Discrete preamp - seeking opinion and advice

[audio_tony]

Today I decided to rekindle an old project I completed about 4 years ago.

After various tweaks in my Spice software (Tina Pro), I arrived at the circuit below.

Can anyone offer any suggestions for improvement, without making any significant changes to the basic topology, as I already have a couple of boards made for this.



And the frequency response.



And here are some other (so far discarded) examples..







Thanks,
Tony.

[analog_sa]

I would immediately short circuit a few of the coupling/NFB decoupling caps. Once i had this built with an LM394 at the input and offset was negligible after trimming.

[carlmart]

Fixing the input and regulating the supply should let you eliminate C1 at least.

[audio_tony]

Originally Posted by analog_sa

I would immediately short circuit a few of the coupling/NFB decoupling caps. Once i had this built with an LM394 at the input and offset was negligible after trimming.

Thanks analog_sa

DC offset in my simulator is currently 2.83mV on the output and -130.28mV on the input. I suspect I'll have to retain the input coupling cap, as there will be a pot there. I tried to get the input impedance as high as possible, so that I can use a low value poly cap (or similar) there.

I removed the feedback cap (C1) and the DC offset rises to -1.09V (which was to be expected), although as you say, I could probably trim that out.

Thanks,
Tony.

[audio_tony]

Originally Posted by carlmart

Fixing the input and regulating the supply should let you eliminate C1 at least.

"fixing the input"? In what way? Do you have any suggestions?

It will be powered with a regulated PSU, probably LM317 / 337 or similar.

Thanks.

[mjf]

hello.
have you tried out to build in some small emitter res (47......100 ohm)?
or a constant current source (j-fet) for r3?
greetings............

[Bonsai]

You might want to try loading the diff pair with a mirror and then also degenerating the diff transistor emitters (typical value of 22 to 100 Ohms). You should get a dramatic drop in distortion doing this.
You are using a 47k/5.6k feedback network. I think you could easily drop this to 4k7/560 Ohms - this will reduce your offset as well

Another step to improve this circuits performance would be to look at an active buffer on the output like a diamond biffer or a class A NPN/PNP putput stage (especially if you drop th e feedback resistor values).

good luck with your circuit!

[Bonsai]

Yes - replacing R3 and R8 with current cources will also improve things dramtically.

You should be shooting for around 10ppm distortion - easily acheived with this type of configuration.

[tiefbassuebertr]

Originally Posted by Bonsai

You might want to try loading the diff pair with a mirror and then also degenerating the diff transistor emitters (typical value of 22 to 100 Ohms). You should get a dramatic drop in distortion doing this.
You are using a 47k/5.6k feedback network. I think you could easily drop this to 4k7/560 Ohms - this will reduce your offset as well

Another step to improve this circuits performance would be to look at an active buffer on the output like a diamond biffer or a class A NPN/PNP putput stage (especially if you drop th e feedback resistor values).
good luck with your circuit!

This are the right steps, if I need a preamp with voltage gain between 5-times until 20-times. But in most cases, unity gain is for the most currently power amplifiers enough (cd players and currently RIAA head amps supplies output voltages above 1Vss in most cases). So I would use only a buffer after the volume control for preamp applications. If you use your preamplifier for driven long cables or for driven more power amplifiers (bi-amping resp. multi-amping), please note, the importance of high values of idle current arround 1A.

I even believe that this high quiescent current is to be selected in a pre-amplifier ever, although I don't had a technically plausible explanation for this. This opinion I had only through own audible impressions by compare of different values of idle current
I choice 1A by many users. To reduce the idle power, I have downscale the power supply to only 9V directly at the POWER MOSFET IRF520 without audible disadvantages (for approx +/-1V3ss undistorted output voltage) . This preamplifier outperform all commercial products, that I know (because the present mostly very high voltage gain - I think), independend of the used cables to the following power amps (a 25-year-old oil filled 5Kpotentiometer from penny + giles also helps). So it is amazingly for me, that there are no such commercial preamplifiers, even not from Mr. Pass.

The chosen basic circuit are the "Power Follower" from Andrea Ciuffoli and it was downscale of me for 9V supply voltage and 1A idle current (18W loss idle power for both channels). The basic circuit of "Power Follower" you find by this weblinks:
the Power Folllower
Project 83 - MOSFET Power Follower (as preamp in the mentioned manner same results)
Power follower 99c Vs Zen (more about Andrea Ciuffoli's circuit)


Holistic view is quite another question to ask: Where do I create the voltage gain: in the pre or power amp?

[analog_sa]

Originally Posted by tiefbassuebertr

please note, the importance of high values of idle current arround 1A.

Very interesting post but obviously not what this thread is about. Would you consider starting a new thread about your preamp?