Lets discuss amp tone and how it changes w/volume and why

[nickb]

I perceive it as a loss of LOW mids, (400-800hz) but a lot of hi mids, like i said sounds like 1k. The mids that create that smooth creamy sound i always think as just below that, around 700-800Hz. I get all these beliefs from playing with EQ's and hearing what those bands add. The biggest issue however more than what frequencies are changed is the loss of any sag and that brittle hard attack which is nite and day different and totally loses that whole feel thing that makes it so much more playable at low volume.

Also, note that i don't think the difference between the low volume goodness and the harshness i'm talking about when turning the master up is anywhere near that difference in gain, IE:20% to 80%. I'm talking turning the knob literally about 1/16" further. Maybe 8:00 to 8:15 to go from smooth and nice to that harsh state.

Well, a small rotation as that implies we've hit an inflection point of some kind; clipping & instability are two that spring to mind. I hate to bring up measurements again but it really is the only way to make any structured progress.

[Chuck H]

Agreed! That small a rotation and that profound a perception imply that something is hitting the wall (so to speak). But I would doubt a parasitic because the turnover is so specific and doesn't continue to manifest with other changes, as such things always do IME. Clipping? Perhaps. This "sag" effect that daz talks about is the especially confusing part for me. Measurements are indeed to only way to know what's happening here because it isn't actually "sag" but something that is implying that effect to daz's ears.

Also worthy of note is that speakers do all kinds of whacky things impedance wise. With the coil jumping up and down on the core plus the inductive properties and all. Perhaps the speakers are doing something at this particular drive level. That would certainly explain why the phenomenon seems to be so volume specific. On a related note, since speakers are dynamic in their impedance, it could be a reflected load on the OT altering operation and tone. Especially WRT the NFB loop.

daz, what is the nature of your speaker cabinet (speakers, impedance, etc.) and how do you have your impedance selector set for that load?

[cooldude666]

Just record the freaking effect so we can HEAR what's going on. Instead of trying to imagine a hard and brittle attack we could HEAR a hard and brittle attack, etc

[daz]

daz, what is the nature of your speaker cabinet (speakers, impedance, etc.) and how do you have your impedance selector set for that load?

1-12 combo with EV12L, and i also use a celestion 100watter in another cab at times. they both do it

Just record the freaking effect so we can HEAR what's going on. Instead of trying to imagine a hard and brittle attack we could HEAR a hard and brittle attack, etc

thats totally a feel thing, you aren't going to notice anything aside for a tonal change as i described.

[trobbins]

Were you able to compare the 'more amplified' sound character via alternate paths than turning up your amps volume control - such as DI the output, and/or mic it and amplify and reproduce the signal via a different 'clean' path ? The aim would be to try and discern if the character really is from inside the amp, and related to the stages from and subsequent to the volume control.

[Chuck H]

You actually didn't answer my question so I'm just going to come out with it... If you have your impedance selector set properly for the load you're using, try setting the impedance selector for half the load you're using. As in, if the speaker load is 8ohms set the impedance selector to 4ohms. This will raise the power tubes primary impedance and could minimize some speaker impedance related issues if they exist. It may also even out dynamics in the NFB loop. If the dynamic speaker impedance is in any way responsible for the problem this experiment will have an affect.

[daz]

You actually didn't answer my question so I'm just going to come out with it... If you have your impedance selector set properly for the load you're using, try setting the impedance selector for half the load you're using. As in, if the speaker load is 8ohms set the impedance selector to 4ohms. This will raise the power tubes primary impedance and could minimize some speaker impedance related issues if they exist. It may also even out dynamics in the NFB loop. If the dynamic speaker impedance is in any way responsible for the problem this experiment will have an affect.

I'll have to wait till i can. (can't crank it at the moment) But i CAN tell you it doesn't make a major difference because it was on 4 yesterday when i removed the 6v6 i was using and put el34's back to see how they compared as low vs loud goes. I forgot to set the selector back to 8 and i didn't notice anything. So i will try it and listen closer but i don't think it matters much if at all. I really don't think theres anything wrong here. It's just in the design as you eluded to, but it may be that with MY particular speaker, tweaks, etc, it's just a better recipe for exaggerating this issue.

