Help recovering volume loss (super fuzz)

[DiscoVlad]

Going by the voltages you posted, and assuming you've got the pinouts right... Q2 doesn't look right, and there's something definitely wrong around Q6. All of the collectors should have the highest voltage, and the emitters lowest, with the bases in between.

This is with Q1 -> Q6 going left to right, and ignoring the wah section.

[StevenJM]

I think the collector of Q2 is connected directly to the power line, so the collector should be the same voltage as my battery... I'm not sure about the other poles tho.  I'm at work and dont have it in front of me... I can recheck my numbers when I get home.

[DiscoVlad]

If you've only got the B and E measurements backwards on Q2, it's probably ok.

Q6 though, I'd take out the two capacitors on the base and collector, and check that it biases up properly.
A quick and dirty spice simulation (using 2n2222 transistor) of just the Q6 part of the circuit gives me:

V(c):    3.97 V - This seems to vary most with the transistor gain, so could be anywhere between 3V and 6V.
V(b):    1.14 V - Depends on the base resistor tolerances - look for somewhere +/- 0.5V of this. MUST BE HIGHER than the emitter voltage.  
V(e):    0.51 V - as long as it's more than 0V, and less than V(b) it should be fine.

If your voltages come right, then replacing the capacitors should be enough to fix it. If they're wildly off from those or the previously posted voltages, the transistor may be bad.

[StevenJM]

I see what you're saying... Q2 is a little messed up then... I'll re-check everything in a few hours when I get home... I got thrown off by the way these transistors are, so maybe just my numbers for Q2 are flipped around.

Q6 is definitely messed up... I have some 10uf caps at home, I'll try swapping the ones around Q6 and seeing if that changes anything.

This stuff gets so confusing haha, especially with 6 transistor, plus the 2 for the wah, and the top of the board is just brown, so it's constant flipping the board over... So I could have easily measured the wrong poles.  Plus I was pretty whiped out last night from a long day.

I contacted the guy who sold this and he claims there was so no volume drop when he tested it (music shop with 700+ 100% feedback on ebay) and he packaged this thing for a space mission, so Im giving him the benefit of the doubt, so I'm going to double check all the solder joints when I get home and maybe it's something loose right under my nose.

[joegagan]

Remember, fuzz pedals' perceived volume can be all over the map. Especially volatile with tube amps when there is a lack of headroom. But many factors affect it. It may have sounded 'normal' to the seller in his rig.

[StevenJM]

He said it was an older man that brought a box of old pedals, a lot of MXR stuff stuff, they agreed with me in thinking that this thing has been siting in a basement for the last 20-30 years.  He was probably doing some cleaning and stumbled upon a box of his old pedals.  I want to replace the input/output jacks while I'm at it too. 

Anyways... I'm off in an hour, I'll re-check my numbers when I get home.  I'll swap those 10uf caps and see if that makes an impact.

[StevenJM]

alright... I double checked my numbers and Q2 stayed the same, but Q6 is drastically different.  I'm not sure what i was doing last night, maybe i just needed to take a break and come back to it the next day.

But Q6 is as follows...

C=0.30
B=0.84
E=0.19

I swapped the 1K resistor off of the emitter of Q6 for a 220ohm instead to try and boost the volume, except I actually think it made it even quieter.  So i think the volume issue is something other than that resistor haha.  Maybe its time to switch the caps and maybe order some replacement transistors just to be safe.

[StevenJM]

another update:  I just replaced the 10uf electrolytic cap right after Q6 and got a lot of my volume back.  I don't have enough to do the whole board right now, but I do have 5 more, so i'm going to continue swapping some out and do the rest when I get them in the mail.  Or maybe I'll cave in and go to radio shack again tomorrow and pay $1.60 for 1.

[DiscoVlad]

Good to see changing the cap(s) worked!
Not sure about those newly measured voltages though... How did the Q6 voltages look if those two 10uF capacitors are removed?

[StevenJM]

Man, I got too excited that it started working that I totally forgot about measuring Q6 without the caps.

