Inverting OpAmp Feedback Path - Help!

[dubspecialist]

Hello, I've been experimenting with different pot values at the inverting opamp gain-stage of a custom pedal. As you can tell from the screen shot the opamp is a TL074. I am really digging the different compression/saturation I can get from a 500k log pot in the feedback path. When the pot is static it sounds great, however there is unfortunately a pretty severe crackle/noise created as the pot is swept in either direction. Doesn't matter where the wiper is in the pots range, the noise is present when it's being swept. With a 100k pot the noise is just barely audible (in headphones) across it's full range, sounds really nice but obviously I don't get the same level of drive/saturation.

I know this isn't news to most, I even saw a comment in another thread that said TL074s are good for any pot value less than 100k in the feedback path... I'm struggling to find a solution to this though. I have tried adjusting the value of the filter cap used across the pot, I raised it up to 440p but this had little impact on the noise apart from some high end roll off. Can anyone offer some advice or assistance here? Perhaps a different quad opamp is better spec'd for handling a variable resistor of this size?

Any assistance is greatly appreciated 

[anotherjim]

I want to see what's on the left-hand side of your schematic, specifically the feed to the inverting input.
The pot doesn't have to crackle.
1M pot isn't a problem.
High resistances "in series" with signal increases hiss noise, but 10k is the "hi-fi" low-noise limit, not 100k.
On our side of the mixing desk, hiss is something we sometimes have to live with so long as the signal sounds good.

You can't get the feedback cap to be right for all settings of the pot. It forms a low-pass RC filter with the pot resistance.
If you want to remove this interdependence, you fix the feedback resistor and vary the input resistor to set gain instead and use the non-inverting input circuit.

[ElectricDruid]

+1 what Jim said. We need to see more of the schematic, particularly the input and the biasing. There's nothing visible that explains it, but it suggests all sorts of potential problems off the sides of the image!

[dubspecialist]

Thanks so much for the response fellas, yes of course... sorry realised after posting that the crop was certainly not helpful but figured it may have just been a general issue with TL074 when opting for relatively large biasing ratios. There are two 15k resistors ahead of the opamp that follow the mix section. I have experimented with raising and lowering their values (was curious to see the effects on gain structure) to mixed results. in the end 15k seemed fine. The RC filter created by the feedback cap is not an issue, in fact the results are pleasing from a tone perspective. There is, of course, a certain amount of "hiss" in the circuit but again this is not an issue, the signal to noise ratio when the pot is static is fine.

The only issue is the artefacts heard when sweeping the pot... a harsh crackle that sounds similar to a "scratchy" pot. See below for a more detailed schematic and thanks again for offering to help.

[Vivek]

Try adding a cap after junction of R9 and R10.

My Logic : pots make scratchy noise when there's DC across them

If both the + and - pins of an Opamp are DC connected to Vref, there is a very slim possibility of extremely small DC imbalance at the input pins due to different voltage drops in the path to Vref. This ultra minuscule difference in voltage could get magnified by the Opamp.

[ElectricDruid]

Seems reasonable to me, Vivek.

Given the resistor ratios (500K / 20K, with the mix pot in the middle) there's a fair bit of gain for any offset (x25), so it could well produce a significant effect.

[anotherjim]

  Well, Vref should be on the input via the mix pot and that's the same as the Vref the opamp wants to bias at so no difference in DC & no crackling unless opamp offset error is significant... unless... something to the left or R7 or R8 is connected to a voltage other than Vref?

How much of a scheme should you show? As far as the coupling caps in and out of the stage in question.

[ElectricDruid]

Well, Vref should be on the input via the mix pot and that's the same as the Vref the opamp wants to bias at so no difference in DC & no crackling unless opamp offset error is significant.

The two inputs have quite different resistances to that Vref, so that doesn't help. Plus the gain. Could it be enough? I dunno, maybe. Like you said, is that offset error significant?

.. unless... something to the left or R7 or R8 is connected to a voltage other than Vref?

Well, yeah, that'd do it!!

How much of a scheme should you show? As far as the coupling caps in and out of the stage in question.

More is more, right?!? I want to see the whole thing!

[antonis]

IMHO, amplified DC offset for FET inputs amp shouldn't be enough for crackling (for MIX pot midway, gain should be less than 50)..
Of course, VR is just a symbol on schematic which should be further interpreted..