I would seriously consider a 4th tube if i thought a recovery stage would do the trick. I wonder if any production or boutiques use that method to even out tone across the volume range.

[Gingertube]

Just a few shotgun blasts at the barn door.

Above there was a suggestion about looking at it with an oscilloscope rather than listening to it - May I suggest that the absolute best piece of test equipment that you are ever likely to posess (at any price) is one set of bog standard ears (and the processing circuitry between them).
Those who wander over to DIY Audio HiFi section occasionally will note my Baby Huey HiFi design. This design was done using ears ONLY as a deliberate policy, not because I did't have access to spectrum analysers, distortion meters etc., but because I was primarily interested in how it sounded not how it measured.

The pre PI Master Volume influence - look up TUT whichever (maybe someone can fill in a reference) and implement the bootstapped Master Volume circuit. That will address any loading of preceeding circuitry issues.

As the Power Amp starts to work harder(as we turn the volume up) we start operating the tubes over a wider part of their transfer function. That will change the harmonic content of the sound, at higher volumes the odd harmonic / even harmonic content skews more toward the odd. That changes the "timbre" of an instrument, if you equate "timbre" with "tone" that may be significant.

Getting tubes to operate over the same part of their transfer function at different volumes is why the guys who design Powerscaling circuits use tracking bias regulators (to track the anode/screen voltage) in an attempt to keep the tone the same at lower power settings.

There is also power compression in the speaker itself, this is a "design feature" of MI speakers, HiFI speakers exhibit much less of this. In MI speakers, it (primarily) has to do with part of the voice coil extending passed the ends of the magnet on normal voice coil excursions.

Not sure if these are all misses or there are now a few pellet marks in that Barn Door. You can however be sure I attempted to aim high enough to miss my foot.

Cheers,
Ian

[daz]

You may be onto something there !I read this then soldered 2 47k resistors across the 2 1M grid resistors on the PI. It's a bad time to crank it, but I did for a couple seconds from what i just heard it may be the fix to some degree. What i found was it still did it to a bit lesser degree, but with the variable NFB and tone controls i was able to get the tone back very close and even when i then turned it up to regain the volume that i lost making those adjustments, the tone seemed to remain intact quite well. I need to really consider doing this, but i will need to wait till i can crank it a bit.

Heres my only problem.....i don't get why 47k resistors, which in parallel to 1M make a total of 44k, would not be very low volume. i turned the pre pi master up a little more than normal, but it wasn't quiet at all. I could do a loud gig with it like that i think ! If thats the case, wouldn't a 1M pot do nothing for 99% of it's travel? Oh, and if i DO try this i only have a 500k dual pot, but i think that'll be good. Of course thats assuming theres a reason the 47k resistors didn't drop the volume to near nothing. Can you explain that? Maybe i missed something else that goes with the pot part of the mod?

Just a few shotgun blasts at the barn door.

Above there was a suggestion about looking at it with an oscilloscope rather than listening to it - May I suggest that the absolute best piece of test equipment that you are ever likely to posess (at any price) is one set of bog standard ears (and the processing circuitry between them).
Those who wander over to DIY Audio HiFi section occasionally will note my Baby Huey HiFi design. This design was done using ears ONLY as a deliberate policy, not because I did't have access to spectrum analysers, distortion meters etc., but because I was primarily interested in how it sounded not how it measured.

The pre PI Master Volume influence - look up TUT whichever (maybe someone can fill in a reference) and implement the bootstapped Master Volume circuit. That will address any loading of preceeding circuitry issues.

As the Power Amp starts to work harder(as we turn the volume up) we start operating the tubes over a wider part of their transfer function. That will change the harmonic content of the sound, at higher volumes the odd harmonic / even harmonic content skews more toward the odd. That changes the "timbre" of an instrument, if you equate "timbre" with "tone" that may be significant.

