understanding unintended/unwanted oscillation in guitar amps.

[mikepukmel]

Oh, sure, after we spouted off theoretical stuff, ask for what to really DO.


What you're after here is the One True Ground. You must construct it. This will be on big, burly lump of solderable connections that are as conductive to each other as you can possibly make or afford. ...


I would clear up other issues and return to this "clot" later if it's already built. The general idea is to clump, for instance, the ground returns for one or two tubes which all do the same work on the same signal at a time. Then run a ground wire back to the One True Ground for that lump.

Hi RG, So it was built, but its a hobby thing, and I have a couple of gifted Scholtz Rockman's to practice with (to my families great happiness ) So, I decided to take it mostly apart, get the RIGHT board, with the right dimensions so the wiring would be as short as possible, and I could clean up the clumps. After all this, it wouldn't make sense not to fix the 100R heater center tap resistors. Taking a little more time on this round, hopefully the amp will be more quiet and oscillation free.

The Valve Wizard recommended larger grid stops, not sure how high to go. I have extra 68k, maybe can use those.

[pdf64]

Grid stoppers in early stages are a significant noise source, so their value needs to be big enough to be beneficial to stability, EMC etc, but not so big that the noise floor is raised significantly.
Remember 'Grid Stops should be at least 8 divided by the triodes gm...', so as low as ~5k might be used for a 12AX7.

[soundmasterg]

[ATTACH=CONFIG]46973[/ATTACH]

HP 200CD ... I bought it 35 years ago for $15. Only thing that doesn't work is the leather handle. One of the things that will be leaving soon.

I've got one of these too, that I need to replace some e-caps in but it works ok. It has a nice, strong, clean sine wave...I can a sine wave from my function generator, but it isn't as nice since it is derived from a triangle wave.

Greg

[soundmasterg]

Hi RG, So it was built, but its a hobby thing, and I have a couple of gifted Scholtz Rockman's to practice with (to my families great happiness ) So, I decided to take it mostly apart, get the RIGHT board, with the right dimensions so the wiring would be as short as possible, and I could clean up the clumps. After all this, it wouldn't make sense not to fix the 100R heater center tap resistors. Taking a little more time on this round, hopefully the amp will be more quiet and oscillation free.

The Valve Wizard recommended larger grid stops, not sure how high to go. I have extra 68k, maybe can use those.

I have built several amps using the same grounding scheme and they are all very quiet, especially the one I used DC filaments on the first preamp stages for each channel. Two amps used a board like a Fender, and the third was true PTP. All of the amps used local filter caps located near each stage they supply, and there was a ground buss on the board with each ground going to the buss near the cap that was supplying that stage, then the buss grounds to a chassis ground point (RG's One True Ground) near the first filter cap. The buss had everything up to and including the PI on it. The output jack and input jack are switchcraft jacks (the longer shaft ones) with fiber isolating washers, (Requires a slightly larger hole) and the output transformer and other high current grounds like heater center tap etc., go to the first filter cap, which has a wire that goes to the main chassis ground point also. Dead quiet all of the amps with that setup, though the PTP amp had an oscillation with higher gain settings that was solved by moving a wire between the PI grid and the previous stage about 1/4" and zip tying it in place. I found that particular problem by chopsticking and moving wires around when the amp was oscillating, and yes it took awhile to narrow down the problem on that particular amp.

Regarding the grid stops that Merlin recommends....highly recommended. I planned my layout with them in place and started with a 10k and didn't need to go larger than those. I think they make the amp sound better overall, plus it helps with some layout things to include them.

Greg

[bob p]

Slugging the Poles

If anyone knows of a book, maybe an old EE book, that would be great.

PM your email address to me.

On these old circuits, were these amps really that close to oscillation? Id like to understand what path(s) the feedback is occurring, and how to fix it properly, so that no matter what the setting of any controls, the amp won't go into sub or super audio band oscillation and cook the tubes (and whatever else tends to get cooked).

None of the old circuits were really that close to oscillation, IMO, but that changed when designers started pushing the feedback loop's closed-loop gain higher and higher. I'll try to explain the function of the NFB loop in a qualitative way that doesn't involve much math.

To understand where the FB is occurring: if you're noticing HF oscillation in the power amp when the controls are max'd then you need to consider how feedback is generated and why.

