Questions about FF leakage/gain, clipping...

[Earthscum]

I went through and measured a bunch of my Ge transistors the other night, and didn't really come out with much for a FF. Most were in the 60Hfe range, with an OC7(?) around 82 and a funky numbered one about 89, with a leakage of only 10uA (HP4111 something or other... black metal case)

BUT, I have a RCA SK3003 that measured out at 120 gain... WOOT! Wait... it leaks about 470uA. So, I kinda remember the OC leaking about 150uA or so. This will be getting used for Bass, possibly after octave up and mixed back with the original signal, so ultra gain isn't key... I can get that with a voltage sag, anyways.

Any rec's? I think I also have a 50uA leaky 77 Hfe that is same number as the 89/10uA.

Also, if I should steer clear of leaky tranny's in Q2, any good fuzzes I could use that leaky RCA in? (I already have a set of transistors put aside for a Fuzz-A-Tort).

And, yes... I've read about all the FF articles. I still have yet to determine which transistor is actually doing most of the clipping. I see people using Ge's in Q1 and Si in Q2, as well as Ge in Q2 and Si in Q1. I would assume that with Joe's diode trick, Q1 wouldn't be clipping, and thus Q2 would be the better candidate for getting the Ge in the case of a hybrid? Anyone have an explanation on this, or opinion? I figured a hybrid would be Si driving Ge, and was confused to see the other way around in so many designs.

[mac]

For a fuzz face you should use a non leaky device at q1, and whatever leakage at q2.
q1 is the tricky one, if too leaky you will need a bigger-than-8k2 collector resistor at q2, and small changes in temperature will have great impact on q2 collector. A tiny leakage means a q2 collectror resistor closer to 8k2 and small thermal runaway.
Go to my gallery, you'll find a pdf about FF bias, and a small app to estimate bias.

The pair {89, 10ua} and {120, 470ua} looks good.

If you have some low gain silicons and some leaky germs you could try a si-ge with a 2k fuzz pot, good sounding and stable.

mac

[LucifersTrip]

Isn't tons of the fun of DIY trying out all the combos and hearing for yourself which one you like better?  You only mentioned about 5 candidates, which means you could pretty much try every combo on a breadboard in about 15 minutes.

I've made a bunch of ge Fuzz Faces, and honestly, in the end, I've rarely gone with the one that looked the best on paper.

But yes, low leak for Q1 is a good place to start if you are going through a bunch. You can easily try a low leak 60 hfe as Q1 and the 89 in Q2....you never know what you'll like the most.

For the leaky ones, I've had a ton of luck with gains as low as 25-40 hfe as Q2 in an Orpheum (Fuzz Rite)

good luck

[Earthscum]

Thanks a bunch on input. Most of them I have are low gain, low leakage, but I have a few that are low gain, <300uA leakage (I only remember because of 300 being my mark upon R.G.'s suggestion in his FF article). Ironically, the only working one with leakage over 300uA was the highest gain one (still in the box... "SK-3003 RCA Top-of-the-Line Replacement Transistor pnp Type AF Driver and Output Stages (9 V Supply)")

Oh yeah... Mark Hammer just did a chunk on the Orpheum vs (?). I'll have to go recheck that thread... I remember finding a couple good links that mentioned some "mojo" transistor ranges.

Isn't tons of the fun of DIY trying out all the combos and hearing for yourself which one you like better?  You only mentioned about 5 candidates, which means you could pretty much try every combo on a breadboard in about 15 minutes.

15 minutes my AZZ!!! I was on a FF kick last night for 3 hours straight... another hour before and after, off and on, and I came home tonight and jammed it a bit. That was only swapping a 5088, BC108, BC109, 3904, and a 2SK117  (I liked it the best out of the 4 types I have. More gain, nice crisp highs. Only took me swapping each in, checking bias, and playing each once to decide. MPF102 and the 5457 may sound better for a fat low gain if I had an NPN around the 30-50 gain range for Q1).

