AD797 and Super Regulator

[Onvinyl]

Hi,
I have the problem not to have a problem. Errm...
I tried to run a simple (inverting) preamp with a AD797, fed by superregs. It is absolutly stable *without any decoupling caps*. (When using small caps over the feedback -R as per datasheet)
Does this mean I don't need any? What worse case scenarios can I try to check out?
Rüdiger

 

[pinkmouse]

I've moved this to a new thread so it gets the attention it deserves.

 

[Onvinyl]

Thx Al,
Rüdiger

 

[SY]

Bounce the current around a bit (you can do this crudely with two resistors or more rigorously with an added transistor and square wave generator) and see what the recovery looks like. Inject some noise, both to input and output and see how it responds, then do it again with a few small values of capacitance across the output to simulate load wiring. Put some inductance in series with the load and bounce the load current around a bit with that extra bit of torture.

 

[peranders]

AD797 has a rather small phase margin at 0 dB gain. How much gain do you have now? The reason for having this cap is to compensate the input capacitance which can cause unstability if the feedback resistors are made too large in value.

 

[jan.didden]

... and there is also some info here:

http://www.linearaudio.nl/Documents/Regs_for_High_Perf_Audio_3.pdf

Jan Didden

 

[Onvinyl]

Thanks for your valuable input!

@Jan: I wasn't aware of your paper, I appreciate it!

@Per-Anders: at the moment, the Feedback R is indeed large, 60k. I have a gain of -30 now. I'm not wanting it as a line amp, but I'm examining working conditions for this chip in a riaa stage (yes, phonoclone...) Maybe I could and should go lower in values for the resistors. OTHO, the datasheet indicates larger cap values with lower resistor values.


@SY:
I'm not sure what you mean by 'bounce the current a bit' (language prob, I guess).

thanks so far,
Rüdiger

 

[peranders]

If you plan to use AD797 in an inverting RIAA with only this opamp I can say that it's rather unsuitable especially for a MM cartridge.

It's better to have a "pre amp" made of AD797 + an inverting RIAA amp with a regular noise opamp with good characteristics around 0 dB gain. You can check my QSXM2 for the basic set up. Replace my discrete opamp with a real one. Schematics can at the moment be found here

 

[Onvinyl]

Hi PA,
care to elaborate?
I use it as the second stage here
It works really good, but now I want to struggle for optimum performance (had a really bad scope back then...)

What fundamental problems do you see with this chip in this place?
Rüdiger

 

[jackinnj]

I'm not sure what you mean by 'bounce the current a bit' (language prob, I guess).

thanks so far,
Rüdiger

You "bounce the current" by subjecting the regulator to a sudden shift in load -- you can do this with a MOSFET switch and resistor -- observe the output of the regulator and look for instability on a scope.

National Semi, Linear Technologies and Texas Instruments have some good papers on regulator care and maintenance.

fwiw, I have not had oscillation problems with the AD797 in practice -- and I have used the AD825 and LT1115 as well with good results.

 

[SY]

That's exactly it. If you want to bounce between two currents, connect a resistor sized to pass the lower one across the reg output. Parallel that with a FET or bipolar that has a resistor in its collector circuit sized to (in parallel with the first resistor) pass the higher current. Drive the base/gate with a square or rectangle wave, hit it with pulses, torture the poor reg and see if it screams.

 

[jackinnj]

hit it with pulses, torture the poor reg and see if it screams.

in other words, beat it like a red-headed step child.