SOT: Attempt at "tubish" MOSFET amp

[brett]

Hi.
I've drafted up a class A MOSFET amp that I'm going to experiment on, with the aim of getting a wall-wart powered, tube-sounding amp capable of about 2W to 8W power.

My questions are: 1.  Is there a demand for such an amp?
2.  Am I making any mistakes in the design?

thanks for any comments or help.

Here's the design so far;
http://members.optusnet.com.au/~jethro.dog/mypic156.jpg

[Connoisseur of Distortion]

1. ABSOLUTELY! if you can get MOSFETs to sound like tubes, you'll make friends with all kinds of amp builders!

2. I am by no means an expert... but it looks like it should work. i don't know anything about MOSFETs (used them, but never really learned about them) or speaker loading (nothing in that section yet) so you should probably wait for actual advice.  :wink:  

you should make a different output optional! a push-pull setup, in case we want it louder!  

[brett]

in case we want it louder!

The main problem with higher power is that wall-warts aren't feasible as power supplies.
In fact, I have wondered whether the MOSFET should have a Vcc of 15V, so that (a) it draws less current, and (b) it clips harder.  For some reason, a heavily clipped 5W sounds louder to me than a clean 10W.  RG has commented on this, too.

[runmikeyrun]

isn't that what the Tubeworks amps were trying to accomplish?

[jrc4558]

How many watts do you think it will produce?
Also, note true class A design!

[toneman]

looks workable.
i would regulate the MOSFET separately.
that way, when drawing large currents, won't affect the
input fet biasing.  The OA probably would B more tolerant of
supply fluctuations, but, wouldn't hurt.  So, i say run 2 regs.
U have potential for overdrive in input stage *and* output "stage".
i've been thinkin bout just a single MOSFET, TO3, powertranny
for a fuzz box.  I gotta bunch of 2N6766's that i've been dying 2 try.
stayMOSed
tone

[brett]

I looked at the Tubeworks amps.  They are high-power designs, and people seem to like them.  I'm aiming for something slightly more powerful than a Ruby (0.5W) or the excellent (but expensive to build) tube amp the Firefly (~1W).

As for power output, the MOSFET is only a voltage follower for the op-amp.  The op-amp output can swing nearly rail-to-rail (+6V, -12V).  If the MOSFET manages to get within 3V of the rails, it might swing (+3V, -9V) into 8 ohms, for about 8V RMS x 1 amp, or 8W RMS.  Maybe??  

I know very little about MOSFETs and power devices, so I'm making some uneducated guesses here.

Oh, yeah, credit is due to others that inspired the "components" of the amp.  The input section is based on a Stratoblaster (General Guitar Gadgets).  The tone section is from a big muff, with values substituted that provide scooped mids.  The buffer for the tone section is kinda like a tubescreamer "drive" section, but with less drive and biased so that it turns the MOSFET on.  The MOSFET voltage-follower design is mine (well, something had to be!).

[brett]

Thanks toneman.  Yeah, 2 supplies sounds good.  I'll run the JFET and op-amp at 18V, the MOSFET at 15V.  That'll lower the power demand, too.

[R.G.]

See:

http://www.passdiy.com/pdf/zenamp.pdf

http://www.passdiy.com/legacy.htm

http://www.passdiy.com/amps.htm

and

http://sound.westhost.com/project36.htm

[brett]

Thanks RG.
Also found this at the ESP site, concerning using a MOSFET as a voltage follower for audio output.  It seems promising.
http://sound.westhost.com/project83.htm

As he says"How does it sound? Wonderful, regardless low or high volume. Entire spectrum from bass to high is perfect. It only needs a good preamplifier."

The question that I ask myself now is whether the zener-Darlington combination froviding resistance on the source of the output MOSFET is better than using a fixed resistor.  It seems great in theory, but would it be better in practice?  RG?  Anyone?

[gez]

If you can get hold of a back issue, Elektor had a 'FET Amp with Valve Sound' feature in the 11/2003 issue.

MOSFET output in a Mu-type configuration.  Never built it, but interesting schematic (liked the power arrangement)  

There's a schematic of a mark I type circuit which used a choke to help shunt DC past a speaker directly connected to the drain of a single power MOSFET.  Personally, I wouldn't chance it, but clever idea (about as simple as it gets)...

