Potentiometer in feedback loop and offset currents

[ramonmaruko]

Good day. I am looking at a schematic for a guitar distortion pedal
(BluesBreaker clone). I read[1]  that it is not a good idea to have DC flowing
through a potentiometer, so I am trying to prevent that from happening. The way
the circuit is now, any bias currents from the opamps could flow through the
potentiometer.

1. I need pointers on how to do this without audibly changing the frequency
response of the circuit. The attempts I've done basically mutes the output
(-60dB to -120dB). What I've done is putting a capacitor in series with the
inverting input of U1 and the nodes to C2, R2. I also tried putting a capacitor
in series with the output of U1 and the nodes to C2, VR1.

2. How do I properly simulate the offset currents for the purpose of testing DC
flowing through VR1? What I am doing now is by turning off the AC input `in`
and putting a DC voltage source to the inverting input or the output pin of U1.
I then place current probes at the three pins of VR1. In another simplier
circuit with DC properly blocked, I still get a femtoampere; but I don't think
that will damage or cause noises when flowing through the potentiometer. Am I
doing it correctly?

[1] from "Small Signal Audio Design" by Douglas Self

[mdszy]

Why does the book say it's a bad idea to have DC through a pot? Does it mean high currents, or any current? Low currents are 100% fine through a pot... but you don't want to go putting a few amps through one!

[Benta]

Why does the book say it's a bad idea to have DC through a pot? Does it mean high currents, or any current? Low currents are 100% fine through a pot... but you don't want to go putting a few amps through one!

This is old knowledge for audio guys. Just a little bit of DC current through, say, a volume or tone pot will give scratching noises (in the audio signal) when moving the slider/wiper. Try it yourself :-)

[mdszy]

Why does the book say it's a bad idea to have DC through a pot? Does it mean high currents, or any current? Low currents are 100% fine through a pot... but you don't want to go putting a few amps through one!

This is old knowledge for audio guys. Just a little bit of DC current through, say, a volume or tone pot will give scratching noises (in the audio signal) when moving the slider/wiper. Try it yourself :-)

Huh, interesting. I didn't know that! I guess that leaves me confused... how does the pot operate then? On the actual audio signal itself, so it's AC and not DC?

[Zero999]

Why does the book say it's a bad idea to have DC through a pot? Does it mean high currents, or any current? Low currents are 100% fine through a pot... but you don't want to go putting a few amps through one!

This is old knowledge for audio guys. Just a little bit of DC current through, say, a volume or tone pot will give scratching noises (in the audio signal) when moving the slider/wiper. Try it yourself :-)

Huh, interesting. I didn't know that! I guess that leaves me confused... how does the pot operate then? On the actual audio signal itself, so it's AC and not DC?

The audio signal is AC of course. Capacitors can be used to block DC and prevent it from flowing through the potentiometer.

[ramonmaruko]

Why does the book say it's a bad idea to have DC through a pot? Does it mean high currents, or any current? Low currents are 100% fine through a pot... but you don't want to go putting a few amps through one!

This is old knowledge for audio guys. Just a little bit of DC current through, say, a volume or tone pot will give scratching noises (in the audio signal) when moving the slider/wiper. Try it yourself :-)

Something I've also come across is this document by ALPS.

[Direct Voltage]
    When direct voltage is flown through this part, terminal to terminal insulation may deteriorate depending on the use environment. This is due to a migration phenomenon. Contact us if you are planning to use this part under direct voltage.

[ramonmaruko]

I just noticed that I forgot to attach the schematic. I just fixed it now and attached the schematic to the first post.

[Zero999]

I just noticed that I forgot to attach the schematic. I just fixed it now and attached the schematic to the first post.

There will be no DC in the potentiometer, except for the tiny bias currents required for the op-amp, because both sides of the pot. are at a DC virtual earth.

[Benta]

Why does the book say it's a bad idea to have DC through a pot? Does it mean high currents, or any current? Low currents are 100% fine through a pot... but you don't want to go putting a few amps through one!

This is old knowledge for audio guys. Just a little bit of DC current through, say, a volume or tone pot will give scratching noises (in the audio signal) when moving the slider/wiper. Try it yourself :-)

My apologies for imprecise wording.
A DC current running from resistance track through the wiper/slider will cause scratching. If the wiper is AC coupled to the next stage, DC through the track will not cause problems.

[ramonmaruko]

How do I prevent the bias currents from flowing through the potentiometer? The
op amp I am planning to use is the NJM4580 which has an input bias current of
(typ) 100nA. In an example circuit of Self's book (Fig. 11.6), he says:

The practical tone circuitry contains more capacitors than the
simplified circuits in Figures 11.4 and 11.5, and this reflects a disadvantage
of low-noise bipolar input op-amps. The bias currents are non-negligible and
must not be allowed to flow through potentiometers if crackling noises are to
be avoided when they are moved.

I don't know if I am allowed to post the mentioned figures in this thread. For Self's
circuit, he used a NE5532 which has a (typ) bias current of 200nA.

As mentioned in my first post, I tried following some of the patterns he used
to block the offsets; but all severely changed the frequency response of the
circuit.

[Zero999]

How do I prevent the bias currents from flowing through the potentiometer? The
op amp I am planning to use is the NJM4580 which has an input bias current of
(typ) 100nA. In an example circuit of Self's book (Fig. 11.6), he says:

The practical tone circuitry contains more capacitors than the
simplified circuits in Figures 11.4 and 11.5, and this reflects a disadvantage
of low-noise bipolar input op-amps. The bias currents are non-negligible and
must not be allowed to flow through potentiometers if crackling noises are to
be avoided when they are moved.

