Tube emulation with FETs

[puretube]

no - unfortunately....

that man Pr*tchard is a maniac...  :shock:
...absolutely far out...  :roll:
~20 patents...
(been thru all his stuff/google/pritchard amps/PRS)  :?

[rubberlips]

Guys,
Thanks for all the input. At least I have a few nights reading ahead of me.
Yes FETs will probably never sound exacly like tubes, nothing will. However in the semi conductor world FETs are probably the closest part to a tube, so it makes sense if you're trying to get something close using semi's - you'd use a FET.

But it's fun to see what sound comes out of tube circuits when you chuck a FET in there and that's the main thihng

Cheers

Pete

[Arno van der Heijden]

Can somebody explain 'grid conduction' and how to emulate this with a diode?

[puretube]

"roosterstroom trekken"  :lol:

(should be explained somewhere in one of Pritchard`s elaborations...)

[Eb7+9]

Can somebody explain 'grid conduction' and how to emulate this with a diode?

... it's a current function which takes on the profile of an Ig=exp(x) function with Vgs as a variable ... so it limits to zero for progressively negative input voltages and high for positive ... to the point where grid current levels approach that of the preceding plate circuit current levels (few mA) and proportionately robs current from that circuit and consequently squashes the rise of plate voltage ...

... you can't bypass a FET with any kind of diode without killing the input impedance and resulting circuit gain ... plus the non-linear current has to add to the Drain current to produce a non-linear Source sum to fully mimmic the Cathode node current in a triode, unless you only want input loading modeling ... actually, there's been a few pretty Tubey sounding FET preamps amps out there so far ... I don't think the power amps can be done THAT well, but whatever ...

I spoke to Erik a few times in '94 after he was featured in EDN ... his "tube amp" emulator circuit was in the news and I tracked him down for a conversation about what exactly he was trying to achieve ... in the end he confided what that "amp" circuitry aimed to do was imitate Tubey-ness through a "low-mix" of dry signal and clean second harmonic, and compression ... the harmonic part became the basis for my Harmonic Multiplier circuit - something that occured to me while working on cloning the EMS VCS3 (it's burried deep in the service notes) ... I brought this up once before but got into it with Mr. Snowjob ... notice that Erik's basis, as well as all other well-quoted researchers on the subject of tube emulation and tube sound "quality", focuses exclussively on Steady-State analysis ...

[amz-fx]

notice that Erik's basis, as well as all other well-quoted researchers on the subject of tube emulation and tube sound "quality", focuses exclussively on Steady-State analysis ...

And that is one of the big hurdles to overcome.  The "feel" of the amp must be duplicated...  not just because you are trying to copy a tube amp sound but because the dynamics and response as you pick the note must feel right to the player.  But not everyone is limiting their testing to a static system -- several of the patents on the subject deal with the dynamics and changing frequency characteristics.      The digital emulations, in particular, are working to model the varying harmonic spectra and are ahead of the published analog circuits in some regards.

regards, Jack

[Joe Davisson]

The depletion area near the drain is thicker than at the source in a typical JFET stage. The gate-source diode can only go to about 0.3v before beginning to go into conduction. If it goes too far, the JFET is cut off prematurely. A large enough positive input pulse (coming from the previous stage) would cause hard-clipping (from cut-off) in the second.

The catch is that you can't divide-down the level between stages without reducing gain. You usually wouldn't want 20 JFET stages. The diode-compression trick provides a drastic signal reduction for the stronger peaks that would forward-bias the gate-source diode in the JFET.

The BJT doesn't get cut off in the same circumstance, more the opposite effect. Here, instead of preventing cut-off, the diode prevents hard saturation. The overall effect on sound is similar in the end.

[rubberlips]

... you can't bypass a FET with any kind of diode without killing the input impedance and resulting circuit gain ... plus the non-linear current has to add to the Drain current to produce a non-linear Source sum to fully mimmic the Cathode node current in a triode, unless you only want input loading modeling ... actually, there's been a few pretty Tubey sounding FET preamps amps out there so far ... I don't think the power amps can be done THAT well, but whatever ...

Wouldn't have any examples would you - makers, models to search for??

And that is one of the big hurdles to overcome. The "feel" of the amp must be duplicated... not just because you are trying to copy a tube amp sound but because the dynamics and response as you pick the note must feel right to the player. But not everyone is limiting their testing to a static system -- several of the patents on the subject deal with the dynamics and changing frequency characteristics. Smile The digital emulations, in particular, are working to model the varying harmonic spectra and are ahead of the published analog circuits in some regards.

Hey Jack, with your mini-tube projects etc. can they be used in place of a preamp?? Do they have enough gain? What would the noise be like if you increased the gain of it?

Cheers

Pete