J-Fet folded cascade headamp, please critique

[aparatusonitus]

Hi word,

Schematic attached represents my first full discrete high power headamp (for mine 32 and 300R headphones), also I would like to use it as a small power amp during summer (already have a Class A SE mosfet amp for speakers, but it's way to hot to live with during summer). I want to know if there is something wrong with design, particularly if there is a better way to apply compensation. Power supply will be fully regulated and separated for input stage (few more volts then 24V shown) and triple EF buffer. Mind you, I am not an EE, I just keep an aye and take notes from big boys around this forum, so be nice, thank you!

[TzeYang]

is it possible for you to post some square waves driving a 32ohm load?

Thanks.

[aparatusonitus]

Hi TzeYang,

Thanks for replay...first image 1kHz step response just before clipping occurs driving 32 Ohms.

[aparatusonitus]

Same condition 10kHz step response

[aparatusonitus]

...100kHz step response

[AndrewT]

Hi,
does the current from D3 defeat the current mirror?
C1 looks a bit small, unless you have a very high source impedance.
What's the dissipation in each of Q17 & Q18?

[aparatusonitus]

Originally posted by AndrewT

does the current from D3 defeat the current mirror?

Hi Andrew,

Thanks for replay,...yes it does, but distortion without cascade are 30db higher (second harmonic)

Originally posted by AndrewT
C1 looks a bit small, unless you have a very high source impedance.

Should I put something like 100pF instad?

Originally posted by AndrewT
What's the dissipation in each of Q17 & Q18?

It's about 220mW, a bit high for TO-92, but I would eventually put small heatsinks on them.

[AndrewT]

Originally posted by aparatusonitus
yes it does, but distortion without cascade are 30db higher (second harmonic)


Should I put something like 100pF instad?



It's about 220mW, a bit high for TO-92, but I would eventually put small heatsinks on them.

can the mirror cascode be applied outside the mirror voltage sensors?

1k0 and cap ranging between 330pF and 1nF would be more effective.

4.6mA in the drivers seems a bit light for a 32ohm ClassA output stage.

[aparatusonitus]

Originally posted by AndrewT
can the mirror cascode be applied outside the mirror voltage sensors?

Probably yes, but I've not tryed that...but I will try that and report back.

Originally posted by AndrewT
1k0 and cap ranging between 330pF and 1nF would be more effective.

Ok, thank you.

Originally posted by AndrewT
4.6mA in the drivers seems a bit light for a 32ohm ClassA output stage.

But standing current is twice as that?

[aparatusonitus]

Gues you are referring to Q9/10/12/13 predrivers not Q23/24 drivers?