My Tube Distortion Pedal Circuit

[Paul Marossy]

For anyone interested, I made up a little page for my first working tube distortion pedal of my own design. WARNING: The soundclip is loud. It is my first attempt at a sound clip from a guitar effect, and I know the levels were a little high... anyhow, it was recorded thru a ZOOM 9030 for a little reverb and then thru my Bulldog speaker cabinet simulator to the sound card.

http://www.diyguitarist.com/DIYStompboxes/PJM-Overdrive.htm

Any suggestions or thoughts on the design are welcome. The main thing that I think could be tweaked is the tone control. The way I have it now works pretty well, but I was thinking that there may be something that might be better. There is some latitude with the LM386 at the output, so it could stand a little signal loss at the tone control section.

[petemoore]

Lotsa Drive in there. Nice job on the case fitting !!! Looks good Sounds Great !!!

[Lonestarjohnny]

A Lil to many lows for me on the full cords , but just right on the lead tones, I liked the way the highs were articulated, very interesting, and a good lookin design on the box,
Great lookin and sounding pedal,
JD

[will]

Hi Paul,

Any suggestions or thoughts on the design are welcome. The main thing that I think could be tweaked is the tone control. The way I have it now works pretty well, but I was thinking that there may be something that might be better. There is some latitude with the LM386 at the output, so it could stand a little signal loss at the tone control section.

Your design looks interesting. I haven’t heard the sound clip yet but I wonder about a few things in the circuit.

I don’t think you need an input cap before Q1; however I believe you do need one before your gain pot.
Also no input cap is need for the input of the 1st tube V1A.

I’m not sure of the Bias trimpot but the 10K bias resistor seems low. Looks like you are throwing away some gain, and some good tube clipping, with the two 10K resistors connected to the Bias. The Bias will be pretty close to the V- anyway.

Your 1st major loss of gain is due to the diodes. Your tube is clipping around 20 volts, then your diodes are clipping less than 1 volt. You could stack some more diodes to recover some gain.  Or you could add some variable resistance in series with the diodes. You could dial in how much of the of harsher diode clipping you want.

Personally I would do the harsher clipping before the tubes, so the tubes can sweeten the sound. You wouldn’t have the big gain loss, In fact you would be trying to get rid of excess gain.

Your next major loss of gain is the tone and volume control combo. When you cut highs with the 50K pot you cut volume with its 5K pot, you actually swamp the 5K volume pot.  I think the tone and volume controls should be swapped. You will retain more gain and the tone control should work much better. If you want to cut more highs you could use a 10K tone pot. Also connecting the .01 uF to the wiper should work better, leaving one end free. Also drop the .1uF cap on the input to ground next to the LM386, I think this was doing the real work in cutting the highs. Problem is it changes with the volume and the tone control setting, not very consistent.

Not sure if the LM386 would be needed. At worst case with the tone control suggestions you would have similar output to a MXR Dist+.

I hope that helps.

Regards,
Will

[Skreddy]

Sounds dead-on to me, Paul!  Excellent job!  I can hear some Weezer tones in the chord work and some Jimmy Page tones in the sliding stops in the lead work.  All in all, a definite winner and certainly more versatile than any solidstate overdrive/distortion/fuzz I've heard.  It's got a refined/rough thing going on.  Even though you've got clipping diodes in there.  How does it sound without those, BTW?  How does the LM386 react if you hit its input really hard?  Is that a pleasing distortion or to be avoided?

From a DIY standpoint, I like the single PS transformer.  

Did I read that right; 220uf output cap?  Perhaps a bit of overkill there. hehe

[Paul Marossy]

Thanks for the encouragement everyone.  


Will:

To preface my response, keep in mind that my design is roughly based on the Shaka Tube, but a hot-rodded version of it. The volume/tone control arrangment is basically the same as the Shaka Tube.

"Your 1st major loss of gain is due to the diodes. Your tube is clipping around 20 volts, then your diodes are clipping less than 1 volt. You could stack some more diodes to recover some gain. Or you could add some variable resistance in series with the diodes. You could dial in how much of the of harsher diode clipping you want."

Yeah, I thought that those diodes at the end were killing my signal, but I still had a very, very low output even with them disconnected... so, I have to conclude that that is not the reason for the low output.


"Your next major loss of gain is the tone and volume control combo. When you cut highs with the 50K pot you cut volume with its 5K pot, you actually swamp the 5K volume pot. I think the tone and volume controls should be swapped. You will retain more gain and the tone control should work much better. If you want to cut more highs you could use a 10K tone pot. Also connecting the .01 uF to the wiper should work better, leaving one end free. Also drop the .1uF cap on the input to ground next to the LM386, I think this was doing the real work in cutting the highs. Problem is it changes with the volume and the tone control setting, not very consistent.
"

That might be an improvement. I used a 50K because that's what I had at my disposal. The Shaka Tube uses a 100K tone and a 5K volume. It seems to work fine for the Shaka Tube, but probably not the optimum for my design. That 0.1uF cap to ground at the LM386 is needed, IMO. It is way too bright and harsh for my taste without it there.

