"Swirly" Distortion Decay Sound Fix !

[HaroldBrooks]

Some of you might remember a while back I had a weird high frequency swirly type distortion present on the decay of high gain power chords in my modified Gretsch 6156 amp. That amp uses two 12ax7s up front for the preamp and PI, and 6973 tubes at the end. The fix was to totally remove the 220uf cathode bypass cap that I added to the shared cathode resistor (there was none there before). That fixed the problem 100%, but some gents here indicated that removing the big cathode bypass cap might have just masked some imbalance that was already present in the power tubes / output phasing, or perhaps an undetected oscillation.

Well it turns out my latest project amp (and RCA with no schematic) has the exact same high frequency swirly decay sound, and without any power tube cathode bypass cap in place. I have to say this sound is only present when I am driving the gain sky high to ridiculous thrasher metal levels on the decay of a long power chord, and I normally don't play like that, and would use a distortion box to get that effect on rare occasion as opposed to whipping the amp to a frenzy.

I switched guitars, and systematically eliminated three boxes (Compressor, EQ, Boost) from my test rig to make sure one of the pedals weren't giving the amp the offending tone to amplify, and they weren't. I also switched speakers, and the swirly decay sound is most evident when I am using a rather bright and trebley Jensen 12" speaker with the new amp, less with other speakers. I also suspected it might be the speaker early on, but swapping speakers didn't fix it entirely, just muted the sound with darker voiced speakers.

I started to experiment with changing values of a single grid resistor on either of the grid leak resistors of the 6L6 tubes, and Voila ! Lowering the grid resistor on just one of the tubes got rid of the offending noise ! I think I know how this worked, but I am looking for some experienced input. I had previously lowered the value of the 6SN7 PI tube to 6L6 coupling caps from .1uf to .047uf. The grid resistors are both marked as 270k. I got a clue to a possible solution, by dialing back bass input to the amp with an EQ, and that helped some but not entirely. Generally the higher the gain, the less bass frequencies I use on input to clear up the tone of an amp.

Correct me if I am wrong, but the coupling cap and grid leak resistor form a high pass filter to the next tube, and working out the original values the corner freq should be 1/(2*3.14*RC) = 1/(2*3.14*270000*0.0000001) = 5.9hz. The new resistor I chose for the single tube and the new lower size coupling cap would yield =1/(2*3.14*30000*0.000000047) = 113hz.

So I believe with the new lower sized grid resistor and coupling cap arrangement on the single tube, I am dumping to ground a portion of the low bass with a corner frequency starting at 113hz, and below for that tube only. This nearly eliminated the Swirly ugly distortion, but I also realize it is perhaps a band-aid to mask another problem.

All comments welcome !

[Chuck H]

I think the swirly distortion must be crossover distortion resulting from the bias voltage increase at the cathode resistor with conduction and/or grid loading. Which is also (and probably more properly) called grid conduction.

By reducing the value of only one of the power tube bias resistors, which are also (and probably less properly) called grid leak resistors, you not only change the knee of the coupling cap, which reduces grid conduction, but also reduce the time constant which also reduces grid conduction. You also add imbalance to the operating conditions of each half of the push/pull circuit. And in my experience crossover distortion sounds less buzzy and pronounced with an imbalance here.

What you've done isn't a BandAid really. An intentional imbalance is pretty common on harp amps and others have done it with guitar amps, though much less common. But I think you can do better. Lower both grid bias resistors to something like 150k and target 70Hz for your knee frequency. This will put the balance back in the circuit (such as it is) and should perform better overall for that. If that still produces too much of the ugly distortion then please report on the idle current. Cathode bias amps are generally run hot to compensate for bias shift with signal conduction.

[Helmholtz]

I would think of OT core saturation caused by PI signal asymmetry (means net DC in the primary) and low frequency content. Only a (preferably dual channel) scope could tell.

[HaroldBrooks]

I think the swirly distortion must be crossover distortion resulting from the bias voltage increase at the cathode resistor with conduction and/or grid loading. Which is also (and probably more properly) called grid conduction.

By reducing the value of only one of the power tube bias resistors, which are also (and probably less properly) called grid leak resistors, you not only change the knee of the coupling cap, which reduces grid conduction, but also reduce the time constant which also reduces grid conduction. You also add imbalance to the operating conditions of each half of the push/pull circuit. And in my experience crossover distortion sounds less buzzy and pronounced with an imbalance here.

