My 200W/4 class AB MOSFET amp.

[sparkybg]

Hi all.

Here's my first MOSFET amp. This is how it looks on simulator:

200W/4ohm (4V P-P input)
100W/8ohm

0.005% THD at 200W

Look at attached picture for schematrics.

Any comments are wellcome.

[Giaime]

ehm.... schematics?

[palesha]

This is July not April.

[sparkybg]

UPS!!!.

Sorry, the image was very big. Here it is.. (i hope)

[palesha]

It is still not here. U can send it to me.

[sparkybg]

Sorry guyz! It seems I cannot attach files. Maybe because I am a newbie.

I will upload my files alsewhere and put a link to them.

Sorry again.

[sparkybg]

Hope this will work:

200W MOSFET AMP SCHEMATIC

[Mr Evil]

A few comments: Blue LEDs are not the best to use for voltage references, since they are more noisy than other colours.

Your frequency compensation is quite complex. Does it have to be like that? The more complex it is, the more likely it is that the sim will differ from reality, requiring more work prototyping. Using an output stage with gain will make it that much more difficult.

The 0.22 resistors are on the wrong side of the MOSFETs. They should be in the source.

[bocka]

The 0.22 resistors are on the wrong side of the MOSFETs

The more complex it is, the more likely it is that the sim will differ from reality, requiring more work prototyping. Using an output stage with gain will make it that much more difficult.

I think that you'll get also problems with thermal stability because the thermal voltage tracking of your Vbe multiplier is multiplied by 10 (R4/R18) which then controls the Vgs of your MOSFETs.

R29 and C16 makes more sence in the collectors of Q16 and Q17 anthough i would recalculate the values

[sparkybg]

Thanks for yor reply.

I didn't know for the blue LED noisyness. The upper one is for a second stage cascode voltage reference only. I can be substituted with other type of LED or anything else then. For the other one - D7, it can be green or red or anything, just R1 and R7 must be recalculated, so thanks for sharing your knowledge.

About the frequency compensation: r12, r13, r14,c6 and c7 at the input are only to prevent RFI. They can be replaced with ordinary input filter - one capacitor in series with the input and one resistor for input impedance, so they are not critical. I've experinted with them and their impact on the frequency and phase characteristic is very little. C6 especially can be removed without any harm. R29 and c16 in the LTP too are not critical even with double values (all combinations) the impact is not so great. There is very little voltage amplification in this stage, so they are there only to prevent oscillations, but I think they are pretty ordynary approach. I've seen many amps with such compensation, and some of them are pretty good ones.
About c12 - again pretty standard approach. R29, C16 and c12 can be removed too, provided that the miller capacitor C5 is changed to 80-100pf, but this would affect the slew rate.

The amplification in driver stage is only to allow the output swing to be comparable with a BJT amp. This maybe will add some instability, I absolutely agree with you. Thats why the above frequency compensation is choosed. C11 cap too are placed for this reason - to swamp the frequency response above the audible frequencyes.

The output filter network id pretty standard i thing.

Why do you thing the 0.22ohm resistors are on the wrong side? Please explain this for me. They are only for better current sharing of the output MOSFETS? If they are properly matched I don't see a problem.

I am not very experienced in making audio amps, so all the above may be wrong. I will be glad to see any comment.

Thanks again.