Jfets and Tube emulation - more ideas

[stratcat]

Hi, I am a guitar player who is also very interested in the technical side of making music. I have modded and built a few pedals (from kits), modded a few amps and I have run a small recording studio on and off. Electronically speaking I am a hack (with a capitol H) but I'm really interested in furthering the pursuit of tube distortion with SS.
So....

Everyone here is familiar with the ROG fetzer valve and blueprinting amp circuits with Jfets.

http://www.runoffgroove.com/fetzervalve.html

Here's a poor translation of some Russian ideas that look very interesting as well and may help those of you who understand this stuff to get even closer.

http://www.google.com/translate?u=http%3A%2F%2Fwww.sugardas.lt%2F%7Eigoramps%2Farticle68%2Farticle.htm&langpair=ru%7Cen&hl=en&ie=UTF8

Furthermore, check out the info and patents by Eric Pritchard of pritchard amps. He has alot of info about the Tubes Vs. transistor subject
http://www.pritchardamps.com/distortion.cfm
http://www.pritchardamps.com/xgpa.cfm
http://www.pritchardamps.com/audio_satori.cfm
http://www.pritchardamps.com/tech_talk/db_magazine.pdf
http://www.pritchardamps.com/audio_myths.cfm#tubevstrans
http://www.pritchardamps.com/patents.cfm

Peavey also has a few patents involving their TRANSTUBE technology which might be helpful

When one of you finally cracks the code please give me credit and forward my royalty checks

[JDoyle]

Tilting at windmills if you ask me. There is a lot of great information there that can be used to learn how to really control the distortion out of a lot of stages...

...I just think it's trying to squeeze apple juice out of an orange; your experiment will always fail to reach it's goal because Mother Nature is one strict wench, because it is impossible to bend her math, and there is a good chance that the outcome was really good in it's own way but is a failure because the goal was untenable.

I think more success comes from making what you've got sound the best it can, rather than trying to make it sound like something else.

My thoughts,

Jay Doyle

[mojo_hand]

My own observations, YMMV:

It is really easy to get a JFET to sound like a triode, in a very linear and squeaky clean range.  Getting one to sound like a clipping triode is harder.  However, if you've ever listened to a single triode clipping, you know it doesn't always sound that great anyway.  In no way does it sound like push-pull pentodes driving a saturated output transformer.  So you may get something close to the sound of a "master volume" tube amp with the gain turned up and the volume turned down, but you will never get close to the sound of a tube amp with the volume cranked.

I'm tinkering with a single-ended JFET input stage, based on traditional guitar (tube) amp design, followed by one low voltage triode.  This gives me the FET's sweet nature, reliability, and cheapness, but I still get the sound of a real triode at the end, where the gain's high enough to clip at will.  Naturally, this sounds *exactly* like a clipping triode!  Is there enough difference to make it worth the bother and expense?  Not sure yet.  But I'm pretty sure it will never sound like any amp's semi-distressed output stage.

Since I'm using a twin triode anyway, I will probably try using it in push-pull with a totally unnecessary output transformer at the end, in the hope of getting a bit closer.  But that drives up cost and complexity that much more.  Like the triode, I'm unsure whether it will be a big enough difference to pay off.  We will see.

Anyway, those are my thoughts on the subject.  I've tried translating some tube amp schmatics ala ROG, with tweaks to follow theories proposed by Danyuk and others, and I get great cleans, but the dirty is always fizzy and unsatisfying.  I am definitely not giving up on the idea yet, but going to JFET/tube hybrids is already halfway to giving up, eh?  If anyone's really happy with the way a JFET emulation of a tube amp has worked out, I'd love to hear from them.

[brett]

Is grid conduction one of those things that only valves can do, and which makes valves sound so good?
Is the equivalent of grid conduction *really* difficult to do in SS?  What about the other major differences?

In no way does it sound like push-pull pentodes driving a saturated output transformer. 

True.  But, that sound isn't just the sound of valves.

I suggest having a listen to a pair of output BJTs, such as TIPs, driving a small output transformer (as done in the Deacy amp).  Although it's 100% solid state, it sounds much more like a valve amp than your typical SS output stage.  Why?  It's probably due to factors such as distortion cancellation, "ringing" and "saturation" (a complete misnomer, but that's another story) in the OT, and more. 

So the "valve" sound consists of at least three parts, due to: (i) the valves, (ii) the transformer, and (iii) the push-pull or SE configuration.  The last two "valve" sounds can be had without any valves at all. 

I suspect that we don't think about this more because we aren't in the habit of building stompbox projects that involve winding transformers, etc, to take advantage of these tones.  In the case of amps, there is at least one major exception: the SS push-pull "Deacy" amp, as used by Queen.
cheers

[tcobretti]

I suggest having a listen to a pair of output BJTs, such as TIPs, driving a small output transformer (as done in the Deacy amp).  Although it's 100% solid state, it sounds much more like a valve amp than your typical SS output stage.  Why?  It's probably due to factors such as distortion cancellation, "ringing" and "saturation" (a complete misnomer, but that's another story) in the OT, and more. 

This is what I was thinking, too.  Would it be interesting to drive the sh-t out of a small x-former with a push-pull booster?  I don't necessarily mean TIPs; could you use a very small x-former so the booster could use low voltage BJTs?  I lack the know-how to make it happen, but do you smart guys think it would result in tube-like power amp compression?  If so, it would be the perfect compliment to the ROG simulators.

[mojo_hand]

If you really want to know, it would be easy to find out.

Go to Radio Shack, buy a 386 and 2 of their audio transformers.  They're 1000 ohms into 8, rated at some small number of milliamps.  And very cheap.

Take 'em home, set up a 386-based amp (DO remember the volume control, you'll need it to control transformer saturation), and drive an 8 ohm secondary with it.  Put the 1000 ohm output into the second transformer's 1000 ohm taps.  Drive speaker off the output of the second transformer.  Does it sound tubey? (If you wanted, since the transformers give no gain, you could even add an "A/B" switch for comparing the direct output of the 386 with the post-transformer version.)

[dschwartz]

those are really great articles!!!!!!!
i´m one of the heretic ones that think that tubes are overrated, i have played a lot of tube amps, fenders, marshalls, my own peavey triumph pag120, vintage farfisa amps, single ended, P1, high octane, etc..etc...

Many jfet distortions and solid state amps oposers just don´t give them a chance cause tubes are really into their minds as THE only way to go..just because they are used by their heroes or something..or, on the other hand, build a jfet distortion, and plug the into their guitar amps main input, not considering the strong filtering and voicing preamps have..ALL fet emu´s sound harsh if you plug them to your normal amp channel..they´re suppossed to be preamps, so the power amp is the way to go there...


i´ve build (and still am) tube preamps, CMOS distortions and jfet preamps, not looking for tube tone..just looking for GOOD tone..and i have to tell you, from my experience,that i dont find any "magic" on tubes..a well designed jfet preamp can kick the s**t out of a standard tube preamp..maybe i have tin ears, but i just can´t hear the real advantahge of tubes..and anytime i hear a great tube sound, the next day i hear a great SS sound..

i continue to build tube preamps for my tube -fan customers, just because i can charge them a lot more for the job.. and man, i just think big amp companies do the same.. "buy tubes they sound better!!(and we get a lot more revenue from them, but shhh dont tell anyone)"

but you have to see their faces when i compare the tube preamp with my fet preamp designs.."damn!!! i should ordered the fet thingy instead!!"..

for my setup...jfet and cmos...cheap, great sounding, versatile, and low maintenance marvels...and all for the same cost of one tube...

well.. don´t hate me...i´ll bet in 40 years in the future, when cars will work on other kind of fuel (cheaper, cleaner and available) there will be a lot of people thinking that gasoline is the only way to go, like yesteryears...
(by the way, have any idea of the pollution caused by tube manufacturers?)

[km-r]

actually i was planning to build a mosfet push-pull output stage through an OT... just to study the effects of the OT [saturation, etc...]

[mac]

Similar post:
http://www.diystompboxes.com/smfforum/index.php?topic=59911.20

I used the diode-from-gate-to-source idea for a booster. I was dissapointed first until I used a schottky diode. Sounds nice for a booster.

IMHO, the very best of a tube amp starts at the power tubes driven at +300V.

mac

[Ed G.]

I guess I'm biased on this, but I think jfets can be tweaked into something that sounds just beautiful.
I think in the scheme of things, they sound more 'tube-like' than overdriven diodes. Care has to be given to the design, because they tend to run out of headroom fast in the bottom end, and can 'mush out' fast. So it's a balancing act, trying to get a good bass response, while still keeping the sound 'tight'


When I was tweaking the BSIAB2, I started with the BSIAB, which was basically a two-stage minibooster with a BMP tone stack tacked on the end. I tweaked it on a breadboard and used the neck pickup to gauge how 'mushy' it was.
I tried various approaches to reduce bass, including reducing the coupling cap, reducing the top-gate-to-bottom-drain cap, and what seemed to do it was to reduce the overall gain of the stage with a larger source resistor, and bypassing that with just enough capacitance to where it sounded tight, but not too thin. I was basically unhappy with the sound of the circuit before I thought of that. It was one of those ideas that came to me while I was away from the breadboard, and when I tried it, I was very happy with the results.