I don't know if I am allowed to post the mentioned figures in this thread. For Self's
circuit, he used a NE5532 which has a (typ) bias current of 200nA.

As mentioned in my first post, I tried following some of the patterns he used
to block the offsets; but all severely changed the frequency response of the
circuit.

The simplest way to overcome this is to switch to a FET input op-amp, such as the TL072, which will have tiny bias currents, around the same order as the leakage current through a capacitor.

[StillTrying]

I can't see why you can't just put a 100nF in series with the pot's wiper or left side, and then a 2Meg2 resistor from the opamps output to -ve input to carry the bias current.

If you're not very careful a few nA of DC current flowing through a control pot in a guitar distortion pedal can cause all kinds of imaginary problems!

[ramonmaruko]

I can't see why you can't just put a 100nF in series with the pot's wiper or left side, and then a 2Meg2 resistor from the opamps output to -ve input to carry the bias current.

If you're not very careful a few nA of DC current flowing through a control pot in a guitar distortion pedal can cause all kinds of imaginary problems!

Thank you! How did you come up with 100nF and 2Meg? I tried 1uF and 200K, which
should have the same frequency cutoff, but that showed a different frequency
response.

[Zero999]

Adding a capacitor will quite likely affect the frequency response.

Have you actually built this circuit to see how good/bad it is? This isn't something which can be simulated.

Have you considered a FET input op-amp? They really aren't that bad and with the input impedances you have, the noise will be similar to a bipolar op-amp in the same price range.

[Buriedcode]

The simplest way to overcome this is to switch to a FET input op-amp, such as the TL072, which will have tiny bias currents, around the same order as the leakage current through a capacitor.

That, along with cost, is why so many guitar pedals use the TL072/TL074.  Guitar effects generally don't require >4MHz GBP bandwidth, but require low noise opamps if there's significant gain (ie: distortion pedals).

As for modding the original design.  I don't think much improvement can be made (in terms of design, not "tone"). But making tweaks to max gain, and/or the tone stack in pedals is a great way to experiment/learn and end up with guitar effects you actually like.

[Zero999]

I forgot to mention the difference in noise between BJT and FET input op-amps such as the NE5532 and TL072 respectively.

With a low input impedance, the NE5532 will have lower noise than the TL072, since it has a lower noise voltage. At higher input impedances, the TL072 will be the least noisy, since it has tiny bias currents and the large bias currents of the NE5532 and their associated noise, will generate a corresponding noise voltage in whatever impedance is at the inputs. From memory the crossover point when the TL072 becomes better than the NE5532 is around 30k of input impedance, which is not far off what this circuit has.

Unfortunately I didn't see a specification for the input noise current on the NJM4580, which doesn't inspire confidence with that characteristic. I doubt it's better than the NE5532 and even if it is, then TL072 will still be better because the lower bias current through the pot. is worth it.

[ramonmaruko]

The simplest way to overcome this is to switch to a FET input op-amp, such as the TL072, which will have tiny bias currents, around the same order as the leakage current through a capacitor.

Have you considered a FET input op-amp? They really aren't that bad and with the input impedances you have, the noise will be similar to a bipolar op-amp in the same price range.

I forgot to mention the difference in noise between BJT and FET input op-amps such as the NE5532 and TL072 respectively.

With a low input impedance, the NE5532 will have lower noise than the TL072, since it has a lower noise voltage. At higher input impedances, the TL072 will be the least noisy, since it has tiny bias currents and the large bias currents of the NE5532 and their associated noise, will generate a corresponding noise voltage in whatever impedance is at the inputs. From memory the crossover point when the TL072 becomes better than the NE5532 is around 30k of input impedance, which is not far off what this circuit has.

Unfortunately I didn't see a specification for the input noise current on the NJM4580, which doesn't inspire confidence with that characteristic. I doubt it's better than the NE5532 and even if it is, then TL072 will still be better because the lower bias current through the pot. is worth it.

The input impedance of my guitar amplifier is rated at 1Meg ohms. I'm assuming
that they were expecting the source impedance (guitar) to be at max 100K ohms.
Does the circuit that I posted have around 19K ohms input impedance? (R5 ||
R1). That really makes using a FET input op amp more correct in this case. I
stuck with using the NE5332 since that was what's on the original schematic,
but it looks like I'll have to use a FET input op amp to easily avoid the
offset currents that I don't want.

Another (unrelated?) question that I have, how do I calculate the current
consumption of an op amp circuit, for power supply sizing? Would it simply be:

((expected max Vo of the op amp) /
 (max impedance of the feedback network)) +
quiescent current of the op amp?

[Zero999]

The input impedance of the circuit you posted is equal to R1 + R5 = 519k, at audio frequencies.

I don't know that much about guitar pick-ups. I believe they're inductive so the impedance increases with frequency. If the amplifier has too lower impedance, then it will cut the high frequencies, leaving the low. I don't think a guitar produces much in the way of treble but could be wrong. Hopefully someone else here will provide more insight.

In this circuit the static supply current is simply equal to that taken by the op-amp because there's no DC current through the resistors in the feedback loop. With a signal, the peak current through the resistors in the feedback loop adds to the current draw but in this case the resistors have fairly high values so you can probably ignore that too and concentrate on the load.

[ramonmaruko]

Thank you very much for answering my questions and being patient with me.