Tubes to FET

[blacKtearZ]

I am not sure whether i am posting in the right section as i am sure threads like these must exist but cannot seem to find them.
I have a genral question regarding conversion process of tubes to jfets. For now using the example of dr boogie. Comparing the schematics of the actual mesa amp with dr boogie, tube stage 1 & 4 have 220k resistors from supply to plate. While in Dr boogey we still have the same 100k pots on all of the stages for biasing. Would this not make the gain of the stage different to the one in the original amp?

[alanlan]

Yes, it's not an exact model.  I think the idea of copying the topology of a well known circuit is the main drive behind this sort of design.

The gain you end up with is dependent on the actual trim value required to obtain the desired bias.  There are other ways to do it i.e. without trimmers and with more predictable bias and gain, but trimmers do work and people seem to like it that way.

[R.G.]

I think the idea of copying the topology of a well known circuit is the main drive behind this sort of design.

It seems they do, no matter how many times it's explained to them that FETs are not tubes, the gain and frequency response is different, and the resulting sound is different.

I've pretty much given up on delivering the lecture. They don't listen.

[Cliff Schecht]

I offended someone at ROG last time I said anything about how lame it is to straight up copy a tube design with JFET's. The guys at ROG did it as more of an experiment from what I could tell, but they DO know what they are doing and aren't putting out circuits without knowing anything about how the circuit operates. Borrowing a topology, gain stage/section and or borrowing a tone-stack really isn't cheating in my book, but taking somebody elses design and literally replacing the tube elements with JFET's and trimmers without doing any calculations or other real engineering work is cheesy.

A 12AX7 breaks up differently than a JFET to my ears and even pentodes don't sound exactly like JFETs. They are two technologies that are completely different and give different sounds when pushed into distortion. A quick look at the waveshapes produced by each will confirm that FET's don't act like just like tubes, especially triodes like the 12AX7. Really though, some waveshaping and EQ can make just about anything can sound tube-like .

[B Tremblay]

Cliff, I don't recall that conversation, but maybe it wasn't me.  Anyway, you'll notice that we stopped doing those straight swaps of FETs for tubes based on the exact reasons that always pop up.  Lameness was certainly one of them.

However, our development of "second generation" circuits inspired by amps wouldn't exist without our first phase of FET-by-numbers circuits.  I think the Thor, Supreaux Deux, Pith Helmet, and Omega are pretty great.  Are the Thunderchief, original Supreaux, and Orange Peel worse?  They're clumsier attempts, but some builders prefer the sound.  And we know it all comes down to whether someone likes how it sounds.  Maybe someone at this forum will present the third generation...

[blacKtearZ]

I didnt think there was a forum at ROG otherwise i would have posted this topic there.
I dont own a fender tweed but i have to admit my bread boarded prof tweed sounds very much similar to the pod 2.0 emulation of tweed. Particularly at lower gain levels. As i understand its mainly the similar filtering stages that makes it sound similar to the original.
If i can partly emulate the sound of a very expensive amp which i cannot possibly afford, i dont think i would be very much bothered by the 'lameness' of the process  .
I came across a schematic of a Randall Solidstate amp, and the FET preamp stages looked very similar to high end tube amps. Perhaps manufacturers do use this topology in their budget solidstate amplifiers?
B Tremblay...by 'second generation' do you mean like the further attempts on Fetzer valve? Is there any other info available on this conversion process?

[Brymus]

Well I dont find anything about the ROG or Homewrecker projects to be "Lame"
I am sure thats not what you meant either Cliff. I am just saying-
They have been a tremendous help and inspiration to me as well as countless others.
I  would like to see more simulated FET circuits from the ROG.
I just wish I understood it enough to design my own FET circuits.

[R.G.]

It's not that JFET circuits can't sound good - they can. It's not that using the frequency rolloffs from some vintage tube amp can't be good - it can. It's not even that the kindly folks at runoffgroove say that their FET transpositions are solid state replicas of tube amps - they don't, certainly not of late when I looked. In fact,

...you'll notice that we stopped doing those straight swaps of FETs for tubes based on the exact reasons that always pop up.  Lameness was certainly one of them.

But one problem we all face as experienced builders is that the newbies are both hungry for knowledge and experience, which is great, but also for the "inside secret trick" where all you have to do to make something that's both new and great is [insert super-special secret of the day here].
Seeing an effect with a neato sounding name and everything that's almost part for part a sub of a JFET into a tube amp schemo gives a lot of the uninitiated the idea that alls you gots to do is...

Then they're back here complaining, or worse yet, contributing to the foggy miasma of "... I read somewhere on the internet that...".

This is bad because then to get real education, they have to unlearn what they think they know, and that's always harder on them.

However, our development of "second generation" circuits inspired by amps wouldn't exist without our first phase of FET-by-numbers circuits.  I think the Thor, Supreaux Deux, Pith Helmet, and Omega are pretty great.  Are the Thunderchief, original Supreaux, and Orange Peel worse?  They're clumsier attempts, but some builders prefer the sound.  And we know it all comes down to whether someone likes how it sounds.  Maybe someone at this forum will present the third generation...