JD:

That sound clip was recorded with the tone control almost totally to the bass end. It can get quite a bit brighter and less bassy.

Skreddy:

Those clipping diodes seem to do a lot less than it would appear. It doesn't change the volume/sound of the circuit that much with them in there. I suppose that using silicon diodes instead of the germanium ones might sound more different. The LM386 DOES NOT like having its input hit really hard. Unpleasant distortion and oscillation can occur. Maybe the cap at the gain control as suggested by Will could help with that? Yeah, a 220uF output cap. It seems big, but it's right off of the LM 386 data sheet. It seems to work OK.

[Peter Snowberg]

Looks great Paul!

Take care,
-Peter

[PaulC]

I did notice you didn't have a grid leak resistor  drawn in for  V1A.  It needs to have a ref to V- feeding the cathode.  Leave the cap there because the gain pot is grounded, and add a resistor from grid to  V- to bias the stage.  PaulC

[Paul Marossy]

"I did notice you didn't have a grid leak resistor drawn in for V1A. It needs to have a ref to V- feeding the cathode. Leave the cap there because the gain pot is grounded, and add a resistor from grid to V- to bias the stage."

Ahhh.... maybe that is the reason for the low output withou that LM386 at the end!? How would I size that resistor?[/i]

[will]

Hi Paul,

"I did notice you didn't have a grid leak resistor drawn in for V1A. It needs to have a ref to V- feeding the cathode. Leave the cap there because the gain pot is grounded, and add a resistor from grid to V- to bias the stage."

Ahhh.... maybe that is the reason for the low output withou that LM386 at the end!? How would I size that resistor?[/i]

You could put the coupling cap before the gain pot and connect the bottom of the pot to the V- and the pot becomes the grid leak resistor for the 1st tube. Currently the gain pot looks like its dropping approx 4.5 volts across it which will make it quite noisy as you change the gain.

I think your sound clip sounds pretty good.

Regards,
Will

[PaulC]

Move the .01 cap to the drain of the jfet, and have the gain pot going to V- instead of ground.  This should do it.

Also you might try bumping the voltage up on the jfet for more drive.  Have the gate/source going to V- instead of ground to make the jfet think it's seeing a 28v supply.  

Later - PaulC

[Paul Marossy]

Will:

I'm glad that you like the sound clip.  
It's kind of my rough draft so to speak.

"You could put the coupling cap before the gain pot and connect the bottom of the pot to the V- and the pot becomes the grid leak resistor for the 1st tube. Currently the gain pot looks like its dropping approx 4.5 volts across it which will make it quite noisy as you change the gain."

Actually, the gain pot doesn't make any noise when you turn it. It does have a limited travel, though. Most of the response is in the last 1/3 or so, going clockwise. Making the pot the grid leak resistor kind of makes sense, but I'm not sure how it would affect the sound. Most of the distortion seems be controlled right at this point. The tube itself doesn't seem to have much distortion to add, but the LM386 will add some, but it can be ugly sounding...

Paul C:

"Move the .01 cap to the drain of the jfet, and have the gain pot going to V- instead of ground. This should do it.

Also you might try bumping the voltage up on the jfet for more drive. Have the gate/source going to V- instead of ground to make the jfet think it's seeing a 28v supply."

Hmmm... that's an idea. I think that the pedal has plenty of gain already, but it might help it to have a little more headroom, I suppose.

Maybe the next step is to do a Rev. 2 schematic with these ideas incorporated into it and see how that works.

[Doug H]

Looks nice, Paul.

I like the look of your box and how you mounted the tube. I'm surprised you drilled all those holes. I assumed it came like that when I saw the picture. Nice work!

Doug

[Paul Marossy]

Thanks, Doug.
It only took about 5 or 10 minutes to drill all of those holes. I just stuck it in a vise a drilled away. Aluminum is easy to drill....
I am happy with the way I mounted the tube socket on this one. I got the socket from www.hoffmanamps.com and it was a leftover from my Firefly build.
It's still difficult to fit all of these parts inside such a small box. I suppose if I designed a PCB for it, that part could get a lot smaller in size.

[will]

Hi Paul,

I still don’t fully understand the Shaka Tube tone control that you used. It only acts as a LP filter. I would expect its effect is quite subtle at first. It sacrifices a huge amount of gain, unless used with a very large volume pot afterwards or a high impedance buffer. As you increasingly cut highs you also proportionally cut output volume. Bass boost is really extreme treble cut.

