discrete op amp

[unclejed613]

thought some here might find this interesting, has a lot of similarity to a power amp design. original design came from the Apt-Holman preamp, except that the Holman had a fet and a bipolar in the diff amp, but that causes a lot of offset, so i replaced the bipolar with another jfet. also replaced the RIAA network with just a resistor in the feedback loop.

-3db points are 2.5hz and 35khz

distortion: -60db (as-is with no tweaking) mostly even harmonic.

[AndrewT]

Hi,
try moving C3 from the drain of J1 to the gate of J2.

Does it still work?

Will the hi Freq response change if R13 is reduced substantially?

[Gigapod]

unclejed613:
Nice design, simple to understand.
A few questions:
- Did you just simulate it or have you actually built one?
- Temperature stability? PSRR? Output impedance?
- Why do you think one should/would use a discrete op-amp vs a modern audio opamp (2604, etc)?
Very interesting posting, hope you cna show us a picture soon of the working prototype.

AndrewT:
Can you explain your suggested changes/questions?

[AndrewT]

Hi Giga,
moving C3 takes out the Miller compensation and gives NFB over two stages rather than one. The Miller comp cripples the LTP.

I think the rather narrow 35kHz bandwidth is due to the gain being set to near +60db. I suspect that this discrete will never be used like a true opamp and it would be fairer to set the gain (and see the CL BW) to nearer where it will be in practice.

[Tim__x]

Your regulators don't regulate any ripple beyond Vbe-Vce(sat).
Your PNP transistors are backwards.
As Andrew said, the gain is unrealistically enormous.
Other than that the design should work, though it is a little lackluster.

[Gigapod]

Thanks AndrewT,
I consider your posts some of the best, most informative in this forum.
Keep'em coming!

[AndrewT]

Hi Giga,
thanks, but not everyone agrees with you.

[audiohead]

Q5 is backward as shown, and would not function as a current mirror.
art

[Leolabs]

Yes,Q5 is a drawing error.

Can the BJTs/FETs be subsituted with Japanese types???

[unclejed613]

tnx...... it's what i get for working late hours..... q5 and q7 are backwards............. i'll redo it and rerun it.....

actually the circuit in it's original form (with a 2n3904 in place of j2) was a very successful phono preamp in the Apt-Holman preamp

tnx for the heads up out there...... unfortunately sims rarely tell you that you just let the "magic smoke" out of components..... those 2 backwards pnp's definitely would have given a nice smoke and light show.......