Any reason this wont work?

[imaradiostar]

I kinda got lost after reading a bit...but a few quick thoughts.

You can change to amount of negative feedback by varying the value of the 5k pot in a traditional marshall presence circuit- it forms the bottom half of the voltage divider.

You can vary the value of the cap to change the frequencies effected by the feedback loop.

You can place a cap, an inductor, and/or a resistor in parallel with either resistor that forms the NFB voltage divider. This technique is used to help give the bogner uberschall its extremely versatile tone circuit and was used to increase high frequency stability in many classic "example circuits" for hi-fi and PA amps from the golden age of tube audio. I think this could really make for an interesting "mid scoop" tone in an appropriately designed amp because it generates an active control rather than the normal passive control.

Properly set up you could probably insert an LC network that would allow you to boost or cut and (based on changing C or L values) alter the frequency of the boost or cut. A resistor in parallel with the LC network could function as a "Q" control to prevent the mid scoop/boost from being too narrow to be usable.

If you apply the same concept to the original poster's idea you could probably end up with two "presence" knobs- on for each of the frequencies represented. Connect opposing ends of each presence pot so that by setting the knob in the middle there is equal signal to and from the output into the PI (no boost or cut). By rolling the control either direction away from center it either shunts the treble NFB to ground and increases treble or creates a more direct NFB path, increasing feeback and decreasing treble.

It seems like this is the concept behind most active EQ circuits- they use some sort of phase inverting buffer that allows you to pan between adding to or subtracting from the original signal. Because the different frequencies are blended resistively you should be able to boost or cut somewhat independently- a shelved EQ is still a shelf but you should be able to boost or cut the low freq and the higher will be boosted or cut above or below the setpoint of the first.

Wow...that was overly technical and long. I can't wait for someone to point out that I'm wrong or didn't describe something correctly! sorry, it's late and I'm rambling.

jamie

[Steve Conner]

I don't think it's rambling at all, it's the same line of thought that led me to the Ninja Toaster parametric EQ. Turning the EQ knob one way shunts signal away from the tube grid, causing a cut. Turning it the other way shunts the cathode to ground, which causes a boost.

I wussed out of actually connecting it to the power amp's NFB loop, though. I was worried that it might react badly on something, and make the power amp go unstable.

[daz]

You can place a cap, an inductor, and/or a resistor in parallel with either resistor that forms the NFB voltage divider. This technique is used to help give the bogner uberschall its extremely versatile tone circuit and was used to increase high frequency stability in many classic "example circuits" for hi-fi and PA amps from the golden age of tube audio. I think this could really make for an interesting "mid scoop" tone in an appropriately designed amp because it generates an active control rather than the normal passive control.

Properly set up you could probably insert an LC network that would allow you to boost or cut and (based on changing C or L values) alter the frequency of the boost or cut. A resistor in parallel with the LC network could function as a "Q" control to prevent the mid scoop/boost from being too narrow to be usable.

jamie

That sounds VERY interesting, tho i must admit it went over my head as to the wiring you are describing. Any schematics you can point me to that are designed like that aside from the ubershall which i looked for but seems not to exist on the net? also, when you say "with either resistor that forms the NFB voltage divider" are you talking about the 100k from the tap and the 4.7k? If so, are you describing a "resonance" cap? I've tried that but it basically just adds low end and really hindered more than helped, even when used with a pot to vary it. i always ended up leaving the pot down. Mainly because my amp's low end isn't tight enough for a lot of low end.

[Chuck H]

Daz, you sleep with your soldering iron, don't you?

When the bug bites it's hard to stop. I still own the first amp I ever "tweaked to perfection". And even though at this point, some 16 years later, I know there are some pedantic or questionable circuits in there, It still sounds great when I plug it in. Well actually it's in need of new caps and tubes right now. But if it weren't it would still sound great. I WORKED ON THAT AMP NON STOP FOR AT LEAST A YEAR. I had the time and the urge so that's what I did. What a great learning experience.

If your amp is sounding too brash on the top and bottom (and you don't mind a little soldering ) I have a suggetion. I noticed on your schem that you were using a .022 cap for the PI input. Even though the tonestack is fed by a cathode follower the signal is substantially loaded by the time it gets to the PI input. You could try a .0022 cap for the input of the PI. This should still give you your bottom end but the mud factor would be lower. And even though "bleeder circuits" are considered a sin you could try a very low pf cap on V1 from the plate to the grid. Probably 5 or 6pf. Make sure the cap is rated for the plate voltage there. I use little ceramics for this. These two mods are reasonably non invasive to your tone. I fully expect they will improve your OD tone, but if they detriment the clean and in betweens too much you can always put it back the way it was.

