Companion Fuzz with contour & blend controls

[dthurstan]

Hi

I've been working on this pedal for a little while now. It's the Shin Ei Companion Fuzz with added contour and blend controls. The blend circuit is taken from the BYOC 5 knob compressor circuit http://byocelectronics.com/5knobcompschematic.pdf. Here's the schematic:


But I've run into some problems. It starts to oscillates as you mix in the dry signal (as the pot moves towards the R13/R14 side of the circuit).

I've been through the vero layout and can't find any bridges, here is the layout and a photo of the underside.


I did breadboard this before I built and didn't have this oscillation issue. Just hoping someone could shed some light on it, should I increase the blend/balance pot to say 25k or is there a fundamental flaw in the design, i.e. is the blend circuit always going to introduce oscillation?

Here are my transistor voltages
Q1
c 4.00
b 0.65
e 0

Q2
c 0.9
b 0.6
e 0


Q3
c 6.9
b 0.87
e 0.2

Q4
c 6
b 0.87
e 0.2

Thanks
Dave

[Mark Hammer]

My firs inclination would be to lift each end of the Blend pot, one at a time, to see if there is anything "bleeding" from one side into the other.  Perhaps R12 and C10 should change their order, such that the Blend pot provides a true attenuator for the clean signal.

[duck_arse]

Q2 collector volts should be .... higher?

[dthurstan]

Thanks Mark, I'll have a go at lifting each side of the blend pot. The r12/c10 order does seem strange, that was how it was done in the 5 knob compressor.

Thanks duck, yeah it is a little low, does increase a bit when you turn the fuzz knob. I have probed at the fuzz circuit output r7 & does sound ok there. I did use a 1m instead of 1m2 for r4.

[antonis]

Dry / Wet signals are out of phase (180⁰)..

That's probably the reason, togheter with Q2 low working voltage, for oscillation..

[dthurstan]

Thanks Antonis, doesn't the wet signals phase depend on the position of the fuzz pot?

[antonis]

Strictly speaking, No.. 
Part of signal comming out of Q2 Collector is fed back to Q2 Base, via Fuzz pot, thus subtracted from incoming signal..)
Fuzz pot sets Gain of Q2 stage (via NFB action) and not phase..
(there is a small phase shift trough C5, like in any other capacitive element  but it should be considered negligible..)

So, in general, a positive input signal comes negative out of Q3 and positive out of Q2 (due to already phase reversal through Q1)

[dthurstan]

Hi Antonis, thanks for the explanation . I was going off what I heard and my assumptions  .

So I'm trying to raise Q2 collector voltage, I've tried increasing R4 to 2M2 & reducing R5 to 33k, I get to 1.2V I'm aiming for around 2V. Not sure about increasing R4 much more. Should I stick with reducing R5?

I'm then gonna try lifting either side of the blend pot and check the bleed as Mark suggested.

[antonis]

I should leave R4 & R5 as they are and lower enough R3 Value..
(or try to halve it with the mods you've already made..)

[duck_arse]

Thanks duck, yeah it is a little low, does increase a bit when you turn the fuzz knob. I have probed at the fuzz circuit output r7 & does sound ok there. I did use a 1m instead of 1m2 for r4.

it shouldn't change at all, so this indicates a fault. you've missed a track cut under C5. I think. also, it is polite to show us the component side.

[Mark Hammer]

The Contour pot really doesn't need to be any higher in value than a few kilohms before you don't hear any difference.  So a 5k or even 2k pot is sufficient.  If you want to go "nuts", you can have a more flexible tone control by doing the following. 

1) Use a 10-25k linear pot for the "Contour".
2) Wiper of the pot goes to ground.
3) One outside lug of the pot goes to the 100nf cap you show,
4) The other outside lug of the Contour goes to a 22-47nf cap from the junction of R6/C7 and the Fuzz pot. 

In one direction, the pot's resistance is inserted between that additional cap and ground, minimizing any treble-bleed-off, while at the same time increasing the depth of the "scoop" between bottom and highs.  In the other direction, the scoop is reduced, while at the same time increasing the treble bleed-off to ground, yielding a more rounded-off throaty sound.

Because the suggested Contour pot value is actually higher/larger than you need to cancel the scoop, feel free to place a fixed resistor in parallel with the leg of the Contour pot between the wiper and C8, such that there is   more resistance on one side of the pot than the other.  In theory, this should get you a range of treble-cut after you've reached the point where the scoop is eliminated. Aim for something that gets you 2-5k at most on the "scoop" side of the Contour pot, no matter what the resistance on the other leg might be.

[dthurstan]

Thanks Antonis, R3 is how to increase the collector voltage, I missed that.

it shouldn't change at all, so this indicates a fault. you've missed a track cut under C5. I think. also, it is polite to show us the component side.

Thanks duck_arse, yep your right and C6 was in the wrong position. I've corrected this now. Here are the track and component side. Sorry I should of added this before, but I thought I wouldn't get a clear photo of it and the vero layout drawing would be better. I did make a number of corrections to the broad before I posted.



