PI trimmer and tube mismatch

[greiswig]

Okay, I managed to find a 91k resistor (only a half-watter..hopefully that'll be fine), and a 10k/10 turn trimmer I'd forgotten I got from a salvage place locally. So I re-wired that section of the amp.

I think I found the spot, using that "Tele in the middle" on the clean channel method. But it's pretty subtle. It seemed to me that getting the AC balance the same was about 90% of the issue. Then when I got a spot where the feedback was most pleasant (i.e. the octaves, not a strange multiple of the frequency). There were a couple of spots where it seemed like the wrong harmonics were being accentuated, so I'd get a "funny" note hanging over things. So I'd move it back to the where the harmonics sounded better. I left the trimmer there and went back to the 1kHz signal test to see how far off I'd adjusted things to get to this point.

I'd started with the AC identical at output, and ended up with the sweet spot at .02V difference.

Soundwise, the amp seems more alive with this setting. It is most noticeable when you hit a chord with just 1 and 5s in it on the OD channel with a bridge pickup. The sustained chord has this organ-like quality to it, with the harmonics kind of swirling around as the chord rings. The 5th harmonics are there in more abundance than before. The phenomenon I think people are calling "note flipping" is happening more easily now.

The amp also just feels more muscular: the low end feels tighter, but maybe a bit more abundant. It still is hard to coax it into feedback at low volume levels, but it does it more easily now for sure.

I would still love to have someone who knows how show me what to listen for, to make sure I'm not missing something...maybe at Portland Dumblefest next month. The trimmer is mounted to the back of the chassis so it doesn't have to be removed to adjust it. For example, the way Gary was describing it early, it almost sounded like it didn't matter which side of balanced you adjusted it (using the spectrum analyzer), there was a spot there where the magic happens. Is that really true, or is it only going to really happen on one side, and do you know which one? The other thing about Gary's post was that it implied that I should be avoiding the first harmonic, the octave up. The way I did it, I almost wonder if what I ended up with was exactly that spot where it was maximized, not minimized.

But this seems like a really worthwhile thing, given that even just getting it AC balanced made it sound considerably better.

Thanks again, all. Really helpful. Hope this helps somebody else, too. [/i]

[bluesfendermanblues]

Okay, I managed to find a 91k resistor (only a half-watter..hopefully that'll be fine), and a 10k/10 turn trimmer I'd forgotten I got from a salvage place locally. So I re-wired that section of the amp.

I think I found the spot, using that "Tele in the middle" on the clean channel method. But it's pretty subtle. It seemed to me that getting the AC balance the same was about 90% of the issue. Then when I got a spot where the feedback was most pleasant (i.e. the octaves, not a strange multiple of the frequency). There were a couple of spots where it seemed like the wrong harmonics were being accentuated, so I'd get a "funny" note hanging over things. So I'd move it back to the where the harmonics sounded better. I left the trimmer there and went back to the 1kHz signal test to see how far off I'd adjusted things to get to this point.

I'd started with the AC identical at output, and ended up with the sweet spot at .02V difference.

Soundwise, the amp seems more alive with this setting. It is most noticeable when you hit a chord with just 1 and 5s in it on the OD channel with a bridge pickup. The sustained chord has this organ-like quality to it, with the harmonics kind of swirling around as the chord rings. The 5th harmonics are there in more abundance than before.

The amp also just feels more muscular: the low end feels tighter, but maybe a bit more abundant. It still is hard to coax it into feedback at low volume levels, but it does it more easily now for sure.

I would still love to have someone who knows how show me what to listen for, to make sure I'm not missing something...maybe at Portland Dumblefest next month. The trimmer is mounted to the back of the chassis so it doesn't have to be removed to adjust it. For example, the way Gary was describing it early, it almost sounded like it didn't matter which side of balanced you adjusted it (using the spectrum analyzer), there was a spot there where the magic happens. Is that really true, or is it only going to really happen on one side, and do you know which one? The other thing about Gary's post was that it implied that I should be avoiding the first harmonic, the octave up. The way I did it, I almost wonder if what I ended up with was exactly that spot where it was maximized, not minimized.

