Tubes to FET

[Cliff Schecht]

I was running a JFET in the input stage of a tube amp I was working on for a while. The supply voltage was around 350V and the drain of the FET (MPF102) was at about 40V without any issues. Then again, I don't think they make any JFET's that can go much about 45 or 50 V.. This probably has to do more with non-existence of application for high voltage JFET's than it does with the process limitation. You'll have to go with MOSFET's which probably won't sound good distorted. The excessive gate capacitance of a typical high voltage FET I would think could become a problem as well.

[B Tremblay]

Something that will merely plug into a "tube" schematic without requiring any thinking and "works" has made some people wrongly conclude or suggest that it works "like a tube" ... seeing that nobody here really understands how NFB works in single gain stages, or won't say (see the jFET biasing tail-chasing thread) it is no surprise that this has not been addressed properly in this context ... gain in it's generalized definition (gee, what's that ?) is the main factor here, noise also comes into it as well ... I've posted about that - the main sign that the ROG guys didn't apply any serious thinking beyond showing that topologies exhibit a characteristic sound - which obviously is an interesting and valuable concept in itself - but one that many of us amp techs have long known and exploited in our work

best thing, go to school and study from a real text (try Sedra's) - you're not likely to get a real answers from the resident "know-it-all" cause he don't know anyway and why should you or anybody, in a  few minutes, understand something that most engineering students don't get after a few years of small-signal analysis studies (as a TA in university I learned to figure out why) ... the realm of non-linearity is certainly not well understood and university curricula does a very good job of avoiding it's mention ... similar to music school programs - they teach you plug-and-play (paint by numbers) but not to go deep in understanding in such areas ... audio researchers are unique in that this area matters much to them - so we are on our own in some respect ... I wrote about this analysis technique in my tube amp book - I see Douglas Self is now doing the same thing in his Power Amplifier design book ...

You seem to take offense from these circuits' existence and popularity.  Rather than issue judgmental comments about others, why not share a circuit that succeeds where you see other designs fall short?

In an attempt to provide some perspective for this discussion, please consider the following points:

-runoffgroove.com does not make any claims about amp-inspired FET-based circuits sounding "just like the real thing."  We don't use the terms "emulation" or "simulation" because we don't want to make false claims.  Others may use those terms and they are free to do so, even if I disagree.

-We are freely sharing those circuits.  I could understand the criticism if we were fleecing unsuspecting buyers after we'd issued some "FETs-can-sound-just-like-tubes" claims, but we've done neither.

-There's no dispute in matters of taste.  What is gained from demeaning certain circuits?  Nothing other than making fans of that circuit feel foolish for having chosen poorly.  If it sounds good to an individual, then it has value.  If you find no value in it, then don't bother with it.

Just because I don't like (insert your least favorite food or beverage here) doesn't mean that I have to mount a crusade against people who enjoy it.  That's simply what I consider to be mature civility and restraint.

[mac]

i found only this, hitachi 2sk373, vgs=100v, ten times over a 9v batt.
i hope to find one...
high gate capacitance is not a problem for me, i like dark amps 


mac

[space_ryerson]

A bit OT, but does anyone remember a product in the 90's that was a transistor-based 'tube replacement', so you could turn any tube amp into solid state? I know it sounds absurd, but it was a tube-shaped unit, that would be plugged into the tube socket. I can't seem to phrase the right google search query to find out any info, and going through a decade of my old guitar magazines would take a while.

[Droogie]

A bit OT, but does anyone remember a product in the 90's that was a transistor-based 'tube replacement', so you could turn any tube amp into solid state? I know it sounds absurd, but it was a tube-shaped unit, that would be plugged into the tube socket. I can't seem to phrase the right google search query to find out any info, and going through a decade of my old guitar magazines would take a while.

I have an early 80's Mesa Boogie Mark II that has the optional use of a "Fetron" which seems to be what your describing. It fits into the V1 12ax7 socket; there's a switch that allows for either tube or Fetron use. It has a very interesting midrange characteristic that sounds great.

[markeebee]

I'm a mild-mannered man, and I like to think I can usually maintain some kind of prespective on trivialities.

I'm surprised by how annoyed I am by Eb7+9's post.  I can't understand the motivation behind it.

[markeebee]

Ooops, sorry, I meant "perspective on trivialities".

I know how pedantic some people can be.

[stm]

I'm surprised by how annoyed I am by Eb7+9's post.  I can't understand the motivation behind it.

The motivation is clearly stated at the beginning of his post, as you can read below.  First, JFET implementations of tube amps as we know them are demoted to "audio toyland", then... well, just read what follows (my emphasis in both quotes):

... otoh, the tube-to-jFET emulation as seen in audio toy land is merely based on graphical emulation (it "looks like" a duck sitting in a duck pond) - and the fact that it "works", does something, etc. does seem to provide some sort of backup in that thinking ... but in fact, that's a very limited framework to be working in - jFETs "in situ" do things that tubes don't and vice versa ... so, at best it's a limited emulation - why, I can't tell you right now because I have a product coming out that exploits the correct idea ... in a nutshell there is something there to be exploited for sure - some researchers have indeed come close
...

