Here is a 1W mosfet class A amp. Uses a laptop power source (no high voltages).
(http://www.diale.org/img/wamp.png)
Some oscilloscope pics:
(http://www.diale.org/WAMP/waveform3.png)
(http://www.diale.org/WAMP/waveform5.png)
Overdrives nicely with a booster in front.
Here's the poweramp only:
(http://www.diale.org/img/1w_mosfet_amp_f.png)
Have fun!
Nice work Tiago! I remember you playing with this a while ago. I will put this together after I move in a few weeks. The clipping does look pretty good.
Now you need some tone controls :)
Thanks!
The 22 Ohm should be at least rated 10W. I've used a 25W wire wound resistor.
Obviously the mosfet needs a heatsink... and if in contact with the metal box it also needs the typical insulator.
Try different mosfets, not critical (maybe the IRF7* will ring to much).
About temperature: you should be able to comfortably put your hand on the mosfet heatsink - 60 degrees Celsius (coffee drinking temperature).
Cheers.
I'd suggest a little mod. (Not actually tested, so from my memory only).
Namely, use a 2N7000 or a BS170 in driver/preamp stage and a heftier IRF (like 730 or 630) with bigger gate charge (gate cap.) in power stage. This combo will up the overall gain of the circuit and compensate for the fizziness.
I played with such pairings ... many moons ago, so from memory only.
600ma quiescent current for a 1W amp? (i.e., drawing 600ma when not doing anything at all)
Commendable gedanken experiment, but what is the sonic advantage, compared to an LM-386-4 with a 9V battery or sextet of AAs?
> Commendable gedanken experiment, but what is the sonic advantage, compared to an LM-386-4 with a 9V battery or sextet of AAs?
Harmonic distortion (clean and overdrive), great cleans and lovely overdrive sounds, high output impedance, stability, no NFB, low parts count, ...
Because it has so few active parts you can identify any hardware value change with its sonic consequence.
It seems a simple amp, almost naive, but as any other good amp you'll need to learn how to use it and play it.
It is a good project for learning about amplifiers and biasing active devices and the sounds that can be obtained.
Besides all that a very fun amp to play!
Just build the poweramp section and use a OD pedal to overdrive it, you'll be suprised ;)
Cheers.
I ran an .op simulation in LTSpice with the suggested mosfets.
I'm getting 4.5v at the drain, independently of the value of R5 and R6.
R5 controls the quiescent current, (18v-4v5)/22 ohm = 610ma
I guess other mosfets will bias a bit higher or lower, so R5 should be adjusted.
mac
The amp is a version of a well-known Hi-Fi schematics by Nelson Pass (Nelson Pass Labs).
Hi-Fi and even Hi-End enthusiasts say the circuit has a very pleasant and soft sound.
So, we can expect good guitar sound, too. Why not?
>The amp is a version of a well-known Hi-Fi schematics by Nelson Pass (Nelson Pass Labs).
> Hi-Fi and even Hi-End enthusiasts say the circuit has a very pleasant and soft sound.
> So, we can expect good guitar sound, too. Why not?
Yep, class A amps are 60 years old... there are a lot of versions floating around ;)
Cheers.
tca - there will be some (not me, that's for sure) that will ask "what is a 1m capacitor?" as shown at C5. or similar wording. perhaps you could clarify for their benefits.
> tca - there will be some (not me, that's for sure) that will ask "what is a 1m capacitor?" as shown at C5. or similar wording. perhaps you could clarify for their benefits.
Really... ok!
Q: What is a 1m capacitor?
A: 1m=1mF=1000uF
8)
Nice! How does that quiescent current compare to a lm386 power amp with a similar drive capacity? Any chance for a gutshot? How do you go about mounting the Mosfets to the enclosure while keeping them isolated?
I used to experiment with such mosfet amps and found designs like this doesn't make any advantage over typical bipolar amps; like TDA20..
The key to get the tube-like sound out of a mosfet is high output impedance and that means an "Output Transformer", and that's the problem; if you're going to use an OT why not build an actual tube amp?!
If you are really interested in this subject, search for "KMG's mosfet amps" and you'll find some interesting designs using mosfets with OTs.
