my first amp, please could someone check it???

[Si992]

This is my first amp design, is the any chance someone could check it to make sure it would work please?
Si

[cunningham]

Looks like a typical audio design. At first glance it looks good. Build it and see if it works.

[jaycee]

looks like a Self Blameless with VAS protection and a MOSFET output stage.

1000pF for miller compensation seems a bit large to me. Also you have an RC on the VAS base for more compensation - is this really neccesary ? If it's that unstable the sound won't be good.

Personally I would use driver transistors on the MOSFETs but that's just me. The 649/669's should work well for that.

[jcx]

coupled bias reference for the ltp and vas is not good, self fought asymmetric slew rate caused by this false economy in his early blameless designs - completely separate ccs circuits are too cheap to bother fixing up any interactions in your circuit

higher mirror degeneration R values would reduce the noise contribution of this stage - they should at least equal the diff pair emitter degeneration R values

2 pole compensation is interesting but the 1000 pF, 1 K loads the vas, eliminating some of the potential extra audio frequency loop gain, putting the network around a buffered vas might work, or you can rework the T network values; ~= C reduces output loading on the unbuffered vas

power mosfets are not particularly high input impedance at high frequencies, buffering the vas will reduce distortion appreciably - although the supply V suggests this is just a line amp?, if the fets are small enough your T network could be the biggest vas load

gate series R may be useful to prevent parasitic RF oscillations in the mosfets

the output fet source series R seem pointlessly small for mosfet outputs

[ilimzn]

...And the 2SJ162 body diode connection is drawn wrong suggesting a MOSFET with reversed D and S terminals...

[sam9]

1000pF for miller compensation seems a bit large to me. Also you have an RC on the VAS base for more compensation - is this really neccesary ? If it's that unstable the sound won't be good.

That just two pole compensation. Reduces THD at higher frequencies. Yes, it can be unstable but that usually only happens if you make the cap too small. 1000p is probably large enough to avoid that and once everything is working well it may be worthwhile reducing it in steps until signs of instability show up and then back up a step or two.

[sam9]

Continuing, 1000p is appropriate as long as the other is 200p. Typically they have a ratio of 1:5 or 1:10 or somewhere in between. The 1k resistor is also worth thinking about. Usually, 1k to 10k is workable range. Increasing it to 2-5k may add some stability is the caps are decreased.

All this may sound like I'm just quoting Self, which is true, but I've built a few of these and experimented with the values and confirmed for myself trhat it works out.

[richie00boy]

The schematic is in need of dots where the connections are -- especially around the LTP current source, which looks a little wrong to me but I'm having to guess without any dots, and the transistors given IDs e.g. TR1. That way we can tell you about which bit of the circuit we mean

As others have mentioned the MOSFETs look wrongly connected/drawn.

[sam9]

...And the 2SJ162 body diode connection is drawn wrong suggesting a MOSFET with reversed D and S terminals...

More than that. I think you wany the 2sk1058 on the top (+rail) and the 2sj162 on the bottom (-rail). The source side should face each other while the drains connect to the rails. It's very easy to loose track of which way around these things go --at least it is for me.

[bobo1on1]

Reverse the diode, also the + connection of the 220 µF elco must be connected to gnd, this is because you used pnp transistors in the ltp.