[markeebee]

Exactly.

I'm sure JC's new product will be great, and respect is undoubtably due, but I thought it was a bit shoddy to take a pop at ROG (and others) in order to promote it.

[amptramp]

A FET may be usable as a substitute for a tube in some designs, but keep these things in mind:

1. The variation of gain with bias in a tube follows a 3/2 power law whereas in a FET it is square law (power of 2).  Square law tends to create strong second harmonics with less harmonic content at other harmonics.  This is from the trig identity of 2(cos²u) - 1 = cos 2u where u = 2pi f t. (f = frequency, t = time and u = 2pi f t is the instantaneous phase angle)

2. There is a relatively small interelectrode capacitance in a tube which is reasonably independent of operating voltages and currents.  In a FET, the capacitances are large and vary significantly, for example, the output capacitance is high at high current and low voltage but low at high voltage.  Harmonics other than second would probably be from capacitance variation rather than the DC characteristics of the FET.  The Miller capacitance (gate to drain) is particlarly significant because it is multiplied by the gain.

I have no argument with people replacing tubes with FETs as an experimental exercise, but the designer should be aware of the differences in harmonic variation between the two designs and not expect them to sound alike.

[d95err]

A bit OT, but does anyone remember a product in the 90's that was a transistor-based 'tube replacement', so you could turn any tube amp into solid state? I know it sounds absurd, but it was a tube-shaped unit, that would be plugged into the tube socket.

You may be thinking of the Fetron:
http://www.slack.com/misc/Fetron.pdf
(I think some early Mesa Boogies actually were shipped with one of those as the first preamp tube.)

And guess what - everything old is new again:
http://www.youtube.com/watch?gl=SE&hl=sv&v=URvfywUxWzk

[space_ryerson]

A bit OT, but does anyone remember a product in the 90's that was a transistor-based 'tube replacement', so you could turn any tube amp into solid state? I know it sounds absurd, but it was a tube-shaped unit, that would be plugged into the tube socket.

You may be thinking of the Fetron:
http://www.slack.com/misc/Fetron.pdf
(I think some early Mesa Boogies actually were shipped with one of those as the first preamp tube.)

And guess what - everything old is new again:
http://www.youtube.com/watch?gl=SE&hl=sv&v=URvfywUxWzk

That sounds about right. Thanks for the PDF; interesting reading!

[slideman82]

You may be thinking of the Fetron

I wish I could get some of those...

[puretube]

Anybody cross-reading?

[frequencycentral]

Anybody cross-reading?

Hahahah.................

[Cliff Schecht]

Anybody cross-reading?

Yes.

[Eb7+9]

You seem to take offense from these circuits' existence and popularity.

I nowhere suggest these circuits don't sound good or shouldn't be popular - on the contrary, like I said, I know they sound good in their essence - in great part for topological reasons ... when DougH came out with the Meteor I made mention of the Meteor's surprising tonal character, I still have mine ...  since that time, it's been noted that jFET-emul designs generally tend to exhibit poor S/N

I posted about making these circuits less noisy through a shift in biasing (the data sheets support this and offers a bonus at the same time) - I wouldn't bother offering that info if I had no respect for jFET Amp-Emul circuits ... all I do is suggest experimenting outside the confines of verbatim emulation as far as biasing is concerned, that's where the hold is IMO ...

as far as the vague concept of "tube emulation" is concerned I think lots of misunderstanding is coming from the fact that two problems are being discussed conjointly, and neither is well defined ... one being, what is it that gives tube amps their dominant "character" ? ... there's a link to a paper on my website that takes a new approach to answering that question ... the other question is whether "that" thing can be emulated in a low-voltage circuit ... well, one case was devised by Bob Moog and used in the Lab-Series guitar amps - the other case is the one I sussed out ... there are other possible questions: for example, why emulate exactly ? why not go further and produce exaggerations ??

I have a circuit that will provide that effect in a booster format - the idea is not to emulate the total Gestalt of a tube amp, but to ... er, well I can't say ... I've brought it to jams and gigs, the thing rocks - I'm still tweaking the design to see how much range to give it ... just waiting for new pots from Smallbear

[Eb7+9]

x

[markeebee]

Anybody cross-reading?

Yes, I did get a bit cross, but I'm over it now.

[mac]

1. The variation of gain with bias in a tube follows a 3/2 power law whereas in a FET it is square law (power of 2).  Square law tends to create strong second harmonics with less harmonic content at other harmonics.  This is from the trig identity of 2(cos²u) - 1 = cos 2u where u = 2pi f t. (f = frequency, t = time and u = 2pi f t is the instantaneous phase angle)

iirc there is a way of making a fet closer to a 3/2 law... rog fetzer valve article???

mac