[Brinkman]

Here's a link.

In summary, biasing op-amps to operate Class-A eliminates crossover distortion in the chip with the tradeoff being decreased efficiency (re: increased heat dissipation and decreased chip life-span).

I am wondering if anyone here has tried this and what results they may have had (especially Jim). Were the improvements audible?

Thanks.

[sendler]

Here's a link.

In summary, biasing op-amps to operate Class-A eliminates crossover distortion in the chip with the tradeoff being decreased efficiency (re: increased heat dissipation and decreased chip life-span).

I am wondering if anyone here has tried this and what results they may have had (especially Jim). Were the improvements audible?

Thanks.

.
I tried the class A bias mod in my pre which is a stepped attenuator into an opamp for a gain of +9 db, followed by a class A mosfet output stage configured within the feedback loop. I used a Jfet configured to 2.2 ma.( a little excessive I know but this is still dissipating less than half of the rated power of the opamp and since this is followed by another stage I would expect that the drawbacks I experienced would be further magnified in the more common situation where the opamp is the final output device. ) from the output of the opamp to V-. I listened back and forth several times but found the sound to be better without the mod. It seems that if you already use a great opamp such as Linear Tech LT1360, this mod may be a step backwards. The midrange improved slightly, taking on an additional ease, but the bass became slower,slightly mushy and seemed to lag. The highs lost some microdynamics and sparkle lending to a more boring sound. Audition for yourself but don't assume this mod will be an automatic improvement. I was rooting through my bins the other day and found the bag of jfets if you want them to play with.

[Brinkman]

Thanks for sharing Scott.

Maybe as I learn more in my engineering tech classes I'll be able to experiment with this.

As it is, sounds like there needs to be an op-amp designed to be run specifically Class-A, assuming elimination of crossover distortion is a worthwhile pursuit.

Best.

[sendler]

My notes were from 2002 so it's not real fresh in my mind but I assume I was using my then and current favorite the LT1360. I didn't spend a lot of time with the concept as I quickly decided it didn't sound better with that amp and wasn't interested in running through all of the opamp comparisons again.

[emac]

I had tried biasing the OPA2107 I use in my USB+ Constantine DAC.  I thought at first that it was an improvement, but I didn't notice any changed when I reversed the mod.  So, take it for what it's worth. 

[reuben]

I tried it once, and liked it in the audio stage of a Kenwood KT-7500 tuner. Using matched FET's, I set the bias at about 2.1mA (more current than the op amp would ever need to source, thus keeping the output in class A).

The OPA2604, already my favorite in this application, seemed a little smoother after the mod. It would be a major pain to switch back and fourth, so I left the mod in. Despite the FM frequency response cutoff at ~15KHz, 'the top end seems to go on forever' (as Jon Risch so aptly described the OPA2604).

Results will vary with application, I'm sure. Keep in mind that this mod only forces the output stage of the op amp into class A operation.

Regards,
Tom