I want to bias lf412 with a 10k resistor from neg. rail to output. The lf412 output to a opa2134 buffer that is not in feedback loop.
Need advice. thanks
Need advice. thanks
my advice is to not use the lf412, certainly don't load it with a resistor - loop gain drops and distortion rises rapidly in these old op amps with resistive loading
I have measured low distortion from the very similar tl070 with a buffer in the loop and no loading of the tl070 output
I have measured low distortion from the very similar tl070 with a buffer in the loop and no loading of the tl070 output
jcx said:
I have measured low distortion from the very similar tl070 with a buffer in the loop and no loading of the tl070 output
What kind of chip for the buffer?
I was using a lm6171, like Walt Jung I think that buffers with gain are superior to just using unity gain buffers but it does require some attention to loop gain/phase shift issues
I had a overall closed loop gain of +4 and the lm6171 was providing a gain of +4 inside the loop, this means the tl070 was seeing the equivalent of unity gain closed loop conditions with the added phase shift of the lm6171 <~10 degrees from the ~ 20-25 MHz bandwidth of the +4 gain buffer:
tl070 GBW ~= 4MHz, lm6171 80-100 MHz GBW/4 ~= 20 MHz; (4 MHz/20 MHz) * 45 degrees ~= 9 degrees added outer loop phase shift
the lm6171 is nominally unity gain stable but at higher gain is less twitchy - tolerates higher C load, bypass C needn't be quite as carefully implemented
look for Walt Jung's website with his composite/multiloop amplifier articles
I had a overall closed loop gain of +4 and the lm6171 was providing a gain of +4 inside the loop, this means the tl070 was seeing the equivalent of unity gain closed loop conditions with the added phase shift of the lm6171 <~10 degrees from the ~ 20-25 MHz bandwidth of the +4 gain buffer:
tl070 GBW ~= 4MHz, lm6171 80-100 MHz GBW/4 ~= 20 MHz; (4 MHz/20 MHz) * 45 degrees ~= 9 degrees added outer loop phase shift
the lm6171 is nominally unity gain stable but at higher gain is less twitchy - tolerates higher C load, bypass C needn't be quite as carefully implemented
look for Walt Jung's website with his composite/multiloop amplifier articles
- Status
- Not open for further replies.