Chris Hornbeck wrote:


To clarify/correct my previous post: An input noise level of
-140dBu (presumably unweighted)

Actually I was thinking weighted so that gives a bit more
room.

is 77.5nV, the noise made by
a resistor of less than 20 ohms. This can certainly be done,
but I don't know how using only phantom power and common
devices. No doubt somebody smarter than me could though.

Not me, for sure. What will the phantom resistors
contribute?


If we can relax the spec 4dB to 125nV and an equivalent
resistance of 50 ohms, we can use a selected pair of 2SK369's.
Today I rough-checked a random batch of BL group 2SK369's
and found four pairs of .5% matching I-sub-DSS, out of sixteen
transistors. All four pairs were within 9.75 thru 11.1 mA,
requiring a small source resistor.

That would be fine. The -140 dVu figure was off the top of
my head as about where the sublime would begin.


Or, if we could relax the spec 7dB to 180nV and an equivalent
resistance of 100 ohms, we can use the 2SK389 GR duals.
Checked two today and found I-sub-DSS matching of 1-2% and
*ideal* absolute numbers of 5.7 and 6.1 mA.

If the spectrum is fairly flat, the A-weighting would make
that about -135 dBu which is still a whole lot better than
many of the boxes it could front.


Do you have any thoughts about a desirable input impedance?
We can have pretty much anything we want. What would be a
good load for the transformers in ribbon mic's, I wonder?

Is Scott Dorsey around? He could address that a lot more
knowledgably than I could.


Bob
--

"Things should be described as simply as possible, but no
simpler."

A. Einstein