this amp work?


2007-03-01 7:33 pm


  • aaaaa6.gif
    13.4 KB · Views: 369


2007-10-10 5:21 pm
This one is at least... unique. I'd keep a fire extinguisher handy - my first impression is that the output stage is gonna run away and you'll need to go catch it. Does it *simulate* ok?

I guess the idea was totally golbal-NFB-free, and not just a tweak on an existing topology? But four stages of current gain - with no Rbe's and no bias tracking is likely to get stupid on you. Might also drift around quite a bit unless Q3-6 were very well matched and epoxied together.

What have you got to lose - 60 volts won't bring down the building and there's nothing expensive in there.

I think the DC bias would be OK, but there would need to be a Vbe multiplier or some temperature compensation if this is class AB. However, as for the current mirrors used as the VAS, linear gain might be questionable. The problem is when the circuit is powered up, C1 and C2 will require current to charge up. This current will come from the mirror source transistors and be reflected in the mirror devices. This will cause the outputs to burn up because their bias is dependent on the VAS current. I would not even try this circuit unless you have a secret vendetta against some poor output transistors. :devilr: :hot: :flame: