Parallel igain stages like matchless and marshall.

[John Lyons]

I have a Matchless "Chieftain" clone amp that I built and I really like the clean tones as well as the distortion amount.
It's a got a good medium gain distortion. The Amp uses parallel triodes stages which I've heard fatten up the tone and double the current.

Here's a thread about Parallel triode stages
http://ax84.com/bbs/index.php?id=257313

I wanted to try a FET version of this amp so I laid out the parts on the breadboard as in the schematic. The problem is with FETs at 9v the input clipped before I get any real gain out of it. I put a 1M voltage divider before the FETs (J201) drain but this just make the first gain stage unity gain. If I used a 2N5457 this would drop the gain even more. To get a clean loud signal it seems that I would clip the FETs.
Other than using 18v etc what am I missing here?

Heres a schematic of the amp. I removed the reverb section for clarity.



My main questions are:
How to overcome the FETs clipping at the input with only 9v?
Is the Paralleled FET set up going to work similar to paralleled triodes, fatter sound, less noise etc etc?


Thanks

John

[alextheian-alex]

Fancy meeting you here.

When you parallel two triode stages, you are effectively halving the rp, but mu stays constant, and gm is doubled.  Since you are doubling gm, the noise can theoretically be 3dB lower... 6db lower for 4 paralleled triodes, but that is all debatable since in real life you can't get there because of other factors.

A fet does not HAVE  rp, bu you will be doubling the gm, which will gove you the same THEORETICAL 3dB noise drop...in reality it will be a bit less, but it is there.  As far as gain... the gain for a common source fet circuit is gm(Rdrain||Rload) so with parallel fets, you are increasing the ability to swing voltage since you are doubling the gm, which makes more gain available, but you can EASILY bump into the 9v rail headroom with one fet, so it is a moot point... you don't NEED more gm.

The small increase in gain in triodes has more to do with the decrease in rp, reducing the upstream loading effect, so depending on the circuit, you will get varying results.  In a fet circuit... the output impedance will be equal to the Rdrain more or less, so that issue is a moot point as well.

As far as the "fatness"  is concerned, if you parallel devices, they won't be absolutely identical, so you will have all kinds of different freq bouncing around as they react differently, but to the same input signal.  That will be even more pronounced if you split the source resistor and bias them up differently.  how that translates into sound... i dunno.  If it sounds fatter then it is fatter, but you may also get phase cancellation, which may make it thinner depending on the frequency.  Try it and see.

When changing a triode circuit to fets, you have to scale down everything, and scale the tone controls to accommodate the new zo or the freq response will be different.  The biggest issue is the headroom of a 9v supply... it does not leave much voltage swing available.  What you'll have to do is figure out the headroom you have available, and then plot out a load for your fet based on the input signal.  If you want to match that to a Matchless, just figure out what % of the rail voltage it has available for the input swing, and figure out how that compares to the sensitivity, and then you can scale down the fet stage accordingly.

honestly, that is a lot of math, so i would recommend simulating it if you can

[John Lyons]

Yeah, small world alex!

I've built a few FET based circuits but never designed one or build one from scratch.
I tried the Run off groove "Fetzer valve" just to see the gain increase with a single J201 FET as an experiment and something to build on.
I clipped the input with the guitar (which has not so HOT humbuckers in it...hmmm ) I put a 1M divider before the fet as ROG illustrates but the gain was just about unity... I figured I'd get some clean gain but maybe not.
Correct me if I'm not getting what I should... is clean gain available at 9v from a FET front end?

John

[alextheian-alex]

Yeah, it is available... but only a few dB.  Bumping it up to 18v would increase your margin of error.  Guitar pickup voltages vary so wildly that it is hard to judge with so little headroom before you hit the rail. If you want it clean, then be careful about your bias point.  Without resorting to load lines, i'd say just keep the drain resistor low... like under 5k... and adjust the source resistor to set the drain voltage a round 50-60% of the source-to-rail voltage, and then djuust each sussessive gain stage's voltage dividers, etc.

Try a 5k or 10k trimmer on the drain and a 1k trimmer bypassed with 10uF on the source, and tweak for a while.

And don't forget that since the gain is so small, we don't have the huge miller capacitance that we do in high gain tube circuits, so to get the same high end response, you'll probably have to use snubbers or caps to ground.  if you figure that the average miller capacitance that we see in typical guitar amp gain stages floats around 150p, then you can also consider strapping a 47-100p or so capacitor from gate to ground before each stage to make the interstage designing easier, then just raise or lower it to trim off the high end harmonics

[John Lyons]

Ok, thanks Alex. That gives me something tangable to work with.

Ideally i'm looking for a pentode type sound with some compression that will get a nice clean sound without being too stiff sounding.
And when pushed will break up nicely. The FET sound is pretty stiff sounding and it would be nice to have a little "bloom" to the dynamics.
Maybe I'm looking for more of a mosfet at the front end.

John