Channel Isolation Resistors Fender/Marshall - how low can you go?

[TubeGuy]

I'm talking about the dropping/channel isolation resistors commonly found in old Plexis and Fenders, the ones that occur just before the channels "meet up" at the input to the cathode driver (marshall) or phase inverter (Fender).

Marshall uses 470k, Fender 220k, anyone have an idea how low you can drop these before you start to get the channels interacting or other issues? You lose a fair amount of gain with both those resistors.

I'd think you could go a whole lot lower since you're only using 100k typically to isolate your audio signal fom coupling through the power supply, but I've never experimented with it to be sure.

For an example of using 0 isolation:

http://music-electronics-forum.com/a...1&d=1160335874

Thats something that Bruce at Mission posted RE adding effects to both channels on a Fender, and he isn't using any isolation at all. Thats a bit different than how I do it, I'd think that no isolation would cause issues but hey I've never tried it. Maybe issues with the plate voltage supply as well.

Reason I'm asking if anyone cares is I'm pondering making the "normal" channel on a DR into a blues junior. I could get my 3rd stage ala BJ by coupling the two channels back up a stage early at the 3rd stage of the "effects" channel post 3M resistor, but the BJ uses only a 10k dropping resistor before its 3rd stage and I'm not sure that wouldn't cause isolation problems if I do it that way.

BJ also doesnt use a cathode cap on that stage which the DR does, but eh. Thats an easy enough experiment both ways.


Any input would be appreciated. Thanks!

[Enzo]

"Mixing" resistors. The amopunt they attenuate the signal is determined by their relationship to a resistor to ground from that point. Well, an impedance to ground anyway.

Get a pair of small trimmer pots - 220k, 100k, whatever you like - and tack one in place of the mixing resistor you have. Now you can dial them up to the max of the trimmer or down as low as you feel like and get instant feedback on the difference.

How high or low ANY resistor can be all depends upon the circuit it is in and the job it is doing.

The 100k resistor from plate to B+ in an amp stage is not to isolate the signal from the power supply. The decoupling filter cap does that. That resistor is the plate load resistor, and it is there to give the current through the tube something to develope voltage across so you'll have a signal there.

I am not sure I see how mixing resistors would affect the B+ supply.

[TubeGuy]

Mixing resistors yes, sorry for the terminology.

The B+ issue I'm referring to is, if you look at Bruce's schematic, he puts the plate voltage source for both V1 stage 2 and V2 stage 2 in parallel with each other....I'd think that would cause problems.

He also uses no mixing resistors whatsoever, which I'd also think would cause problems.

In the specific case of those Marshall Plexi and Fender BF/SF circuits, have you experimented with those values to know what would work?

RE what you said on the ground reference being the contolling factor, the only ground reference I see from those 220ks in a Fender is the 47 ohm resistor to ground at the phase inverter section. I'd think the 3M would be plenty to isolate ch 1 from Ch2 , issue would be ch1 and the verb return on what I was talking about.

What kind of things am I going to see if I'm not getting enough isolation from those mixing resistors?


Thanks!

[Enzo]

Have I experimented with mix resistor values? No, I never had a reason to. My trimmer pot idea in post #2 is easy to do and in a couple minutes of twiddling you would know far more about their effect that I could reason out in a half hour.

You didn;t specify a circuit, so I am looking at the AB763 Deluxe Reverb drawing. There is a 220k mixing resistor from each channel. They join and through a .001 cap they enter the phase inverter. From that grid is a 1meg, a 22k, and THEN your 47 ohm resistor to ground. SInce the signal passes through the cap, each 220k makes a voltage divider with itself and the 1,022,047 ohms to ground, and really that is close enough to just 1meg that we will call it that.

SO whatever signal we start with will be diminished to 1meg/1.22meg times the original signal - .82 - about an 18% drop. If you changed them to 470k the signal would drop further. The numbers would then be 1meg/1.47meg times the signal - .68 - about a 32% drop. BUT ONLY IN THIS CIRCUIT. You can't look only at the mix resistor value and decide what it does to the signal level without looking at the whole circuit.

The B+ issue I'm referring to is, if you look at Bruce's schematic, he puts the plate voltage source for both V1 stage 2 and V2 stage 2 in parallel with each other....I'd think that would cause problems.

You mean he is connecting the plates together? They are already on the same power supply node. If Bruce says it works, I'd trust that it works. Unless you plan to play through both input channels at the same time, it shouldn;t be an issue.

And in general, I am not sure what kind of "isolation" you need.

[TubeGuy]

I see what you are saying.

Basically what I'm looking for is the channels not to 'interfere' or feedback into each other, or end up having the controls on one affect the other.

Maybe not so much an issue on the Fender circuit since there is a tube stage between the volume/tone controls and the mix point, more of an issue on some other amps where the channel volumes occur just prior to the mix point; if they werent isolated by those mix resistors turning one volume down would directly affect the other channel in an audible way.

I'll try the pot idea and see what works.

Any idea what I'll hear start to happen when I go too far?

I have no doubt that what Bruce is doing "works", in terms of putting both channels through the effects (although they then share the coupling cap). Whether it causes potential issues is more a concern, like if you were to jumper the channels as you described.

Thanks!