Simple 85W amplifier

[keantoken]

BUT, you can still try the scheme in my schematic. I'd really like your feedback on that.

I think the "local feedback VAS" idea is very misleading. I think far more people have been confused by this concept than helped.

IF the VAS has higher distortion than the output stage, and IF the input stage contributes negligibly to the total distortion, then local feedback at the VAS may decrease distortion. But this is unlikely, and furthermore, miller compensation is the wrong way to apply this because it will simply conduct more distortion from the output stage. Degeneration is a better way IMO, if such a situation ever arises.

 

[GregH2]

I have to admit that I have always found that in simulation large miller caps always increase distortion. But I had attributed that to losing OLG and hence decreasing overall global feedback...

I think I should stick to SET amplifiers. I think I understand those.

I will try your compensation scheme at some point, thanks Kean.

 

[keantoken]

Remember the degeneration resistor R16, that's part of it too. Size this resistor to have at least 100mV or so across it (it can also be sized to make clipping symmetric if desired). In my example it is actually larger than it needs to be.

 

[keantoken]

I have to admit that I have always found that in simulation large miller caps always increase distortion. But I had attributed that to losing OLG and hence decreasing overall global feedback...

That is right. OLG is a multiple of VAS gain, so if you halve VAS gain you halve OLG. So adding local feedback to the VAS decreases both VAS gain and OLG by the same factor. Likewise, if you halve any other stage's gain, total OLG will be halved.

The difference is that you are describing the amp in terms of total OLG, and I am describing it in terms of individual stage gain. In context, what happens to the VAS happens to the rest of the amp as well, since all stages are inter-dependent.

 

[GEirin]

Greg: any progress of your amplifier?