Q about PI trimmer design

[greiswig]

Okay, this may be a dumb question, but you've come to expect that from me.

The way most people do the PI trimmer is to put it with the wiper attached to the HV supply, and the other electrodes going to the 100k-ish resistors feeding Pin1 and Pin2.

Since my board wasn't designed with that trimmer in mind, there aren't turrets to mount to. So on the feedback side, I put a lower resistor value in in series with a trimmer acting as a variable resistor. Part of my goal was to size the trimmer and resistor such that I'd have better "fine tuning" control, so I could find the elusive sweet spot. That I'm still not convinced I can find.

I can't see any real drawbacks to doing it the way I did, and they seem similar in principle, but...gotta ask. Am I doin' it wrong?

[Structo]

I can't say if that will work or not but doesn't the Dumble PI trimmer act as a voltage divider between the two plate resistors?

Here is how I addressed installing the PI trimmer.

I drilled three new holes and installed eyelets in them.
I can't remember how I crimped the back of the eyelet.
Possibly with pliers or something, but I know I didn't completely pull the board.

110K/ 120K , 10K trimmer

Hope this helps.

[vibratoking]

I don't think that your modification is the same as the original implementation.

My understanding of the PI trimmer is that it adjusts the relative gain between one side of the PI and other.

If I understand your implementation correctly, I might not, it seems to me you are changing the amount of feedback which is applied from the output of the tranny to the input of the PI? If so, this is very different.

[greiswig]

I don't think that your modification is the same as the original implementation.

My understanding of the PI trimmer is that it adjusts the relative gain between one side of the PI and other.

If I understand your implementation correctly, I might not, it seems to me you are changing the amount of feedback which is applied from the output of the tranny to the input of the PI? If so, this is very different.

...but I thought gain was being adjusted by adjusting the plate voltage. Whether you do that with a divider or via varying the series resistance on only one side, the effect is the same, isn't it?

[vibratoking]

Maybe I misunderstood. I re-read your post and drew this up. Is this what you implemented?

[ChrisM]

Add a variable resistor in series with R6 then you have what HAD did. But then your using two pots wired as variable resistor, seems silly. Just use one pot and wire it like HAD did.

[greiswig]

Maybe I misunderstood. I re-read your post and drew this up. Is this what you implemented?

Bingo, that's the idea. It's just that it was easier to do this way given the way my board is laid out. I don't see how this way is effectively different from running the B+ into the wiper and then out through either side, but I've been wrong many times.

[ChrisM]

Maybe I misunderstood. I re-read your post and drew this up. Is this what you implemented?

Bingo, that's the idea. It's just that it was easier to do this way given the way my board is laid out. I don't see how this way is effectively different from running the B+ into the wiper and then out through either side, but I've been wrong many times.

See my post above yours.

[greiswig]

Chris,

I did see your post. I don't understand how the two are really very different: they both change the plate voltages relative to one another. The original D'Lite had no trimmer, just resistors of differing values to get the feedback side a few volts higher. How is my method different except that it is variable?

The more I write about this the more I convince myself of the answer to my own question... but I'm still open to being shown the real difference.

[Bob-I]

The more I write about this the more I convince myself of the answer to my own question... but I'm still open to being shown the real difference.

You're missing the point of the trimmer. It's not to adjust plate voltage, it's to adjust the gain of the 2 sides of the PI. Sometimes they're closely related which caused some folks confusion, but if the PI tube is not closely balanced the plate voltage may be quite different on the 2 plates.

That said, your method may work just fine.

[greiswig]

You're missing the point of the trimmer. It's not to adjust plate voltage, it's to adjust the gain of the 2 sides of the PI. Sometimes they're closely related which caused some folks confusion, but if the PI tube is not closely balanced the plate voltage may be quite different on the 2 plates.

That said, your method may work just fine.

Okay...but how is the gain adjusted with the standard trimmer?

I'm confused about something else, too, now. As I understand it, you're supposed to adjust the trimmer so there is higher voltage on the feedback side of the inverter. But as I look at the schematics for #124, for example, it looks like the higher resistor value, even with the trimmer, goes to the feedback side. Higher resistance should mean less voltage, shouldn't it?

So on the attached schematic, where is it that I should be reading "about 10V higher for the feedback side?

[martin manning]

The input (inverting) side, pin 6 in your schematic, should have the lower plate load resistance and higher voltage. Changing the plate load changes the gain, so the trimmer is just trading plate load resistance from one triode to the other to balance the gain of the two sides. If you put the trimmer on one side only, then you are only adding or subtracting resistance from one side. That will be less effective unless you use a bigger pot, and the average gain of the PI will be varying too.

[greiswig]

The input (inverting) side, pin 6 in your schematic, should have the lower plate load resistance and higher voltage. Changing the plate load changes the gain, so the trimmer is just trading plate load resistance from one triode to the other to balance the gain of the two sides. If you put the trimmer on one side only, then you are only adding or subtracting resistance from one side. That will be less effective unless you use a bigger pot, and the average gain of the PI will be varying too.

Martin, thanks for this. The point about the average gain changing is one I hadn't thought of. I'm not sure how critical this is, though: changing the series resistance on the side with the trimmer still manages to change the voltage on the other side. Still, I think I'm convinced that I need to just do it the "standard" way and judge for myself.

I'm now really puzzled by Tony's #124 layout that I got from the files section here, where the notes on the side say "on the order of 280VDC for the non-feedback side of the phase inverter plate, and about 10 volts higher for the feedback side." I think that is my source of confusion about which side should be higher. Maybe I'm confused as to which is the "feedback side," but to me that looks like V3a above. So can someone please help me understand where the extra gain is needed?

Since it's negative feedback, the side that gets the NFB signal (V3a in the schematic above) should have overall lower gain, correct? So that should be the side that needs extra gain? Now, if more gain associated with more voltage on the plate, that would lead me to the conclusion that Pin 1 (V3a plate) is where the higher voltage needs to be. But that's the opposite of what you, Martin, said. So where am I wrong?

[phsyconoodler]

An imbalance in the PI can sound quite good. 'Exact' balance is not always the best when it comes to tone.That is why Dumble used a trimmer so he could fine tune for TONE not necessarily perfect balance.
Remember a guitar amp is not having to reproduce a very wide frequency spectrum like a home stereo so precision is almost a waste of your time.
Adjust the trimmer til it sounds good to your ears.
The imbalance from negative feedback does not always need to be balanced.Try it for yourself:use two 100k plate resistors instead of a 100k and an 82k.It may or may not sound 'right'.You have to realize that tubes are not going to react like a mathematical equations.That's the real beauty of a tube amp.

[greiswig]

An imbalance in the PI can sound quite good. 'Exact' balance is not always the best when it comes to tone.

I realize that, but my question isn't about finding exact balance. Either Tony's description on his #124 layout is wrong or I'm not understanding the relationship between plate voltage and gain, or I'm not understanding the way GNFB affects gain on the one side of the PI. Then there is also the question I started with, which is whether a single series variable resistor on one side accomplishes the same thing as the way everyone else implements the trimmer.