[PStamler]

On Sep 12, 5:27*pm, Mark wrote:
On Sep 12, 2:30*am, PStamler wrote:



On Sep 11, 8:54*pm, (Scott Dorsey) wrote:

Mark wrote:

true enough,

but if you put an equal amount of effort into the design and impedance
using a low noise FET, it will beat the tube hands down...

Not really. *There's a really nice paper from Marshall Leach on the subject,
too. *It's probably referenced on his website.

Tubes and FETs come out about neck and neck, with similar transformers.
Transformerless bipolar input stage comes out hands down, though, assuming
600 ohm source impedance.

When you're looking at attainable noise figures of 2 dB over the
inherent noise of the microphone, which you can do with a tube or a
FET and the right transformer, there's not a lot of hands-down
improvement possible. In fact, there's only 2dB of improvement
possible, max, unless you dip the mic into liquid helium.
Realistically, more like 1dB, which is barely noticeable.

Modern designs, with any of the three technologies, are scraping
pretty close to the theoretical limits.

Peace,
Paul

I agree that an FET can achieve a 2 dB noise figure and that means an
FET can be 2 dB away from as good as it gets at room temperature...

Can you achieve a 2 dB noise figure with a 12AX7? *Or are you talking
about some special tube?

It's a little hard with a 12AX7, but it can be done with a 12AU7 or a
6SN7, neither of which is exotic. The main thing is to use a
transformer with low noise, like the Jensen 13k7 I mentioned earlier.
Oh, and run the tube at a highish current to maximize gm (lower noise)
and minimize the value of the cathode resistor.

Peace,
Paul