op amp bias voltage divider - resistor values?

[thunderaxe]

hi all,

i've been playing around with the sour grape (keeley rotten apple) circuit, which is based on the op amp big muff ("dream fuzz"). so in comparing and contrasting the two, i noticed something interesting about the original op amp big muff -- the clipping stage op amp is biased by a voltage divider consisting of two deliberately mismatched resistors - 1M on the ground side and 820K on the V+ side, creating a slightly asymmetrical offset and, presumably, a slightly asymmetrical clipping as a result. looking at the equivalent voltage divider in the sour grape, it consists of two symmetrically matched 10K resistors.

so my first question is -- putting aside the asymmetry for now, is there any particular reason why the original uses 1M and the sour grape 10K? that's a difference of two orders of magnitude. secondly, if i wanted to play around with a bias pot on the sour grape, can i just replace either the ground or V+ side resistor with a 10K pot, or a 5K fixed resistor and 5K pot? or is there any reason why i should increase the resistance values on either side, say to 100K or 1M?

 

[Paul.Ruby]

This is likely because the op amps used don't have a symmetrical common mode voltage allowed at the inputs. Here's the spec range for a TL072.

This means the input voltages for op amp function are from 1.5V above the negative (ground in your case) but can go all the way up to Vcc. So, the ideal DC bias for maximum operating range is midway betwee (ground+1.5) and Vcc. For a 9V Vcc, that's 5.25V. Using a smaller pull up to Vcc is pulling the bias voltage up toward that more optimal point. The common mode input range is different for other op amps... This is just an example.

 

[Paul.Ruby]

Oh, and the choice of 1M or 10K depends on the load that will be put on the bias voltage supply. For jfet input op amps, the input current is very low and 1M ohm is fine and won't drain a battery.

But, if you are running from a non-battery power source and have other circuits drawing from that bias (like clipping diodes) the 10K may be needed to maintain a stable bias voltage.

 

[thunderaxe]

i should mention, the sour grape uses 3 sections of a TL074, whereas the original op amp big muff uses a 741 for the clipping stage that has the asymmetrical bias voltage

 

[thunderaxe]

Oh, and the choice of 1M or 10K depends on the load that will be put on the bias voltage supply. For jfet input op amps, the input current is very low and 1M ohm is fine and won't drain a battery.
View attachment 87242
But, if you are running from a non-battery power source and have other circuits drawing from that bias (like clipping diodes) the 10K may be needed to maintain a stable bias voltage.

so the only other thing that the bias voltage is connected to is the unused op-amp in the TL074. the clipping diodes are in the negative feedback loop of the op-amp in question so i'm not sure what you mean about other circuits drawing from that bias

 

[JTEX]

The unused op-amp could have been put to good use as a rail splitter, I suppose.

 

[thunderaxe]

is that just a buffered bias voltage supply?

i was planning on using it for an active baxandall tone control. but i also might want to add a clean blend which would also require at least one more op amp

 

[Paul.Ruby]

The unused op-amp could have been put to good use as a rail splitter, I suppose.

It could be used to buffer the bias voltage, too.

 

[Paul.Ruby]

so the only other thing that the bias voltage is connected to is the unused op-amp in the TL074. the clipping diodes are in the negative feedback loop of the op-amp in question so i'm not sure what you mean about other circuits drawing from that bias

Show me a schematic...

 

[thunderaxe]

fair! here's the sour grape (rotten apple):

and here's the dream fuzz (original op amp big muff):

 

[darwin999]

The Sour Grape's two 10k resistors do (in principle) allow Vref to symmetrically supply much higher currents, although nowhere is it used - as noted, the opamp has a FET front end that only requires <<1µA. So I guess they just wanted to keep the low frequency filter on it at ~1Hz despite using a smallish 10µF capacitor.

 

[thunderaxe]

if the 10uF capacitor on the sour grape's bias voltage is smallish, isn't the 220nF one on the dream fuzz tiny by comparison?

The Sour Grape's two 10k resistors do (in principle) allow Vref to symmetrically supply much higher currents, although nowhere is it used - as noted, the opamp has a FET front end that only requires <<1µA. So I guess they just wanted to keep the low frequency filter on it at ~1Hz despite using a smallish 10µF capacitor.

 

[JTEX]

Just wanted to point out that the R7/R8 divider will dump whatever noise/ripple there is in the AC power supply straight into the op amp input. Why not simply tie pin 5 to Vref through a resistor?

 

[thunderaxe]

Just wanted to point out that the R7/R8 divider will dump whatever noise/ripple there is in the AC power supply straight into the op amp input. Why not simply tie pin 5 to Vref through a resistor?

i'm very new at this, but is C2 not there to take care of noise or ripple?

 

[JTEX]

That's on the Dream Fuzz. I was talking about Sour Grape. My bad, I forgot to mention it.

 

[thunderaxe]

That's on the Dream Fuzz. I was talking about Sour Grape. My bad, I forgot to mention it.

haha yes, we were both talking about the sour grape. i see C2 in the dream fuzz directly attached to the voltage divider, but i was referring to C2 in the power section of the sour grape. i assumed it was there to get rid of any noise or ripple. do we need another one attached to the ground side of each voltage divider?

 

[JTEX]

C2 in Sour Grape will help reduce ripple to some degree, depending on the output impedance of the power supply. The "stiffer" (lower impedance) the power supply, the less effective is C2 at filtering. If you look at R17 in Dream Fuzz, its main purpose is to help with PSU filtering when combined into an RC filter with C12.

To me, the main purpose of Sour Grape C2 is as a local power reserve, which you'll appreciate if you have an even slightly intermittent/dirty DC jack and it loses power for a few milliseconds when you accidentally touch or wiggle the power cable. Then C2 is there to "fill the gap" so you don't get a big pop or scratchy sound as the power gets briefly disturbed.

However, you'd get much better noise rejection (and save a $0.02 resistor ) if you got rid of R7/R8 and replaced it with just one resistor from VREF, which is much cleaner than VCC. Something like this:

 

[Paul.Ruby]

The sour grape has a vref buffer and doesn't use it! The output of this opamp is a perfect Vref supply to the rest of the circuit. And the resistor divider could be bumped up to 1M and only need a 0.1uf filter cap on it since the op amp will draw only a couple hundred pico-amps. That said, it's not going to make any difference in the sound of the circuit. But, I would certainly change this if running on a battery (and change to a high brightness LED to drastically reduce it's current increasing R2 to 68k or higher).

 

[Paul.Ruby]

And, the sour grape has a vref supply and then doesn't use it as the reference to the first op amp???

 

[Paul.Ruby]

Could also save a component and do this...