Why can't I get this clean?

[Mark Hammer]

Here's a little thing I whipped up today.  On another forum there had been a spate of threads about "low-gain" pedals, so I was intending this to provide just a hint of grit, and the capacity to provide a bit of sparkle, as well as more conventional TS-like tones.

The first stage can be set as low as unity gain, and the second stage has a fixed gain of 15.  Gain is adjustable in the first stage and in tandem with the fixed 15x gain in stage 2, can ratchet the total gain up to over 350x.  The 4k7 series resistor is supposed to provide softer clipping in stage 1. 

The 50k tone control is a little different, and pans between a treble-boost pre-emphaiss in stage 2, at one extreme, to a warm rolloff at the other.  At the treble end, the 22nf cap goes through the 6k8 resistor directly to the inverting input to provide a little more gain at just over 1khz.  At the bass end, the resistance in series with the 560pf cap is zero, effectively placing it in parallel with the "default" 100pf cap, yielding 660pf, and a treble rolloff around 730hz.

The diodes are a melange, and I didn't specify them in the drawing because I wasn't committed to anything.  I used a schootky and 1N914 in series + a 1N4001 for the first stage, and a 1N4001+Unlabelled Ge in each direction for stage 2.  Forward voltage for stage 2 is around 900mv each way.  There's a JRC4558 in the op-amp slot. and the output  is a 4u7 cap and 100k pot. (need to revise the drawing)

What I cannot for the life of me figure out is why I simply cannot get this thing clean.  I can certainly make it cleaner and dirtier, and the tone control works very nicely, but even with a simple 15x gaiin stage 2 and unity in stage 1, it never gets  clean.  A gain of 15x should not be enough to push those diodes into clipping, should it? 

So what am I doing wrong?

[amptramp]

The TS is set up as a non-inverting amplifier, unlike your first stage, so that even when it is clipping, the original signal is mixed in with a gain of 1.  The inverting amplifier here does not do that and its input impedance is very low, 22K in the midband.  The 4.7K will give a gain of 0.213 for the incoming signal.  You will get dirt once the input stage 1N4001 starts to conduct and this will be at about 0.6 volts (especially in hot weather where the diode drop is lower).  Check your guitar output and see if the signal is near that, but the clipping is probably in the second stage.

The gain of 15 in the second stage will provide clipping once the input exceeds 60 mV.  Reduce the gain of the second stage and you will get a cleaner signal.  Of course, you could swap out the Schottky diodes to get a higher voltage drop - say normal silicons or LED's.  And remember, the gain increases when the tone control slider moves to the left in this diagram.  The gain is:

=-330/22||6.8 at maximum treble
=330/5.19444
=-63.529

with this gain, clipping begins at 14.166 mV.  Some guitars can put out 500 mV (although not into 22K!), but even a light touch on the strings can give you more than this.  Not surprising that you clip all the time.

[Mark Hammer]

Good analysis.  Thanks.

The pickups aren't all that hot (40mvAC for the unwounds), and at max guitar volume it breaks up whether I use my guitars with SC pickups or with HBs.  It actually does clean up and produce what I was originally aiming for if I back off the guitar volume, which is a clue I suppose.  It'd just be nice to get it without having to dial back on the guitar.

The 22k input resistance on stage 1 was simply chosen so I could use some nice compact .1uf caps I had, and everything else flowed from that choice.  I figured that if I was going for at least a little bit of coloration that the loading of a higher input resistance would work in my favour.

The boost at the treble end was anticipated, and part of the goal, actually.  I figured it might add some upper midrange bite to an otherwise clean tone.  It adds the expected bite, but I may want to aim for a little less than 63x at max treble, by changing 22nf/6k8 for 15nf/10k.  That will yield largely the same treble boost, just not quite as much of it.

I lifted the diodes in stage 1 and that helped a bit.  I still can't get entirely clean, even at min gain, though.  May be time to swap the Ge diodes in stage 2 for Si, although that won't raise the forward voltage all that much.  These diodes, for some reason, measure around 370mv forward voltage (higher than typical Ge, lower than Si), so a 2+2 of 1N4001s will only raise the threshold from around 900mv to around 1150mv.  That may make the difference, though.  If I lower the 330k to 270k, that will help too, and will also raise the treble rolloff from 4.8khz to about 5.9khz.

The behaviour of the tone control is interesting.  At low gain settings, it makes a difference, but much less so at high gain.  The reason would seem to be because it is a pre-clip tonal change, rather than post.  At max gain, the circuit behaves pretty much like a 2-knobber, but that's okay.

[TimWaldvogel]

i would REALLY  like to see\hear the end results of your endeavors