Calculating resistor value for negative feedback mod.

[greengriff]

I'm interested in making a switchable negative feedback mod for my Marshall JVM410 so that I can toggle between stock, no negative feedback at all, and the amount of feedback employed by the 2203 master volume. A permanent negative feedback mod has been done countless times by others but I wonder if their calculations are off? The JVM comes as stock with an 82k resistor connected to the 16ohm tap, whereas the 2203 has a 100k resistor connected to the 4ohm tap. The suggested mod by all the JVM guys is to change out the stock resistor for one valued at 176k, but is this right? Unless I'm wrong (which someone here will surely tell me quickly enough) the voltage at the 16ohm tap is double that at the 4ohm tap? In which case you would need a 200k resistor at the 16ohm tap to deliver the same amount of negative feedback as that of a 100k resistor at the 4ohm tap?

Now I don't intend to swap out the stock resistor. I intend to splice the wiring and add an on-off-on switch to give me the choice of either bringing in another resistor in series to raise the total value, or to have no negative feedback at all.

[Helmholtz]

Unless I'm wrong (which someone here will surely tell me quickly enough) the voltage at the 16ohm tap is double that at the 4ohm tap? In which case you would need a 200k resistor at the 16ohm tap to deliver the same amount of negative feedback as that of a 100k resistor at the 4ohm tap?

Both schematics please. Amount of NFB also depends on the LTPI tail impedance

You're right about the doubled voltage at the 16 Ohm tap.

I doubt there's a noticeable difference between a 176k or a 200k series resistor.

[greengriff]

Thank you - as always - for the quick response. Schematics attached.

[Helmholtz]

Thank you - as always - for the quick response. Schematics attached.

Thanks.

The LTPIs look identical. The JVM uses a speed-up cap (C24: 120pF) across the NFB series resistor which causes a phase lead of the NFB signal above 16kHz. I'd leave that alone, it might be required to avoid HF oscillation with high NFB.
No problem wiring some resistor in series to reduce NFB.
I don't see a justification for a special value of 176k total. You might go for 200k as well.

Why do you want to modify the amp?

[greengriff]

Thanks. Since I want to keep the stock feedback level as a switchable option I will wire a 120k in series the stock feedback resisitor, through the switch. The reason I want to make the NFB levels switchable is just curiosity really. I'm perfectly happy with the stock performance of the amp. The master volume is linear and predictable and makes the amp very versatile, as you can get a good sound at everything from home practice levels to very, very loud gig levels. However I also have a 2203 (well a 2103 actually) and it has a particular character, which I think is imparted by the lower levels of NFB. I'm interested to see how close I can get to the tone/character of that amp with such a small - and easily implemented - modification.

[greengriff]

Also do you have any thoughts on what wattage rating is right for the resistor?

[eschertron]

Also do you have any thoughts on what wattage rating is right for the resistor?

There's no power flowing through the resistor, 1/2W is fine if that's what you have. If you want some unsolicited advice, I'd suggest putting a 250k pot in that line (or higher, like 1M if audio taper) and dial to suit. Then replace with the (measured) value you find most effective.

[greengriff]

Thanks. Great idea re the pot.