how about LME49720 driving LM3886?

[mondogenerator]

im considering using LME49720, inputting line level signal, preamplifing it, before power amplification using LM3886.

DOes anyone out there think this is worth it? and also are there any line level pre amp circuits out there for the LME49720, as ive seen mainly balanced line drivers and phono stages.

Any help out there?

thanks

 

[linuxguru]

The Yuanjing buffered LM3886 (and its successor) does exactly that, with a socketed NE5532 instead. There's no problem rolling an LME49720 in place of the NE5532 - I've tried it and it works well.

 

[peranders]

I have used an AD8620 (expensive) to feed an inverting LM3886 with only 1 kohms inimpedance. The the other half of the opamp is used for a DC servo.

 

[twest820]

Depends on your gain budget through the signal path and output requirements. Generally the numbers come out better with a power amp gain in the vicinity of unity than with the more conventional 25 or 30dB. As the 3886 has a minimum gain of 20dB sticking a 20dB attenuator in front of it makes sense. For unbalanced input that's just a couple resistors, but a 49710's not a bad choice for an attenuating balanced to single end conversion.

Circuits wouldn't be any different from any other voltage feedback op amp. As you're mentioning the 49720 I assume there'd be a pair of 3886s.

If you don't mind the phase shifts associated with the capactive compensation needed for stability you'll get better performance using 497x0s to form a composite amp---performance of the 3886 is input stage limited so putting an attenuator in front of it isn't exactly ideal. 'bas recently posted a composite amp schematic that's a good starting point. I was looking at something similar a while ago but abandoned the approach as I had a requirement for flat phase response (some of Walt Jung's Avol limiting circuits offer flat phase but implementing them seems to require resistances of a megaohm or more, which I preferred to avoid).