I propose that in the Shaka Tube the tube clipping is not all that harsh and since you need to drop the volume considerably it works as a mild tonal change for the 1st half of the pot rotation.

Your hot rodded circuit is different; the diodes will provide more high order harmonic distortion and you have already dropped the volume. To soften this harshness you need a less subtle LP filter. You actually created this when you put the .1 uF cap after the volume control. I’ll bet if you removed the .01 uF cap and connected the tone pot to ground you wouldn’t hear anything different.

If you moved the .1uf to the wiper of the tone control, insted of the wiper of the volume pot, you would essentially get the same response as you get now however the changing volume will not change the frequency balance. Increasing the volume pot to 500K or 1 M would recover much of the lost volume. The volume will be mostly set by the clipping diodes you use.  It would also make the output level less affected by changes to the tone.

Regards,
Will

[Paul Marossy]

"I still don’t fully understand the Shaka Tube tone control that you used. It only acts as a LP filter. I would expect its effect is quite subtle at first. It sacrifices a huge amount of gain, unless used with a very large volume pot afterwards or a high impedance buffer. As you increasingly cut highs you also proportionally cut output volume. Bass boost is really extreme treble cut."

Yes, it is a simple low pass filter, but it has quite an effect on how it sounds througout the pot rotation. It can get quite harsh sounding near the treble end, and I did not notice a signifigant drop in volume as the treble is increased. I'd have to play with the controls some more. Maybe it does drop some, but I don't remember it dropping...

"I propose that in the Shaka Tube the tube clipping is not all that harsh and since you need to drop the volume considerably it works as a mild tonal change for the 1st half of the pot rotation."

Yes, the Shaka Tube is a starved plate lower gain circuit with not a whole lot of clipping. There is a soundclip of it here: http://www.diystompboxes.com/pedals/sounds/shakaTubeDemo.mp3

"Your hot rodded circuit is different; the diodes will provide more high order harmonic distortion and you have already dropped the volume. To soften this harshness you need a less subtle LP filter. You actually created this when you put the .1 uF cap after the volume control. I’ll bet if you removed the .01 uF cap and connected the tone pot to ground you wouldn’t hear anything different."

Maybe so. The tone control as shown isn't that subtle in any case. That .1uF cap at the input of the LM386 is there because without it, it sounds way too harsh and it grates on my ears. This smooths out the tone quite a bit.

"If you moved the .1uf to the wiper of the tone control, insted of the wiper of the volume pot, you would essentially get the same response as you get now however the changing volume will not change the frequency balance. Increasing the volume pot to 500K or 1 M would recover much of the lost volume. The volume will be mostly set by the clipping diodes you use. It would also make the output level less affected by changes to the tone."

I would normally agree with you about the clipping diodes, but even if I remove them from ground, I still have a very low output without the LM386 at the output. I don't understand why this is, but it is. I would think that the gain pot after the FET would have to do more with the low output than the diodes based on my experimenting. Changing the tone cap from .01uF to .1uF may help to create a better tone control. About the 5K volume pot, IIRC, when I made the pot larger in value (10K), it actually decreased the volume. I am totally baffled by that one.

[will]

Hi Paul,

I would normally agree with you about the clipping diodes, but even if I remove them from ground, I still have a very low output without the LM386 at the output. I don't understand why this is, but it is. I would think that the gain pot after the FET would have to do more with the low output than the diodes based on my experimenting. Changing the tone cap from .01uF to .1uF may help to create a better tone control. About the 5K volume pot, IIRC, when I made the pot larger in value (10K), it actually decreased the volume. I am totally baffled by that one.

Maybe you are not getting the gain you think you are from the tube sections. You should be getting tons of gain! Missing the grid leak resistor the 1st tube section may self bias for next to nothing gain or even loss. I'm not sure. 1 Meg & 100K are typical values used.

Normally the tubes auto bias, in your 2nd tube stage you are forcing the tube bias. I take it's used to mis-bias the tube to increase distortion. If it is adjusted way out, the tube may not provide any gain.

Regards,
Will

[Paul Marossy]

"You should be getting tons of gain! Missing the grid leak resistor the 1st tube section may self bias for next to nothing gain or even loss."

Yeah, I was thinking that I should be getting tons of gain, too. I bet that arrangement at the gain pot might be the problem with the low output. Even so, I kind of like how it sounds with it the way it is right now. I guess I'll have to do some tweaking and see what I come up with.

[MarkDonMel]

I'm not knowlegeable enough to offer any sort of technical opinion, so all I'll say is that it is bitchen.

Great job.

Matt.