ARRL EME CONTEST 2001 Relative ERP power levels.
(Mars 22 2002)
Look here for info about the raw data Recordings made with the big antenna of SM5FRH

With version 00-40 Linrad has a peak level meter. When applied to the practically noise free recordings made with the big X-yagi system of SM5FRH it is possible to determine the relative ERP powers of many active (Nov 2001) EME stations.

When the signal level is estimated from the peak height in the spectrum, the peak height depends strongly on the keying. There is much more power in a sequence of O's than in a sequence of R's.

The peak power is independent of what morse code is being transmitted.

The qsb characteristics at 144 MHz is such that the largest peak power level within a one minute period does not vary by more than about 2dB RMS. It is a statistical distribution however, so taking the maximum during a single minute may give a value that is far from representative. When taking the largest value from 3 minutes or more one has a large probability to get a result that does not deviate by more than about 1dB from the value one would get as an average of the peak level during many different 3 minute periods.

Below is some data extracted from the FRH1135.BZ2 (213032622 bytes) recording. Each signal level is from one transmit period. The periods vary in length but the length is usually around 1 minute. The first value, before each call sign is the relative ERP power in linear power scale taken from the largest value for each station, normalised so F3VS, the strongest station is 1000.

       Relative power levels
linear
power  call    max dB during one period 
scale  
1000   F3VS    58.1 56.4 62.5 61.7 60.7
 269   KB8RQ   56.5 55.2 54.3 49.8 56.8
 233   I3DLI   54.6 56.2 55.9
 109   HB9Q    52.9 51.6 51.6
 107   G4LOH   49.8 49.8 49.7 51.7 52.8
 107   F1FLA   52.8 50.9 51.9 48.9
 105   K5GW    52.7 50.3 51.0
  55   I2RV    48.9 49.7 49.9
  47   WA2FGK  47.5 49.2 (2 min periods)
  25   K7ND    44.1 46.5 (2 min periods)
   7   IZ1BPN  39.1 38.5 40.8
  0.5  5 sek peak value of noise floor in 20Hz
Each of these linear scale power levels is uncertain by about +/- 1dB but the error may be larger in some cases if for example moon tracking was inaccurate during the minutes when the peak level was measured. The peak power level is fluctuating at random and there is no guarantee that some maximum power level is incorrect by a much larger amount than 1dB. There is also a gain difference between H and V pol that amounts to about 1dB. The effect of different gain is not taken into account although it would not be difficult to do so.

The recordings cover several hours. By extracting many more peak power levels one can get a much better insight in the statistical variation of the peak value during a time span of a certain length. It is also possible to improve the accuracy this way by forming average values or by picking the peak value over more minutes.

Listening to over one hundred stations during 30 minutes for each one would take a lot of time....

Maybe, some day, someone else is interested to do this work.