Need help tailoring boss ds-1 w blend circuit (math/theory encouraged!)

[Keeb]

So I thought I would build a distortion pedal for the bass player in my band. Decided to go with a boss ds-1 type pedal modded for bass guitar.
I also thought I'd add a blend function and a diode switch.

I based my build on the Tonepad TL072 version of the curcuit (http://tonepad.com/getFile.asp?id=78) but changed caps suitable for bass.
The blend function was the one proposed here: http://www.diystompboxes.com/smfforum/index.php?topic=83869.0;prev_next=prev.

The pedal works; it distorts and all the pots do what they're supposed to, the diode switch as well. Volume drops progressively between no diodes, 2xLEDs, LED+1n914 and 2x1n914 - but that's just the forward voltages of diodes at work...right? All diodes are wired in parallel.

The problem is the blend control. When turned all the way CCW the sound is clean (some distortion can faintly be heard) but the sound is gated and lacks sustain.
In full CW it's all distortion but the volume is lower than bypass. I guess that's the 100K from the wiper attenuating the signal?
When I place a jumper between the tone and volume pot (with the blend pot fully CW, therefore jumpering the blend circuit) the signal goes way above bypass level so I assume the pedal circuit works as intended. I know the tonepad schematic differs from the one posted in the blend-thread here at DIYSB but I don't know in what way that should matter (converted for true bypass?).

Here's what "my" circuit looks like (I omitted the op amp parts and the last transistor for clarity). I added the 1uF cap going to the dual ganged pot so the DC bias of Q2 wouldn't be affected.
Q1 and Q2 are 2N5088 and Op amp in blend circuit is TL071.

http://i.imgur.com/SdMIugu.jpg (image was huge so I decided to link to it rather than embed it).

How can I make the clean signal sound more musical and less gated?
Any way to keep the blend circuit from attenuating the signal?

[WaveshapeIllusions]

An active blend would be more effective. Instead of mixing before the volume pot there, take the output of the DS1 before the tone pot and the clean line into an inverting opamp. Choose the feedback resistor and two resistors (one for clean & one for dirty) for what you want the max gain of either side to be. I think a gain of 2 is usually pretty good, so I'd say use a 100k feedback resistor and two 50k input resistors.

So you have two 50k inputs then, but what of the blending? Wire up both parts of the dual pot as rheostats, reversed from each other. Dirty goes in one, clean in the other. Both feed to their respective input resistors. At center point both pots should be about 50k; add that to the other 50k input resistors and that's a gain of unity each side. Turn it so the pot shows no resistance to the dirty side and it has a gain of two, while clean is attenuated a bit.

Alternatively, you can use a single pot and just vary the level of the dirty signal, while the clean stays at the same level. Using the same values as before, clean has a gain of 2 all the time and dirty goes from 2x to 0.75x. I personally prefer that way for bass, since the dirty signal tends to seem louder anyways.

Actually, after looking at the schematic (5thumbs) you linked a bit more, I think I noticed something else. The clean signal is tapped off right after the JFET switch, which is before a common emitter amplifier that inverts the signal. The two opamps in the DS1 look non-inverting, so the dirty signal is inverted relative to the clean at the mixing point. That would cancel a lot of signal and quiet things a lot. Try taking the clean signal from right before the first opamp and see if that fixes it first.

Hope this helps.

[Keeb]

Thank you for your reply.

I decided to go with a simple fet buffer parallel to the input and using a single pot to blend the output from the distortion and the output from the clean buffer.
Some difference in loudness/gain from the blend pot but it works and this will be a kind of "always on"-pedal so it should be easy to dial in a good setting anyway.

If it doesn't work out, I'll try your suggestion.

[WaveshapeIllusions]

I would suggest using the FET as common source rather than a source follower. As noted earlier, the DS1 looks like it inverts the signal, which would explain the volume drop. A common source stage has clean and dirty at the same polarity and allows you to boost the signal.

[Keeb]

Thank you for your input!

I took your advice and fed the buffer the inverted signal instead - it works much better.
I added a trimpot after the buffer to be able to tame the signal a bit, and I think I'm done with this one!

Thank you for looking at this for me, it would have taken me ages to sort this out myself.