Great project, tca! :)
Thank you! ;)
Quote from: Mark Hammer on July 02, 2015, 09:30:43 AM
600ma quiescent current for a 1W amp? (i.e., drawing 600ma when not doing anything at all)
Starts to sound like vacuum tubes :P
> Nice! How does that quiescent current compare to a lm386 power amp with a similar drive capacity?
LM386 quiescent current is about 5mA. The WAMP takes > x10 more!
> How do you go about mounting the Mosfets to the enclosure while keeping them isolated?
Using insulators:
(http://www.diale.org/img/insulators_mosfet.jpg)
and thermal grease (https://en.wikipedia.org/wiki/Thermal_grease) to increase thermal contact.
Cheers.
> I used to experiment with such mosfet amps and found designs like this doesn't make any advantage over typical bipolar amps; like TDA20...
You'll need to justify that claim about the advantages of the TDA* ;)
There are a few advantages in using a single mosfet driving a speaker. Just to name a few:
- Aesthetic simplicity, just one active device
- no NFB
- no severe rail sticking when overdriven
- overdrives nicely
- high output impedance
- natural short-circuit protection (all current is already flowing ;) )
- low parts count
Disadvantages:
- lots of heat
- needs a very regulated power supply
- (...)
> The key to get the tube-like sound out of a mosfet is high output impedance and that means an "Output Transformer", and that's the problem; if you're going to use an OT why not build an actual tube amp?!
ah, "tube-like-sound". I think one should get rid of such an obsession when playing with ss devices. There are a few ways to get high output impedance in audio amplifiers, common source and class A configurations is just one of them.
Similar amps with a OT works nicely but the bias gets a little tricky, also with a inductor. Mosfets can drive speakers directly without transformers and the quadratic dependence of drain current as a function of the input voltage gives a pleasant harmonic distortion.
If the obsession goes to far just repeat the math of the Fetzer-valve-triode-emulator for mosfets...
> If you are really interested in this subject, search for "KMG's mosfet amps" and you'll find some interesting designs using mosfets with OTs.
I've seen those, they are class AB. Don't like how they sound but that's irrelevant, matter of opinion.
Cheers.
The drain to gate resistors are feedback
> The drain to gate resistors are feedback.
Correct... but local.
whats negative (pun) about negative feedback?
Great thread by the way tca!
> whats negative (pun) about negative feedback?
I can not explain it better than Nelson Pass so here's the text: Audio distortion and feedback (https://passlabs.com/articles/audio-distortion-and-feedback)
P.S. (edit)
Although is about hi-fi the physics is the same.
nice. thanks!
Gus said it like it is. Biasing resistors ARE the NFB. These are "Schaded" circuits. The biasing arrangement makes them IRF's to sound more "triode-like" ... if you appreciate such buzzwords.
But ... if you desperately crave "a little more" of NFB it's so easy to stick a small .1 - 1 Ohm wirewound into the output IRF's source ... and up the volume knob a notch ... or proportionally increase the biasing/NFB resistor to compensate ...
> Gus said it like it is. Biasing resistors ARE the NFB.
Agreed.
> These are "Schaded" circuits. The biasing arrangement makes them IRF's to sound more "triode-like" ... if you appreciate such buzzwords.
Sounds better that way comparing with the fixed bias arrangement.
> But ... if you desperately crave "a little more" of NFB it's so easy to stick a small .1 - 1 Ohm wirewound into the output IRF's source ... and up the volume knob a notch ... or proportionally increase the biasing/NFB resistor to compensate ...
I know what you mean, but I'm getting more and more the feeling that when it comes to these simple circuits the decision to include of an extra component should be taken after a long period of listens tests.
Because I'm running this with a 18V laptop powersource and limited to 1A max (heatsink restrictions) that 100k drain to gate resistor is on the verge of its acceptable value due to the drain-gate capacitance. Drastically increasing, or only for a small amount, that resistor will require to alter the input and output caps values and, at least for my ears, the pre and pos filtering will change dramatically the sonic character of the stage.
Obviously one can fine tune all the components values for a hand chosen device.
Cheers.
IRF's are fun