I'm not learned in the art of electronics by any means. I'm basically a hack with a breadboard and enough parts to try out stuff until it sounded how I wanted, but I find it interesting that while there are several ways to skin the theoretical cat, the results vary.

[jpm83]

Is there a fetzer valve type tube to FET based amp (power MOSFET at power amp)? Because that would be kind of cool as an idea. It would be almost like having a working scaled model of an classic tube amp.

Janne

[Eb7+9]

Is grid conduction one of those things that only valves can do, and which makes valves sound so good?
Is the equivalent of grid conduction *really* difficult to do in SS?  What about the other major differences?

no, jFet's do a similar thing because the gate is a reversed biased diode until Vgs starts getting positive, and when it does it acts like a brick wall the same way a std diode does ... triodes have grid current rise much slower and more progressively as Vgc goes positive - that's the big diff between the two as far as input current sinking is concerned ... hence jFet's make better/sharper clippers

btw, it's not tubes that sound good, but the carefully crafted topology that makes use of them ...

the answer to the second question is again no ... if you try to provide external means of slowing down the current response it won't do much for you because the mechanism has to lie within the device's gain/feedback path, which it can't if it's external ... in that sense the Peavey approach is bogus as far as copying triode response - not that it doesn't sound interesting

I talked to Eric Pritchard in the early 90's ... he's put lots of work into his inventions, and much of it was based on the idea of combining the AC response based on exaggerated 2nd harmonics (which lead to my experimenting with multiplying octavers) and Transient response of dynamic compression ... you can get neat fake-imitation but that's all you'll get because part of the picture in a real tube circuit has to do with bias shifting ... Eric talks about bias shifting in the context of clipping but the same does occur in pre-clipping contexts to a subtle degree ... when you cascade these stages it becomes less and less subtle ... and then there's the output stage as Mojo points out ... it's interesting that Eric posts clips of his amps and, well to my ears anyway, they all pretty much sound like SS to me ...

then there's also the semantic question of "which" tube amp does "it" sound close to, which nobody seems to be asking ... and within that great category some tube amps sound like crap, so ... stop beating yourselves with this, there's things tubes do in context that nothing else can come too close to - the reason is meta-mathematical ...

much of this is just inexperienced people falling for bogus lingo ...

[stratcat]

Well its nice to see my post generated some buzz... 

I guess I'm biased on this

Another electrical pun

Anyway, I had the pritchard amp for a trial run. It sounded OK but it did not sound like a good tube amp when it broke up. I still think the guy is on to something and he certainly has investigated how tubes distort more thoroughly than anyone else I know of.

I happen to think the superior tonal quality of DISTORTED tube amps is FOR REAL. That said how many of the famous songs and tones we know and love are PEDALs into the front end of a tube amp. Jimi, Jeff beck, Toni Iommi, EVH, Satch, Vai, Gilmour, Eric Johnson, Richie Blacmore etc., all use(d) pedals to craft their tones. I think that many pedals capture some of the magic of distorted tubes and can actually sound better than poorly designed amps. Sometimes a pedal-amp combo is greater than the sum of its parts and a new animal is created. And pedals are undeniably more convenient than cranking your 100 watt marshall.

The current state of the art of jfets for tubes (the circuits we are seeing from ROG, Menatone, Ed Guidry etc. seem to be an innovative use of an old concept) begs for someone to revisit the hybrid amp concept. To me a good tube amp is a system where the desirable distortion is a sum of the distorted preamp, phase inverter, power amp, output tranny and even the speaker. It seems possible that if you substituted a well designed SS preamp (based on one of these excellent jfet tube amp sims) into a tube power amp/ output tranny you may get VERY CLOSE to the real thing with greater reliability and lower cost. There is a clip on runoffgroove of the Thor pedal run into a DI with a cab simulator that sounds so much like a real Plexi I think it would fool MANY people if they were never told it was a pedal. Imagine if that circuit was driving a real tube power amp through a real cab. Previous attempts have been lame for the most part (musicman and marshall hybrids come to mind). The new Line 6 spider valve is the latest but the digital sims still dont cut it for live use IMO. They don't feel or react right even with a tube power amp. Discrete jfet sims actually respond to voltage/current changes (i.e. volume control and picking nuances) much like an amp would...

