Characterizing station performance

---------

From: howie cahn (howi@world.std.com)
Date: Fri Dec 17 1993 - 21:07:57 EST


    I've often been asked questions like "How much will .5 dB in coax
losses hurt me?" or, "What can I work with two watts into a dipole?".
There are a lot of anecdotal answers floating around but I think many of
them are wrong.

    To try to get some better answers I did a little data analysis
exercise using results from the recent CQWW cw contest. Whether or not
you're interested in contests, a contest is a good place to get data;
lots of stations with different configurations making lots of QSOs.
Results and station descriptions are usually posted to Internet so I
had lots of information to play with. I made a graph with one axis (x)
as the effective radiated power (in dB relative to 1 watt into an
arbitrary dipole), ERPd. The other axis (y) is the total number of
QSOs made. Both scales are logarithmic. On this graph I plotted four
points -- representing the winners in each of three categories: high
power, low power (150W), and QRP, and the results for my station. I
assumed the three winners used the maximum power allowed in their
class. I estimated their average antenna gain by taking the gains of
their antennas on each band and weighting them based on what percentage
of the total contacts that band accounted for. For example, AA2U has
stacked 6-element tribanders on the upper bands, a 2-el on 40, and a
loop for 80/160. I estimated that this averages out to 10 dBd gain.

    The resulting data points were:
        Call QSOs Effective Power
                          (ERPd watts)
        ------------------------------
        K1KI 2927 24000 high power
        K2ZJ 1292 2000 low power
        AA2U 734 50 QRP
        WB2CPU 234 3 me

    The points are surprisingly colinear. The only anomaly was that
AA2U, the QRP winner, was a bit over the line that connects the other
three. Since Randy often places in the top ten in the LOW POWER category
while running QRP, this is not unexpected. The line that resulted had a
slope of about 4 log power/log QSOs, i.e., it takes 4 dB of effective
power increase to increase your QSOs by 1 dB's worth. I won't try to
draw it here using ASCII characters, but you can start with a point
representing 1 watt, 300 Q's and draw the line to the right from there
with a slope of four.

    From this data I'm concluding the following:

        a. Getting rid of losses in coax, antenna switches, etc.
improves things somewhat, but not that dramatically. The line's slope
of 4 dB power per 'QSOs dB' says that reducing your power loss by 1 dB
would have increased your QSO total by about 10 ** .025 ==> 6%. A 2 dB
increase would give 12% more Qs, a 4 dB increase 26%, etc. While the
numbers are only specifically applicable for this contest, they
probably can be generalized to imply a success rate for working
non-contest DX or making readable contacts in general.

        b. While you're not going to be competitive with the big guns,
you can make a reasonable number of DX contest QSOs with QRP power and
simple antennas. In this contest the graph suggests that you could have
made over 400 QSOs with five watts and dipoles, and, 300 QSOs with just
one watt and dipoles. Since these numbers were derived mostly by
analyzing the contest winners, they represent the upper bound of what
could be done.
    It might be surprising that the line was linear over a very wide
power range, about 5 decades; from about 1 watt to tens of kilowatts.
Actually, my experience indicates that it may be valid even lower than
that, say down to .1 watt effective power.

    I hope this information provides a way of estimating the
performance effects of changing station characteristics and encourages
people to try working QRP. To save bandwidth here, I've described
things quickly. I'd be glad to elaborate or discuss other people's
thoughts on the subject.

        73,

        howie, wb2cpu
        howi@world.std.com


Search QRP-L Archives

[ QRP-L Archive | ]
[ 1993 | 1994 | 1995 | 1996 | 1997 | 1998 | 1999 | 2000 ]

---------

This archive was generated by hypermail 2b29 on Fri Jun 02 2000 - 11:26:27 EDT

kd4ab@kd4ab.org