W3FPR:
>Current tests do consider MDS to be the minimum signal that can be
demodulated with no other considerations. Right or wrong, that is the
way it is for now - something must be defined for lab measurements to be
valid - we just trust that they are adequate to be useful in actual
operating conditions.
GM4ESD:
>MDS as you appreciate means 'Minimum Discernible Signal' not a signal at the
noise floor. IIRC the term was first intended to quantify the signal which
could be discerned by a human or device connected to a receiver's output,
which certainly in the human case introduces a variable. Some people can
hear a signal well below the noise floor thanks to the filter between the
ears, others with impaired hearing may not hear the signal until it rises
well above the noise floor.
ARRL uses the following definition of MDS in their
test procedure (page 35 below):
http://www2.arrl.org/members-only/prodrev/testproc.pdf
5.1 NOISE FLOOR TEST
5.1.1 The purpose of the Noise Floor Test (also known as "Minimum
Discernible Signal" or MDS) is to determine the level of signal input
to the receiver that will produce an audio output where the power in
the signal is equal to the power in the noise (S + N = N + 3 dB). The
test is conducted with the receiver in the CW mode using the 500 Hz,
or closest available, IF filters (or audio filters where IF filters
are not available. For receivers that have appropriate IF filters,
all audio filtering is disabled.) Set the AGC to the OFF position if possible.
This test is meant to provide a repeatable, quantifiable
means of measuring sensitivity, that takes subjective
human elements out of the picture (i.e. my ear's MDS may
be better or worse than your ear's MDS). The key issues
above are to use a standard 500 Hz bandwidth (to remove
noise bandwidth variations for other BW choices) and to
use a True RMS voltmeter to determine the 3 dB point for
(S+N)/N. Since noise is present in both S+N and N, the
True RMS voltmeter is critical to the measurement.
MDS is one of the least meaningful parameters to me
since I spend most of my time on the low bands where the
natural noise floor is far above the noise floor of any
modern receiver. In fact I am more likely to be using
attenuation rather than a preamp. If I were a VHF type,
MDS might be more important due to the very low galactic
noise floor at those frequencies.
As Geoff said earlier, a well-trained human
"ear/brain DSP" is capable of effective bandwidths of
about 50 Hz. In the case of published MDS specs for
a standard 500 Hz bandwidth, you can typically subtract
another 10 dB from published figures (i.e. improvement
for 50 Hz ear/brain DSP versus 500 Hz radio DSP yields
10 dB improvement). This puts the actual MDS level at
(+3 -10 = -7) 7 dB below the 500 Hz noise floor, which
is quite believable since some people are reportedly
able to copy signals as much as 10 dB below the noise floor.
73, Bill W4ZV
_______________________________________________
Elecraft mailing list
Post to: Elecraft@mailman.qth.net
You must be a subscriber to post to the list.
Subscriber Info (Addr. Change, sub, unsub etc.):
http://mailman.qth.net/mailman/listinfo/elecraft
Help: http://mailman.qth.net/subscribers.htm
Elecraft web page: http://www.elecraft.com