NFB is defined as feedback that is 180* out of phase with the feedback loop's input signal. Provided that the fed-back signal is always 180* out of phase then the feedback loop will operate properly to clean up the signal at the amp's output. If the phase shift were allowed to deviate from the 180* value then the closed-loop gain will begin to increase as the effect of NFB is reduced. When the phase shift approaches exactly +/- 90* then the NFB loop becomes non-functional and there is no effective NFB at all. If the phase shift is allowed to exceed 270* or become less than 90* (ie: approaches zero from the 90* null point) then the feedback becomes positive and reinforces the input signal. When that happens the closed loop gain increases above the level of the open loop gain and oscillation results. Ringing will occur (producing oscillations that are not sustained) if the closed loop gain is not high enough for the input signal to trigger sustained oscillation. Continuous/sustained oscillation will occur if the feedback signal becomes greater than the input signal and the amp achieves infinite gain.

These sorts of problems typically occur only when selected controls are set very high. One can make an argument that the design of the controls should be changed so that they can't result in settings that are high enough to allow feedback to occur. (To circumvent this problem some people will keep a treble control from having zero resistance when turned to 10 by inserting sufficient resistance in series with the control to prevent oscillation, though there are better ways to accomplish this goal.)

The typical NFB amp that brings speaker output back to the phase inverter has 3 areas in the NFB loop that contribute a phase shift to the signal. These three phase shift points are known as "poles" in the feedback loop's frequency response. They are the areas in the amp's design that need to be controlled to eliminate HF oscillations.

Pole 1 : HF losses in the output transformer
Pole 2 : input capacitance of the power tubes
Pole 3 : input capacitance of the phase inverter

The problem we're faced with is that the dominant pole of your typical NFB tube amp is going to be the output transformer. It imparts an HF rolloff that is difficult to deal with, as the OT is effectively a black box that cannot be modified without changing the OT. This is the main reason that you want to be sure not to skimp on budget by buying a cheap, low quality OT.

Because you can't really address problems in the OT short of changing out one OT for another, the "textbook" method of dealing with the problem is to force the other 2 poles contained within the feedback loop to become more dominant in determining the feedback loop's frequency response than the OT, and to use those poles to impart a HF rolloff in the feedback loop's frequency response to prevent HF oscillation.

The textbook technique for controlling oscillations in an NFB circuit is known as "slugging the dominant pole." The dominant pole is the pole in the NFB loop that has the lowest rolloff frequency. That's normally the OT. But the other poles can be "slugged" by adding capacitance to them in order to push their rolloff frequency downwards, thereby making them dominate the frequency response in the NFB loop.

This type of "slugging" is typically performed by adding capacitance to the inputs of the power tubes or the phase inverter. The classic way to accomplish this is seen in the later SF Fender amps that place additional pF-rated capacitors from grid to ground on the power tubes. This lowers the frequency response of the feedback loop so that the offending HF content that is likely to cause oscillations is rolled off at a high rate of dB/octave by combining the rolloff effects of 2 or 3 poles within the feedback loop.

Clear as mud?

[mikepukmel]

Getting much more clear. I need to get off my lazy and do more reading. Fund to sling solder, but harder to get back into reading that might involve brushing up on (eek) calculus.

Im a consummate tinkerer. Character flaw. I get so far on a project, read a great post by one of you guys, then think "Wow, that would be great to put in this amnp", so I stop, think a bit, order some parts, and take something apart, ... etc. I have to fight that alter ego to, as Leo says "Just get the amp built, then!" so I can use it for its intended purpose, rather than as a very large, 1,500.00 bread board.

[mikepukmel]

. . . These sorts of problems typically occur only when selected controls are set very high. One can make an argument that the design of the controls should be changed so that they can't result in settings that are high enough to allow feedback to occur. (To circumvent this problem some people will keep a treble control from having zero resistance when turned to 10 by inserting sufficient resistance in series with the control to prevent oscillation, though there are better ways to accomplish this goal.) . . .

Clear as mud?

Oh bingo! So, the tinkerer me did a whole bunch of experiments, that's exactly what I found. Test number 62, Turn all pots on 10, except the treble. All pots on the guitar on 10. Move the guitar around and back and forth to minimize the hum from positioning of the pickups. Turn off all fluorescent lights, etc to minimize external influence. Then, slowly turn the one treble pot up, strum the guitar, let it decay, turn the treble up ... At some point, increasing the treble a tiny bit, pouf, the output tubes went into police search light mode: bright bright orange, and the amp sounded hideous. Turn the pot back down below point X whatever that was, and the tubes immediately stopped screaming.

I can hear the Fender engineers talking at the water cooler, its 1964, ... "Nobody but NOBODY will ever turn all controls on 10. They just won't do it... so don't even worry about that, Jenkins, you're wasting your time . . ."

So, back to the original original question: If I put a few good quality, metal film 6.8k stop resistors, on pin 2 and 7 of a few of the tubes, would this most likely damp out this rare case "all pots on 10" oscillation?

As an aside, I asked around various blogs and sites, and could not get an answer why most folks don't put stops on the phase inverter.

