Well, lets see what you all think about this one

[12AX7]

I have endlessly tweaked my marshall style homebrew forever. I will leave it alone for months, almost a year at times, then get the bug and start trying new things. And I almost never find anything that i immediately find to be a great improvement. Today i seemed to have found a somewhat odd one. Odd in that i never see this done....at least not till i noticed it on a marshall schematic. They generally have 1M grid leaks on both sides of the PI. This one i was looking at showed a 330k on the input side and a 120k on the NFB side. Seemed odd to me but i had to try it. It sounds great. It gave the tone a clarity, fullness, and more natural or organic cleanness to it. To a degree that i felt so good about i changed out the 1 M's for those values which normally i'd never do, but wait till i felt i was sure. (i tried it by paralleling resistors so i didn't have to unsolder the 1M's)

Any thoughts on why this is such a little used value on other LTP's? The unbalanced nature of it i suspect has something to do with it, as even when i do see less than 1M it's usually the same value on both sides unlike this one. Also, not being a tech or having much theory, any thoughts on what exactly this does? I assume it just drops the gain on both sides, one obviously more than the other. Or is that over simplified?

[HotBluePlates]

It doesn't drop the gain, it changes the apparent input impedance on each side of the PI.

The reduced input input will roll off bass a little more than the stock 1MΩ resistors.

Each side is really a much higher input impedance than you'd think, due to the bootstrapping effect of returning the "ground side" of the resistors to the top of the tail resistor.

[12AX7]

If it rolls off bass it's not apparent. Not saying you aren't right, but just that unlike other things i've done that roll off bass, the tone doesn't sound at all thinner or less fat/full, something that usually happens when dropping coupler values and other such tone bass reduction tweaks. Maybe because i'm usually doing those types of things in the gain stages where you're dealing with shaping distortion. Whatever it is tho it doesn't sound at all less full or bassy, just clearer and purer. Thats the best i can describe it. Less garbage in the tone too maybe. Still wonder tho why since i got this nice change which isn't such a subjective thing (because the change is in ways i think most anyone would prefer) that you almost always see the same value on both sides, and usually 1M.

[HotBluePlates]

If it rolls off bass it's not apparent. Not saying you aren't right, but just that unlike other things i've done that roll off bass, the tone doesn't sound at all thinner or less fat/full ...

I didn't say it would be thinner or less full, just that there is more roll-off than stock.

Ever wonder why you might not like going below 0.022uF for a preamp coupling cap, but the blackface Fenders often use a 0.001uF leading into the phase inverter?

It's because due to bootstrapping a 1MΩ resistor behaves as though it was a 5MΩ or bigger.

So you can drop the value of those grid reference resistors, and even with a smallish coupling cap, the bass roll-off is still well below the guitar's range.

A real analysis is beyond me at this time of night, but since you're also making changes inside the feedback loop, there has got to be some amount of interaction there.

[phsyconoodler]

What Marshall schematic did you see this on?

[12AX7]

The 2205, it's the JCM800 channel switching head. Oddly, if you look at the channel switching 1x12 combo (4210)the schematic is very different in a number of ways including the combo shows the typical PI values. By the way, any reason why I should follow the rest of the PI values by changing the 10k tail to the 22k shown there? Just wondering if theres any reason changing the grid leaks would then make the 22k a better choice.

[tubenit]

They generally have 1M grid leaks on both sides of the LTPI. This one i was looking at showed a 330k on the input side and a 120k on the NFB side.

The 2205, it's the JCM800 channel switching head

http://www.el34world.com/charts/Schematics/files/marshall/jcm800_splitch_50w_2205.pdf

[Geezer]

Here's a video demo of that amp >
 http://m.youtube.com/#/watch?v=7wTLFB-QgxA&desktop_uri=%2Fwatch%3Fv%3D7wTLFB-QgxA

Very RUSH/Lifeson sounding to me at the lower gain settings.....a favorite tone from my youth. The metal tones, not so much. 

The schem is a bit bizarre to me....is IC1 used for channel switching? And it seems that v2 is the "boost" and it is simply added in parallel to the "normal" or clean signal, correct?

I've thought about trying that before (parallel OD) imagining that it would leave a bit of clarity and transparency to the OD.

Geezer

[12AX7]

The schem is a bit bizarre to me

It's VERY bizarre to me. Whats really odd besides the fact most of the circuitry is unlike any marshall is that it appears to be so different than the 1x12 channel switching combo it's nite and day. They look like they could be 2 different manufacturers, yet they are both from the same series and are supposed to be head and combo versions of the same amp. Or at least thats what i always thought and how it usually works.

[12AX7]

I was reading a post elsewhere in which a guy was asking how to clean up a PI with changing R values. Someone said to drop the 470R cathode resistor to 270R ala a silverface twin. Can someone explain this to me....is the PI different in how the cathode resistor affects gain that in a regular gain stage? Because if you drop the cathode resistor in a typical preamp gain stage it INCREASES gain, and if someone wants a cleaner PI i would associate that with lower gain. What am i missing?

[Willabe]

I'm just guessing but it might alow a bigger voltage swing on the PI plates?


                Brad