So far I'm digging on the BC108/109 combo with a .01 feedback cap on Q2, and a 5089 driving a 2SK117Y (which just happens to be a direct drop-in, just flip it around from your normal position). If I were using a battery, MF resistors, and had this in a box, the 5089/117 would be ultra quiet, even at high gain on the BB with a badly grounded PS it was quieter than most other distortions. If I can get those Ge pair biased up and sounding good, I am putting 3 FF's in a single box.

[LucifersTrip]

Isn't tons of the fun of DIY trying out all the combos and hearing for yourself which one you like better?  You only mentioned about 5 candidates, which means you could pretty much try every combo on a breadboard in about 15 minutes.

15 minutes my AZZ!!! I was on a FF kick last night for 3 hours straight... another hour before and after, off and on, and I came home tonight and jammed it a bit.

hhaha!  man, there's a big difference between practicing and testing...lol. How are you testing each pair to take that long? I usually start with a Q1 that looks good on paper, then swap out Q2 every minute or two. I play chords, then hit a few single notes on each string and wait  for the decay.  When you get one that sounds way better, put it aside and have a shoot-out at the end. I usually go thru 20-30 different Q2's in about an hour.  The only time that it takes longer for me is when I  have 2 that are so close I have to swap back and forth numerous times to decide. 

anyway, have a blast...if it's 3 hrs of fun, it's worth it!

[petemoore]

  Just a bit too much gain, and most any of those transistors'd make cool sounds.
   The '89' is the 'higher for Q2' choice, or lowish gain Si transistor...
  Still swappin' after all these years...for all the good it really does.
    Each swap either matters or not, and to find 'it' IME it helps to try a transistor set for a good time after tweek-dialing in the transistor set, ie adjust gain, guitar volume...
   When using ballpark gain/low leakage transistors, the important thing seemed to be the voicing/gainset/bias...they all work similarly and...try it for days...tweek something...either way has set of compromises...etc.
   It's easy to read ad copy or other and get caught up thinking there is FF beyond normal FF...I started giving up on the notion and have tried some 'others' FF's in the mean time..play devil and just say..there isn't.
  Low leakage/ballpark gains, set the thing up so it's not over the top but gets good fuzz and cleans up at the guitar, there's the voicing [easily insterted cap/tweek right off of the circuits DPDT lugs, other]..otherwise it sets up pretty simple and getting the gain to go 'over/under' where a good bit of cleanup range is using guitar volume.

[LucifersTrip]

There are some "shortcuts" I've found with certain fuzzes to know when you've got a "good" set of transistors.  For the Fuzz Face, if you built to spec and you get the 4.5v with Q2's collector resistor set very close to the 8.2K (7K - 9K), you've most likely got a "good" set.

[Earthscum]

Well, I tried out the OC77 in Q1 (about 87hFE today), and the SK3003 in Q2. I must've been hasty in measuring the leakage. Today it settled right at 260uA, with about 120 gain. It started out around 380uA.

So, I have an apparent lackluster for Ge, lol, but I'm definitely bagging these together. On Bass, the contrast in clipped tones vs. dry tone is so great that I can't really tell one from another. The sounds I was getting were almost right on par with my Bazz Fuss. But, when I plugged in the guitar, it was a night and day thing... which is why I'm bagging them for a future project. Maybe I'll just populate one of my boards (I made a couple extras when I made my buddy's FF... NTE158's, luck, a bit of tweaking, and I made a really nice sounding unit!). I'm probably gonna try the same collector resistors I used in his (gotta pull out the notes, but I think I biased Q2's Rc to 4.5V wiith a 18k Rc on Q1... it really smoothed out the tone over stock values, I think by limiting the bandwidth with the increased collector current?).

Anyways, I'm having fun. I'm gonna go hit it back up later... probably have it boarded for the weekend messing with different values, and maybe an even lower gain device in Q1. Then, of course, gonna have to try a buffer and tone stack... (tonebender, right?). 

[LucifersTrip]

Well, I tried out the OC77 in Q1 (about 87hFE today), and the SK3003 in Q2. I must've been hasty in measuring the leakage. Today it settled right at 260uA, with about 120 gain. It started out around 380uA.

So, I have an apparent lackluster for Ge, lol

lackluster? isn't it good that the leakage was lower than you initially measured?