I have 2 more caps that still need changing in the fuzz circuit... The fuzz kinda sounds like its "fighting" to get out, almost like its not sustaining properly now that it's louder.  I'm gonna spend the $3 or so and get some more caps at RadioShack and then I ordered 10 2SC828 transistors online, I'm just not convinced that Q6 is working properly.  So I'm thinking I'll finish replacing the caps and then just swap out all the transistors for new ones so I won't have more issues in a year or so.

[StevenJM]

I took the caps out before and after Q6 and the readings were the same as I posted last night.  When i get the new transistors in the mail, I'll replace them all.  I replaced the corroded input and output jacks today too (even tho they were clean inside the pedal).  The wah pot is pretty scratchy too so im going to look into getting a new one of those too.  Hopefully one of the shops online sells them.

I also replaced the 100k resistor after the last 10uf cap, before the switch to a 47k, but I don't really notice a difference.

[LucifersTrip]

I really think you might be doing too much since except for the swapped Q2 and Q6, the voltages are not much off.

remember, you can audio probe the signal after each transistor to see what you have...

After I breadboarded it, I wrote a report on the other forum.   I was using this schematic and talking about the evil scooped tone:
http://3.bp.blogspot.com/-cBIzyse_wes/T0EHt4QmTvI/AAAAAAAAA88/Y1fbps_nMsQ/s1600/univox_superfuzz_schematic.gif
It may help you.  An important note I made below: " This is really an easy fuzz to get to sound good since you are able to adjust the collector voltage all the way from 2.1v up to 7.2v and still get a killer fuzz throughout with no gating or poor decay."


========================
I finally breadboarded this and screwed around for a while....and excuse me if this was covered, but after scanning all 15 pages of this thread (other forum), there seemed to be a lack of reported voltages, and even more surprisingly, a lack of Q6 discussion.

Firstly, I'd like to thank solidhex again for reporting the voltages of an original:
http://www.diystompboxes.com/smfforum/i ... ic=64068.0

C B E
Q1 6.01 .66 .117
Q2 8.88 6.01 5.41
Q3 6.23 3.27 2.72
Q4 3.1 1.7 1.11
Q5 3.1 1.67 1.11
Q6 5.75 .98 .361

There is nothing more beautiful than starting with info from an original, excellent sounding box.

I started with BC1708's with hfe ~ 150-185 and just for the fun of it, after I breadboarded the first 2 transistors, I plugged it in at the 10uF right before the expander. It turned out that it sounded (like an overdrive) far better with a higher Q1 (a 2N2222 with hfe 235). If I used a 2N2222 for both Q1 & Q2 , it sounded worse, so I stuck with a higher Q1 and lower Q2.

I finished the breadboard off with 4 BC1708's and measured the voltages. I got so close to solidhex's above I was amazed...except for Q6, where I got 3.85v on the collector. I checked all over the place and it seems many have reported voltages in that range. Here's the one set of voltages in this thread:
Q1: 5.63 0.71 0.14
Q2: 9.76 5.69 5.08
Q3: 7.46 2.79 2.31
Q4: 3.61 2.00 3.60
Q5: 3.60 4.35 3.59
Q6: 3.71 1.21 0.59

I tried about 20 different transistors in the 125 - 700 range and the highest collector voltage I got was 4.6v. A 2N5088 with around 600+ gave me around 3.5v and a 2SC1815 with 125 gave me the 4.7v. So, I dropped an old Heathkit transistor in there with a gian of 55 and hit the mark almost right on the money, 5.8v. Big difference...it's less sharp, with a little less pronounced octave, but smoother, richer and seemingly fuzzier. The 5088 was the worst...much more harsh and not as smooth.

...but, of course, you can also adjust the resistors around Q6 to change the voltage ....my favorite is the 15K from base to ground. This is really an easy fuzz to get to sound good since you are able to adjust the collector voltage all the way from 2.1v up to 7.2v and still get a killer fuzz throughout with no gating or poor decay. The big difference is the tone. Down at 2.1v it's fatter and richer and 7.2v thinner. The volume also increases as you head towards 2.1v. If you go a little below 2.1v or above 7.2v, it drops volume quickly then sharply cuts off.

