Quote:
Originally Posted by drtechno  Ive used 6386 when building true balanced line stages and in reality there are several tubes out there that can be used, but it all boils down to the fact is that the dual triode sections need to match. I remember that there was trim adjustments in the Fairchild 660/670 for that reason.
what would be really cool is to come up with an auto bias scheme maybe with some op amps that balance the gain between the top and bottom tiode, therefore no calibration required. |
One of the problems with bias is that in a compressor, the bias (and thus, operating point) changes with signal - so the bias to get the identical amount of gain in the top tube may be different than what is required in the bottom tube (for best distortion characteristics). Fair enough in a straight gain stage. But changing the bias point with operation (ie. when you change the CV in a compressor) is tricky because if you do that, you get a thump (CV feedthrough) if both the DC and AC gains don't match throughout pretty much the whole curve. You can tune for minimum thump and distortion between, say, 0dB and 10dB of reduction, but then as soon as you hit 15dB you motorboat. So tweak it in to increase distortion and you get 20dB without thumping but the THD at 0dB GR is about 15%. It's a no-win situation.
Although that Altec thump can be a neat effect on drums, that's what both tube selection and trims are for - to get rid of it. Some variable mu comps (based on 6SK7's and such) have plate load trims, cathode trims, and screen grid trims, to try to warp the curves to be useful - to get rid of both distortion and CV feedthrough.
In theory, an 'auto bias' circuit for a compressor would optimize the bias between the two halves, and also modify the application of the bias to minimize feedthrough. Trivial to do that with some kind of microcontroller, but isn't the simplicity of a tube compressor part of its beauty? Putting software in there just seems wrong to me.