Chuck

[daz]

Hey Chuck, listen to this. Two things have happened in the last few hours. Oh, and no....i don't sleep with the iron, but only because i know i'd get burned. Anyways, i tried what you said but it didnt really help. I've tried that before anyways, tho not with that small a value. anyways, two things have happened. One good one bad. the good, i stuck a 33uf cap in place of the 22uf bypass cap on one of the el34's but left the other one at 22uf. the result was great. the tone just got very nice in several ways. Less harsh in the highs, more sponge factor, less mud, more articulation. So i really don't need to lift the caps to get that tone and lower volume anymore. It sound even better than it did lifting the caps i think. Definatly better in some ways.
the bad, and this started BEFORE i did this so i don't recall what i had done that could have caused this, but i'm getting intermittant motorboating. In fact, thats what led me to try the 33uf....i was trying to replace the 22uf on that side because i suspected it might be bad after lifting it to see. When i lifted it the motorboating stopped, so i thought the cap was bad. Turns out it wasn't, it was just that for some reason it stops when i lift the bypass caps. so feeling that cap may be bad i replaced it with a 33uf because i didn't have another 22uf. Anyways, what can cause this motorboating? It only happens either when i change a cap value that just seems to trigger it for some reaon, ot when i put it on standby or turn the power switch off it will do it as it's powering down. any ideas?

Anyways, I think i'll try an even bigger cap on that one side....maybe it'll sound even better ! I guess it's just throwing the signal balance off thats giving me richer tone.

EDIT: no cigar. But still it sounds great with the 33uf on one side. I wonder also if the voltage rating could have anything to do with it. Both are rated high enough since theres only 30v there and the 22uf is a 50v. But the new 33uf is a 160v. Wonder if somehow that could have anything to do with it. I think not, probably the imbalance of AC signal with the different UF ratings huh? Gott say tho, this has made a truly big improvment. I was pretty sure the needed improvments were in the PA and now i know thats true. the thing is just so much sweeter. the #2 strat position just quacks like crazy, which is one of my tests for great tone. This may be the best it's been so far. also fogot to mention....i put the 220k back on V1A's plate. I now know it just sounds much better like that and that the 100k was partially responsible for that thinness i was getting.

[Chuck H]

Anyways, I think i'll try an even bigger cap on that one side....maybe it'll sound even better ! I guess it's just throwing the signal balance off thats giving me richer tone.

I suspect that's just what's happening. You could also try an imbalance at the PI. Imbalance has been discussed here as a possible componant of good overdrive. I'm still on the fence since I try to balance everything and I'm getting good results. It could be in the "different strokes for different folks" category too.

As far as the motor boating I have a clue. Motor boating is an oscillation. When you lift the caps it stops. That tells me that when the amp is putting out less power that whatever is causing the motor boating no longer has the strength to induce the oscillation. So consider lead dress on any circuit that would increase in voltage when the amp is at full output. Anything in the NFB loop and OT secondary leads are the most suspect. Be sure to move all preamp leads as far from the output and NFB loop as possible. Likewise, move all leads related to the NFB loop and OT secondaries as far from the preamp as possible. After that you would need to consider your board layout and even your grounding sceme.

Proximity... Consider that word... It's the biggest reason proponants of hand wired amps tout for the inferiority of printed circuit boards. I'm still on the fence on this too. But it's true enough that all the componants in your amp, even or especially lead wire, has a radiant field. If two componants radiant field are sharing space that can induce a reaction in each other. Two componants of like phase in too close proximity results in an induced positive feedback. This can cause squealing or motorboating. Depending on the size of the radiant field and the sensitivity of the effected circuit these two enemies don't need to be that close to each other for there to be a problem. Two radiant fields of opposite phase causes NFB. It's these small NFB's that occur with the tight spaces on PCB's that give them their bad rep. I once modded an amp for a friend that had incredible OD but occasional fits of oscillation. I moved the OT on the chassis as I felt it was poorly located. The oscillation stopped but the OD tone slipped from hairy and grand to mediocre. I put the OT back where it was and the OD tone was back. As were the fits of oscillation. The OT location was introducing positive feedback that was increasing the preamps gain in a very specific way. He ended up with another amp and I now own this odd ball. Whenever I get around to it I hope to solve it's issues and learn a little about how to intentionally induce minor instability in a good way. The late Ken Fischer was said to use this practice. But it seems like such an obscure art that I can't imagine ever being able to do it at will with a new design or even repeat it consistently considering componant tolerances and all. Anyhoo, there's a place for you to start and some info to digest.