Results with base to ground resistor (15K) altered.

Q6 (hfe 55)
23K (2.1v)
12K (7.2v)

Q6 (hfe 235)
17K (2.1v)
10K (7.2v)

A big note: The 235 hfe transistor at 2.1v sounds different than the hfe 55 transistor at 2.1v and the same for 7.2v
The hfe 55 was still smoother, a bit fatter & fuzzier with a little less octave effect.
========================

[joegagan]

great post, LT. wish i had taken bias readings on this original after i fixed it.

can you tell me, does this one fall into more of the OD category, based on your above descriptions? i didn't think it was superfuzzy, but i liked nonetheless.

[joegagan]

this original sounds really good. never mind that the guy was running it into a fulltone OCD before the amp, oh well. ( skip in a minute or two two skip the talking)
also the watson electronics replica he compared to was almost as nice as the original. far gnarlier than the one in my video.

[StevenJM]

Well I think i found my problem... all thanks to the handy audio probe... i always feel like i'm doing it wrong, so I don't trust myself using that technique... but on my fy-2 build, if i would've just trusted that I was doing it right, I would've pinpointed that problem immediately... and now the same with this one.

The 10uf cap BEFORE Q6 sounds awesome, full fuzz, full volume, sounds awesome, then the next thing in line is the 10K resistor thats also connected to the collector of Q6 and the sound just dies. 

I checked the resistor and the resistor still reads at 10k, so i'm assuming its the actual transistor thats dead.

[DiscoVlad]

I finished the breadboard off with 4 BC1708's and measured the voltages. I got so close to solidhex's above I was amazed...except for Q6, where I got 3.85v on the collector. I checked all over the place and it seems many have reported voltages in that range. Here's the one set of voltages in this thread:
Q1: 5.63 0.71 0.14
Q2: 9.76 5.69 5.08
Q3: 7.46 2.79 2.31
Q4: 3.61 2.00 3.60
Q5: 3.60 4.35 3.59
Q6: 3.71 1.21 0.59

I tried about 20 different transistors in the 125 - 700 range and the highest collector voltage I got was 4.6v. A 2N5088 with around 600+ gave me around 3.5v and a 2SC1815 with 125 gave me the 4.7v. So, I dropped an old Heathkit transistor in there with a gian of 55 and hit the mark almost right on the money, 5.8v. Big difference...it's less sharp, with a little less pronounced octave, but smoother, richer and seemingly fuzzier. The 5088 was the worst...much more harsh and not as smooth.

Very interesting! Q6 shouldn't have too much effect on the tone as that section is gain recovery (approx +18dB over audio range) after the "tone" section. However, every time I've tried the calculations or done Spice sims, the collector would always hover around 3.7V...

Of course both of these methods assume "ideal" devices. Never thought to use a... shall we say "less than ideal" transistor because they tend to be noisier, but in this case, it seems to be the err... ideal for the job.

I think the 2.1/7.2V limits are from the transistor getting saturated at one end, and not having enough base current to turn on at the other. Being near these limits should slightly clip the signal compared to the "standard" 3.75V collector, which is most likely right in the middle of the transistors linear region where the output signal is cleanest.

[LucifersTrip]

great post, LT. wish i had taken bias readings on this original after i fixed it.

can you tell me, does this one fall into more of the OD category, based on your above descriptions? i didn't think it was superfuzzy, but i liked nonetheless.

thanx...I wouldn't say OD, but as you kind of answered in your next post, it's not really gnarly...and it's thinner with less pronounced octave.
with a superfuzz, you don't have to go up to the 12th for heavy octave!

either way, it still sounds very cool and probably, no 2 sound identical

[LucifersTrip]