What you've done isn't a BandAid really. An intentional imbalance is pretty common on harp amps and others have done it with guitar amps, though much less common. But I think you can do better. Lower both grid bias resistors to something like 150k and target 70Hz for your knee frequency. This will put the balance back in the circuit (such as it is) and should perform better overall for that. If that still produces too much of the ugly distortion then please report on the idle current. Cathode bias amps are generally run hot to compensate for bias shift with signal conduction.

Thanks Chuck, I will play with the values of the 6L6 tubes as you mentioned, I just may not use 150k owning to the fact of the .047uf caps I newly put in... But those could be changed as well. Question, Is there any material difference in using a higher value resistor (as in 47k vs 30k) with the same .047uf caps, to lower the corner frequency ? I realize it can be done with the cap or with the resistor and achieve the same results frequency wise, just wondering if it matters which element is changed, and are the limits or negatives.

Thanks !

[HaroldBrooks]

I would think of OT core saturation caused by PI signal asymmetry (means net DC in the primary) and low frequency content. Only a (preferably dual channel) scope could tell.

You may be onto it as well Helmholtz. I am running a "20W" OT, but I would think based on the sturdy power transformer, and the high plate and screen voltages I am running for the pair of 6L6s, that at high gain they would easily exceed that. What I've found in actuallity is the amp is only modestly more powerful than my other 6V6 amps. The original OT was missing from this amp, So I just used the one I had on hand, a 20W 6.6kohm push pull with leads for 8 and 4 ohms.

I think I will be buying a 35W transformer very soon, as this amp deserves it !!!

(I need a scope...)

[Chuck H]

I started to experiment with changing values of a single grid resistor on either of the grid leak resistors of the 6L6 tubes,..The grid resistors are both marked as 270k.

I think I may need some clarity. 270k would be very high for a "grid resistor" but very typical for a "grid leak resistor". And "grid resistor"s are not typically found on "either side of the grid leak resistor". The term "grid resistor" usually describes a resistor directly feeding the tube grid. Anything behind the "grid leak resistor" would qualify as part of a voltage divider.

So... What did you change, what was it's value and what value did you change it to?

[HaroldBrooks]

I think I may need some clarity. 270k would be very high for a "grid resistor" but very typical for a "grid leak resistor". And "grid resistor"s are not typically found on "either side of the grid leak resistor". The term "grid resistor" usually describes a resistor directly feeding the tube grid. Anything behind the "grid leak resistor" would qualify as part of a voltage divider.

So... What did you change, what was it's value and what value did you change it to?

I think it would be considered a "grid leak resistor as it appears directly after the coupling cap, and is tied to ground. However...

The "other" tube I didn't change the resistor for is probably not a grid leak resistor, and I don't know what to call that one ! The 270k resistor is between the pin #4 grid of the 6SN7 tube grid and the grid #5 pin of the 6L6 tube. Not sure why the two grids would be connected in that way, but there is another 18k resistor from the 6SN7 grid that goes to ground, so I guess that is a voltage divider, but not with input from a PI plate. and I would be grateful if someone could explain that arrangement to me. Dropping the ohms on that resistor increases gain. The coupling cap on that tube goes from the 6SN7 pin #2 plate to the grid #5 of the 6L6, as one might expect, but with no resistor dumping to ground directly.

And yet, it all works quite nicely !

I made one other change in the position and grounding of the primary filter capacitor (a new independent electrolytic in place of the failing can cap), as I felt the amp had too much hum that came from my repairs and the gain boost I added, particularly at idle with the bass pot turned up. There is a rail that all the grounds are tied to running in the direction of the length of the chassis, so after trying a few locations to ground the main filter cap, I decided to try the very end of the ground "bar" closer to the rectifier tube, and what a difference !!! The hum is 1/4 what it was before. I have a lot of respect for people who know how layout effects hum, and perhaps any modulation to the sound that would be caused by hum, and I still need learn a lot in this area. The hum reduction was a trial and error affair, but what a difference this position made !

Pick of original chassis (no changes)

[g1]

The 270k resistor is between the pin #4 grid of the 6SN7 tube grid and the grid #5 pin of the 6L6 tube. Not sure why the two grids would be connected in that way, but there is another 18k resistor from the 6SN7 grid that goes to ground, so I guess that is a voltage divider, but not with input from a PI plate. and I would be grateful if someone could explain that arrangement to me.

