balancing a LTPI ?

[12AX7]

It's like a JCM, typical 82k/100k plates with 1M grids and 470R/10k from the cathode. NFB, the whole ball o' wax. Questions...
1-can a balanced PI make for smoother better transition from low volumes on the master up to typical small bar gig volumes? If not, what WILL balancing do?
2-Is it the A/C signal to the PA that you are trying to balance? (Stupid Q i know)
3-to balance it how would you do it? I'm assuming a trimmer at one plate then a constant signal injected and turn it till both power tube grids show the same signal? If not, how?
4-if i were to try and achieve balance with say a 47k resistor in place of the current 10k tail, should i change the 82k plate to 100k?

[HotBluePlates]

3-to balance it how would you do it?

How do you know it's not balanced now?

I'd assume you'd need 2 meters, one for each plate output, and a test signal. Watch both plates to see a roughly-equal a.c. output voltage.

Or you use a dual-trace scope and measure the same (for either method, you could measure at the output tube grids instead of PI plates).

... what WILL balancing do?

Assuming the PI tube sections are matched, the output tubes ae matched under all conditions of operations and your OT primary winding halves are exactly equal, balance in the PI will yield the most clean output power for your particular amp.

Perfect a.c. balance ensures cancellation of even harmonic distortion generated in the output section.

That's pretty much it.

4-if i were to try and achieve balance with say a 47k resistor in place of the current 10k tail, should i change the 82k plate to 100k?

No.

The tail resistor does most of the work in balancing the outputs.

The plate load resistors are purposely different values, because that's what's required to get balanced outputs when you drive only one grid of the long-tail pair. The other grid is seeing a negative feedback input.

Some Vox amps that don't use NFB drive both grids of the long-tail, and use a 100kΩ resistor for both halves.

2-Is it the A/C signal to the PA that you are trying to balance?

This "balanced" is not that "balanced".

The "balanced input" of a PA refers to the impedance to ground from the hot and neutral inputs on the jack.

On an unbalanced guitar amp jack, the hot might have a 1MΩ resistor to ground; the "neutral" contact is directly connected to ground. So the hot and shield of the guitar cord/jack have "unbalanced" impedances to ground.

On a balanced input, the hot and neutral are separate from the grounded shield, and those connections have equal ("balanced") impedances to ground.

Those balanced impedances might be halves on an input transformer winding, equal resistors to ground, or opamp inputs with equal impedances to ground.

[12AX7]

Thanks. And no, i don't know that it's balanced. I just wanted to see what tonal differences i might get if it was. I guess i can't really know just by balancing the PI out. I do have dual bias on the PA if that matters.

[tubeswell]

As a general RoT, as you increase the resistance of the tail resistor, you can generally achieve more balance with the plate resistors on each side. E.g.; 47k tail with 100k Ra each side will yield pretty good balance for a 12AX7 LTP. The trade-off is you lose plate-to-cathode voltage (all other things  - like the HT voltage - being equal), which amounts to reduction in gain.

[SILVERGUN]

Do you have the dual trace scope and 2 probes?
There's no better way than to see it with your own 2 eyes.....

Next thing you know you'll want to have a scope screen showing out of the front of the amp......like me

[HotBluePlates]

... And no, i don't know that it's balanced. I just wanted to see what tonal differences i might get if it was. ...

Re-read what I wrote:

How do you know it's NOT balanced now?

[PRR]

> I just wanted to see what tonal differences i might get if

So lift one end of one plate resistor and stick another 47K in there. That will change the balance, probably somewhat un-balanced.

I would not expect any real difference except at the edge from "clean" to "flavored".

> "balanced input" of a PA

Abbreviations suck. "PA" could be Public Address or Power Amp. I suspect you are not talking the same thing.

[HotBluePlates]

"PA" could be Public Address or Power Amp. I suspect you are not talking the same thing.

Yep, power amp makes more sense on a 2nd reading.

2-Is it the A/C signal to the PA that you are trying to balance?

For a push-pull stage in general, there is a.c. balance and d.c. balance.

A.C. balance includes the things I mentioned earlier: phase inverter output signal balance, output tube gain/Gm balance, output transformer winding balance. A balanced signal into balanced output tubes yields the maximum clean output power. Me on one end of a 2-man saw with 1980's Schwarzenegger on the other end would be unbalanced...

D.C. balance is mostly about idle current of the output tubes.

Output transformer intended for big unbalanced d.c. through them have air gaps and big cores for a given amount of power at some low frequency limit. Push-pull transformers assume no unbalanced (unequal) d.c. in the primary, so they can use no air gap and have smaller cores for the same power at the same low frequency limit.

If the output tube idle current is grossly mismatched, you will hear hum in the output, will likely not get the same full power output at as-low a frequency. If the imbalance is due to radically different bias voltages or Gm for each tube, you will have less clean power output and more distortion on top of the other issues.

Most people make more of the balance thing than it deserves, at least in a guitar amp.

[12AX7]

The trade-off is you lose plate-to-cathode voltage (all other things  - like the HT voltage - being equal), which amounts to reduction in gain.