But this seems like a really worthwhile thing, given that even just getting it AC balanced made it sound considerably better.

Thanks again, all. Really helpful. Hope this helps somebody else, too. [/i]

Hi G,

Great post about your findings.

We have to use both the analytic /technic approach AND fine tuning by ear.

Since the amp is build like this

Preamp V1 => OD V2 => PI V3 => Output tubes => Output transformer

Balancing PI/V3, perfectly in regards to measured AC voltage, will only take you part of the way to feedback nirvana.

Output tubes are never fully matched/balanced and neither are OT's

Hence, you have to fine tune the trimmer by ear in order to compensate for pwr tubes and OT.

This takes the voodoo out of PI Matching.

[glasman]

For example, the way Gary was describing it early, it almost sounded like it didn't matter which side of balanced you adjusted it (using the spectrum analyzer), there was a spot there where the magic happens. Is that really true, or is it only going to really happen on one side, and do you know which one? The other thing about Gary's post was that it implied that I should be avoiding the first harmonic, the octave up. The way I did it, I almost wonder if what I ended up with was exactly that spot where it was maximized, not minimized.

The first harmonic increases on either side of the sweet spot. It never goes to a max and then decreases again.

The delta in AC balance you saw is probably due that measuring the output of the PI does not take into consideration the output tubes, bias level, feedback loop or transformer. All of these add into the overall equation.

Bottom line, this adjustment sets the THD of the power amp. By the nature of its design a push pull power amp stage should eliminate the even order harmonics, the trimmer helps to zero out the variables.

[glasman]

Bottom line, this adjustment sets the THD of the power amp. By the nature of its design a push pull power amp stage should eliminate the even order harmonics, the trimmer helps to zero out the variables.

I didn't state this as clearly as I should have... by the nature of the push pull power amp it should not generate even order harmonics.......

[greiswig]

Thanks, Gary. So the more I hear this, the more I think that I may be setting the trimmer to emphasize the first harmonic (octave up from a played string), rather than finding the sweet spot between a pair of peaks that emphasize the first harmonic. But the reason I did so is that this was that subjective point where the guitar seemed to feed back most easily. It kind of makes sense, given that the first harmonic is going to be the easiest to excite, isn't it?

Is the attached crude graphic showing what you look for?

So a couple of more questions:

1. Does this sweet spot lie on one side of the AC balance point only or the other? In other words, do you want to start adjusting away from AC balance by making the feedback side go higher? Or is it that you'll find a spot going in either direction? (guess that's what I'd expect) Or is it that you don't know which side it will be on, because it is dependent on the OPT, but once you find it you'll know for that amp which way you need to adjust away from AC balance?

2. Could a guy use your method, Gary, with something like TrueRTA and just use a microphone on the speaker, honing in on that narrow bandwidth and watching the amplitude changes, then seeing if you can catch the spot between the two peaks?

[glasman]

Thanks, Gary. So the more I hear this, the more I think that I may be setting the trimmer to emphasize the first harmonic (octave up from a played string), rather than finding the sweet spot between a pair of peaks that emphasize the first harmonic. But the reason I did so is that this was that subjective point where the guitar seemed to feed back most easily. It kind of makes sense, given that the first harmonic is going to be the easiest to excite, isn't it?

Is the attached crude graphic showing what you look for?

So a couple of more questions:

1. Does this sweet spot lie on one side of the AC balance point only or the other? In other words, do you want to start adjusting away from AC balance by making the feedback side go higher? Or is it that you'll find a spot going in either direction? (guess that's what I'd expect) Or is it that you don't know which side it will be on, because it is dependent on the OPT, but once you find it you'll know for that amp which way you need to adjust away from AC balance?

2. Could a guy use your method, Gary, with something like TrueRTA and just use a microphone on the speaker, honing in on that narrow bandwidth and watching the amplitude changes, then seeing if you can catch the spot between the two peaks?

The graph would look like a deep V, and not like Dolly Parton lying on her back .

