[ARC5] S+N/N ratio results.

[Geoff]

>     For the most part I agree with your remarks. Except
> for:
> Audio meter; the meter may be calibrated for some db reference or it may 
> not, however, we are concerned only wtih ratio, not absolute level so the 
> impedance of calibration is of no consequence. One can make the 
> measurement on a 4 ohm output or a 600 ohm output.


** As long as whoever is taking the measurements understands the difference 
between voltage and power
and computes accordingly. A meter calibrated in dB and used properly is 
often less confusing and causing errors.


>     As far as RF matching, its more complicated.  The critical thing is to 
> know what the voltage is at the receiver terminals. You are quite right 
> that the only way to be sure is to measure the actual impedance presented 
> at the antenna terminals with a bridge.  OTOH, one can just feed from a 
> low impedance and assume the voltage will be close. I mean a terminated 
> output from a 50 ohm generator is 25 ohms and will not be much affected 
> when connected to a receiver with 200 or 400 ohms input.

** I already stated that using a 6dB pad is the ideal way. It eliminates 
errors without having to think about calculations or using more equipment.
Actually measuring the load is more an engineering exercise or for curiosity 
and I listed those ways for those that may be interested in the way the 
design engineers went about it.



At any rate close
> enough for the sort of rough measurement we are making.


** As long as "rough" is understood and it probably suffices for the simpler 
military and commercial receivers.  OTOH when a document states a specific 
dBM or SNR level I want to be sure Im not off by a factor of 2 or more by 
oversimplyifying things.


>    Leakage is a real problem.  Laboratory grade generators generally have 
> very low leakage

** Not really until the SS years came around. The HP-606 family, URM-25D, 
Measurements 60 & 80, etc, leak like a sieve as built but can be cleaned up 
with a bit of work. I use a pair of 606A's for most of my vintage work as 
they are easier to use and more reliable than the pair of 8640B's that are 
mostly gathering dust. The 606A's are modified for counter use and with 
leakage in mind.



but the connections to the
> receiver must be well shielded and that presents a problem. One can have 
> ground loops at RF that screw up any sort of sensitive measurement.

** There is no such thing as RF ground loops. RF susceptability is a more 
commonly acceptable description but it all goes back to common mode leakage 
which can be a design or user induced problem.


  The connections to a receiver with
> screw terminals will probably always be leakier than a coaxial connector 
> but can be made decent by using coax connected to the terminals with 
> almost zero length leads. Another problem with the lead from generator to 
> receiver is that it may pick up external signals including noise which 
> throw off the measurement.


** As Ive already stated


  This becomes difficult without
> heroic measures like screen rooms.

** Having used them at work since the 60's for both ingress and egress 
including in the TS CIA/DOD Tempest program at Sanders Associates and Wang 
Labs Ive developed a great respect for what really happens when things leak.


>     At any rate, getting an approximate measurement on a boat anchor is 
> possible. A meaningful measurement on a receiver with fractional microvolt 
> sensitivity is quite difficult.

> Richard Knoppow
> Los Angeles
> WB6KBL

** Not really. It is only as difficult as you want to make it. Whether or 
not it makes any difference in use is up to the individual user/owner. Those 
antiquated tables and graphs showing useable sensitivity by frequency have 
been proven wrong time and time again by many other studies.

Carl