"Frankenstein'ed" circuit has ground problem(s). Please help, if you can

[KarenColumbo]

In the course of some experimentation I combined the AMZ Mini Booster (image left side) with a parametric EQ circuit from Rod Elliott (on the right).



Only thing I changed are C3 and C4 in Rod Elliott's part of the circuit: I took 2n2 instead of 4n7 to lift the center a bit. And I happened to have only OPA2134s instead of the 5532.

Principally this Frankenstein monster of a circuit sounds fine, but I have some hum which gets very loud when I touch the double pot. It seems I got a real ground problem, but I can't see where I failed.

Could you please point out the mistake(s) I made here?

(I should remark that I want to stick a bit of overdrive after the output)

[Kipper4]

Please post links to the original circuit snippets.
I might spot something.
Thanks
Rich

[KarenColumbo]

Here they are, Rich

http://www.muzique.com/amz/mini.htm
http://sound.whsites.net/project152-1.htm
figure 7 @ rod elliott
thx!

[Tubebass]

The Mini Booster circuit has a pretty high output impedance, and the parametric EQ looks like it has a low input impedance. Try a buffer between the two and see if that fixes your problem.

[Kipper4]

The buffer in rod Elliot's fig5 u1a will do.
Probably put a 10k series between the mini booster and buffer.
And you'll will need a cap on the output.
Plus there's no biasing of the op amps in the eq.
In Rods the biasing is probably supplied by a previous stage I haven't read the whole thing.
I'm not sure R1(100k) is needed too.
Rich

[samhay]

Are you powering the op-amp half of the circuit with a bipolar supply as intended by the designer?

[KarenColumbo]

Ouch. Should have held the horses. Bipolar! Would this do the trick?
http://gaussmarkov.net/layouts/bipolar/bipolar-schem.png

[samhay]

^yes, but you could alternatively tweak the circuit to run on single supply - connect all of Rod's ground points to Vref at half supply (e.g. made with 2 10k resistors and a large cap).

[KarenColumbo]

Yeah, I thought of that, too - but this would shrink the range of the EQ's gain (+/- 12 dB @ +9/-9V), wouldn't it?

[samhay]

It doesn't change the gain, but it does reduce the headroom.

As the tone stack has gain > 1, I would put a level control between the Minibooster and tonestack. I would also use the other half of the op-amp as a buffer between the level control and the U1B. That way you can turn the signal if it is hitting the power rails and the buffer will ensure the tone stack behaves itself.

[Kipper4]

Yer The mini booster will have shed loads of gain and likely clip the eq.
Why not put the Eq before the booster.
It (The EQ) will still need to be biased at 4v5 and decoupled from the booster with a cap.
I hope this helps
Rich

[KarenColumbo]

Thanks, guys, much appreciated!

I thought of putting the EQ before the booster, too. But I wanted to make some sort of "deluxe version" of the well-known tube screamer. Since the Mini Booster makes any guitar sound more lively, I wanted to split the signal, further isolate these lively frequencies and then destroying them with some overdrive. In the end I wanted to mix both the undistorted and the distorted signal together again.

Btw., I tried that already and I still had major difficulties with hum and such. Damn physics getting in my way again Seems one cannot simply mash together several circuits without thinking of what electrons actually DO in there.

But I will try Signal -> EQ -> Mini Booster. Sounds sensible.

[KarenColumbo]

I ripped this apart and took heed. Thanks again for the creative inputs.
I came up with this (I hope you're not sick of my groping in the dark):


It's more or less the AMZ Mini Booster into an opamp buffer into some sort of mid boost, then a switchable cap to cut some bass, if so desired, the ROG UBEScreamer with the tone stage left out. The clean signal leaves the circuit right before the mid boost and mixes into the last buffer stage on the 4049.

Principally it sounds nice (LOTS of gain - probably R19 is way too small?), smooth, creamy overdrive.

Still: I got some hum when idle. I think this Frankenstein drive has too much gain somewhere so it amps up dirt from the breadboard, the guitar and whatnot. It's not extreme but I get some pops when I touch the gain pot.

Now I have questions to the community of experts:

1. Is there a grounding problem? Did I miss some earthing or whatnot?
2. I think I should roll off some treble in the end to get rid of the (relativeley small amount of) hiss. Where and how would you put a low pass?
3. I put pin 3 of IC1A to Vb (4,5V) via 1M resistor. Should I do the same to pin 5?
4. I'm not happy with the "CLEAN MIX" thing. I'd love to really BLEND the distorted and the clean parts, but after trying several mixing circuits I gave up. How would you do this?
5. Can I safely "noiseless bias" Q2 by decreasing the size of R3 and R4 to something less noisy? Like ... 47k or so?
6. I thought about putting in another 5532 instead of the IC2A buffer stage - better specs. But does this make sense? I could always do something with the second half But is it necessary?

Thank you all anyway

[Edit 1] Pin 1 of the 4049 now gets his 9V from the filtered +9V output of the PSU circuit AND is grounded via a fat 100uF pol cap after the 270 R resistor. No change - but I feel safer. Pins 5 & 8 of the 4049 are grounded as demanded by ROG.

