How would I compare a FET to a tube spec wise?

[EricKnabe]

What about making your design as close as possible to the original..??

Well... I could use a high voltage, with MOSFETS that can take the high voltage. Or I could just build a tube amp. If that's what you're getting at.
But where's the fun in that? I've learned all this stuff, why give up now?

[printer2]

How about this?

http://milas.spb.ru/~kmg/jcm800fetLnd150_en.html

[EricKnabe]

Something like that would work well enough. Back when I posted this thread I didn't wanna redesign the stages on my own, but I've finally admitted to myself that
I'll have to do that. Now it's a matter of figuring out a few things.

1. Do I use eMOSFETS, dMOSFETS, or BJTs? I kind of like BJTs because they're easy to bias. Or at least they don't change as much if I change the emitter resistor as a mosfet does if I change the source resistor. That's important if I decide to make the cathode resistor switchable on the second stage. BJTs are kinda cool because they have that vintage 60s overdrive sound. You know, before they started using diodes... but obviously MOSFETS have softer clipping and thus will sound more "tube like" so to speak...

2. In the simulator I'm using to test some circuits the long tail pair phase inverter doesn't bias correctly. And using dMOSFETS I know it should have the same general biasing principles as tubes because in a normal class A stage with the right threshold voltage they bias the same.
I think the software just doesn't recognize that type of circuit.

3. How to make the power amp clean. If I can make the power amp stage completely clean that would be ideal. That way I could add the PI gain and output stage gain in the form of one or two extra preamp stages, and that way my overdrive sound won't be dependent on the power amp volume at all. That's important. Ideas I have for that are more negative feedback or higher voltage on the output stage. Or I could just make the maximum volume of the amp exactly what I need out of it.

[EricKnabe]

OR I could put the entire amp (including the class A/B output stage) in a 9 volt pedal format and then just use a clean power amp for volume. The thing is I just don't want to overdrive that at all.

[Gus]

BJTs with feedback
https://www.its.caltech.edu/~musiclab/feedback-paper-acrobat.pdf pages 14 to 23

[Eb7+9]

you could make use of Jack Orman's work as a starting point for "building up" an isomorphic preamp since there is some tube-like curvature to the transfer of those mosfet circuits ... no need to replicate different sections of the output stage either, just think of some overall non-linear equivalent - or do what Doug Hammond did in the Meteor ... if you wanna go with soft output-stage clipping, you could use an OTA like Bob Moog did (and wrote about) ... if you wanna go hard, a symmetric clamp using bjt's ...

see for example fig.8 in this paper:
https://www.dafx.de/papers/DAFX02_Moeller_Gromowski_Zoelzer_measurement_nonlinear.pdf

[swamphorn]

... use an OTA like Bob Moog did (and wrote about)...

I had to look this up myself; one source is U.S. Patent 4,180,707¹ of which the distortion circuit is the highlight. It consists of an OTA driven either soft or hard, and a portion of the signal optionally modulates the bias current for even-order distortion. Lucky for us, the patent expired in 1997. I suspect you can get equally good results with a discrete long-tailed pair with the signal modulating the tail current. Matching shouldn't be necessarily, since any control feedthrough would be of the signal itself.

¹ https://patents.google.com/patent/US4180707A

[teemuk]

Yes. Differential stage without negative feedback will have very soft clipping characteristics. In fact, such circuit clips softer than any generic tube circuit so for realistic emulation of what tube amps actually do it may not be the best option. 
The transfer characteristics are such that one can input a triangle wave to such "waveshaper" and get nearly perfect sine wave as output. OTA-based overdrives typically exploit this softness. Clipping will also be symmetric so it is closer to that of push-pull power amps than single-ended preamp gain stages.

The downside is that sans feedback the headroom of signal input is limited to millivolt regions. The differential may overdrive already at 15 - 25 mV peak inputs. Yes, signal-to-noise ratio will be very much a concern.