> Hi Raimund, this is indeed a very telling test, but the the imbalance is =
not as much as you perhaps think. i.e. by leaving the 150R across the pins,=
from the 48V you have 6K8 to the 2k2 direct and 6K95 to the 2K2 via the 15=
Subject: i.e. a 2.2 percent imbalance.
> And its supposed to be a pure DC, not contaminated with system noise!
Sure, but it is often a challenge to generate such a perfectly pure DC volt=
age out of a low battery voltage.
> Anyway, the point I take issue with is the "poorly designed (unbalanced) =
microphone". The only poor design appears to be the R44 phantom supply, eit=
her a poor return path, poor decoupling or both.
I beg to differ here! The phantom power specification demands perfectly bal=
anced currents on both rails. For that reason, the two internal 6.8 k resis=
tors must be matched to within at least 0.4% (see http://en.wikipedia.org/w=
iki/Phantom_power). The microphone itself should therefore also be perfectl=
y balanced. Otherwise, the matched 6.8 k resistors within the recorder woul=
d be meaningless.
> In principle there is absolutely no reason why one couldn't short either =
of the two signal pins to ground, save that the common mode rejection would=
of course negated.
Hmmm, I'm not sure if that would be a good idea, given that it is quite dif=
ficult to design a perfectly ripple-free phantom power supply.