Could happen. But I think we need to arm the newcomers with the right stuff. With that under their collective belts, we're more likely to see the third generation emerge. I've always felt a strong sense of giving the beginners the unequivocal right info, even when they get the wrong stuff by accident.

[B Tremblay]

I agree completely, RG.  Very good points for all to consider.

[petemoore]

  EWizards ['n wannabe's like me] choose components which produce preferred sounds.
 
  There exists a seemingly infinite variation of possibilities.
 
  Experience is consistantly better known: after.
 
  'Magic' words work best before experience transforms them to knowledge.
                                       Sounds just like Tubes.
  Guiseppe came across a 3 legged creature, it looked like an example of something he'd heard could be magical. The creature was then brought to life and thoughts we're ''it is a very impressive component''.
  He found more words of the valueable sort, related to the tri-legged-creature [using the links, GEO and AMZ^] and the 'magic experienced soon transformed to knowledge registered, most of the garbage magic disappeared instantaneously as it was replaced by useful knowledge.
  It's nice that an individual can merely wave a finger and have voluminous, encyclopedic textual or graphic knowledge at the fingertips which is extremely accurate, seeming to be almost intuitive at times.
  Also available from similar hand-movements are scads of magical fairyland tales of things we'll never be able to understand. Guiseppe confided he still spends time sorting through the magic-garbage-info to something out, and still experiments occasionally. 
  The search for the ever illusive amalgum of magical particles which enrich the sound-listening experience goes on, our considered opiniion is: ''Jfets are activated pretty easily, and can be quite lively, may require additional attention to their bias''.

[mac]

I believe that ROG fet emulations are valuable circuits for the DIY community, in particular for newbies. They are simple builds and can be a good solution for those with low budget. And for those who already have a tube amp they are nice pedals. I built a 45v clone of a Valve Jr, sounds better than at 9v. Not close to the real thing but drives my Laney in the right direction.

mac

[R.G.]

I believe that ROG fet emulations are valuable circuits for the DIY community, in particular for newbies. They are simple builds and can be a good solution for those with low budget. And for those who already have a tube amp they are nice pedals. I built a 45v clone of a Valve Jr, sounds better than at 9v. Not close to the real thing but drives my Laney in the right direction.

Read what I said.

It's not that JFET circuits can't sound good - they can. It's not that using the frequency rolloffs from some vintage tube amp can't be good - it can. It's not even that the kindly folks at runoffgroove say that their FET transpositions are solid state replicas of tube amps - they don't, certainly not of late when I looked

I did not mean to, nor did I express, anything about the ROG stuff not being worth looking at and even building if you're inclined that way. No testamonials about how good the ROG stuff is are needed. They can be interesting things to do if you want to build different kinds of pedals.

However, JFETs are not tubes, however good they may sound on their own. As you said:

Not close to the real thing

Which was exactly my point. We should not tell or imply to the newbies that JFETs are "just like tubes!" because they don't have the learning to translate that to the real conditions.

[Isaiah]

Cliff, I don't recall that conversation, but maybe it wasn't me.  Anyway, you'll notice that we stopped doing those straight swaps of FETs for tubes based on the exact reasons that always pop up.  Lameness was certainly one of them.

However, our development of "second generation" circuits inspired by amps wouldn't exist without our first phase of FET-by-numbers circuits.  I think the... Pith Helmet...

I've never heard of this circuit before, or found any mention of it.
Have you added it to the ROG website yet?
I'm intruiged!
Keep up the good work:)

[caspercody]

I am a newbie to this community, and I love the ROG builds (which you would come out with more). I wish I knew more of why some components are here and there in circuits. So for me I enjoy just looking at different schematics and trying o understand what this is for or that. And taking segments of this circuit, adding this tonestack, and seeing what I get.

But, is there any way one could read up on what the different components do in circuits? To me it looks like some of these amp to pedal designs are the same. Different tone stacks, different values on the Jfets, but same layout. What really makes a Marshall a Marshall sound? Or a Fender?

[grolschie]

I believe that ROG fet emulations are valuable circuits for the DIY community, in particular for newbies...  

I have recently re-discovered the enjoyment of the ROG Eighteen (after some tweaking for my own tastes).  I am grateful that ROG did not consider such an idea/project to be lame.

[Eb7+9]

Would this not make the gain of the stage different to the one in the original amp?

First, some background - for example, in analogue circuit design the "op-amp simulation" of a discrete circuit is a well-known problem that has its mission clearly defined (ie., matching incremental drive/load parameters) and is mathematically based - that's a std problem they teachin analogue circuit design to drive home the small-signal concepts ... 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

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 ...