There is minima point (sweet spot), the first harmonic grows and continues to grow (within the range of the trimmer) on either side of this minima point there is no peak on the first harmonic (you would need a tuned circuit to get that effect). The minima point corrresponds to the point of lowest total harmonic distortion, this is what you are attempting to acheive. All of the even harmonics increase on either side of the minima point. As I remember, someone (maybe Max) posted a comment that HAD recommended using Ampeq procedure for adjusting the SVT. The approach is almost identical to what I use.

As far as using TrueRTA, it should work, BUT you would need to measure at the output to the speaker. Using a microphone would introduce too many variable (room, mic, preamp etc etc etc).

You would better served to build a small interface box that would allow you to hook your sound card directly to the electronics to measure the spectrum. I used this approach for a couple of years before I got my first real time analyzer.


You would need

1. Signal out to amp from sound card
2. Reference signal to the left sound card input, this should be sampled at the input to the PI.
3. Output from amp (I like the FB resistor) to the right sound card input.

You will need a sound card that is full duplex capable and the interface box must have some form of attenuation (potentiomenter 1M works) and a cap to keep any DC off of the sound card.

I used to use SiaSoft for this application (original bought to use with my PA system).

Gary

[greiswig]

Gary, muchas gracias! Okay, so rather than Dolly, it should look more like this?

Also, you didn't answer this little group of questions, and I'd be curious as to your response:

1. Does this sweet spot lie on one side of the AC balance point only or the other? In other words, do you want to start adjusting away from AC balance by making the feedback side go higher? Or is it that you'll find a spot going in either direction? (guess that's what I'd expect) Or is it that you don't know which side it will be on, because it is dependent on the OPT, but once you find it you'll know for that amp which way you need to adjust away from AC balance?

Knowing the answer to this might help a guy know when to stop looking for the sweet spot if he's on the wrong side of the AC balance point. Unless there's a sweet spot on both sides...you get the drift.

Thanks again!

[talbany]

The PI balance, tranny balance and output tube transconductance and NFB and reflected impedance all effect the harmonic distortion of the output stage. There is a direct relationship of the PI and the output tube matching as well so as your tubes age (become mis matched) you begin to drift away from the preset operating point so it's never exactly perfect for long (depending on the hours of coarse)..Consider also wall voltage shifts and varying amounts of NFB and speaker configuration..Thought I'd mention
Carry a variac,add seperate bias pots use the same speaker cab and don't touch the presence control.

Tony

[greiswig]

The PI balance, tranny balance and output tube transconductance and NFB and reflected impedance all effect the harmonic distortion of the output stage. There is a direct relationship of the PI and the output tube matching as well so as your tubes age (become mis matched) you begin to drift away from the preset operating point so it's never exactly perfect for long (depending on the hours of coarse)..Consider also wall voltage shifts and varying amounts of NFB and speaker configuration..Thought I'd mention
Carry a variac,add seperate bias pots use the same speaker cab and don't touch the presence control.

Tony

...or set to AC balance and call it good enough, not worrying about the extra 9% of harmonics that you'd gain by finding the sweet spot? That's what I gather some people do.

[talbany]

The PI balance, tranny balance and output tube transconductance and NFB and reflected impedance all effect the harmonic distortion of the output stage. There is a direct relationship of the PI and the output tube matching as well so as your tubes age (become mis matched) you begin to drift away from the preset operating point so it's never exactly perfect for long (depending on the hours of coarse)..Consider also wall voltage shifts and varying amounts of NFB and speaker configuration..Thought I'd mention
Carry a variac,add seperate bias pots use the same speaker cab and don't touch the presence control. :D

Tony

...or set to AC balance and call it good enough, not worrying about the extra 9% of harmonics that you'd gain by finding the sweet spot? That's what I gather some people do.

IMHO.. nothing wrong with injecting the 1k signal into the amp doing the analysis and figuring out how and why the balance trimmer works.. I encourage it and also intrigues me (This is a wonderful discussion)..It does confirm that the output section is clipping evenly at the 1K with an analyzer attached however we don't play at just 1K so even the spectrum analyzed method gets you close but never perfect.. By the time you add in the real world elements it all changes somewhat.. If you know what to listen for and can adjust it by ear with your current setup and settings you use, this always to me seems to be the most logical method ..I would file this under impedance matching rules as more of a guideline than an exact never changing science..