Getting tubes to operate over the same part of their transfer function at different volumes is why the guys who design Powerscaling circuits use tracking bias regulators (to track the anode/screen voltage) in an attempt to keep the tone the same at lower power settings.

There is also power compression in the speaker itself, this is a "design feature" of MI speakers, HiFI speakers exhibit much less of this. In MI speakers, it (primarily) has to do with part of the voice coil extending passed the ends of the magnet on normal voice coil excursions.

Not sure if these are all misses or there are now a few pellet marks in that Barn Door. You can however be sure I attempted to aim high enough to miss my foot.

Cheers,
Ian

[Gingertube]

Daz,
I would need to look at a schematic but if your PI is a "standard" diff splitter" then the joined cathodes sit at 1/2 the input signal in which case those 1M are bootstrapped to 2 M and when you add the 47K in parallel for 44K they are bootsrapped to 88K. Given the Master Volume is a audio taper (log). you won't notice the difference of adding those 47K until say 90 to 95% Master Volume setting.
Cheers,
Ian

[loudthud]

Above there was a suggestion about looking at it with an oscilloscope rather than listening to it - May I suggest that the absolute best piece of test equipment that you are ever likely to posess (at any price) is one set of bog standard ears (and the processing circuitry between them).

OK, post some MP3s.

Daz, have you tried a simple resistor between the tone stack and Master Volume, perhaps 1 Meg?

[NateS]

If it's not already been mentioned, I think one aspect that is really difficult to capture/quantify/emulate is the physical vibration. I have noticed, that working with a seat shaker allows me to get by with vastly lower headphone volume. I'm not a great fretless bass player, but I would get totally lost on intonation playing through headphones, unless I had the shakers going. Just getting that vibration helped not only with intonation, but timing too.

[daz]

OK, post some MP3s.

Daz, have you tried a simple resistor between the tone stack and Master Volume, perhaps 1 Meg?

Yes, in fact today i tried a 470k after reading one of chuck's post which gave me that idea. I was surprised to find very little change of any type. Weird. I have done that before (long ago) and i could swear it made some sort of considerable difference of some sort.
Anyways, i really think this bootstrapped master may be the ticket because it not only seemed to work when i put those 47k's there, but the tone and feel improved overall and the entire feel of the tone circuit changed in a good way.It felt really good. So i wanna try it but the question is, what value? I have a 500k dual but i think thats it. may have a dual 100k but i'll have to look around.

[pdf64]

Sorry, I've not got the TUT books, could someone summarise the rationale behind this bootstrapped master volume idea?
TIA - Pete

[daz]

Sorry, I've not got the TUT books, could someone summarise the rationale behind this bootstrapped master volume idea?
TIA - Pete

I can't, but i can tell you this...it works ! I just tried a single pot on the input side using a 250k. I'll be damned....it retains the tone when turned loud ! Granted, it does get brighter and you need to compensate. But the basic tone and saggy feel are still there. Not as good as lower, but then thats the expected munson curve and room acoustics etc etc. The things i always knew were there, but i also knew were not responsible for the night and day change.

Anyways, problem solved, tho i still have work to do. I need to try the dual pot method because i seem to be losing NFB with the single pot. Tho i would think the dual would be worse in that regard but who knows. But in any case i need some NFB to counter the brightness along with the tone controls. Then get an appropriate pot value which i think is likely to be 100k, plus determine the best taper. I also need to use a grid stopper because when cranked up theres ringing here and there, and i saw that typically this mod IS dome with stoppers which i believe are implemented for that exact reason.. So i will have to get all the details worked out, but bottom like is this is a revelation to me ! Only worry is there seems to be more OD like this and i have to lower the pre gain which i didn't want. I wanted all drive to come from the pre. I can't figure whats causing more OD because i would think the PI would be clean with the grid R low, and the PA shouldn't contribute till it's louder than i ever use it at home.