If schematics have been posted elsewhere, please post them in this thread.
Without schematics, I don't think anyone will be able to explain the arrangement to you so you will have to draw some up.

[Chuck H]

If schematics have been posted elsewhere, please post them in this thread.
Without schematics, I don't think anyone will be able to explain the arrangement to you so you will have to draw some up.

I think this is the RCA amp? True PTP and a real PITA (if not impossible) to try and decipher from photographs. Not much easier on the bench But there's not a lot in there so maybe one focused hour working on a drawing of the power amp, PI and perhaps the power supply.?. That would help a lot.

[g1]

Yes, I think the business about the grid resistors for the power tubes would be at least easier to contemplate if we even knew what kind of PI arrangement was being used.

[HaroldBrooks]

Wife had me doing the furniture store tour today... once that important and life enriching task is completed, I will have more time to get back to what matters, and I think you all know what that means, amps !

It's not fair to ask you questions without at least an attempt at a schematic, and I will get to it soon, I promise.

One related question : How many what's can you specifically expect from a fairly high voltage push pull 6L6 amp, with an ample power supply, and a non-limiting power transformer ? The reason I ask is I see on other sites OTs for 6L6 amps with a 4.2K primary and rated as high as 50watts. The transformer I am using on the RCA currently is a Classic industries 20W version with a 6.6k primary, and I have a feeling it's holding back the amp. I have plenty of amps with a lot of limiting squish courtesy of the smallish output transformer. I want this amp to be different !

Thanks !

[nosaj]

I think this is the RCA amp? True PTP and a real PITA (if not impossible) to try and decipher from photographs. Not much easier on the bench But there's not a lot in there so maybe one focused hour working on a drawing of the power amp, PI and perhaps the power supply.?. That would help a lot.

The very first amp I started with was an RCA MI- something (MI for Master Index number) It was a 40's amp 2 6v6's
I think I was told it was a parallel SE amp. It was a hot little amp just the way it was. Never found a schematic online but I did find a guy who had one in a storage shed an scanned it for me. When I get back home I can check the model. I thought I posted it here or at ARF but my search-FU skills seem to be sucking today could be the arithritis in my fingers.

nosaj

[nosaj]

The very first amp I started with was an RCA MI- something (MI for Master Index number) It was a 40's amp 2 6v6's
I think I was told it was a parallel SE amp. It was a hot little amp just the way it was. Never found a schematic online but I did find a guy who had one in a storage shed an scanned it for me. When I get back home I can check the model. I thought I posted it here or at ARF but my search-FU skills seem to be sucking today could be the arithritis in my fingers.

nosaj

This guy has a schematic lists as
RCA VICTOR PTU-1 PROJ.TV AMP MI-6599 '48 SER $22.00

http://www.agtannenbaum.com/r_cat.htm
Now if we could figure out the exact model of tv we could pull the Sams folder to get to the audio portion of the tv.
Sorry for the tangent found a rabbit hole
https://www.earlytelevision.org/pdf/...n_receiver.pdf
https://www.earlytelevision.org/rca_9pc41.html
nosaj

[Chuck H]

When you say "Classic industries 20W" do you mean ClassicTone 20W Deluxe Reverb transformer? Because if you do then a pair of 6L6's with high-sh voltage from an ample power transformer will be enough to actually jeopardize it without limiting safeguards built in. In which case you're looking at something like 25W to 30W. I've used Hammond 16XX and Stancor transformers rated at half an amps intended watts before without any problems or limitations. But those are full range transformers and guitar amps aren't full range. In fact a tech at Hammond once told me that choosing from their standard transformers it's better to go with a half rated model for predictable tonal results in guitar amp builds. But the amp clone transformers made by Hammond, ClassicTone, etc. are rated for guitar amps. So if you have a ClassicTone 20W OT then it's good for a 20W amp and not much more. Not that using an underrated OT hasn't been done. Hell, it's been done by major brands. But the results aren't usually very satisfying from a power or fidelity angle. So if you want something different than a smallish, squishy amp I think you're going to need bigger iron.

[nosaj]

If your ready to hunt here is a list of televisions produced by RCA listed by years which helps to limit scope.
nosaj