[Kipper4]

Does the noise go away if you take the amz booster out of the equation. Make c5 the new input.

I suspect the noise might be from the Amz.

If it's on the breadboard it would be easy to try.

Rich

[Cozybuilder]

A couple of thoughts-

It looks like you are trying to introduce a lot of filtering for the power supply. Look at the Low Pass Filter calculator: http://www.learningaboutelectronics.com/Articles/Low-pass-filter-calculator.php You can get a nice knee at about 34 Hz with 47R and 100uF, or 22R and 220uF etc. By setting the knee significantly below the AC source frequency, you will get the maximum benefit. However, electrolytics aren't great at removing high frequencies, so its a good idea to stick another cap in parallel with the electro that is, a ceramic. 100nF does a good job, and doesn't change the knee frequency (100uF compared to 100.1uF)

If you really want to split off the power for the different chips, OK, but for the source to IC2, change your resistor R22 to 47R. For biasing the mini-Mu, simply run a 1M from Vb to the gate Q2 rather than setting up the bias network drawn. Also just run "9V circuit" to the drain (Q2).

Do you really need C9 and C10 in series? I would think the 100nF alone would be sufficient.

C11 and C12 in parallel- again I think either one alone would be fine.

Maybe R16 should be connected to the output of the inverter amp, ICD2 at pin 10, or alternatively wire the wiper to lug 3 also. It shouldn't be popping when you touch it, unless the pot has an internal short to ground, perhaps from buildup of the wafer dust? If its a new pot, it could be defective.

[KarenColumbo]

Thank you, Rich and Russ,

Will try this tomorrow and report back Can't wait

[KarenColumbo]

Does the noise go away if you take the amz booster out of the equation. Make c5 the new input.

I suspect the noise might be from the Amz.

If it's on the breadboard it would be easy to try.

Rich

Hei, Rich, I just tried. I think the noise is there already, it's just amped up by the AMZ.

Here's two sound samples recorded with my Zoom H4 recorder:

Here's input @ AMZ: http://www.mrscolumbo.com/uploads/stomps/Noise_Pre.mp3
Here's input @ C5: http://www.mrscolumbo.com/uploads/stomps/Noise_Post.mp3

The hum is obviously made by improper shielding on the breadboard. It's there even when I omit the circuit completely. What I'm worried about is the hiss. Can it be "introduced" by something in this circuit and fed into, I dunno, the ground maybe?

And I'd love to keep the AMZ - he adds some "bell-like" quality to the sound, like an old strat sounds played through a good tuba amp But maybe it's too much here.

I will try other variations and Russ's suggestions on tuesday.

Thank you!

[KarenColumbo]

A couple of thoughts-

It looks like you are trying to introduce a lot of filtering for the power supply. Look at the Low Pass Filter calculator: http://www.learningaboutelectronics.com/Articles/Low-pass-filter-calculator.php You can get a nice knee at about 34 Hz with 47R and 100uF, or 22R and 220uF etc. By setting the knee significantly below the AC source frequency, you will get the maximum benefit. However, electrolytics aren't great at removing high frequencies, so its a good idea to stick another cap in parallel with the electro that is, a ceramic. 100nF does a good job, and doesn't change the knee frequency (100uF compared to 100.1uF)

Do you mean C20? I paralleled it with a 100 nF now and will do some A/B tests.

If you really want to split off the power for the different chips, OK, but for the source to IC2, change your resistor R22 to 47R. For biasing the mini-Mu, simply run a 1M from Vb to the gate Q2 rather than setting up the bias network drawn. Also just run "9V circuit" to the drain (Q2).

I did that now and it sounds more open, but drives a lot more. Very pleasant surprise, but the need for some treble roll-off is very obvious now

Do you really need C9 and C10 in series? I would think the 100nF alone would be sufficient.

But they ain't in parallel - or did I overlook/misread something in the schematic?

C11 and C12 in parallel- again I think either one alone would be fine.

I can't seem to find the schematic now, but I think it was a mid booster by Jack Orman (?) that specified a 3.3uF cap there. I don't have one so I tried to come close. I will try and keep 'em single. Is paralleling caps bad practice btw? I run into this problem from time to time. Ordering specific ones wouldn't be a problem but it takes some days ...

Maybe R16 should be connected to the output of the inverter amp, ICD2 at pin 10, or alternatively wire the wiper to lug 3 also. It shouldn't be popping when you touch it, unless the pot has an internal short to ground, perhaps from buildup of the wafer dust? If its a new pot, it could be defective.

I did that according to http://runoffgroove.com/ubescreamer.html. They say if it's wired this way the pot controls gain for both IC2D and IC2E (in the schematic shown on ROG's websites these are U1B and U1C). I will try your suggestion of course!

Thank you Ross!

[Kipper4]

And if you change the 1n5 ?
What does it sound like? more hiss ? less hiss ?

[KarenColumbo]

And if you change the 1n5 ?
What does it sound like? more hiss ? less hiss ?

Which one do you mean? Can't find it