Ok, I'm done rambling.

[Gus]

Did the varidrive use a fet to drive a tube?  I seem to remember that

I have played with transformers on small amps.  Radio Shack had 70volt PA transformers.  Using under a 5watt amp not to sat the small transformer use the transformer as an autotransformer.

  I like to use single supply amps and use a cap to keep DC out of the autotransformer to help it out, as well as limit low frequency's by changing  the uF value not to overdrive a small speaker if used.  With a PA transformer you could drive the 16ohm tap and connect a speaker to the 4 or 8 ohm tap or drive the 4 ohm and go out the 16 or 8 or.........


Fets for tubes is not that new, look at Neumann microphone schematics over the years.  Look at the early 80's Harvard amp that fender made.  It was in a thread here not long ago.   

[DougH]

I've got an '82 Yamaha solid state G50-112 amp (50W 1x12 combo). It's selling point at the time was that it used JFETs instead of BJTs to give it a "tube sound". It has 2 channels, clean/dirty with 3 band eq, pull-fat, pull-bright and a parametric EQ (on the output IIRC). I've had this amp for 25 years and I did something the other night I can't believe I've never done before.

I cranked it.

It is really loud with the volume at 25% and I guess over the years I assumed there was no reason to turn it up louder than needed, being SS and all. Plus, I don't really use it much anymore. It mainly gathers dust.

So (1-10) I put the treble on 7, mids on 4, bass on 5, pulled the Fat switch (sounds too scooped without it), and did not use the bright switch. I set up the para with a 100hz narrow peak, level maxed. This makes the combo cabinet "kick" like a 4x12 closed back. I started on the clean channel with the volume set at 25%, 50%, 75%, and 100% and then repeated it on the dirty channel.

And it sounded good. I mean really good. It completely floored me how good this amp sounded.

I never imagined that, in addition to the preamp "distortion channel" stuff, it gets the power amp breakup thing going, coming on gradually as you turn it up. It is not abrupt. It is not nasty or fizzy. It sounds pretty smooth, actually. Color me surprised, in the least... I never would have expected this from this amp. And this was with the stock Yamaha speaker in an opened back cabinet. I could control it with the guitar volume, the whole deal. It does not have the cranked-out-the-wazoo gain of modern amps, but definitely gets into vintage Fender/Marshall territory no problem. I feel like a fool that I never tried this before.

I really hate to say this, but I almost like it better than my tube amps. It has a smoother timbre. This may be due to the higher-wattage, the speaker, etc, i.e. the "bigness" of it. The tube amps have a little more "feel" to them, however. In comparison, this amp feels a little sterile, not quite as organic. But it is definitely usable, and for a loud gig would be an awesome amp.

A long time ago I quit worrying about "flavors of technology". I could care less about which device types are used to create sounds. This experience reinforced all that. Tubes/transistors who cares? Why worry about making transistors sound like tubes. Instead, concentrate on making it, whatever it is, sound good. Because if it sounds good it is good. And the path taken to get there is irrelevant...

[brett]

Look at the early 80's Harvard amp that fender made.

I never trealised they used JFETs for better tone, but I bought one coz it sounded good.  It was my first amp. Early 1984.
(My first longboard that year, too - 8' T&C Chris Brock for Lennox.)
cheers

[Gus]

Brett

   I don't know why fender used fets in the harvard.  I have not heard one but the schematic of it in one of the groove tubes book looks interesting. 

[WGTP]

Wow Doug, that took some balls.  I feel much cooler about using Peavey SS amps now.  I was suprised, when I had learned enough to understand schematics a bit, to find that the PV I had been using for 20 years had the same 4558/Dist+ type distortion circuit in it that we use around here.  More refined and integrated into an amp, but to my ears great for low bedroom volumes and not so bad when cranked to club levels.  I have found that tubes can sound great, but they are not the only way to go and I may prefer non-tubes.  Even non-Jfet distortions sound good.  I have also found that I prefer the op-amp/BJT/Mosfet circuits more.

Reqardless of the clipping "engine" I end up "voicing" my distortions the same, making them all sound similar.   

As R.G. etc. have said, the EQ may be more important than the "engine" 

[frankclarke]

There is a G50-112 yamaha service manual on ebay if someones hungers for a schematic.

[km-r]

http://www.google.com/translate?u=http%3A%2F%2Fwww.sugardas.lt%2F%7Eigoramps%2Farticle68%2Farticle.htm&langpair=ru%7Cen&hl=en&ie=UTF8

so, has anybody tried building victor kempf's drive lead?