I'll give you a hint : the main problem, as always in audio, is in being satisfied exclusively with the steady-state perspective ... the recent tube screamer analysis page quoted by Aaron is another dud in this respect ... until people get serious and stop being lead by the blind then they'll stay blind to these ideas - it's tough because audio methodology has long mis-applied these techniques to give pseudo-support to heuristic (low level math) arguments ... 

if you really want to know why, best thing is you can start by asking yourself why Spice is structured the way it is ... or, if you don't have time to do all that work, read my work on non-linear signal processing when it comes out next spring

[Cliff Schecht]

Would this not make the gain of the stage different to the one in the original amp?

...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

This is a pretty bold statement and I feel there is a lot of condescension in your post.. I have a strong feedback background and understand the topic thoroughly but at the same time, there are some topics that don't get covered for various reasons. Feedback, be it local or overall, can quickly become a tedious topic that one could write an entire book on and still not cover the breadth of the topic. Plus this is only one of the considerations a designer would make in adapting a tube amp to a JFET simulation. I think that half of the reason people like these tube simulation circuits so much is that they can do a similar job of coloring the sound as the copied amplifier would do. The JFET's are acting as the active components in the circuits and while they aren't the same thing as a triode or pentode, they will still add some distortion to the sound. As we all know, FET's are very popular for their distortion characteristics and although they aren't tubes, they do a similar job of adding harmonics to the input signal. The (not-so) subtle differences between a JFET and tube can be heard if compared side-by-side, but the goal seems to be achieving a certain tone coloration more than anything else. The fact that FET's sound good in these circuits is just because FET's themselves have a pleasant distortion sound in a common-source configuration. I haven't done the A/B tests as of now (nor have I ever built a tube amp simulation with FETs) but I bet that if you look at the distortion being produced by each device, the variations will be more subtle than outstanding. Especially when you're going after a high-gain sound, it's all about getting a nice mix of second and third+ order harmonics with the proper tone shaping.

With all of that said, I completely agree that the method people use to copy these amps (especially with the drain trimpots) without doing any calculations or really trying to have any understanding of the basics of biasing are doing themselves and the community a disservice. I don't know how much this applies to the ROG guys (we can look at them as instigators ) but I still don't approve of people posting circuits as finished products without having at least a basic understanding of what's happening.  This is also the case with stringing together random circuits. While it can be a great way of getting into design, it's only really useful if you apply what you are doing in a methodical manner and start learning about things like impedance matching/bridging, biasing, etc. and learn how to do at least some basic calculations. This community excels in that people can post their questions and get answers from more knowledgeable individuals without a lot of bickering. And there is even help for those who do dare to delve into the deeper math, but you have to be willing to do a lot of "self-learning". It's the readers job to take what they've read and apply it to their situation.

As far as school is concerned, I can honestly say that most of my learning as far as being a "good" designer came from the amount of work that I've put in outside of school. Going to school for an EE degree and making a 4.0 doesn't automagically make you a good circuit designer (audio or other), more than anything it can hinder this (taking History, Humanity and English classes are a waste of my time!!). I know that I've spent a lot of time in the past reading about things that were actually interesting to me than doing my homework or studying for tests and while my grades aren't as high as some of my piers, people still come to me before anybody else to ask why their circuit doesn't work or how they can squeeze more out of their design because they know that I know my sh*t. IMO school only provides you with the tools to be a successful engineer, it's up to you as the student to decide how you plan on using those tools. For me, my grades aren't as impressive as my peers but my resume is much stronger because of the breadth of working knowledge and experience that I have compared to the typical engineering student.

Edit: Also, I agree completely about the lack of coverage on non-linearities in, well, everything that I've read in a school text book. This topic really only gets covered once a student gets past their math and stats classes and can start understanding why it's important to be able to predict and model non-linearities. Also agreed with wanting to see more on steady-state response. You really can tell a LOT about a circuit just by looking at how it responds to different input types. I always like to look at a square, triangle and ramp wave when I'm looking at circuit response, especially when I'm working on a synth filter where how the ringing looks can tell more about how a filter sounds then actually listening to the damn thing!

What tube amp book did you write? I own more than a few books on the topic is why I ask..

[mrslunk]

I'm intrigued...
As someone who has taken a few electronics course during my undergrad(which i am now waiting patiently to see if i graduate...), i can definately agree with the "Plug and play" nature of university electronic engineering.
The courses i've done dealt mostly with ideal op-amps, basic BJT's and diodes, and active and passive filters.

So, by a real text, did you mean something like this:
http://www.amazon.com/Microelectronic-Circuits-Electrical-Computer-Engineering/dp/0195323033/ref=sr_1_2?ie=UTF8&s=books&qid=1257899270&sr=8-2

I'm still searching for electronics textbooks that aren't going to dumb down the math...

[Cliff Schecht]

For books, I've found that the older the book the better (within reason). Most of the basics of what we do around here is based off of ideas that were thought up in the 20's, 30's and 40's. These books tend to be much more theoretical in nature but can become cumbersome quickly, so be warned. For tubes books I really like Spangenburg's book and anything by Seely. For transistors, I can usually find what I need with google unless the topic is highly specialized.

[mac]

Do you know of a high voltage commercial fet amp, +200v?
I ask because I want to try some ROG circuits at voltages similar to tube amps.

mac