A Stupidly Wonderful gain-shape control?

[Mark Hammer]

stopstopsmile was asking about adding some simple tone-shaping to a DOD250, and while pondering what might be easily added to a stock unit/board, I had an idea (uh-oh, better back up a few steps and put your safety goggles on!  We may also end up needing another Timmy. ).

The feedback resistance in the DOD250  is set via a 1M fixed resistor, with gain adjustable via a 500k pot to ground.  In some instances, there is also a 25pf feedback cap just to keep things stable and less hissy.

Some background, first.  In general, one needs to think about op-amps as wishing desperately to produce their maximum gain.  Give them half a chance and that's just what they will do.  The amount of negative feedback finding its way from the output back to the inverting input determines how much the op-amp is prevented from doing just that.  Think of it like stepping on the brakes.  More negative feedback is stepping on the brakes harder.

How much feedback gets applied depends on how much is fed back, and how much of what is fed back gets used.  So, even if the feedback resistance is large, a small resistance to ground means that much of even that small amount of the negative feedback is "dumped" to ground and the op-amp is allowed to move closer to max gain.  If there is a cap in the feedback loop, that allows high-frequency content from the output to find a more efficient path to the inverting pin and move the chip even further away from maximum gain for those frequencies that can best take advantage of that low-impedance path. If there is a cap in series with the resistance to ground, that makes it easier for higher frequencies to get dumped to ground, but not as easy for lower frequencies.  That's why the TS-9, Dist+ and DOD250 tend to have a wimpy low end; the small-value cap in the ground leg means there is more gain for higher freqs than for lower ones.

The RAT takes an interesting approach to this configuration and provides two paths to ground; one with a cap that provides a basic amount of gain for all frequencies, and a second with a cap value that provides additional gain for higher frequencies.  The Rat has the latitude to do that because it adjusts gain using the feedback loop, not the ground-leg as the 250 and Dist+ do.  Can we do something like that for the DOD250 and Dist+?  If I'm right, yes.

Here's how.  Let's use a pot in the feedback loop, wired up like the SWTP, where the wiper is connected to a cap that goes to ground.  Imagine I have a 250k pot.  It forms part of the feedback path and is situated between a 510k and 240k fixed resistor.  So, the output of the chip goes to a 240k resistor, then to one outside lug of the pot, and the other outside lug of the pot goes to a 510k fixed resistor and then to the inverting pin.  Add them up and you get roughly 1M .  So, our gain adjustments via the normal pot should be unaffected.

Okay, lets run a 220pf cap from the wiper of the pot to ground and move the wiper of the pot over so that it's close to the lug where it meets the 240k fixed resistor.  The 240k resistor and the 220pf cap form a lowpass filter that rolls off content at a 6db/oct slope for content above roughly 3khz.  However, since that means there is less negative feedback in that zone, what it implies is that we are "reducing the brakes" for that content, which implies that we should be getting more gain for those higher frequencies, by having bled off high-frequency feedback.

As the wiper gets moved along, more resistance is added to that basic 240k. Add in 100k of the pot's total resistance, and the point where gain boost kicks in moves down to just over 2.1khz.  By the time you've added in the full 250k, the point of high-freq gain boost is moved down to 1.6khz.  Note that we have not really changed the basic gain for content below that point since it continues to face a 1M path from output back to input.  Although there is the risk of runaway oscilation for very high frequencies, happily the 25pf cap is still there and keeps things in check for the insect noises.

Essentially, what we've done is the inverse of what the Rat does, except we've also made it variable.  In theory, this change can add variable amounts of extra treble drive which one is then free to either keep or shave off.  The component values are not critical, as long as there is some fixed resistance on each side of the pot, and the values selected impact on frequency content which is useful to you.  I just happened to hit on a useful range with these values just by luck.  Note that the pot effectively covers a one-octave range because the pot-value is approximately equal to the value of the first resistor (double the resistance before the cap and you halve the corner frequency).  You could cover other ranges and wider or narrower ranges by adjusting the pot value compared to the resistor and the cap.  For instance, a 100k fixed resistor, followed by a 250k pot and 680k resistor, gets you a range of 1.4-4.8khz boost-point, using a 330pf cap instead of a 220pf.

