building jfet preamps from tube preamp schematics (mesa D-180)

[runmikeyrun]

Hey all,

I've seen before that people will build pedals using jfets in place of triodes, mostly from "proven" guitar amps (Mesas, Marshalls, etc) with mostly good results.  I also know that this is not a "plug and play" procedure but does require some fiddling- trim pots on the jfets for one.  I have a preamp schematic from an old Mesa D-180 bass amp,  just for V1 and V2.  I am not sure what all it would take to do the conversion.  I plan on putting trim pots on the jfets for one, but I'm not sure if I should run it at 12v or 9v.  

I'll assume that running it as a starved plate design, maybe 12v or 24v would require changing some resistor values (plates, etc) and is probably not worth the trouble.

Any tips would be appreciated, as spring break is coming up next week and I'll need something to do

Thanks!
Mike

[smallbearelec]

I've seen before that people will build pedals using jfets in place of triodes
Any tips would be appreciated, as spring break is coming up next week and I'll need something to do

Try sorting some devices and breadboarding this:
https://www.smallbearelec.com/HowTos/BreadboardBareAss/BreadboardBareAss.htm

[GibsonGM]

Another small part of the process would be getting on LT Spice and simming up the tube circuit, then trying to come up with something that has a similar frequency response with the FETs.  You'd also need to approximate the gain and whatnot - any parameter you can reasonably transfer over.

The values will be way different between tubes/FETs (and different FETs) due to impedance and miller capacitance differences, etc, but you can find something close if you work at it. 

The smallbear article is awesome and a great way to break into the FET realm.

[teemuk]

I also know that this is not a "plug and play" procedure but does require some fiddling- trim pots on the jfets for one.  

If you want to get consistent results, and results that actually sound like the circuit you try to mimic it's not that put-100K-trimmer-at-drain-and-call-it-done "plug-n-play" either.

It's true that FETs have some characteristics that make somewhat straightforward circuit conversions possible. However, they, and the circuits they require, do not entirely match up to their tube counterparts.

First key difference you encounter are the headroom limits introduced by running tubes at several hundreds of voltages versus running FETs at about 20 - 40 volts, if not even lower. Tubes have tremendously more range in voltage swing where signal remains clean, the FETs will in comparison clip much too early at equal input signal levels.

Tubes are much more prominent to grid conduction where grid begins to draw current and clips one halfwave of the signal. With FETs you usually have to emulate this characteristics with additional diode(s).

A 100K trimmer is a terrible inconsistent biasing method: First of all, by varying drain resistance you not only vary drain voltage and its affects to signal symmetry (or asymmetry) you also vary the stage gain. Furthermore you vary output impedance of the stage and this will have subsequent effects with the following circuitry, which is usually designed for entirely different source impedance, the one of the tube circuit. When each device always requires its unique resistance to bias building with consistency becomes impossible.

Many tube stages are deliberately designed to NOT HAVE a nice "center bias". There are often deliberate attempts to clip the signal more prominently either by saturating the tube or by driving it to cutoff. Symmetric center biasing is rarer because it reduces duty cycle modulation and bias shifts, not to mention their interesting effects to dynamically vary the harmonic patterns of clipping distortion. With perfect center biasing, which is usually advised in these plug-n-play conversion schemes much of this interaction is lost. Things are usually made worse by insufficient emulation of the "grid diode" effect. For accurate capturing of performance of the modelled circuit focus should be pointed in modeling the clipping characteristics of each tube accurately.

These are just few things wrong in "100K trimmer to drain" conversions.

I hate to sound discouraging but personally I feel FETs are way overrated. If you aim for really accurate emulation you soon find that plain OpAmps do all that you ever wanted much easier than trying to force a FET to do the same thing. Not to mention first acquiring that FET, or best to say a whole bunch of them because you have to wade through a lot (they usually are not on-the-shelf items like generic opamps either) and then figuring out how to build anything consistent out of so terribly inconsistent devices.

They are nice switches though.

[runmikeyrun]

Wow, thanks everyone for all the input!  I haven't played with LT spice and I dont know if the learning curve is something I want to take on at this point.  I'm not too afraid to experiment but this project sounds a little over my head right now.  I only have degree in trial and error and diystompoxes lol so I'm not too up on theory, which seems to be a lot of making something like this work. 

I do really appreciate the advice though.  Sounds like I'd be better off just building a 450v power supply and using tubes and the schematic the way it is if I want it to be consistent.

Thanks again!

[GibsonGM]

Wow, thanks everyone for all the input!  I haven't played with LT spice and I dont know if the learning curve is something I want to take on at this point.  I'm not too afraid to experiment but this project sounds a little over my head right now.  I only have degree in trial and error and diystompoxes lol so I'm not too up on theory, which seems to be a lot of making something like this work. 

I do really appreciate the advice though.  Sounds like I'd be better off just building a 450v power supply and using tubes and the schematic the way it is if I want it to be consistent.

Thanks again!

Yeah, you could go that way, but I'd do some real builds and learn a bit, first.   A tube build is actually pretty hard on you, ha ha.  Just cuz of the high voltages present, mainly....when you're ready, I'd suggest you get a kit, like an 18Watt Marshall or something from AX84. 

Teemuk hit the nail on the head...but don't get discouraged, as you can still make awesome pedals - they just won't be the same as that tube amp you've dreamed about - but they can make an amp sound incredible!

[garcho]

It's a good way to learn about gain stages and tone stacks and stuff. I was always disappointed in the FET/tube thing. Just sounds like an overdrive box to me. All that work and then for what power amp and cabinet? Doesn't that end up changing the tone, radically?

[GibsonGM]

Yes and no, Garcho. There are a few VERY nice items out there that really take on the color of what they're supposed to be simulating.  I'm thinking of the Catalinbread stuff, mainly the R.A.H.    To my ears, THAT one is SO close to the real thing, you'd be hard-pressed to do better without having to go buy the actual amp head ;o)   Playing with FETs is 1,000 times easier than trying to build a few of these amps, I think...

Of course you'll get some tone change depending what amp and cab you use, but that would also be true of the tube stuff...there are lots of cabs out there, some radically different than others.     For me, and I think maybe for others, we build this stuff hoping to sound "a lot" like our guitar heroes, but not so close as to be a joke, you know?

I mean, Jimmy Page sounded like Jimmy page on one tour, then a little different on another, then he started flanging everything....see what I mean?   Like taking a snapshot and trying to work with it.