[dubspecialist]

Thanks again to all for weighing in. I'm happy to experiment with a cap after the junction of R9/R10. I am not on my personal computer so do not have a copy of the eagle file on hand today but it is a modified version of the PPCB FV-1 pedal so for the sake of this discussion this schematic will do fine as I had the same exact problem when working with this design. The vref comes after the voltage divider and there is 3.3V being supplied to the FV-1 so perhaps, as suggested there is a DC issue being presented at the opamp inputs?

As a note - I believe the gain is actually less than x25 as the resistors ahead of the mix pot (R4/R19 in this recent schem) also factor into the equation, at least that's the conclusion I came to when measuring/experimenting with the values of those resistors up stream of the "vol" stage.

[ElectricDruid]

No DC is going into the Mix stage, at least on that schematic. It's separated by the two 1u caps C6 and C11. If the problem is DC making the pot scratchy (which does seem to be what the symptoms suggest) then that DC is coming from somewhere else, not being fed in from outside. Unless the caps are leaky I suppose, but that seems a bit unlikely.

[anotherjim]

Some pots can just be bad! Those 16mm plastic-bodied ones are just useless for audio and even Alphas can get contaminated.

[antonis]

I believe the gain is actually less than x25 as the resistors ahead of the mix pot (R4/R19 in this recent schem) also factor into the equation, at least that's the conclusion I came to when measuring/experimenting with the values of those resistors up stream of the "vol" stage.

Indeed, 'cause IC1.2 & IC1.3 buffers outputs are considered ground due to their very low output impedance..

[dubspecialist]

Thanks again to everyone for weighing in!

I have an update but it could be slightly contentious. I decided to complete a third build last night using another 500k pot and a very interesting thing occurred. The pot was dead quiet (well close to) before I boxed it - which made me think "wow so it was just a mechanical issue with the other two recent builds - dodgy pots!". I boxed it up and secured knobs then began testing further... After a few minutes of heavy use of the vol pot the noise crept in, only it was worse than the two previous builds, especially in particular spots along the resistance path.

I am hesitant to open this can of worms as I have seen a couple threads about this at other forums and it's a rabbit hole that seems to lead no where good BUT I am thinking that maybe I got a bad batch of 500k Alphas from Tayda? I will buy a can of deoxit (I don't currently use it ahead of a build, should I be?) and source some 500k Alphas from elsewhere to test. I'll report back.

[antonis]

Never experienced defective/bad/crappy Alpha pot (either 10mm or 16mm) from various batches but you never know..

[ElectricDruid]

Like Antonis, I'm also a bit sceptical about the pots being the problem. While I can see that if you did have a bad batch, that would explain why you've had three bad ones in a row, another explanation would be because the situation you're putting them into is causing the noise somehow. To me, that still seems more likely. I just can't see where the "stray DC" is coming from!

Have you got any other pots you could try? For the purposes of experiment, it wouldn't have to be a 500K value, say 100K. Clearly the gain changes, but if the problem is caused by bad pots, a different pot should fix it *completely*. If it improves it *only slightly* that suggests it might be a DC problem which is affected by reduced gain.

Dunno, just a thought.

[anotherjim]

How are the pots being connected? You can become a pot serial killer if you solder wires into the pot tag rivets.

[dubspecialist]

Pots are PCB mounted. I have a batch of 100k pots and they sound very clean in place of the 500k but you lose the possibility to saturate/overdrive. I should add that the 500k pot sounds good with a guitar signal. The noise is only present when I'm running a relatively hot line level signal through the circuit (from a synthesiser or drum machine). This seems to suggest an impedance issue. I have everything gain staged well, the signal is hot but it does not clip the opamp when the pot is static (up to a certain point of course). However, if I sweep it in either direction I get this shifting "crackling" sound, especially as I approach full attenuation (lower end of the pots "volume") . I should add that one pot is noticeably worse than the other two, which leaves me even more confused and thinking I need to give deoxit a try before sending myself any further down the rabbit hole.

After that I will consider adjusting the values of the input resistors and lowering the pot value. Adjusting the impedance but trying to keep a similar ratio/amount of gain.

[antonis]

Make R4, R19, R20 & R21 3k3 with 100k pot value..