Just another current feedback amplifier

I employed a few tricks such as improved wilson current mirrors, dual BC847/857 packages and CFP output with parallel BJT drivers so I can bias the output stage fully into Class A even into 16R. I also used the DC servo from hpasternack's A4. So far, I'm getting 0.0007% THD with 2Vrms input, ~7.8Vrms output into 16R at 20kHz and I'm quite happy with that result, but I was wondering if there was any way I'm missing that I could optimize this circuit further.

I'll post further measurements later, currently not home as I post this!
 

Attachments

  • image-30-2.png
    image-30-2.png
    40.4 KB · Views: 463
Last edited:
I have decided to call this amplifier the Mirza! I slightly lowered the gain to 4.64 and got rid of the parallel transistors for a single TO220 per side. Quiescent current is ~473mA per device.

1694088084247.png


Power and Vrms measurements, 2Vrms in at 1kHz, 16R, 32R, 64R, 150R, and 600R loads.
1694084451603.png


Gain/phase margin.
1694085413033.png


Loop gain.
1694085268316.png
 
Last edited:
As for the changes I made, I increased R2/R3 to 330R, which has changed the current to 1.25mA and the power through Q3/Q4 to a far more reasonable 20.64mW. I also decreased Q25/Q26's bias to 234mA as I decided I didn't need the output stage biased into class A into a 16R load. I also changed the values in the gain cell C1 || R17 which has had a corresponding effect on the loop gain and phase margin.
1694164512143.png

1694164626165.png

1694164541702.png
 
Could It be used as a headphone amp? If so I wonder if a 4 gang? volume pot could be used and a circuit could be made so that the headphone socket will switch the amp to a lower gain and disconnect the speakers as soon as the headphone is inserted? Sorry for this diversion. I just noticed it could drive high impedance, and it would be a cool feature.
Cheers!