2
\$\begingroup\$

Class B PreampCircuitI'm attempting to design a class AB guitar amplifier with a class A preamp. I'm a novice and this is the first actual project I'm attempting.

What are the possible caveats, as well as my choice in MOSFET? Should I include a feedback loop into my preamp to reduce crossover distortion? I'm going to drive an 8 ohm speaker.

Should I change anything in this circuit if I were to power it off a 24v 2A supply? Edit: I ended up redesigning it with a class B preamp for efficiency, I don't know if its any better, but it seems to work

\$\endgroup\$
3
  • \$\begingroup\$ What is the 33 ohm potentiometer expected to do? \$\endgroup\$
    – Marla
    Apr 7, 2022 at 17:49
  • \$\begingroup\$ It was put there to limit the output in the original circuit as a fixed resistor (to prevent headphone damage), but I wanted the option to adjust it on the fly. I don't see the point of having it there, but I figured it wasn't doing any harm. \$\endgroup\$ Apr 7, 2022 at 17:57
  • 1
    \$\begingroup\$ You're trying to bias your MOSFETs as though they were bipolar transistors. Which they aren't. \$\endgroup\$
    – user16324
    Apr 7, 2022 at 20:30

4 Answers 4

2
\$\begingroup\$

If you choose FETs with Vgs(th)= 2 to 4V then a RED or Yellow 5mm LED will work even with an LM741 as simulated below. A CMOS Op Amp will give more power to the FETs. The FETs must be low RdsOn preferably 1% of speaker load, but will still need a large heatsink with insulators + thermal grease.

enter image description here

This was simulated with an LM741, but there are rail-to-rail CMOS types with more output range. The gain is 1M/10k = 100 which could also be 10M/100k or any ratio you need which affects BW = GBW/Av e.g. GBW=2MHz/ 100 = 20 kHz BW.

Using diodes or a 2V LED with low current to get 1.5V to match Vt threshold of the FETS , more current from the 10k current limiter will raise Vgs and thus Id quiescent current and heat loss in FETs. I used Vt=1.5 Beta=20.

The large triangle wave 100 mVin 10V out resulted in 8.2W avg speaker power and 3.4 W avg. loss per FET thus 8.2Wout/15W total for 55% efficiency with about a 10kHz BW.

Bass-boost or mid-range cut can be added later unless specified.

\$\endgroup\$
3
  • \$\begingroup\$ I ended up modifying the circuit here and it seems to work ok. Not very efficient, but it seems useable, although me using the feedback for gain seems iffy. \$\endgroup\$ Apr 8, 2022 at 22:23
  • \$\begingroup\$ Front end N.G. Look at the current. Do what I said, your front-end experience is lacking on biasing. \$\endgroup\$ Apr 8, 2022 at 23:38
  • \$\begingroup\$ Don't expect much more than AV=50 from a single BJT stage tinyurl.com/y8bmgawb\ \$\endgroup\$ Apr 9, 2022 at 0:03
3
\$\begingroup\$

The most obvious problem I see is that you've got the MOSFETs the wrong way. MOSFETs have a sort of built in diode called the body diode, so you generally want to make sure that it is not forward biased. The way your schematic shows them, this is basically what you have as your output circuit:

schematic

simulate this circuit – Schematic created using CircuitLab

You can see that this will pretty much just short out the power rails.

I see several other issues in the circuit as well, what I would suggest is that you do some more research and look at some existing designs and see how the devices are connected and biased. A good place to start is diyAudio.com.

\$\endgroup\$
2
  • \$\begingroup\$ The problem with the MOSFETs being backward is because I'm unfamiliar with the schematic symbols. This circuit isn't really original, as it is essentially this circuit reviseomatic.org/help/s-push-pull/… without the feedback loop, using this circuit content.instructables.com/ORIG/FA9/H9UL/GJQEBPD0/… as a preamp. \$\endgroup\$ Apr 7, 2022 at 17:25
  • 1
    \$\begingroup\$ @DanielOren-Ibarra Okay, I see that you had the symbols flipped. You can see the importance of this though, if you had actually built that it wouldn't have gone well. That pre-amp stage is using a single ended supply and is probably not going to give you the full voltage swing that the output stage needs, so there's another thing to look at. Also if you don't have one already get a simulator like LTspice and learn how to use it. \$\endgroup\$
    – GodJihyo
    Apr 7, 2022 at 17:57
2
\$\begingroup\$

Your use of LEDs as biasing elements in the power section is iffy at BEST - and almost certainly not going to work. With bipolars you use silicon diodes, whose forward voltage is adjusted by varying the resistors to get a reasonable bias current in the output transistors. If the diodes are well-coupled thermally to the output transistors the diode voltage will track well with the bias requirements of the transistors.

None of this is going to work in your case. Even if you can find LEDs which are nominally the right voltage drop to match the FET gate voltages, the voltages won't track with temperature. Either you'll get increased distortion as the LED voltage gets too low, or excess heat dissipation (or even thermal runaway) if the LED voltage gets too high.

And getting the LEDs to match the FETs at setup is not in any way guaranteed. MOSFET turn-on voltages are notoriously variable from unit to unit. Take a look at the data sheets, specifically Vgs(th).

\$\endgroup\$
1
\$\begingroup\$

Instructables frequently show parts upside down.

Your class-AB output stage has no idling current control. One or both Mosfets might be in class-A or in class-B.

Some Mosfets will draw a huge current all the time depending on their threshold voltages.

Most amplifiers powered from 24V and using bipolar transistors can produce an output of 6.25W into 8 ohms. Your Mosfets might produce only 1.5W depending on their threshold voltages.

ALL audio amplifiers use negative feedback.

\$\endgroup\$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.