Here are the new voltages;
Q1
c 4.01
b 0.64
e 0

Q2
c 3.42-3.55 (between fuzz pot extremes)
b 0.37
e 0

Q3
c 6.9
b 0.87
e 0.2

Q4
c 6
b 0.87
e 0.2

There is still a change in the collector voltage of Q2 when I move the pot, so I'll need to see if I can spot anymore faults in my layout.
It still oscillates as the blend pot is turned to the dry side. I guess its the out of phase thing. It didn't do this when I bread boarded it.

Thanks Mark for contour pot suggestion. I was following your older advice which is why I used the 50k pot. In debugging this pedal I found your previous posts about the contour pot and the rewiring of the fuzz pot. I will try a lower value.

Still need to lift each side of the blend pot to check bleed through, but wanted to raise the collector voltage first.

Thanks again everyone for your help. Apologies that it takes so long to post and make changes, I've been trying to fit it in around childcare so it's little snatches of time here and there. I'm sure other people are in a similar boat.

Dave

[Mark Hammer]

Thanks for the nod, Dave.  I initially suggested 50k because I had little idea of how much resistance was actually necessary.  Of course, as we keep being reminded: what goes up on the internet is forever, so the unnecessarily large pot value is permanently findable; probably on a drawing I posted prematurely.  My bad.  Happily, I'm still here to say it was a mistake. 

[duck_arse]

now your Q2 base has lost 300mV. where have they gone? check your B-C resistor, and the resistor to supply - value and connection. while you are there, please measure the voltage on each leg of the fuzz pot.

[Mark Hammer]

For those who may find their interest piqued by this thread, see the pics that dthurstan has graciously taken and posted, and think they may now give it a try, I will note that I own an original.  Rehoused from the enclosure I purchased it in, but otherwise unmodified.  And every single cap on the board is a butt ugly ceramic disc.

That does not mean use of plastic caps will somehow degrade the sound - hard to know how one "degrades" fuzz - but nobody who would like to make one for themselves should get their knickers in a knot over having wima box caps, polyesters, or whatever.  The cheapest ugliest caps you have on hand, of the indicated value, are quite sufficient to deliver "authentic" Shin-Ei tone.

[dthurstan]

Thanks duck_arse, I had given up hope of finding that 300mV, then I thought why not turn the really big 50k contour pot & violà there it is 0.58v at the base. I'm getting 0v on the fuzz pot lugs btw.

Yeah Mark, I put in what I had, those 100n box caps where cheap on eBay. I'm waiting on some ceramics now, I've never noticed a difference in cap type myself. Those 100n green polys are a pain cause there to big.
Dave

[duck_arse]

the contour pot in this ^ circuit? and that's affecting ANY base voltage reading? that makes no sense. 0V on the fuzz pot looks right, tho. can you read any voltages on the contour pot? what are your Q1 and Q2 voltages now?

[antonis]

There should be something wrong on your stripboard..

None of pots should alter DC voltage readings..
They all are DC decoupled via capacitors..
(Ballance pot included..)

Check for short each one of C5, C6 & C8 at start..

[dthurstan]

Hi guys thanks for the posts.
Q1 voltages are the same.
Q2
C 2.2
B 0.64
E 0

No voltage on the contour pot. No shorts on c5-8.
I've been through the stripboard layout & can't see anymore faults. I'll keep checking. I was gonna swap out the 50k contour pot as Mark suggested anyway.

Thanks
Dave

[dthurstan]

Hi

I thought I'd update on my progress. Still having issues. So I changed the 50k contour pot for a 10k pot which is closer to what Mark suggested.
I lifted each side of the blend/balance pot in turn to see what was bleeding through, there was a lot of oscillation on the dry side. So I decided to remove the Q3 gain section to see if this reduced the oscillation, also as Antonis pointed out this was putting the dry signal 180 degrees out of phase with the fuzz signal. Removed Q3 & C12 and ran a 10k resistor from C9 to balance pot/R15.

Initially I thought this had worked! I test do a lot of pedal testing with a mini amp & a loop pedal (keep noise levels low during naps/sleeping). However, when I got round to trying it out with an amp & guitar it oscillates again. It starts to oscillate close to the mid point of the pot then goes away again when completely dry. It's almost playable, just don't ever stop playing!

Voltages are as follows:
Q1 same

Q2 (either side of the balance pot)
c 1.04
b 0.6
e 0

So the collector and the base do vary as the balance pot is moved to the centre (and also varies with the balance pot in the centre and changing the contour pot).
Collector varies upto 2v and base between 0.55-0.6v.
There are no voltages on the balance, fuzz or contour pots.

I have been checking through the stripboard layout, test shorts, checking for bridges and can't see any issues.
As mentioned before the collector of Q2 is low, I'll try increasing that to see if this resolves the oscillation issue.

Thanks
Dave