[bob p]

I would only put the grid stoppers on tubes that actually demonstrate a need for them. 6k8 seems like a reasonable starting point. I'd go up or down from there, using only what's needed to cure the problem.

[Gingertube]

Sometimes you can get unwanted oscillation from left over bad design. See the Fender Supersonic 60 Burn Channel:
http://www.thetubestore.com/lib/thet...atic-Rev-A.pdf
V1b, V3A, V3B are 3 cascaded gain stages all from the power supply node "X".

I got "motorboating" style oscillation on one of these - not your usual "put put" but a squeal.

I "fixed" it by fitting a 470nF 630V polprop. cap at the top of R44 to 0V locally at V3B to reduce that power supply node impedance to minimum. That treated the symptom (got rid of the squeal) but not the disease (bad design).

I ended up looking for something like this after fitting grid stops everywhere in that channel did nothing at all for the squeal.

Cheers,
Ian

[mikepukmel]

I would only put the grid stoppers on tubes that actually demonstrate a need for them. 6k8 seems like a reasonable starting point. I'd go up or down from there, using only what's needed to cure the problem.

Thanks Bob. I get stuck here, since I don't have a deep understanding of the problem, and don't have a scope to see it either, so not exactly sure. It was more of a question if I could put grid stops in as a preventive measure, and not cause huge changes in the amp character (and of course not damage anything). I don't know the path, but seemed to be preamp V2, through the reverb ckt then through to the output tubes. The last, smaller problem I found was only by fiddling with pots as mentioned above. So, without a scope, and of course more advanced diagnostic skills, I wasn't sure exactly where grid stops were needed. The output tubes have stops, standard value (1K5?) none on pin 7 of v1 or V2 or the 2 reverb tubes. In the rewiring, I put grid stops on pin 7 of V1 and V2, moved the 2 68k stops on pin2 right to the tube, added grid stop on the reverb recovery tube. Didn't touch the phase inverter, wasn't able to convince myself that it would not affect the intended operation.

[mikepukmel]

Sometimes you can get unwanted oscillation from left over bad design. See the Fender Supersonic 60 Burn Channel:
http://www.thetubestore.com/lib/thet...atic-Rev-A.pdf
V1b, V3A, V3B are 3 cascaded gain stages all from the power supply node "X".

I got "motorboating" style oscillation on one of these - not your usual "put put" but a squeal.

I "fixed" it by fitting a 470nF 630V polprop. cap at the top of R44 to 0V locally at V3B to reduce that power supply node impedance to minimum. That treated the symptom (got rid of the squeal) but not the disease (bad design).

I ended up looking for something like this after fitting grid stops everywhere in that channel did nothing at all for the squeal.

Cheers,
Ian

Wow LOOK at all the cascaded gain stages! A whole bunch. I think even more than Dumble had in his magic amps. So, Fender, this really big company, with all its resources, didn't put design checks into the process? Especially since this is such a late amp, Its not like they didn't know what to expect from a tube amp design after 50 years. How does the amp sound?

[bob p]

The SuperSonic is a very popular amp. But don't think of it like a cascaded gain Boogie, it isn't like that.

If you read the various forums, there are a lot of people that swear by it as being the best amp that Fender has ever made. There are a lot of people who buy these and don't waste their time chasing vintage gear or the expensive reissues. As you might expect there are lots of videos on the tube that put the amp through it's paces.

This guy says that he records every one one of his videos on a SuperSonic 60 with the controls kept in the same position, and only switches back and forth between the Clean and Burn channels as he changes guitars and pickups. He's done quite a few videos comparing different guitars and picups, which should give you a good idea what the amp sounds like.

[bob p]

This guy says that the amps get painfully loud while still staying clean.

I've never owned one.

[Justin Thomas]

That wasn't even "clean" on 4... Not by a long shot. As far as "painfully loud," well... pansy.

I think we've come to a complete misunderstanding of what "clean" is, after a few decades of super-saturated cascading gain monster amps. I would consider that initial setting on 4 to be crunchy, especially if you're pounding out full chords.
And jeez, come on - turn it up to ten already!

Also: I think the Prosonic is better...

Justin

[J M Fahey]

¿¿¿¿¿¿¿¿¿¿¿ Clean???????????
Not for one second in the entire video.

And I´m perfectly aware of the SPL produced by 50W RMS, both cranked and clean, into a Celestion V30.
Nice loud and defined but not overpowering volume, perfect to play live Rock or hot Blues at a Club with a regular drummer, but forget Heavy Metal and such.
Unusable full blast in a bedroom, living room, and barely in a garage, but that´s not the point, 95% of amps are designed to be played live.
That they often are not, is not the fault of the designer.

In a nutshell? this guy is both deaf (he doesn´t hear distortion) and a wimp.