AC vs. DC filaments
Date: Sat, 27 Nov 1999 22:57:40 EST
Subject: Re: 2A3 vs 6A3 vs 6B4G
In a message dated 99-11-27 22:38:15 EST, you write:
Turns out the DC filament excites a higher
order term (than the 3/2 law Ep/Ip) leading
to more upper order terms than AC
It must be because of the gradient of voltage along the length of the fil,
effectively changing the shape of emitted electrons–the pattern of
emission in a way. It adds yet another variable to the equation that
equals total plate current… but a strange one?
This is pure conjecture.
I’m still working thru the math, but if you imagine an incremental triode (physically snip a small piece of tube thru the cathode, grid and plate) with AC each incremental piece looks pretty much the same, so the superposition of all the increments is the same as each piece. With DC, the characteristics of each incremental triode is just a bit different, so it’s sorta like paralleling a bunch of similar type tubes. You notice that these defocus just a bit in terms of clarity… same thing’s happening. Structure wise, this causes a slightly larger change in MU with change in bias than with AC. Translated to the frequency domain, the higher order distortion components are subtly intensified, as the curvature around the lower current operating points is slightly higher.
You can also see the same effect by taking a conventional cathoded tube (e.g. 6SN7) and run it at very low plate current (say 50 microamps) and listen to the resulting sound.
Strange but oddly consistent with much seat-of-the-pants wisdom.