Need a bit of help balancing this

[KarenColumbo]

As always, the AMD CD is a real treasure chest of ideas. So I threw together a Mini Booster with some addons to get this clean "sparkle" and some eq knobs to twirl.

Here's what I have breadboarded so far:



It sounds quite fine, I think. But:

• it's not really clean, breakup sounds ugly.
• when I crank up the "TREBLE" pot, it clips quite heavily. "BASS" pot cranking doesn't produce clipping.
• maxing out the output pot has some strange "warping" effect ... dunno how to discribe it, it gets louder, then it stops, in the last 5% of the pot's way it clips. maybe I'm overloading my H&K clean channel?

Here's some details:

1.) I converted the 9V from the wallbug to 18 Volts using the well-known "Road Rage" schematic (I used an LTE1054). 18V, because I wanted as much clean headroom as possible. This is also why I chose 3k3 (R6) and 4k7 (R20) for the lower JFETs' source resistors.
Question #1: If I increase R6 and R20, will this decrease the gain, so I get even more cleanliness? How high can I go?
Question #2: Should I throw in a "GAIN" pot instead of R17/R18 (divider before the second Mini Booster stage)?

2.) The EQ section is a bit of everything. The schematic is something of Rod Elliott's "designing with opamps" pages. I took the idea of tying the first opamp's (buffer stage) non-inverting input to VCC/2 (9V) from R.G.s page about parametric eqs (http://www.geofex.com/Article_Folders/EQs/paramet.htm. Then I did the same with the second half, an idea I got from here: https://www.petervis.com/record_players_and_turntables/baxandall/baxandall-tone-circuit.html. Which brings me to my next questions:
Question #3: Do I have to "haul back in" the center of the swing after the eq section? I mean in the active baxandall the wave swings around +9V, +18V and 0V being the "rails". After this I return literally to earth with no virtual ground. Do I miss something here?
Question #4: The frequency response of the two eq pots changed drastically when I switched from 9V to 18V. Is there some formula which I can use to calculate the new values? Or doesn't voltage matter in this case?

3.) And the final question: How would YOU tweak this thing so it NEVER begins to clip at all?

[KarenColumbo]

What I mean is:

- I don't want full clean (could do that with an opamp and be done with it), but neither a distortion effect, just this teeny bit of "rough" that creates those harmonics I crave

Maybe the Mini Booster is too gainy.

[garcho]

try adding Jack's volume output pot between your first "minibooster" and the tone stack. Then disconnect everything from the second "minibooster" and audio probe after C10 and see if you get the same behavior.

[KarenColumbo]

try adding Jack's volume output pot between your first "minibooster" and the tone stack. Then disconnect everything from the second "minibooster" and audio probe after C10 and see if you get the same behavior.

Thanks, will try this
Edit: Just tried it - it's now extremely quiet and horribly distorted I think I got huge problems with impedance and such

[anotherjim]

Looks to me like bass can get x5 boost while treble can go to x10. Treble content usually is weaker in the first place, but it rides on whatever bass is there - so treble can clip easier. This issue needs taken in consideration of the next one....


Maybe you aren't meant to go to 18v power for the whole thing. Is it not the hope that the FET stages clip rather than the op-amp ones & it sounds nicer that way? In other words, nice clipping at 9v rails means it shouldn't clip in the op-amp buffers. 18v is a worthwhile increase in headroom, but it still only gives you clearance for +6dB or x2 extra gain.

[Digital Larry]

Forgive my relative ignorance on analog things, but isn't the cascode structure (using an active FET load) a method of getting a whole @*!($-ton of gain?  It's in two different places leading me to think you really want a whole ton of gain here.

Also I would drop the resistance of the output pot and AC coupling capacitor should go up as well to maintain your low frequency corner.  You're using an emitter follower to get a low output impedance and then turning it back into high impedance if that pot is turned down at all.

[KarenColumbo]

Looks to me like bass can get x5 boost while treble can go to x10. Treble content usually is weaker in the first place, but it rides on whatever bass is there - so treble can clip easier. This issue needs taken in consideration of the next one....


Maybe you aren't meant to go to 18v power for the whole thing. Is it not the hope that the FET stages clip rather than the op-amp ones & it sounds nicer that way? In other words, nice clipping at 9v rails means it shouldn't clip in the op-amp buffers. 18v is a worthwhile increase in headroom, but it still only gives you clearance for +6dB or x2 extra gain.

Good hint with the bass and treble gain - did not think of this, of course. maybe decouple those two. hm ...

Yeah, i could of course volt up the opamp only - in fact I've seen this done before some times. I will try that.

[KarenColumbo]

Forgive my relative ignorance on analog things, but isn't the cascode structure (using an active FET load) a method of getting a whole @*!($-ton of gain?  It's in two different places leading me to think you really want a whole ton of gain here.

Also I would drop the resistance of the output pot and AC coupling capacitor should go up as well to maintain your low frequency corner.  You're using an emitter follower to get a low output impedance and then turning it back into high impedance if that pot is turned down at all.

Yeah, in principle it's this SRPP phenomenon. I took this because it sounded so good at first - before I messed it all up I'll rip this apart and start one jfet at a time.
Thanks for the hint with the pot - should have thought of that