are related but there you go.

[mac]

High leakage is not a big problem at FF Q2. If you have some low leakage devices, or low gain silicons, you can use them at Q1 since most of the thermal runaway depends on Q1, and the high leakage ones at Q2.

Or you can build a Tonebender MKI which needs leaky germs.

mac

[italianguy63]

Thanks Mac-- that is good to know.  I have heard "some" leakage is desirable in Q1 also.  So, I guess a solution for me anyway could be Q1 (mid gain/mid leak), and Q2 (high gain/high leak).  I'll give it a try once the rest of my parts show up.  I finished my voltage inverter board (charge pump) last night-- so I can keep the positive ground, as well as being able to daisey chain the DC input.

[mac]

I have heard "some" leakage is desirable in Q1 also.

Leakage adds nothing to the sound. Its only effect is to turn your FF into a nice vintage thermometer
I'd say more than 50u = 0.05ma is bad at Q1.

There a re a couple of ways to deal with leakage.
You can add a reverse leaky germ diode, or a 50k-220k resistor/pot from Q1 base to gound to shunt some leakage to earth.
The input impedance of the FF is less than 10k, so you can use a resistor not less than 50k or so.
If leakage is not bigger than 200ua you might bias Q2C from there.

mac

[Thecomedian]

Are you sure about that mac? I've been thinking that the reason people prefer some leakage is because it keeps the conduction "open" where normally the input signal would drop too low to bias it on, such as the "decay" region.

http://www.analogman.com/fuzzface.htm

Since silicon transistors took over from germanium about 40 years ago, almost nobody makes germaniums anymore, and the few that are made just don't do the trick in a fuzzface circuit. Each transistor used in a fuzzface needs to be tested for several parameters including gain, leakage, noise, and tone to make a great sounding fuzz.

Now I imagine that since the DIY or even boutique or boutique-offshoot-of-dunlop versions of Germanium fuzz faces haven't gotten away from using "vintage transistors" to achieve the sounds that are desirable, leakage probably plays a role. Otherwise, we could just be gobbling up current day Ge's at the proper gains to make "vintage tone".

The fuzz face works by dancing around cutoff. In this region prior to cutoff, transistors are very non-linear. Silicon shuts off sooner than Ge because of higher bias requirement for junction. If you get a free Spice program, you can load 2n3906's in the circuit and bias the 8.2k down to 6k for 4.5 across Q2 collector, then measure the voltages for Q1 collector/Q2 base against Q2 collector. You should see that as the voltage rises to it's peak for Q1 collector/Q2 base, its at it's lowest point at Q2 collector. The lines might even cross over, which is clear cut-off region.

Only you can decide that! 

Build it.... Rock it..... Decide on it!

Good Luck!

The numbers used for testing Germ transistors is merely a guideline to assist you. You and your ears will decide what is good or not.

this. Im currently on  a project with 20-40 gain transistors, and if the circuit is developed properly, they can sound loud and nicely fuzzy. Im having a hard time telling the difference between "true fuzz" transistors and my own varying Ge transistors, if all biases are properly tweaked.

[mac]

Are you sure about that mac? I've been thinking that the reason people prefer some leakage is because it keeps the conduction "open" where normally the input signal would drop too low to bias it on, such as the "decay" region.

I'm not saying no leakage at all, which would be imposible as germs always leak a bit.
What I'm saying is that the thermal drift is proportional to Q1 leakage and gain product, see my pdf below,
so IMHO it is better to use a low leakage device.
50ua is still a fair amount of leakage to keep the transistor in open state.
That's the reason why I put a 2M2 resistor from Vcc to B1 in silicon FF, to simulate leakage.

FF DC Bias PDF

http://www.diystompboxes.com/smfforum/index.php?topic=104818.0

...If you get a free Spice program...

... Ebers-Moll simplifications outside the active region... mmhh 

mac

[Electric Warrior]

Are you sure about that mac? I've been thinking that the reason people prefer some leakage is because it keeps the conduction "open" where normally the input signal would drop too low to bias it on, such as the "decay" region.

Guess that's more important with low hfes? Those NKTs mostly had hfes in the 50's and 60's. That usually results in a rather farty decay with low leakage trannies.

Or you can build a Tonebender MKI which needs leaky germs.

Or a MKII (use the OC75 schematic with the 10k/47k bias setup). They don't sound right with too little leakage. These are probably in the right ballpark.

[Thecomedian]

you can usually get rid of farty decay by rebiasing. At a certain biasing point, you can lose the fuzz sound completely and it just sounds like a booster.

[pinkjimiphoton]

just use sockets and your ears. a trimmer added for biasing will go a long way!!

[italianguy63]

Yeah-- I got a bunch of cheap SIP strip sockets from Tayda.  I cut them to a length of 5, and punched out every other pin (Pin 2 and 4).  The remaining strip fits perfectly in the Dallas Arbiter board.  So, I can do a bunch of swapping until I get it right.  Unfortunately, I am using old NOS parts on this build for that "vintage" look... so no bias adjustment.  I will start using bias pots on my next build.

[pinkjimiphoton]

cool, cool...
i use 6 inline dip sockets cut in half, sockets hold up a lot better.

i'd still reccomend doing the bias trim, if only temporarily... then you can swap transistors and set a proper bias... once ya settle on trannys and bias, just replace the trimmer with a fixed resistor at the closest standard value, and you'll have a monster!!

try and get the transistors biased so the fuzz interacts well with your guitar... i can do a whole show with a face on, even playing clean shit. crank up the guitars vol and tone, and you'll find pretty much every classic guitar tone of the last 60 years or so.

good luck bro!!

ps... a trick dave (old school analog) hipped me to with them sockets? once ya get the stuff how ya like it, if ya carefully squeeze the plastic insulators from the sockets, you can break  the plastic off withour screwing with the integrity of the sockets.. a quick kiss with a hot iron and a miniscule drop of solder to each tranny lead, and you're permanent and road worthy.

rock on!

[italianguy63]

Right on!
I learn a bit more everyday.
Does anyone know of a cheap supplier for regular round 3-pin tranny sockets?  Most I've found are over a buck each..

[italianguy63]

Note to self:  Maybe I should start saving the pins I am pressing out of the SIP sockets, and solder them in....

[pinkjimiphoton]