Maximizing headroom in an LTP PI

[roberto]

It's feeding a push-pull stage so filtering probably isn't as critical as you'd think.

Hi Cliff, I was thinking about variation in the Vdc more than noise rejection.

Thanks for all the answers The bias method sounds interesting. By "raw" bias, do you mean connecting to the bias supply before the pot?

Yes.

I don't doubt it works, but won't the negative voltage cause an extremely cold bias that could affect the sound?

You have same plate voltage and the 22k connected to the -50V bias, so you virtually raise the supply of the PI from 300V to 350V, raising its bias and headroom. I suggest you to come back to a 12ax7 (Sovtek LPS sounds pretty good on PI).

use the silverface PI with 12at7, 47k plate resistors, 270k cathode, 15k tail, 330Ks, and around 300v at the plates of the 12at7

Hi Roe, I think you mean 270R, but with a PI voltage of 300V and around 220V on preamp, then you take around 20mA for the PI you'll have 260V on PI and 180V on preamp. You'll obtain the opposite result.

[Cliff Schecht]

It's feeding a push-pull stage so filtering probably isn't as critical as you'd think.

Hi Cliff, I was thinking about variation in the Vdc more than noise rejection.

This shouldn't be an issue either, this DC is more stable than the B+ typically because of sag. Also as the wall voltage drifts all of the voltages in the amp track nicely. I don't see the problem..

[roberto]

Maybe just my impressions, but I feel the amp sounds better if there is more separation from PI to PA supply. I always use more than 4k7.

[VacuumVoodoo]

Just connect the 22k tail resistor tot the negative raw bias voltage, instead of ground.

Very interesting, how would you implement something like this when there is GNFB connected to the 22k? Any side effects besides increased headroom?

Increased headroom is not a side effect, it is the intended effect.
You can apply GNFB two ways:
1. The correct by the book way i.e. directly to the "other PI input" which is at ground DC potential by design.
2. The usual guitar amp way but you need to block the DC from the feedback path.

Maybe just my impressions, but I feel the amp sounds better if there is more separation from PI to PA supply. I always use more than 4k7.

Put a current sink in place of the tail resistor, apply feeback (if you need to) as per 1. above. It will do the trick.

[Roe]

...

use the silverface PI with 12at7, 47k plate resistors, 270k cathode, 15k tail, 330Ks, and around 300v at the plates of the 12at7

Hi Roe, I think you mean 270R, but with a PI voltage of 300V and around 220V on preamp, then you take around 20mA for the PI you'll have 260V on PI and 180V on preamp. You'll obtain the opposite result.

yes a 270R kathode . Could you elaborate on the voltages? I'm running 300v anode voltage in the PI and around 250v at v1 with good results

[erwin_ve]

Just connect the 22k tail resistor tot the negative raw bias voltage, instead of ground.

Very interesting, how would you implement something like this when there is GNFB connected to the 22k? Any side effects besides increased headroom?

Increased headroom is not a side effect, it is the intended effect.
You can apply GNFB two ways:
1. The correct by the book way i.e. directly to the "other PI input" which is at ground DC potential by design.
2. The usual guitar amp way but you need to block the DC from the feedback path.

Thanks Mr Niemand for your response; Side effect, off course; increased headroom is not a side effect; my bad:-) But do you notice side effects: the PA and PI are somehow influencing each other when sharing the bias.

I'm looking for option 2. What is your approach when putting in a cap there to block dc; make the cap as large as chosen for a presence control or do you use a different approach?

[Lynxtrap]

] You have same plate voltage and the 22k connected to the -50V bias, so you virtually raise the supply of the PI from 300V to 350V, raising its bias and headroom. I suggest you to come back to a 12ax7 (Sovtek LPS sounds pretty good on PI).

Right, when you think of it that way it doesn't seem so extreme
I don't know about 12AX7. As far as I've learned and tested, AX7 gives the opposite result (less headroom, less swing).

By the way, does anyone think that mixing the channels before the PI and using one grid only as input, would change anything?

[rdjones]

] You have same plate voltage and the 22k connected to the -50V bias, so you virtually raise the supply of the PI from 300V to 350V, raising its bias and headroom. I suggest you to come back to a 12ax7 (Sovtek LPS sounds pretty good on PI).

Right, when you think of it that way it doesn't seem so extreme
I don't know about 12AX7. As far as I've learned and tested, AX7 gives the opposite result (less headroom, less swing).

By the way, does anyone think that mixing the channels before the PI and using one grid only as input, would change anything?

Of course it will which is what I've been hinting at since the first thread.
Then you can adjust the plate loads for balance and ground the 'second' grid (through a cap) without relying on the driver's plate load for your ground reference path.

All tube based inverters have some degree of natural imbalance.
If you want an ideal balanced drive to the output tubes, use a transformer, but you may not care for the sound.
It's possible to have a well balanced LTP by selecting a matched section tube, and by fine tuning the plate loads.
With a decently matched 12AT7 and standard 82k/100k loads you can get pretty close.

Input a steady state test tone and measure your AC voltages at the 6L6 grids to see how close they are.
(A scope would show how clean they are or aren't.)

rd

[Lynxtrap]

By the way, does anyone think that mixing the channels before the PI and using one grid only as input, would change anything?

Of course it will which is what I've been hinting at since the first thread.
Then you can adjust the plate loads for balance and ground the 'second' grid (through a cap) without relying on the driver's plate load for your ground reference path.

All tube based inverters have some degree of natural imbalance.
If you want an ideal balanced drive to the output tubes, use a transformer, but you may not care for the sound.
It's possible to have a well balanced LTP by selecting a matched section tube, and by fine tuning the plate loads.
With a decently matched 12AT7 and standard 82k/100k loads you can get pretty close.

rd

Yes, I know you've said it, but maybe I am too stupid to understand it I guess I'd like to know the reasons.

The way I understand it, both grids have ground reference to the tail resistor (or reference to a minus voltage with Alex's solution). At least, both halves of the PI work under similar conditions, so I don't see why the balance would be off.

I've also come to understand that the difference in plate load is kind of hypothetical, and many designers simply use identical resistors on both halves.

And how does imbalance have an effect on headroom? I can see how it would affect swing to some degree, but not so much that it would have a huge impact on the sound in a guitar amp.

As far as I know, this setup work in Vox AC30-50 amps for one example, also without NFB.