No sound Thru Marshall Haze 40 FX Loop (that mosfet/jfet cascode thingy)

[waspclothes]

I'm having a hard time fixing this one, mostly because I don't understand the circuit itself at all.

This is a Marshall Haze 40 (schematic: https://www.thetubestore.com/lib/the...-Schematic.pdf )
that works fine, but no sound from the digital FX and no sound through the FX loop. Both go through this mosfet/jfet cascode circuit that I've read about but admittedly do not understand.

I've attached a drawing showing my measurements from the schematic while injecting a 1Vpp signal into the FX input. I'm not sure why it gets negatively biased before hitting the GATE of TR12 (Jfet J202) and I also do not see any signal coming out of it's DRAIN like I would expect (maybe I'm wrong).

If I remove TR10 (mosfet TN2404) all the resulting measurements stay the same.

My first attempt was to just shotgun replace TR10, TR12 and D27 (zener) but that didn't work. Since I couldn't find TN2404, I tried a few replacements ( https://www.mouser.ca/ProductDetail/689-TN5325N3-G and https://www.mouser.ca/ProductDetail/522-ZVN0124A ) so maybe my replacements weren't specced properly. But I'm kinda wondering why I never see a signal out of the JFET, and also expecting to see the mosfet GATE at 24V from the Zener.

Thoughts?

[waspclothes]

Here's the measurements

[Helmholtz]

Check if C59 is shorted.

[g1]

Check if C59 is shorted.

Or D27 if C59 is hard to get at. They are in parallel.

[Helmholtz]

Or D27 if C59 is hard to get at. They are in parallel.

I know, but the OP already replaced D27. Of course parts generally should be checked out of circuit.
Now assuming D27 is good, it measuring shorted in circuit would indicate a shorted C59.

[waspclothes]

Check if C59 is shorted.

I'll try this tomorrow Helmholtz and report back, I was definitely thinking about it while writing my post.

[teemuk]

The "cascode" is combination of common source and common gate amps. Common source amp receives input through gate and feeds amplified, inverted output signal from drain. Source provides bias and degenerative feedback but latter is suppressed by capacitive bypass for AC signals.

The common gate amp acts as drain load for the amp. It receives input from source and outputs non-inverted amplified signal from drain. Gate is DC biased (resistor+zener) and capacitively "grounded" for AC signals. Therefore "common gate".

And yes, that gate should indeed be at DC voltage potential that is zener diode voltage higher than the DC source voltage of the common source amp. It isn't.

There is also negative feedback loop from output (common gate amp drain) to input (common source amp gate) that reduces closed loop gain to about 10x. This negative feedback should form a "virtual ground" to cascode's gate input (so there should be no apparent signal there or one with just very low amplitude) but obviously now that the amplifier does not work the loop is inoperational as well and you can measure a signal at the gate. When the gain stage works the signal should dissapear so don't try to find it there. Measure before the input series resistor, which is the other part of this shunt feedback loop.

[teemuk]

In addition to checking if C59 is shorted, check if R78 is open; the zener can't zener without voltage potential and DC readings of less than one volt could be suspect to measurement tolerances.

[waspclothes]

Check if C59 is shorted.

Yup, this was it! I don't have the old parts any more, but I wonder if the mosfet or zener failed and caused this cap to short out? Either way, interesting.. this is the first electrolytic capacitor I've seen fail as a short. Something to look out for from now on..

Thanks everyone!