[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

RE: [linrad] OH0 June 12: Rx/Tx measurements

Hello Richard,

> A friend of mine(Rod VK6KRG) devised a test using a "comb" of strong 
> signals across the band of interest.
> The mixer performance is affected by the total energy it has to absorb 
> from unwanted signals and a single strong interfering signal may not be 
> as rigorous a test as some HF band conditions, or multiple pager signals 
> on 148 MHz, or even TV line oscillator noise for example.
This is the very reason why I do not think that the close range 
intermodulation tests are so important:-)

Dynamic range is limited by two numbers, the lower one is the 
noise floor and the upper one could be due to non-linearities 
or due to a degradation of the noise floor.

Degradation of the noise floor (most often reciprocal mixing) is best
investigated by a single interference signal.

The upper limit due to non-linearities can be taken as IP3, 
the 1-dB compression point, the "comb" test that you describe or 
by white noise that is filtered through a notch filter or by 
a measurement of intermodulation from an AM modulated signal 
and in many other ways.

In receivers where the signal energy in all the different methods is outside
the selectivity of the first IF filter, all these methos give results 
that do not differ terribly much in which order they rank receivers.

It seems to me that the most valid test is the one with white noise
and a notch filter but I think the "comb" test would give very similar
results. IP3 is more synthetic in the sense that it does not really
mimic a practical situation too well. The worst problem with IP3 is
that the non-linearity of several cascaded amplifiers/mixers may add
in a way that makes the IM3 vs power curve deviate significantly
from the ideal third order law at high levels. Measuring at low
levels is difficult on a good receiver and therefore we see 
IP3 data that are incorrect now and then.

At the current state of the art amateur equipment is far too ill
designed to make a comparision between the different methods of
measuring the upper limit of dynamic range due to non-linearities
meaningful. The reason is that transmitters (mainly due to ALC) 
generate wide signals that makes it uninteresting to go much 
closer than 5 kHz for rx testing. Further, the sideband noise 
in both receivers and transmitters is a more severe limitation.

The subtle differences between the different methods of evaluating
non-linearities may turn out to be very significant in the way 
different receivers would be ranked, but to find out one would 
need information about the practical problems when using receivers
with inadequate dynamic range on real antennas. Maybe the close
in performance is so bad in some receivers that it actually is 
the practical limitation, maybe the wideband input (several MHz)
in some receivers is the problem. I do not know. Once a reliable
practical ranking list is avaliable it would be interesting to
study which way to measure non-linearities that correlates best.
(one would need an attenuator to place the external noise floor
equally in relation to the internal noise floor when ranking 


Leif / SM5BSZ