Thats fine with me ! I feel theres too much as it is, the amp is feakin loud.

[12AX7]

... And no, i don't know that it's balanced. I just wanted to see what tonal differences i might get if it was. ...

Re-read what I wrote:

How do you know it's NOT balanced now?

Right, and i get that. My response is meant to say i don't know if it is or not, but i need to know regardless of which way it is or i won't know if what i'm hearing is a result of a balanced PI or not. Point being, how to go about balancing so i KNOW if it is and then know what a balanced PI sounds like, or if it isn't, balance it to know.

[12AX7]

Oh, and by PA i meant power amp. I just assumed that would be obvious when talking about balancing the PI. Sorry.

[HotBluePlates]

I see.

Measure first with a test signal, using the method I suggested, and find out how balanced (or not) it is now.

If you want to unbalance it without any doubt (and without boogering the phase inverter), replace one of the resistor from grid-to-ground (or grid-to-bias supply) at your output tubes with a volume pot. NOT both sides, like a post-phase inverter master-volume.

Turn that new "volume control" down to about half. You have just seriously unbalanced the a.c. signal driving the output tubes.




My prediction: you'll be under-whelmed by the result.

[spacelabstudio]

It's like a JCM, typical 82k/100k plates with 1M grids and 470R/10k from the cathode. NFB, the whole ball o' wax. Questions...
1-can a balanced PI make for smoother better transition from low volumes on the master up to typical small bar gig volumes?

No.

If not, what WILL balancing do?

Reduce even harmonics.

My question for you is what prompted this line of questioning and what is it you really want?  ;)

My experience is the effect of futzing with the balance in a PI is anywhere from subtle to inaudible.  

Merlin's preamp book does have a 'scale' control that involves using a pot to increase tail resistance of an LTPI, where the goal there is to reduce output signal while still allowing the PI to be overdriven.  It's in the master volume, power scaling class of controls.  Maybe that would be of interest to you.

Chris

[DummyLoad]

the greater the impedance as seen by the cathodes through the tail resistor, the better the balance. with better balance the anode resistors need to be closer in value. at <2% balance they effectively can be the same value.

in some hi-fi plans the tail resistor is replaced with a constant current source. with a constant current source the cathodes see near infinite impedance and balance delta becomes negligible, therefore both anode resistors can be of equal value. to generalize; [i don't like to] the larger the tail R, the closer in value the anode resistor will need to be to maintain system balance.

case in point: doug's 18W stout. it has a mis-wired LTPI. the balance is grossly upset by wiring the non-inverting AC grid coupling cap to the cathode R and tail R node instead of the bottom of the tail. this kicks the balance out significantly and does alter the tone considerably, as in you introduce heavy 2nd order harmonic distortion. in some of my experiments, i've added a switch to move the tail of the cap to take advantage of this tonal characteristic. with the cap switch into the "normal position" (grounded) for clean and more power, and tilted for enhanced distortion. you do lose power, stress the OT (unbalanced DC), but the tone quality outweigh those pitfalls, IMO.

http://www.el34world.com/projects/images/Hoffman%20Stout%2018watt%20No%20Switch%20Schematic.GIF   

notice how the .022 cap is wired from grid (pin 7) of V2-B to 820R and 47K tail node: this is incorrect. for proper balance it should connect from grid to ground, but...it sounds so damn nice that way we leave it mis-wired.
 
--pete

[12AX7]

Stuff like this always interests me, noty to accomplish something i'mlooking to do, but to see if something new and usable comes from it. Can i do this even with a marhall 800 style PI, and if so all i'd do is lift the non grid end of the cap ion the NFB side and attach it to the other side of the 10k tail?

the greater the impedance as seen by the cathodes through the tail resistor, the better the balance. with better balance the anode resistors need to be closer in value. at <2% balance they effectively can be the same value.

in some hi-fi plans the tail resistor is replaced with a constant current source. with a constant current source the cathodes see near infinite impedance and balance delta becomes negligible, therefore both anode resistors can be of equal value. to generalize; [i don't like to] the larger the tail R, the closer in value the anode resistor will need to be to maintain system balance.

case in point: doug's 18W stout. it has a mis-wired LTPI. the balance is grossly upset by wiring the non-inverting AC grid coupling cap to the cathode R and tail R node instead of the bottom of the tail. this kicks the balance out significantly and does alter the tone considerably, as in you introduce heavy 2nd order harmonic distortion. in some of my experiments, i've added a switch to move the tail of the cap to take advantage of this tonal characteristic. with the cap switch into the "normal position" (grounded) for clean and more power, and tilted for enhanced distortion. you do lose power, stress the OT (unbalanced DC), but the tone quality outweigh those pitfalls, IMO.

http://www.el34world.com/projects/images/Hoffman%20Stout%2018watt%20No%20Switch%20Schematic.GIF   

notice how the .022 cap is wired from grid (pin 7) of V2-B to 820R and 47K tail node: this is incorrect. for proper balance it should connect from grid to ground, but...it sounds so damn nice that way we leave it mis-wired.
 
--pete