Greg ODrobinak wrote:
> Raimund, is the R16 available yet to you? Is there a chance you could sna=
g one and do the self noise measurement?
I had to purchase one in order to make the measurements. I would definitely=
do this if this unit would meet the requirement of my customers. However, =
the fact that it cannot provide phantom power on all 8 channels makes it le=
ss interesting.
I agree with you that the limitations on the phantom power are probably cau=
sed by general power constraints. A fully compatible P48 phantom power supp=
ly circuit must be able to provide 10 mA. This means that it would theoreti=
cally draw more than 100 mA from the internal 5V supply for each channel. =
> Also, if someone could post the procedure for doing the self noise measur=
ement on an arbitrary recorder's mic preamp, I would be most grateful. I ha=
ve an idea of how to do it, but I am not certain that my procedure is corre=
ct.
You just need to record a known reference signal in order to get an absolut=
e voltage reference. I use an attenuated 1 kHz sine signal from a function =
generator. Alternatively, one could play a .wav file containing a synthetic=
1 kHz signal through the line output of any recorder (I did this successfu=
lly with an Edirol R-09 and various sound cards). Because the output level =
that is present at the line output exceeds the input voltage range of the m=
icrophone input at the maximum gain setting, it is required to insert an ap=
propriate attenuator between the line output and the microphone input (I us=
e a self-made 56.5 dB attenuator consisting of two resistors). The absolute=
output level of the reference signal at the line output can be measured by=
using a common multimeter. Measuring the reference signal voltage behind =
the attenuator would be more difficult due to the very small voltage levels=
.
After recording the known reference signal, the microphone input must be te=
rminated by a 150 ohm resistor that is connected between the XLR pins 2 and=
3.
To finally get the EIN noise level figure it is required to band-pass filte=
r the recorded sound file (20 Hz ... 20 kHz) and to measure the RMS levels =
(expressed in dBFS) of both the reference signal section and the terminated=
(quiet) section in the sound file. The EIN is then calculated by subtracti=
ng the difference between the two levels from the absolute reference voltag=
e level.
Here is an example:
The absolute output level of the 1 kHz reference signal (at -6 dBFS) played=
through the line output of a sound card is 0.7 V (measured with a common m=
ultimeter). The measured 0.7 Vrms is equivalent to -3.1 dBV [ =3D 20 * log(=
Message: 0.
Subject: 7)] or -0.9 dBu [ =3D -3.1 dBV + 2.2 dB].
Then add the attenuation (-56.5 dB) of the attenuator to the result:
-0.9 dBu - 56.5 dB =3D -57.4 dBu (this is our reference signal level)
Assuming the RMS level of the reference signal was -12 dBFS and the level o=
f the (band-pass filtered) quiet section was -75 dBFS, then the final EIN w=
ould be:
-57.4 dBu - (-12 dBFS + 75 dBFS) =3D -120.4 dBu (unweighted)
For A-weighted measurements one had to apply an A-weighting filter. If the =
noise is equally distributed over the entire frequency range, the A-weighte=
d figures would be about 2 dB below the unweighted figures.
Regards,
Raimund
Regards,
Raimund
|