Diode||Resistor in signal path

[martin manning]

I'm looking at a case where a diode is paralleled with a 10K resistor after the coupling cap out of each side of a LTP phase splitter. There is a PPI master volume next (signal to wiper, one side grounded), and then another coupling cap before the 6V6 push-pull power amp's grid resistors.

Anyone have an idea of what is going on here? Seems like the diode (which is placed to let the upper half of the waveform through) passes the signal that is wanted, and the resistor attenuates the lower half, but... why do this?

MPM

[Don Moose]

Sounds like somebody was adamant about only pushing or pulling in a PP power amp. Seems silly, but if it sounds good ...

Now that PPIMV, it goes PI->coupling.cap->wiper->pot.top->coupling.cap->power amp, with the other side of the pot grounded?

Badly thought out - this gives the PI a highly variable load depending on the MV setting, and at bottom, shorts it to gnd.

Switch things around so that it goes PI->pot.top->wiper->power amp (coupling caps implied) and see how you like that sound.

Hope this helps!

[martin manning]

Yes, I see your point. Next time I have this thing (Marshall 4001 Studio 15) opened up I'll have to look to see if it really is wired that way. I'm thinking maybe it is since both versions of the schematic are showing it like that. Funny thing though, mine has wires connecting the double-gang MV pot to the board. The other controls are direct-mounted, and I've seen chassis pics where the MV is too. At first I thought someone probably changed the pot, but the insulation on the wires (three colors are used, IIRC) matches the rest of the wiring in the amp exactly. Whatever, that makes it possible to swap the wiper and top connections. It just gets weirder and weirder...

[Steve Conner]

I think those diodes are there to emulate the self-biasing effect you get when overdriving the power tubes into grid current. The PPIMV gets in the way of this effect when it's turned down, so they probably added the diodes to get it back.

[d95err]

Here's the schematic:
http://www.schematicheaven.com/marsh...5_15w_4001.pdf

It seems like the schematic above is correct, the PPIMV is wired "backwards". Wierd!

[Enzo]

There are caps in the way of the grid DC. Just a guess, but might the diodes help keep the charge on the caps from the PI from causing artifacts when the control ssetting is changed?

[martin manning]

When the 1 meg MV is turned way down, the diodes and 10K resistors don't matter much, right? I'd think the biggest effect would be when it's wide open, and the diode eliminates the 10K (give or take the junction resistance) on the positive half of the signal.

To Enzo's point, the DC grid current has to come from the bias supply, but it's the voltage seen on the coupling cap that makes it happen, no? Steve, can you elaborate on the "self-biasing effect" and what it does to the sound? I know that at some point grid current will flow if the bias voltage gets less-negative enough.

D95err, that is the earlier schem, the "artsy" one. There is also a redraw that shows it the same way.

MPM

[Enzo]

The diode would only start conducting when the signal voltage ACROSS the 10k resistor tops half a volt. Seems like that is more likely to happen at low volume settings.

[martin manning]

Originally Posted by Enzo

The diode would only start conducting when the signal voltage ACROSS the 10k resistor tops half a volt. Seems like that is more likely to happen at low volume settings.

Ah. Good point. So, if the thing were hooked up "right," the 10K and 1 meg pot element would form a 100:1 voltage divider, and the diode would probably never see enough voltage drop to conduct, basically looking like an open circuit.

But how about this: Wired the way it is, when the MV is turned all the way down the diodes conduct and ground out the signal. Maybe the idea is to kill the signal completely and avoid having anything coming through when the MV is turned down?

MPM

PS: I am from East Lansing!

[Enzo]

Hi


I doubt that, since there would still be potentially half a volt of signal the diode can't shunt, not to mention the other polarity.

I think it is an odd arrangement.

Once it is running, remove the diodes and see if anythig funny happens.

[Steve Conner]

About the self-biasing effect: You can see this if you test a guitar amp on a signal generator, dummy load and scope. When you overdrive the power stage, it temporarily biases itself colder, lowering its gain and adding crossover distortion. The reason is that the power tube grids start to draw current, which charges up the coupling capacitors. It's like a push-pull version of the blocking distortion you get in preamp stages.

I believe this is an important part of the cranked power amp tone (it's almost like a compressor) and hence Marshall put those diodes there to simulate the effect at low settings of the PPIMV.

[Earl]

Som this amp is cathode and fixed bias?

[Steve Conner]

The effect happens regardless of whether it's cathode or fixed bias.

[d95err]

The Peavey Valveking amps have something remotely similar with two diodes working as a very primitive noise gate (see diodes D101 and D102):
http://www.schematicheaven.com/newam...ing_50_112.pdf

I wonder if these diodes could have a similar purpose (although I can't figure out how)?

[Earl]

Steve - Sorry, I was just making a comment about the dual bias set-up. Don't see that very often.

d95err - I think that 'noise gate' deal is called "corring", as it cores out the center section of the signal.