"Thuddy" background noise during pick attack

[wizard333]

design flaw or a repair issue.

Repair issue; I've seen plenty of them before that don't do this, or at least it didn't jump out at me. That said, I don't see any reason for the coupling caps to be as large as they were on the gain channel; reducing them orders of magnitude made no perceptible difference in low end response.

[wizard333]

Also, correct me if I'm wrong, but if the phase shifting in the coupling caps were causing positive feedback to occur through the power rail, that would indicate a problem with the filter cap for that stage(s) correct? A properly working 22uf cap should pass pretty much any AC signal to ground that might otherwise build up that way, unless it weren't working properly.

[Richard]

A looper pedal might be the trick for reamping. It will save you the trouble of hooking you amp up to a computer, etc.

[wizard333]

Great idea!

[loudthud]

Look at the output of IC1A. Run a sinewave through the gain channel. If the input of the IC is driven beyond it's common mode range, the output may go crazy and suddenly slam the opposite rail. The data sheet for the OP249 does not go into enough detail to describe this type of operation.

The pic below shows a TL072 in a Voltage follower configuration with too much input. Top trace is the input, bottom trace is the output. The cure is to zener clamp the input to +/- 10V.

[R.G.]

Look at the output of IC1A. Run a sinewave through the gain channel. If the input of the IC is driven beyond it's common mode range, the output may go crazy and suddenly slam the opposite rail. The data sheet for the OP249 does not go into enough detail to describe this type of operation.

The pic below shows a TL072 in a Voltage follower configuration with too much input. Top trace is the input, bottom trace is the output. The cure is to zener clamp the input to +/- 10V.

Good catch. I didn't dig into the insides. The TL07x series is practically notorious for this. Its common mode range ends 1.5 to 2 V from both rails. Get closer than that and you get a BLAT.

[loudthud]

The Analog Devices OP249 data sheet lists input range with 15V rails as:
minimum +/- 11V
typical +/- 12.5V

The Normal channel has an attenuator at it's output, R9, R10 but the gain channel has a complex network with a tone stack, Presence and Master Volume controls. It might be able to overdrive IC1 depending on control settings.

[wizard333]

I'll check that this weekend. Thanks for the scope pic, that really helps!

If that does turn out to be the issue, wouldn't using the zener result in clipping the signal? Is there another IC you know of that could substitute with a wider range?

[R.G.]

If that does turn out to be the issue, wouldn't using the zener result in clipping the signal?

Yes, it would. However, clipping is far, far less objectionable than that pulse of inversion, assuming that's what is actually going on in there.

Is there another IC you know of that could substitute with a wider range?

Yes, but you're limited by the power supply to the IC in any case. *IF* this is what's happening - the scope work will tell that - then you can get at most a couple more volts of signal range.
The OP249 shares the issue of phase inversion if the inputs get too close to the power supplies, according to its datasheet. The datasheet says the input common mode range is "typically" +/-12.5V with a /-15V supply, the maker's way of saying "It might really be anything, but we think most of them will be close to this. Your mileage may vary." If you take them at their word, the common mode range is 2.5V less than the power supply in both directions.

Next question: what it your power supply voltage? The schematic is pretty vague. It's simply whatever you get by rectifying a 9-0-9 volt AC winding. That's nominally +/-12.7V, but it varies directly with the line voltage and how well your local AC mains matches the intent of the transformer designer, as well as any variations in the AC line. The line is usually something like +/-10% (and there's a long train of speculation and special cases to get to there), so you could easily be getting something like +/-10.5V on the power supplies to the opamp. That number is suspiciously close to the "10V" the schematic notes for the supplies derived from the winding. If it's really 10.5V, the opamp can't be relied on to do any better than a common mode range of +/-8V.

There are opamps that will get you both higher power supply voltage *withstanding* and a wider tolerance of the inputs for the power supply range, although perhaps not in the same IC. For the same power supply voltage, you could pick a "rail to rail input" opamp, and get another 2.5V in both positive and negative directions, plus no phase inversion if the signal gets close to the supply. Not a bad choice, but it gives up the "specialness" of the OP249, which is unusual enough that the designer specified it at a higher price than the TL072. It's not clear what he specified it for, but maybe for its ability to drive 600 ohm loads well.

Generating a higher power supply voltage is another option, but it gets messy pretty quickly, as the power supply voltage to the opamps is simply taken from rectifying the power transformer winding.

IF this is the problem, it might be simpler to redesign the opamp use so that the two sections are used as inverting gain stages, not followers. An inverting gain stage can be set up so that the input common mode voltage is always zero. You could change the follower before the FX loop to be an inverting gain of one to preserve the signal level going out the FX loop (and the reverb, which needs its own set of thinking, for the same reasons) and then inverting with some variable gain coming back in. With an inversion each way, the overall signal phase is maintained to the power amp, and common mode range is not an issue.

[jmaf]

If it were the case the TL is the problem, wouldn't all amps in this series present the same issue?

What's OP doing differently? Overdriving the amp to oblivion?

[wizard333]

What's OP doing differently? Overdriving the amp to oblivion?

No, read previous posts as to operating conditions.

If IC1 were the issue, that should happen only in conditions where there is a high signal level at it's input, correct? That's not what I'm seeing. I'll run it on the scope this weekend, but I can tell you this: I can turn the clean channel up to clipping and run an overdrive for the same drive level the gain channel is getting, and no issue, and LOUD; I can run the drive channel with the master very low and the issue occurs. That seems to indicate that exceeding IC1's vptp is not the issue.

[jmaf]

No, read previous posts as to operating conditions.

This thread is 3 months old, mate. I did scan over your earlier messages for clues but since it's an old thread I thought things might have changed, thus my question.

Just to throw a(nother) wild guess out there: a wrongly biased output stage close to cutoff makes this same ugly distortion you describe, especially on high notes with low intensity since they have lower energy to drive the output stage.

[wizard333]

The tubes are new and biased at 60% PD at idle.

[jmaf]

The tubes are new and biased at 60% PD at idle.

Does increasing PD change the ugly noise at all? Are you able to quickly increase the idle current just to test? Just to rule out this possibility once and for all, if possible?

[wizard333]

I fiddled with it between 50% and 70%, no change. I didn't go outside that range.