The meter is calculated on a single filter with a large signal in its
path and for the rest of the time, it is assumed that noise is
negligible compared to the tones in the calibration. This assumption
appears to break down near the noise floor if your measurements are
correct. This is useful information. I will be in Austin Friday -
Wednesday and we will simply make several measurements to see what
gives. I do clearly recall making the meter measurements at all sorts
of levels with my (calibrated) 8640B (as accurate as that can be with
its analog level set) and it was definitely close but I cannot state
emphatically that it was not (3.1 - 1.8) = 1.3 dB off since we just
didn't have test equipment capable of that accuracy then.
Bob
Clay W7CE wrote:
Hi Bob,
It's certainly possible that I'm doing this wrong, so let me explain
my procedure and then you can tell me if I'm making a mistake. All of
the following is performed on 6M at 50.200 MHz with a 50 ohm
termination connected to antenna port 1 and the bandwidth set to 500
Hz (mode CWU). Also, the multimeter average time is increased to 5000
mS and the Digital Refresh is increased to 2000 mS. The external
preamp I am using is an ARR P50VDG preamp with a spec'd noise figure
of 0.5 dB and a measured gain of 26 dB. Gain was measured on an HP
4396A Spectrum/Network Analyzer and confirmed using the ADC L and ADC
R meters.
First, I measured the receiver noise floor using the RX1 meter set to
"Sig Avg". With the internal preamp turned on and no external preamp,
the average noise level is -125.0 dBm (the value fluctuates some, but
seems to be centered here). This indicates a receiver noise figure of
22.0 dB. Now if I enable the external preamp the average noise level
decreases to -143.9 dBm, corresponding to a receiver noise figure of
3.1 dB. However if I do the math, adding my preamp and assuming
connector, coax and internal relay losses of 0.3 dB in addition to the
0.5 dB preamp noise figure, I should see a receiver noise figure of
2.0 dB. For the receiver noise figure (including external preamp) to
be 3.1 dB the external preamp noise figure and passive losses would
need to add up to 2.2 dB which is much larger than I would expect.
Next, I computed the receiver noise figure by measuring the receiver
MDS using an external HP RMS volt meter at the speaker connector and
the HP 4396A as a signal generator. Without the external preamp the
measured MDS is -126 dBm and with the preamp the MDS is -145 dBm,
corresponding to noise figures of 21 dB and 2 dB, respectively. If I
do the math, adding my preamp in front of a 21 dB NF receiver, should
yield a system NF of about 1.8 dB, so the measurements agree closely
with the math.
Based on the results, my MDS measurements (using the voltmeter) with
the preamp enabled are more consistent with the calculated values than
those using the internal meter. I suspect that measurements this
close to the noise floor will not be as accurate and that I'm asking
too much. I'm certainly not complaining though. The accuracy of the
metering rivals that of my HP spectrum analyzer on measurements that
I've made from -100 dBm to -30 dBm.
73,
Clay W7CE
--
(Co)Author: DttSP, Quiktrak, PowerSDR, GnuRadio
Member: ARRL, AMSAT, AMSAT-DL, TAPR, Packrats,
NJQRP, QRP ARCI, QCWA, FRC.
"It is human nature to think wisely and act in
an absurd fashion.", Anatole France.
Twitter:rwmcgwier
Active: Facebook,Myspace,LinkedIn
_______________________________________________
FlexRadio Systems Mailing List
FlexRadio@flex-radio.biz
http://mail.flex-radio.biz/mailman/listinfo/flexradio_flex-radio.biz
Archives: http://www.mail-archive.com/flexradio%40flex-radio.biz/
Knowledge Base: http://kc.flex-radio.com/ Homepage: http://www.flex-radio.com/