Re: very high gain preamp+mic combination

"Don Pearce" <spam@xxxxxxxx> wrote in message
news:49a06367.1338643125@xxxxxxxxxxxx
On Fri, 20 Feb 2009 06:44:02 GMT, "Paul Stamler" <pstamler@xxxxxxxxx>
wrote:

"Don Pearce" <spam@xxxxxxxx> wrote in message
news:49adc4c0.1298029125@xxxxxxxxxxxx
On Thu, 19 Feb 2009 20:36:22 GMT, "Paul Stamler" <pstamler@xxxxxxxxx>
wrote:

<Don Pearce> wrote in message news:49a509c1.1250158437@xxxxxxxxxxxx
<Don Pearce> wrote in message news:49a16e75.1210401515@xxxxxxxxxxxx
On Wed, 18 Feb 2009 15:12:30 -0500, "Soundhaspriority"
<nowhere@xxxxxxxxxxx> wrote:


Do the sums.A preamp with zero excess noise and a 150 ohm mic gives
an
input noise of -139.37dBV in a 20kHz bandwidth.

How did you come up with that figure? I get -133.04dBV (-130.82dBu) of
Johnson noise for 150 ohms in a 20kHz bandwidth.

Oops - missed a 4 in my calc. Sorry about that.

OK, this is where we stand. Using the NT1-A, a perfect preamp imposes
a noise penalty of 0.9dB. The Gordon Audio preamp identified by Hank
adds 3dB of noise (the same for a cheap Behringer, incidentally) and
an RNP adds 6.5dB. So the choice of preamp is critical, and designing
for ultimate low noise will be worthwhile.

I'm still not following you. The equivalent noise figure of the NT1-A
presumably includes the thermal noise of its output impedance, no? In
which
case, a perfect preamp would add no noise at all.


A perfect preamp has noise - the voltage is given by the equation
sqrt(4kTBR) (k is Boltzmann's Constant, T is the temperature, B is the
bandwidth and R is the total resistance at the input).

No, it doesn't. A perfect preamp is one that adds no noise when its input
is
shorted (in other words, the source impedance is zero). Under those
circumstances, the preamp (a mythical beast) will generate an output noise
of zero volts.


Plug zero into my formula above, and that is exactly what happens. Why
are you quibbling?

Take away the short and put a resistor in its place. The resistor will
generate Johnson noise based on the formula you gave. The resistor -- not
the preamp -- is generating the noise, just by virtue of being a resistor
that's not at absolute zero.


Yes - this is just semantics.

Now replace the resistor with a condenser microphone. It generates noise,
a
lot more noise than a resistor with the same value as its output
impedance.
We know how much noise it generates because the company specifies how
much,
and (assuming they're telling the truth) they specify how much by
measuring
it at the output.


Yes, that is what all my calculations have been about.

And that measure will include the output resistance of the microphone as
one
of the sources of the microphone's noise, along with
electronically-generated noise from the microphone's head-amp.

If you use the microphone's noise figure, and then add to it the noise of
the equivalent impedance, in effect you're counting the noise of the
output
impedance twice, which is wrong.

As for the Gordon or Behringer...as I recall, you gave the noise figure
of
the NT1-A as -121dBV (correct me if my memory's wrong). If the
Gordon/Behringer has an equivalent input noise of -127dBV then the total
noise will be -120.02dBV, almost exactly a 1dB penalty. If the preamp's
noise is less than that, the total noise will be slightly less as well.

Yup, it is possible to get about an extra 5dB of noise performance out
of the preamp by careful design.

Not my point. My point is that adding two noise sources, with one 6dB
lower
than the other, will add only 1dB to the noise level of the noisier
source,
assuming they're uncorrelated.

Do the arithmetic. Say one noise source is 1V and the other is 6dB lower,
or
0.5V. The total noise will be sqrt(1^2 + 0.5^), or sqrt(1.25), or about
1.118. That's about 0.969dB worse than the noisier signal, 1V.

Isn't that what I said? (I quote - almost exactly a 1dB noise
penalty).

No; you said there'd be a 3dB noise penalty. "Almost exactly a 1dB noise
penalty" is what I said.

I'm not just quibbling; I'm pointing out that, in effect, you're trying to
count the output Z of the microphone twice in figuring its noise
contribution, and getting the wrong numbers from doing so.

As for "It is possible to get an extra 5dB of noise performance out of the
preamp by careful design"...well, yes, without question. A perfect preamp,
with a 150 ohm resistor at its input, will give you an EIN of -133dBV. But
combined with an NT1-A, the improvement in noise performance won't be more
than 1dB.

Oh, the following *is* a quibble, and labeled as such. The EIN of the Gordon
preamp isn't -127dBV, but -129.2dBV (spec'd at -127dBu).

Peace,
Paul


.

Relevant Pages

  • Re: Input stage mess
    ... I've finally got doing that low noise preamp: ... >>a low servoed 1V VDS to minimze thermal noise problems. ... >>negative real part input admittance. ... >>I can't insert a damping resistor in series with the gate connexion for ...
    (sci.electronics.design)
  • Re: Microphone preamp advice
    ... can record my amp silently (well, get the best approximation to it ... without speaker / mic / noise). ... round it was) I get more noise and my preamp overloads the soundcard ... read the RNP pairs with the real nice compressor quite well. ...
    (rec.audio.pro)
  • Re: Microphone preamp advice
    ... Due to my circumstances I record direct from the amp head ... without speaker / mic / noise). ... round it was) I get more noise and my preamp overloads the soundcard ... loud enough but to my understanding the sound will loose alot of its ...
    (rec.audio.pro)
  • Re: tascam hd-p2 + gracedesign lunatec v3 + rode nt4 stereo mic ?
    ... significantly less noise and distortion. ... very specialized preamp, and perhaps choose a particularly quiet mic. ... And if you just amplify a low level recording, ... better than what's in the recorder as well (or maybe not - it's a few ...
    (rec.audio.pro)
  • Re: Is this a good way to test pre-amp hiss of a digital recorder?
    ... noise of a preamp with a 150 ohm resistor and a 600 ohm ... of the preamp, and how much is actually due to thermal noise in the ...
    (rec.audio.pro)
Loading