JCM800 Circuit - No Drill Gain Stage - N-Mosfet Source Follower

[tmknight]

The idea is to "repurpose" V2 in that V2b (cathode follower in original circuit) becomes the "added" gain stage between V1b and V2a. This is achieved via N-Mosfet source follower replacing the cathode follower. The goal is to achieve greater levels of gain/saturation/distortion. Beyond what is achieved with the common tweaks sans extra stage and pedals. Looking for technical insight, from those more learned than me, into the design itself...please, I am not seeking pontification on the merits of doing it.

In doing some reading, thinking and simulating (not yet tinkered), I have arrived at what I think might work...at least in simulation.

A second set of eyes would be appreciated to poke holes in the concept and design. Schematic of pre-amp to tone-stack attached (original and modified). I've considered both depletion and enhancement mode fets for the design, and have seen both "successfully" used - I guess I'll have to try both to determine what "sounds" best to me.

I think the design "visually works better" when the output from the fet source is either direct to source or after the first source resistor (R11) - the waveform created in sim more closely replicates the wave of a 12AX7 cathode follower (5 triodes). Placing output at the "traditional" resistor to ground (R13) significantly clips the negative side of the signal. Also, I note the same when deleting the diode clamps...in simulation.

I'm for sure not certain if I'm biasing the transistor correctly. I ultimately tried to replicate downstream voltages that are produced with a cathode follower (5 triodes) by adjusting R11 accordingly...I'm not certain what would be the correct approach.

In all, inspiration largely comes from these sources:
http://milas.spb.ru/~kmg/irf_en.html
https://sound-au.com/project167.htm
http://www.geofex.com/Article_Folders/m ... tfolly.htm

I also found use in recently discovering this thread: http://www.ppwatt.com/node/16864

Thanks in advance for your technical insight on this design...

[Reeltarded]

There are at least two threads in the main tech forum that ended up being all about this very thing. Still searching.

[pompeiisneaks]

I'd think that works just fine, read up on the mosfet follies by R.G. here on the forums, in theory all things being equal, the right mosfet can be a swap in replacement for a triode gain stage/cathode follower etc.

http://www.geofex.com/Article_Folders/m ... tfolly.htm

~Phil

[Colossal]

There are at least two threads in the main tech forum that ended up being all about this very thing. Still searching.

Don't think your heresy has gone unnoticed.

The Guv.gif

[romberg]

In doing some reading, thinking and simulating (not yet tinkered),

I'd be interested if this mosfet cathode follower exhibits similar compression effects described by Merlin here:

http://www.valvewizard.co.uk/dccf.html

I would guess that one would want to bias the mosfet to achieve this as well.

Mike

[Reeltarded]

It would make a direct coupled SLO possible in a 100w Marshall chassis..

[tmknight]

Thanks all. Components expected tomorrow and hope to have a practical run at this in the coming days.

[tmknight]

All has gone swimmingly. I revised the design a bit as I was not happy with the (simulated) waveform. I then remembered (i.e. stumbled upon it in reviewing my notes) another article I read discussing SF wave shape solutions. Sticking with the LND150 (enhancement fets won't work with this design), it definitely brings the wave to near 1:1 with a valve triode (in simulation). The rubber meets the road in that the practical application sounds pretty darn good to me.

[Matthews Guitars]

Rather than add parts to an already good sounding circuit, I recommend starting by adjusting component values to achieve what you want to achieve.

I would, in this case, start by increasing the values of R4 and R7, the plate resistors for V1A and V1B, from their default values of 100K to some higher values, I would suggest anywhere between 150K and 220K. This will add a LOT more drive and saturation and it's literally just replacing one or two resistors with different values.

Just increasing one resistor to 330K will put the amp firmly in metal territory.

[pdf64]

…I revised the design a bit as I was not happy with the (simulated) waveform…

Sorry to be late in the day, but the schematic has a wiring error in the tone stack, ie the mid control is shorted out. So that may be messing up the ‘post tone stack’ bit of the simulation.

[tmknight]

Final schematic updated (final values and corrections as pointed out) in OP.

It's been a very interesting exercise to undertake. Though I still need to implement the switching scheme so I can revert to (near as the SF will there be there) stock, I've learned a lot with satisfying results.

Cheers, all

[tmknight]

Version two of "final" schematic added to first post. Several tweaks added to design and implemented on the physical amp that have greatly improved stability, significantly tightened the bass response, and increased gain/saturation while still retaining articulation and roll-off control. I have also implemented a resonance control, moved NFB to 8Ohm output from 4Ohm (JCM stock), and added a Richmod PPIMV (not shown in the schematic).

I had intended to make all changes switchable to take it from stock to monster, however, I decided that was just silly and would never be used. I retained Hi/Lo functionality via switch that takes out the first triode (as is done in the stock JCM circuit); this provides plenty of flexibility.

All-in-all very pleased with the outcome - I only had to lift one resistor on the board, which was for the sat control; everything else was accomplished (largely - of course socket connections were changed) off-board or in parallel to existing components. Is it still a JCM 800 (clone)? Not really, though I guess as soon as you start modding the circuit, it stops being what Marshall intended. It's starting to approach the other amps spawned from a modded JCM/JMP (Laney VH100, Mesa Rectifier, Soldano, etc.).