converting tube preamps to jfet pedals?

[Doctor Algernop]

what all do I need to take into consideration when trying to do this, I imagine there's a lot but is there a formula I can follow?

[antonis]

Seriously, bro..?? 

[Groovenut]

what all do I need to take into consideration when trying to do this, I imagine there's a lot but is there a formula I can follow?

an in depth study of this place is probably a good place to start
http://www.runoffgroove.com/

[teemuk]

what all do I need to take into consideration when trying to do this

At least:
- Input and output impedance of the stage
- Source and load impedance of the surrounding stages
- Associated RC filters and their frequency responses
- Overdrive characteristics, like symmetry/asymmetry of clipping, overall "transfer characteristics", signal gain + "ratio" of clipping at given signal levels, operating voltages and bias voltages, etc.
- Inherent variance of FETs and related (in)consistency

... is there a formula I can follow?

Nope. But you can actually observe and research how the tube stage in question operates and then design a FET stage that operates and behaves identically or similarly enough.

an in depth study of this place is probably a good place to start
http://www.runoffgroove.com/

Nope. Runoffgroove site is a good example of how it is NOT DONE properly. You can't just throw a random FET into a circuit designed for specific vacuum tube, replace plate resistor with a 100K drain trimmer, bias drain to 4.5VDC, and then expect that things work even remotedly similarly than they work with that vacuum tube.

For instance,
- Output impedance will be different and thus RC filters and associated frequency responses will be different
- signal gain, and more importantly ratio of clipping at signal levels in question, will be different
- overall symmetry or asymmetry of clipping will be different
Add to the insult, the drain trimmer will make all of these characteristics variable. They ALL change every time you adjust that trimmer. That doesn't provide any consistent results because variance of FETs assures you need a unique drain resistance setting for every FET out there. So -everything- you build with that scheme will be different from another and most certainly will be entirely different than that tube stage you specifically try to mimic. On top of that, FETs do not automatically have "grid conduction" characteristics of vacuum tube devices, nor even similar transfer characteristics to triode tubes. So you need some schemes to mimic those characteristics too. Additionally, many vacuum tube stages are deliberately designed for something else than "center bias" of plate voltage, so you can't "center bias" the FET drain and expect similar results.

Unfortunate fact is that what you get with runoffgroove.com -method is not even a close approximation of mimicing a common cathode gain stage.

[GibsonGM]

Also take a look at the topic "Question about diode clipping sound" just below this one.  I started to talk about this a bit, others continued more on the topic of diodes themselves without going too far down the 'variables' road.  But my point was pretty much what Teemuk just stated - to sound like a tube stage, there are a few variables that need to be addressed, or it will NOT sound like tube.

My personal take is that if you want to sound 'just like the original', BUILD the original.  But trying to get there with FETs is a noble pursuit, and certainly is educational.   You definitely can find an awesome SOUND!  But maybe not like what you started out to get.   It would be a LOT of work - every emulator I've heard (Line 6 etc....) stating it is a model of a specific circuit falls short, to me. 

YMMV 

[B Tremblay]

You can't just throw a random FET into a circuit designed for specific vacuum tube, replace plate resistor with a 100K drain trimmer, bias drain to 4.5VDC, and then expect that things work even remotedly similarly than they work with that vacuum tube.

I agree.  That's why we stopped using that approach.

Some people enjoy the results of the "incorrect" approach,  so we've left them online.

[suncrush]

LAWD JESUS WHY?!

[BubbaFet]

With the proviso that I may have ingested too much of a particular locally-legal-substance,  I do declare that B Tremblay's & company's approach was very provocative to my awareness about the power of JFETs, and I am grateful for their sharing. (also Eb7(?) and Joe Davison. Oh! and of course PRR! But I blame RG and JackO for my affliction. It has changed my life for the better. Thank you, one, and all.

JFETs Rock, but they also Roll!  So does Tequilla.

[robthequiet]

I say put some parts on the breadboard and keep tweaking until you get something you like. There are some pedal builders out there that will share the results of their hard work, many represented in this forum. The good news is that a preamp is a helluva lot easier to emulate than the inverter and power stages.

[Renegadrian]

I spent quite a lot of time studying tube preamps, mainly Fender ones, and try to replicate them using the ROG method.
Do they sound close to the original!? not that much.
Do they sound good!? hell yes!!!

[GibsonGM]

I spent quite a lot of time studying tube preamps, mainly Fender ones, and try to replicate them using the ROG method.
Do they sound close to the original!? not that much.
Do they sound good!? hell yes!!!

   The "Sound good" part is probably the most important thing here!       FETs are LOTS of fun for big guitar sound!!

[Groovenut]

This thread has some pretty interesting information

https://www.ssguitar.com/index.php?topic=2006.0

and results

[Frank_NH]

Speaking only for myself, I would like to thank Brian, Sebastian and entire crew at Runoffgroove.com for giving me (a lowly mechanical engineer) the impetus to learn about all about JFETs and MOSFETs - a journey I wouldn't have undertaken had I not built the Omega booster and had it actually work the first time!    From my initial study (using online EE resources), I developed some spreadsheets to predict JFET and MOSFET gain for given bias voltages and/or source/drain resistors. These enabled me to understand how the basic gain stages worked.  I've since learned how to use LTSpice, and have simulated several effects pedal circuits, including the ROG Azabache and Britannia, again in a quest to understand how the circuits perform.

I also agree that JFETs circuits with topologies similar to a tube can yield good results.  I particularly like the Supreaux and others like the Eighteen and Umble are cool in their own ways.  The Dr. Boogey I built is outstanding as well.  And this kind of design has been replicated over and over in the commercial pedal world (e.g. Plexidrive).

I should note that I never expected these circuits to sound ** exactly ** like any particular amp (or preamp), although at some settings with a given guitar and clean amp, they can yield some nice tube amp-like tones.  Of course, if you really want/need REAL tube amp tones, then buy a tube amp!  But playing with JFET/MOSFET circuits is good fun and, well, you never know - you may in the end come up with something that sounds great!