Arny Krueger wrote:

Looking at the Sennheiser spec sheets, I don't see where their RF-based mics
are *that* more quiet than the competition. They are quiet, but so are some
mics built using more conventional technology. I also note that they give
no noise spec for the MKH800.

This might be one of those things that's more apparent to the ear than
to the eye. One of the first things you notice when you connect an
MKH20 is that you don't hear any noise. But on paper, that might only
be a few dB.

Mike Rivers wrote:

One of the first things you notice when you connect an
MKH20 is that you don't hear any noise. But on paper, that might only
be a few dB.

Indeed, yes... I've found that with an MKH 50 also, which is practically
noise-free, even compared to mics that are actually slightly more
sensitive and only marginally noisier in theory (Beyer 803)...

Daniel

Scott Dorsey wrote:
Arny Krueger wrote:
I suspect that the problem is not necessarily due to general cheapness and
shodiness, but rather due to electret mics typically having very small
diaphragms so that they can meet other engineering and product design goals.

That's a lot of it, but using the FET-IC in place of a conventional front
end causes a huge noise increase and an equally huge cost savings.
--scott

What, besides FET's, are used in the front end that are more conventional?
Not sure what you mean by IC. They just use regular FET's (2sk123 for
example) and rely on leakage to get the gate to 0V quiescent. Is it
quieter to use a bulky, high valued resistor to bias it instead?

The diodes that some show in their pictures actually just represent the FET
junction I'm told. It's shown just to imply a usable path for biasing.


Bob
--

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

A. Einstein

Scott Dorsey wrote:
Don Pearce wrote:
On 5 May 2006 10:01:13 -0400, (Scott Dorsey) wrote:
Don Pearce wrote:
If it were not thus, then Sennheiser would be wasting their time
making the MKH series of RF microphones. If mechanical noise
dominated, these would be as noisy as conventionally biased
microphones - they aren't, they are much quieter.
Sennheiser ALSO does some things to reduce mechanical noise in those
capsules too. You'll notice that the way they are designed, the capsule
has a huge midrange peak, then the electronics equalize that peak out.
Sorry - confusion of what I meant here. By mechanical noise, I meant
the inevitable Brownian motion that forms the acoustic noise floor of
the microphone - not stuff coming in through the body. And of course
flattening out the response in the electronics is a good thing, and
hardly sleight of hand.

Right. The Brownian noise is reduced by radically changing the capsule
geometry. Then equalization is used to compensate for that.

Incorrect. That noise is strictly a function of the diameter of the
diaphragm and cannot be reduced in any other way. It's actually the
resistive part of the air load that causes the noise. This becomes larger
with increasing diameter. Sometimes it's called the damping resistance.


Bob
--

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

A. Einstein

Bob Cain wrote:

What, besides FET's, are used in the front end that are more conventional?
Not sure what you mean by IC. They just use regular FET's (2sk123 for
example) and rely on leakage to get the gate to 0V quiescent. Is it
quieter to use a bulky, high valued resistor to bias it instead?

The FET-IC isn't a standard FET at all.... it's a FET with a diode
fabricated on the substrate, and the diode leakage is used for bias.

Diode leakage is VERY noisy. It is much, much quieter to use a high
value leak resistor.

The diodes that some show in their pictures actually just represent the FET
junction I'm told. It's shown just to imply a usable path for biasing.

If you actually look at the die, the ones I have seen have a seperate
junction.
--scott
--
"C'est un Nagra. C'est suisse, et tres, tres precis."

Scott Dorsey wrote:
Bob Cain wrote:
What, besides FET's, are used in the front end that are more conventional?
Not sure what you mean by IC. They just use regular FET's (2sk123 for
example) and rely on leakage to get the gate to 0V quiescent. Is it
quieter to use a bulky, high valued resistor to bias it instead?

The FET-IC isn't a standard FET at all.... it's a FET with a diode
fabricated on the substrate, and the diode leakage is used for bias.

Diode leakage is VERY noisy. It is much, much quieter to use a high
value leak resistor.

The diodes that some show in their pictures actually just represent the FET
junction I'm told. It's shown just to imply a usable path for biasing.

If you actually look at the die, the ones I have seen have a seperate
junction.

This doesn't match the information I've received but I don't have the
source at hand so I really can't argue the point.

The fact remains that if you use a vanilla FET, the GS leakage will bias
the gate to the same voltage it would if a separate diode were there.

Do you know where I could find the picture of a die showing the diode?


Bob
--

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

A. Einstein

Bob Cain wrote:

The fact remains that if you use a vanilla FET, the GS leakage will bias
the gate to the same voltage it would if a separate diode were there.

This is definitely not the case with the 2SK170. It may be the case with
some other FETs.

Do you know where I could find the picture of a die showing the diode?

I'll see if I can find some of the ones we shot when we were doing noise
testing on the things. I surveyed a whole bunch of different FET-ICs for
a microphone manufacturer about ten years back and I probably still have the
file.
--scott

--
"C'est un Nagra. C'est suisse, et tres, tres precis."