Op Amp biasing

[lowell]

I have a built a custom preamp and am a bit confused at the result (doesn't work), but think I may know what's up. I've checked all connections and everything looks right. But there's no sound, and there is 9v on all input/output pins.

I'll try to explain this as best as I can. I bought a Marshall Gov'nor kit at GGG. I have completely redone the circuit though. It is no longer an overdrive but is a preamp for jazz guitar. Anywho, I took a circuit that was designed for op-amps with the same pin outs, however the circuit operated from a dual supply of +/-15v. I'm using that circuit with a 9v single supply and grounding pin4 of the opamps. In the gov'nor circuit, and most other fx circuits I've seen for that matter, there is a 4.5v bias supply that goes to the input of each opamp. I'm thinking this is why my preamp isn't working, as I have left this bias supply out. If someone could enlighten me on whether or not this sounds right, and why this bias supply is need please respond.

Finally there are no DC coupling caps in the new circuit so I'm thinking I'll need to add those?

[Steve Conner]

You got it!

In order for an opamp to work, the input pins have to be biased about halfway between the negative and positive supply voltages. Halfway between -15 and +15 is 0. However, halfway between 0 and 9 is 4.5. So, when converting to single supply, all nodes that were signal ground need to be moved to a 4.5V bias supply.

DC block capacitors are then needed on the input and output where the circuit interfaces with the outside world. Otherwise, your guitar would short out the 4.5V bias to the input, and your amp might not like getting 4.5V shoved up its input jack either. (the op-amp outputs should all sit at 4.5V too)

[J M Fahey]

Funny thing is, the original Guv'nor pedal *was* made for +9V battery power, and *did* have all necessary biasing elements .
Wonder why you used that PCB but pulled functional elements out.
If you wish, post both the GGG/Guv'nor schematic and the Jazz Guitar Preamp.
Thanks.

[lowell]

Thanks guys. I figured that was it. What I don't get is this. Opamps amplify the difference b/t the inputs right? So if the inverting input is at 4.5v and the non-inverting input is at 0v, then why isn't the output sitting at a negative voltage that is multiplied by gainx4.5v of the circuit?

[lowell]

Oh wait I meant the non-inverting input is the 4.5v bias. Why isn't the output gainX4.5v ^-1?

[J M Fahey]

1) Usually the non inverting input is biased at +4.5V (so far so good) but the inverting one is referred to the output (it does *NOT have its own biasing string nor is grounded) and the Op Amp will do whatever is in its power to keep both inputs at the exact same voltage (or within a couple millivolts) so the output will also be forced to sit at 4.5V.
That's the idea behind an Op amp !!!
2) the OpAmp gain string usually has a capacitor to ground, meaning the grounded leg has infinite impedance for DC meaning the DC gain is 1 meaning the couple millivolts offset between inputs will not be amplified.
Or, to be more precise, will be amplified by a gain of 1x (meaning it's the same value).

[lowell]

Thanks for the info JM. Not sure I grasp that.

Here is the schematic for my preamp. As of now the thing works. It is lower volume than I expected and the tone controls do nothing. I just moved the right side of the tone stack to AFTER the output cap thinking maybe DC was being coupled back to the input and was causing an issue, just a guess, but the tone controls do not work. Before I moved the stack to after the output cap the bass control was acting more like a volume, and the mid/treble controls were not working. The sound is also distorted as well BTW.

[J M Fahey]

Well, U1/2/3 are properly biased: you apply Vbias (which I hope is 4.5V) to the "+" input, and have a resistor fom the output to the "-" input.
U4 is not, it has no resistor from "out" to "-" so it probably is slammed against some rail or at least drifts wildly.
U2 needs a resistor to set gain, in series with C9.
As is, it's practically open loop, at least on high frequencies. It probably hisses.

I'd suggest you search for a preamp idea you like, build it on ProtoBoard, test and tweak at will and only when happy with the results, commit it to pCB.
Good luck.

[lowell]

Thanks JM. Actually my schematic was wrong. I have a 2k2 resistor in series with U2's input. I have also move the right side of the tone stack to pre-output cap. So now there is a resistor network feeding back to the - input. The thing works, but I don't hear any change in the midrange when I adjust the Middle. As said before it is overdriving a bit and maybe I need to correct the gain on each stage, or maybe one or more opamps are damaged from my tests. Here's an updated schematic.

[lowell]

In regards to spike protection, should I add back to back diodes across the opamp inputs or from input to supply and input to ground reverse biased? Or both? Also, I should setup the gain of each stage so as to never have each stage's output produce more than 4.5v p-p right? Finally, gain is Rf/Rin right? Ignoring caps.

[R.G.]

In regards to spike protection, should I add back to back diodes across the opamp inputs or from input to supply and input to ground reverse biased? Or both?

Depends. What opamps are you using and what are their input common mode voltage range and differential mode voltage range?

Differential mode is when you force the inputs farther apart than they would otherwise go. This is protected by back to back diodes from + to - inputs. Common mode is usually protected by reverse biased input clamp diodes to some desired clamping voltage. If you insist that the opamp not latch up or oscillate when this happens, you may have to connect these diodes to voltages a couple of volts inside each power supply. The TL07x series, for instance, can latch up if its inputs are forced closer than 1.5-2V to either power supply rail, I think.

Both techniques need some input resistance to limit the current the diodes carry. Some opamps have such diodes internally.

What spikes do you think you'll get?

Also, I should setup the gain of each stage so as to never have each stage's output produce more than 4.5v p-p right?

Yes, if you don't want the stage to be pushed to clipping. This requires knowing ahead of time what the biggest peak input signal that's not a spike is. That guess is the hard part.

Finally, gain is Rf/Rin right? Ignoring caps.

If the (-) input resistor is blocked by capacitors so it must return for DC to either the reference voltage or to the output, then the gain for signals into the (+) input is G = 1+ Rf/Ri. For signals fed into a resistor to the (-) input, G = -Rf/Ri. Both can happen at the same time.

[lowell]

Wow there's so much more design effort that goes into designing around discrete and IC solid state devices vs tubes. Thanks RG ill try to take all that in and put it to practice.

[J M Fahey]

Actual design is simpler, nothing is easier yet more predictable than an Op Amp.
Now, as soon as you start to add safety features, the sky is the limit.

[R.G.]

Actual design is simpler, nothing is easier yet more predictable than an Op Amp.
Now, as soon as you start to add safety features, the sky is the limit.

Dead right! The whole universe is there waiting when you start to think "Now, what could possibly go wrong?"

[lowell]

Hmmm... Well how predictable is it whn u don't know what the input source is?? This begs the question how does any designer know if a singled coil, humbucker, active pickup, or a string of preceding pedals/boosters are the input source? How does one design with that in mind?