I finished the breadboard off with 4 BC1708's and measured the voltages. I got so close to solidhex's above I was amazed...except for Q6, where I got 3.85v on the collector. I checked all over the place and it seems many have reported voltages in that range. Here's the one set of voltages in this thread:
Q1: 5.63 0.71 0.14
Q2: 9.76 5.69 5.08
Q3: 7.46 2.79 2.31
Q4: 3.61 2.00 3.60
Q5: 3.60 4.35 3.59
Q6: 3.71 1.21 0.59

I tried about 20 different transistors in the 125 - 700 range and the highest collector voltage I got was 4.6v. A 2N5088 with around 600+ gave me around 3.5v and a 2SC1815 with 125 gave me the 4.7v. So, I dropped an old Heathkit transistor in there with a gian of 55 and hit the mark almost right on the money, 5.8v. Big difference...it's less sharp, with a little less pronounced octave, but smoother, richer and seemingly fuzzier. The 5088 was the worst...much more harsh and not as smooth.

Very interesting! Q6 shouldn't have too much effect on the tone as that section is gain recovery (approx +18dB over audio range) after the "tone" section. However, every time I've tried the calculations or done Spice sims, the collector would always hover around 3.7V...

Of course both of these methods assume "ideal" devices. Never thought to use a... shall we say "less than ideal" transistor because they tend to be noisier, but in this case, it seems to be the err... ideal for the job.

I think the 2.1/7.2V limits are from the transistor getting saturated at one end, and not having enough base current to turn on at the other. Being near these limits should slightly clip the signal compared to the "standard" 3.75V collector, which is most likely right in the middle of the transistors linear region where the output signal is cleanest.

I'm not sure if you'd call the transistors "less than ideal"...just lower gain. I asked on the other forum about the average gains of the 2SC828 and the reply was, I believe, 150-200...which did give me what you calculated in Spice (3.85v).

I was shooting for 5.75V since that's what Solidhex measured on his bandmate's killer sounding original. Since I doubt anyone screwed with any components, it seems that Q6 had to be lower gain.

According to this, the 2sc828 min hfe is 65, so it's possible:
http://transistor-spravochnik.ru/description/2sc828/17535

In the end, I just wound up putting an external bias pot on Q6 to give me 2.1V - 7.2v

[DiscoVlad]

Ideal as far as transistor models go, that is linear gain, 0 Vce, etc. which despite several years of development is unattainable in actual transistors.

The data sheet I have says Hfe (at Ic=2mA, and Vce=5V) for 2sc828 is 65-700...
However, On vintage Superfuzzes the transistors weren't all 2sc828 - some had a mix of 2sc539 in there as well. These had Hfe of 90-700.

A device with an hfe of 55 is below the minimum spec for either of these (presuming equal Vce and Ic conditions as the data sheet), my guess is a damaged transistor on Solidhex's bandmates' (holy apostrophes Batman!) Superfuzz (Can I just call it the Fu Manchu Superfuzz? I'm pretty sure that's the band Solidhex is talking about. ).

Maybe the transistor in that one is damaged... but luckily in a way that sounds [really, really] good. Anyway, as per your experimentation you can get a similar sound from dropping the bias point of the base so it's all good.

[LucifersTrip]

A device with an hfe of 55 is below the minimum spec for either of these (presuming equal Vce and Ic conditions as the data sheet)

yes, though, if you use one with 65 gain (min spec), you will get very close to the 5.75V "goal". remember, I got 5.8V with the 55

my guess is a damaged transistor on Solidhex's bandmates' (holy apostrophes Batman!) Superfuzz (Can I just call it the Fu Manchu Superfuzz? I'm pretty sure that's the band Solidhex is talking about. ).

exactly
http://www.youtube.com/watch?v=MNBxv4hP184

Maybe the transistor in that one is damaged... but luckily in a way that sounds [really, really] good. Anyway, as per your experimentation you can get a similar sound from dropping the bias point of the base so it's all good.

Not exactly...I did make one "big note"

A big note: The 235 hfe transistor at 2.1v sounds different than the hfe 55 transistor at 2.1v and the same for 7.2v
The hfe 55 was still smoother, a bit fatter & fuzzier with a little less octave effect.