[Steben]

Isn't there a problem in clipping the OA rather than the MOSFET?
(I'm not familiar with discrete poweramps...)
Or is it the MOSFET that really gives the harmonics?

[brett]

MOSFET output in a Mu-type configuration. Never built it, but interesting schematic

Yeah, I've seen those schematics in a couple of places.  One is the Zen amp that RG mentions above.  Some people like it, others dislike it, but most are referring to it in audio/hi-fi terms.  I might build one and see how it behaves when driven hard.

Isn't there a problem in clipping the OA rather than the MOSFET?
(I'm not familiar with discrete poweramps...)
Or is it the MOSFET that really gives the harmonics?

That was one reason why I was going for a voltage follower.  The MOSFET would start clipping (gracefully, one would hope) well before the op-amp.  Hopefully, any nasty op-amp clipping would not be reproduced in the ouput (ie the MOSFET can't swing its output nearly as wide as the op-amp).

I'm going to incorporate some changes and re-post a circuit later today.

cheers

[Tim Escobedo]

Run at low voltage, those single transistor class a amps are not likely to ever run higher than about two watts. You'll need a pretty hefty power supply capable of providing lots of current and good heat sinking.

Here are a few more similar amps:
http://www.rason.org/Projects/transaud/transaud.htm

ISTR a Zen amp of some type that used a light bulb as a current source. Kind of a neat, bizarre idea. Would make a nice pseudo-tube conversation piece.

[brett]

Run at low voltage, those single transistor class a amps are not likely to ever run higher than about two watts

That's what I was aiming for.  Low power, maximum tone (given that it's solid state).

A 24V 1A wall wart regulated to 18V will give me 18W of heat to dissipate in total.  I expect up to 4W might appear in the speaker coils, and about 5W in the MOSFET, 3W in the Darlington pair and 2W in the load resistor (new schematic using a constant current source appearing soon).  So yes, the heatsinking will be considerable!

thanks for the comments, Tim

[brett]

Here's version 2 for those intersted.
Unfortunately it is more complex (now 3 output semiconductors), but I *think* it is significantly better.

I'll start working on pcb tomorrow.
http://members.optusnet.com.au/~jethro.dog/mypic157.jpg

[Steben]

MOSFET output in a Mu-type configuration. Never built it, but interesting schematic

Yeah, I've seen those schematics in a couple of places.  One is the Zen amp that RG mentions above.  Some people like it, others dislike it, but most are referring to it in audio/hi-fi terms.  I might build one and see how it behaves when driven hard.

Isn't there a problem in clipping the OA rather than the MOSFET?
(I'm not familiar with discrete poweramps...)
Or is it the MOSFET that really gives the harmonics?

That was one reason why I was going for a voltage follower.  The MOSFET would start clipping (gracefully, one would hope) well before the op-amp.  Hopefully, any nasty op-amp clipping would not be reproduced in the ouput (ie the MOSFET can't swing its output nearly as wide as the op-amp).

I'm going to incorporate some changes and re-post a circuit later today.

cheers

Still, after the first clipping of the FET (let's say at 80% of the OA clipping) the OA will clip itself, resulting in a "double clipping" (like a fuzz before a Tube screamer). I would instinctively put diodes around the OA that have offset voltage a bit higher than the FET clipping point. This will result IMO in a sweeter high gain setting.

[JimRayden]

And so, the complexity grows and grows, until we have yet another solid state amp to ruin our tube-world. Lol, just kidding. I really like the idea though.

Any sample sound clips anytime soon?

------------
Jimbo

[petemoore]

Something I'd like is a Small Amp, that sounds better than my Little 386-er that I use...it' ok...
 Brett, I appreciate the interest and everyone who's helping out this thread !!! ...And will probably begin construction of whatever it is ya'll settle on here soon.
 Mosfet outputs...some mfr. was making "Mosfet' amps for a time...wasn't that Marshall ? They were all the rage ... I never tried one though...

[Gringo]

Brett, I appreciate the interest and everyone who's helping out this thread !!! ...And will probably begin construction of whatever it is ya'll settle on here soon.

Same thing here I hope i can get the output stage semis (or a suitable replacement) around here  :twisted: