choosing input valve

[Merlinb]

If the imperfections are below audibility, how is that "handicapping"?

OK, so how do you know when it is below audibility (at the design stage I mean, without having to build it and find out the hard way)? As a know-nothing newb, can I only get below this 'good enough' threshold with a 6SN7 running with a CCS? Or would an ECC85 with resistive load do just as well?

 

[SY]

I don't have measurements on an ECC85, so can't say. I've done enough testing on 6SN7 line stages to know that IF they're designed properly, I can't hear them in a bypass test. MJ's versions use mu followers, mine use CCS plus cathode followers, but we end up with pretty darn similar specs (0.00x% THD at 2V out).

 

[TheGimp]

RMB,

the 15,000Hz upper limit is imposed by the FCC to limit bandwidth of the modulated carrier. This was done to allow more stations to be crammed into the available rf spectrum.

 

[DF96]

And then often use transistor CCSs to try and scrub away all the pesky nonlinearities of valves!

A CCS load allows the true linearity of a triode to appear, if it is a linear triode (such as ECC83/12AX7).

The dynamic range problem was introduced by you, when you suggested that we could avoid the inherent non-linearity of BJTs by severely attenuating all signals to below 1mV. This would ruin S/N ratio, so I was surprised that you suggested it. However, you seem to be fixated on non-linear distortion as the only measure of an amp. An amp with 0.000001% THD but 25dB S/N and 300-3kHz bandwidth would sound horrible - a bit like a good telephone line.