Does this sort of control already exist on either a commercial product or on somebody's DIY experiment?  Is it part of one of Jack Orman's marvelous innovations that I may have happened to miss?  Whatever the case, I'm looking forward to trying it out.

[stopstopsmile]

its funny because i was thinking of trying a rat tone control, and while I have no clue at one I am really doing I stumbled upon a flawed missing pieces kind of idea that you seem to have nailed.  Very interesting Mark. 

[slacker]

Cool idea Mark, Tim Escobedo's TMK has something that looks vaguely similar, but I can't think of anyone using it in a "normal" sounding pedal.

[beatstrat]

Great work, Mark.

[Gus]

I like more control over things.  However interacting things can be fun

  Here you go people lets change the dist +

First use a dual opamp.  first stage as a buffer this also lets you use a DPDT with an LED, then a Cap then make the input resistor a pot and a fixed resistor to the 1/2 supply ref.  work out your  F=1/(2 pi RC)  think about noise BJT and FET input might "want" different R.  voltage and/or current noise

The fixed cap and vairable resistor CR variable high pass filter to the distortion stage could even be a range master

Lets get away from clones and band aids

Buffer or low gain stage(what ever you like) then
variable high pass or low pass or muff tone control then
distortion
10-03-2007

[johngreene]

If you make this an inverting amplifier, you are basically describing how the treble control in an active Baxandall EQ works on the boost side only.

So now try to wrap your brain around what would happen if you also paralleled the pot with another capacitor....

--john

[snoof]

Cool, Nice idea Mark.  I'm unclear as to why on the Rat schem they are using two RC networks in parallel from the feedback loop to grnd(560r and 4.7u and 47r and 2.2u)??  Why not roll them into one??I have been playing w/ opamps lately, and this is good timing for me to try this idea!

[johngreene]

Cool, Nice idea Mark.  I'm unclear as to why on the Rat schem they are using two RC networks in parallel from the feedback loop to grnd(560r and 4.7u and 47r and 2.2u)??  Why not roll them into one??I have been playing w/ opamps lately, and this is good timing for me to try this idea!

Because they are two different RC networks. The result is a kind of 'shelving' or '2 step' of the response between them.

--john

[snoof]

I see, thanks for the clarification John!

[Dingleberry Electronics]

Hi.

First of all, a very nice topic!!!
Next:
How about if you replace 47R and 560R resistors in a RAT's rc-network with a pot, and two smaller value resistors.
For example if you take a 500 ohm pot and solder a 33R to lug 3 and 68R to lug 1 and connect the wiper to the feedback loop.
The sum of the resistances woulb be very close to RAT's specs. Just few ohms away.
Now you would be able to shape the RAT's frequency gain too?

Should the pot be smaller? Is 500R too big range? Would it start to affect just nearly in the end of pot's trail?
What about 200R pot and 18R and 390R resistors?
Or even smaller: 100R pot?
Lin/log or rev. log taper???
Probably needs some serious breadboarding..

Might it work something similar like the "lube mod" in the FKR.

What I think it would be very handy if you could tweak the values that, when you set the "frequency pot" in the 12.00 o'clock position you would get standard RAT frequency response and turning  either way, you could raise or lower the corner frequency.

[mac]

What about reducing the 47nf cap to gnd in the Dist+, say to 10nf, and adding a pot and a 82nf in series, both in parallel with the 10nf. When pot is zero the cap is 92nf, more bass; and when it is at max the cap is 10nf, less bass.

mac

[Freddy205]

Dingleberry: www.audigoa.com/extra.php <- Check out that project.