proper voltage at PI input ?

[Dave H]

There are a ton of schematics showing this kind of MV. Google PPIMV then click images. And many are separated from the bias by the cap so they shouldn't change the bias. And yes, i hate PPIMV'S, but thats if they are the only MV in the amp because then the PI is slammed all the time. I have a pre PIMV, so the one i propose is nothing more that a divider to take the signal down To the point where with the MV on 10 the amp is just past clipping. It would be to set the max gain of the power stage wherever i want it then the pre PIMV controls it.

But your grid to ground resistor is not separated from the bias by a cap! It's DC coupled to the bias voltage through the grid stopper and 220k and will divide down the bias voltage by a lot more than the range of the bias twidde pot. If that's not clear I'll draw a schematic but if you prefer add a 10k from grid to ground and watch your EL34s red plate

Edit: I've just checked some PPIMV schematics and the MV ccw pin is either grounded but isolated from the bias by a cap or the ccw pin is returned to the bias supply not ground as your grid to ground resistor would be.

[daz]

But your grid to ground resistor is not separated from the bias by a cap! It's DC coupled to the bias voltage through the grid stopper and 220k and will divide down the bias voltage by a lot more than the range of the bias twidde pot. If that's not clear I'll draw a schematic but if you prefer add a 10k from grid to ground and watch your EL34s red plate

Edit: I've just checked some PPIMV schematics and the MV ccw pin is either grounded but isolated from the bias by a cap or the ccw pin is returned to the bias supply not ground as your grid to ground resistor would be.

Yes, you're right. But my point is to do the PPIMV mod whether you add a cap or however they do it. The point was you can put a mv or voltage divider there. And i will try it eventually. Possibly today if i have time.

[guitician]

I thank that’s a bad idea. It will mess up the bias voltage and the PI can’t drive a low resistance like that.

I think Dave H thought you were going to use 5k6 to ground, or even smaller, when you said you were going to make a voltage divider from the grid stop resistors. Of, course a PPIMV(fixed) works and is a valid solution to the problem.

[Dave H]

I think Dave H thought you were going to use 5k6 to ground, or even smaller, when you said you were going to make a voltage divider from the grid stop resistors. Of, course a PPIMV(fixed) works and is a valid solution to the problem.

I did but even a 220k grid to ground would divide the bias voltage by two.

[guitician]

You connect it with DC blocking Caps... here's an old thread that shows one http://music-electronics-forum.com/t15137/

[teemuk]

There are several ways how to decrease the open loop gain by attenuation. First, output impedance of the PI stage is directly related to its plate resistor values; therefore you already have one potential voltage divider formed by output impedance of the PI and input impedance of the power tube grid circuit. Note that this divider is subject to ordinary rules of “impedance bridging”: You aim for Zout > Zin and Zin > Zout often loads the source stage too excessively. If tampering with these component values seems to change operating conditions and characteristics too much then yes, you can also change any PPIMV –type solution that works two use fixed resistor values. Third alternative is fitting the attenuator right at the PI plates: “Split” each plate resistor into two resistors and take the output from their interconnecting node. By controlling the ratio of the “split” you can control the output level. Now, read the second sentence of this post again and you’ll see why this attenuation scheme may again alter the overall operation of the power amplifier.

But if the ultimate goal is just decreasing gain to prevent signal peaks from triggering oscillation then I think you may be attempting a worst kind of band-aid for the problem. I would tackle eliminating the oscillation first, whether it requires rethinking of layout and current routing or just local high frequency negative feedback to kill closed loop gain towards higher frequencies. After nailing that issue I’d reconsider if interstage attenuation within the PA is required for best possible “tone”.

[Chuck H]

Hey daz.

Earlier Juan suggested that you try a 100pf cap across the PI plates. Now I know you said that you're not trying to "fix" it, but fixing it is probably the best solution. Don't change the PI plate loads, don't use split plate loads, don't change the feedback arrangement you have and don't change the bias resistors. You already attenuated the master to get the power tubes and PI gain structure in line. Keep that and stabilize the amp. A 47pf to 100pf cap across the PI plates is a very common circuit on many well laid out amps with lower gain than yours. So there's no band aid about it. Consider it part of the Marshall mojo if it makes you feel better. 47pf is the standard Marshall value and you probably won't hear it (other than not hearing the ugly oscillation) and it could explain why the tone is better with the odd PI bias resistor. So put that cap in there and try the PI with both bias resistors again. See how it works out.

[daz]

Hey daz.

Earlier Juan suggested that you try a 100pf cap across the PI plates. Now I know you said that you're not trying to "fix" it, but fixing it is probably the best solution. Don't change the PI plate loads, don't use split plate loads, don't change the feedback arrangement you have and don't change the bias resistors. You already attenuated the master to get the power tubes and PI gain structure in line. Keep that and stabilize the amp. A 47pf to 100pf cap across the PI plates is a very common circuit on many well laid out amps with lower gain than yours. So there's no band aid about it. Consider it part of the Marshall mojo if it makes you feel better. 47pf is the standard Marshall value and you probably won't hear it (other than not hearing the ugly oscillation) and it could explain why the tone is better with the odd PI bias resistor. So put that cap in there and try the PI with both bias resistors again. See how it works out.

Heres the thing Chuck....either the larger bias resistor or the cap across the plates removes that oscillation. But the tone of the amp is quite obviously better with the bias upped than the cap. I tried both ways last nite while playing it in a mix of music for an hour last nite and changed them back and fourth several times. That cap just causes the tone to get lost in the mix and the tone is harsh. With the bias R at 2k the tone is much better and the tone cuts better. So that oscillation is definitely not why i like the amp better with the 2k cathode. It's not a subtle thing either. The amp sounds nasty with the 470R in comparison to the 2k. I wish i understood why, but since the cap removes the oscillation yet sounds harsh with the 470R in place then it's not the lack of oscillation that makes the 2k resistor sound better. I also removed the attenuation at the master because it seemed that had a negative effect too. But i'm going to keep screwing around with that because i tried a PPIMV voltage divider and it didn't sound good. Once i get it attenuated in a way i am happy with then thats as good as i think it's gonna get. I believe the main reason for the tonal change is pretty much impossible to fix because i think it's the fact that the master has no treble bleed cap and when turned up it therefore gets bright. problem there is if i put a cap on it then it gets super bright. I have fought that before by putting a bleed cap there then losing brightness in the preamp, but i completely lost my tone. That may be the fix, but it might be another decade before i nailed that down. LOL! Maybe i'll screw ith that idea again tho. But it's one of those things that i hesitate about because of the fact no other amps do that to a master, and that tells me I'm probably barking up the wrong tree.