Chuck

[Steve Conner]

High frequency squeals may be caused by lead dress, but motorboating isn't. The formal explanation for it involves a lot of math (and can be found in RDH4) but what it amounts to is that, hidden in your circuit, is something similar to Fender's old tremolo oscillator.

And what the trem oscillator is, is an amplifying stage with a negative feedback loop around it containing three RC filters. Each RC filter contributes some phase shift, and at some frequency, the three shifts add up to 180 degrees. This flips the negative feedback into positive, and if the tube provides enough gain to make up the losses in the RCs, the circuit oscillates.

In the case of a power amp, the OT already contributes one time constant, so you only need another two RC networks to get your 180 degrees and start it motorboating. The input coupling capacitor to the PI, with the PI grid leak resistor, is your second time constant. (Yes, this is part of the feedback loop.)

So you just need one more RC to kick it off, and I guess Daz's undersized cathode bypass caps are it.

The worst oscillations occur when all three time constants are the same, as they are in the trem circuit. If the amp is only just oscillating, then changing any one of them should help.

Even if the NFB loop isn't actually oscillating, it can still be underdamped, if this is the case you can sometimes see the speaker cone flapping in and out more than it should.

Old hi-fi designs like the Williamson etc. have one more gain stage inside the NFB loop than guitar amps, which adds one more RC, so they naturally just want to oscillate. Tuning all the coupling cap sizes for good bass without motorboating can be a nightmare. (And again, the magic incantations for doing it can be found in RDH4.) I think this is why the audiophools give them a reputation for poor bass.

Preamp stages can also motorboat, again the circuit is just like the old trem oscillator, but one of the three RCs is the supply dropping resistor and electrolytic bypass cap. That's why dried up electrolytics cause motorboating: if the cap loses enough capacitance, the RC time constant gets shorter until it's similar to the preamp coupling caps, and off it goes.

[daz]

I think you nailed this steve. Because while it had already been doing this occasionally just a little, when i tried chuck's idea to change the .1uf cap on the NFB side of the PI thats when it went crazy and began doing it constantly instead of just a few little putts when i turn the standby or power switch off. So i guess i'll just have to change most any component in the NFB or PA in general that can be changed in value slightly w/o affecting the tone but will change the balance of things enough to stop it, eh?

Today i intend to try a few things that i liked before i discovered the imbalanced cathode caps but didn't keep because they contributed to the flabby lows. Now that those are no longer an issue i want to try a few things to beef the tone up. For example, someone elsewhere recommended i try .1uf couplers in front of the el34's in place of the .022's, and it seemed nice but because my lows were muddy this made that worse, tho otherwise it seems to sound really nice an beefy. So i will try that and some other things. I find that when something isn't quite right like flabby bass, many potential improvements don't work when they otherwise would have if that issue didn't exist. So once you fix that issue you can then go back and re-try those things. Thats what is so addictive about the design process.....one little change makes the entire rest of the amp now very different in terms of what tweaks will or won't work. so you can literally make one change that un-does all the things you perfected in the past ! It's crazy stuff and thats why i've been at this amp for a probably a year now !

[daz]

Chuck, how would one go about unbalancing the PI? Right now the plates are 00k/82k, but from what i've read that actually balances it. So i tried adding a 15k r to the 100k so it's now 115k/82k. It seems to cut some more of the brightness which is good because the amp has more than i wat anyways and other methods of cutting it haven't left me with very pleasing results. But is the way i just did this good or are there rules i should go by such as WHICH plate gets the lower resistor and how to know how much difference to use between them, or things like shoul i drop the 82k further or add to the 100k as i did? Any suggestions as to how to do this correctly instead of blindly as i am now.

[Chuck H]

High frequency squeals may be caused by lead dress, but motorboating isn't.

Well I personally never had a problem with it. Nor am I an EE. But I thought it was true because of accounts I have read about and (perhaps wrongly) was trying to relay helpful info. One account in particular that is in print comes to mind. Now I can't speak to the technical competance of Gerald Weber but he does have an excellent reputation for repairing and restoring vintage Fender amps. In one of his books he writes that an early Deluxe model (can't recall which) has a tendancy to motor boat when overdriven. The common fix is to turn the PI input cap 90* on the circuit board. Perhaps this is a rare example but I have read other accounts of PFB due to layout causing motor boating. Like I said, I have never had a personal experience with it so anything I read about may have been a lie or an incorrect assesment of the problem.

It's worth noting that some of the changes Daz made to the amp in this process would not have altered the phase relationships but somehow made the problem worse.

Chuck