Tony

[Dr d]

Tony, would it be fair to say that you could decide on a couple of notes/range of notes and dial in in the feedback effect just for them?

[talbany]

Tony, would it be fair to say that you could decide on a couple of notes/range of notes and dial in in the feedback effect just for them?

Doc..
I find certain notes that feedback during my tuning can very somewhat amp to amp and guitar, (too many variables).. In some amps the feedback might not appear as the most prominent anomaly like most people here mention.. If it does it generally reveals itself in strumming an open string chord lightly in the cowboy position and listen for the bloom to start on the low end..Where I can hear it most consistently is on the low frequency's and G notices a tighter bottom and the amp becomes a bit more muscular as I too hear this and try to focus in on that..If you get that dialed in the bloom will usually follow..At this point some amps will bloom nicely others not so much depending on other factors PI tube gain possibly amount of NFB etc..Sometimes replacing the PI tube and recalibrate will help with the bloom factor but is no guarantee..

BTW.. I've had amps that have a subtle bloom on the bench and take them out to a gig or other rooms and bloom nicely

Hope this helps!!

Tony

[greiswig]

Can anyone address this?

, and I'd be curious as to your response:

1. Does this sweet spot lie on one side of the AC balance point only or the other? In other words, do you want to start adjusting away from AC balance by making the feedback side go higher? Or is it that you'll find a spot going in either direction? (guess that's what I'd expect) Or is it that you don't know which side it will be on, because it is dependent on the OPT, but once you find it you'll know for that amp which way you need to adjust away from AC balance?

Knowing the answer to this might help a guy know when to stop looking for the sweet spot if he's on the wrong side of the AC balance point. Unless there's a sweet spot on both sides...you get the drift.

Thanks again!

[/i]

[greiswig]

Bueller? Bueller?

Can anyone address this?

, and I'd be curious as to your response:

1. Does this sweet spot lie on one side of the AC balance point only or the other? In other words, do you want to start adjusting away from AC balance by making the feedback side go higher? Or is it that you'll find a spot going in either direction? (guess that's what I'd expect) Or is it that you don't know which side it will be on, because it is dependent on the OPT, but once you find it you'll know for that amp which way you need to adjust away from AC balance?

Knowing the answer to this might help a guy know when to stop looking for the sweet spot if he's on the wrong side of the AC balance point. Unless there's a sweet spot on both sides...you get the drift.

Thanks again!

[/i]

[BobW]

The PI balance, tranny balance and output tube transconductance and NFB and reflected impedance all effect the harmonic distortion of the output stage. There is a direct relationship of the PI and the output tube matching as well so as your tubes age (become mis matched) you begin to drift away from the preset operating point so it's never exactly perfect for long (depending on the hours of coarse)..Consider also wall voltage shifts and varying amounts of NFB and speaker configuration..Thought I'd mention
Carry a variac,add seperate bias pots use the same speaker cab and don't touch the presence control.

Tony

...or set to AC balance and call it good enough, not worrying about the extra 9% of harmonics that you'd gain by finding the sweet spot? That's what I gather some people do.

IMHO.. nothing wrong with injecting the 1k signal into the amp doing the analysis and figuring out how and why the balance trimmer works.. I encourage it and also intrigues me (This is a wonderful discussion)..It does confirm that the output section is clipping evenly at the 1K with an analyzer attached however we don't play at just 1K so even the spectrum analyzed method gets you close but never perfect.. By the time you add in the real world elements it all changes somewhat.. If you know what to listen for and can adjust it by ear with your current setup and settings you use, this always to me seems to be the most logical method ..I would file this under impedance matching rules as more of a guideline than an exact never changing science..

Tony

I agree Tony, and nothing wrong with trying out a few test frequencies or even a sweep to find an average sweet spot.