Place the horn in the same location each time, and move it back and forth to peak it, then take your reading. If you have been approximately locating it and not moving it to peak your results, then your aren't exactly in the same spot and 1.5 dB change makes sense.  It has to be level too.

Also, when cables bend, they change phase and change your received power levels. Take a reading, and bend your cable to the antenna to find out how sensitive it is. If it's too sensitive you need better test cables.

Location is usually the largest problem.

kch

You are probably getting ground reflections via the  antenna sidelobes - which add and subtract to the direct signal at the RX antenna.  You can get up to about +3dB/-10dB effects depending on the phase and magnitidude of the reflections.  

I was testing the radiation pattern of an 8-element IFF Interrogator antenna (1GHz) mounted on a planar plate aircraft radar antenna and encountered exactly this.  There are 2 inexpensive ways of beating the reflections, one is to place sawtooth top fences between the 2 antennas to breakup the phase of the reflections and the other was to vary the height of one antenna measuring the signal until an area giving consistent results was located.  You have to be able to steer both antennas to do this so your best course is probably to use the fences.

Add this to what Higgler said and you should be OK.

"...7.5 foot Dish as Source..."

Are you a reasonable distance (50 or 100 feet - maybe ?) away from the source?

 

Even a parabolic dish may have side lobes -- if anything changes there, changing the reflecion it may influence the
received amplitude.

<nbucska@pcperipherals DOT com> subj: eng-tips
read FAQ240-1032