[ICOM] 7800 question, part 2

[mecker at peoplepc.com]

Im wondering if other users in the group have noticed the following.
At first I thought I was nit-picking, but maybe not:

Saw the response to my earlier post and appreciate it.

I have now attached the 7800 to just calibrated peak-reading wattmeter (N8LP vector meter)

for PEP readings.

At first, it reads 200.6 watts into a 5KW resistive load, and the power
meter on the 7800 itself shows 200 plus 1 notch watts on its selectable bar peak hold meter.

However, after tx for a short period, PEP reading begins to cut
back some, not a lot, but some. (this is what I had sensed/suspected
originally) . After a few intermittent 1 min ssb type pre-recorded
cq transmissions, (microphone is the SM20), Mic and RF
controls to max....I get these new peak readings...198, then 196
then 192 and 187. A few more transmissions reveal an occasional
194W, and a few 190s and a few 185s. These readings are different
from when the rig is first turned on.

The SWR on the dummy load, according the N8LP wattmeter, did
not change significantly.

For the skeptics, 187/200 x 100 is still 93.5 percent of spec (200W) but my
concern is more for reliability rather than a 6.5 percent change.

I noticed the V sub D final voltage meter, formerly dead center between
48 and 52 volts, moved back 2 micro-notches. After a 10 minuite break,
the voltage gradually returns to the near 50V center position, ebbing
back the two micro-notches. This meter appears when holding down the
selectable meter button for a few seconds.

Would like to know if any 7800 users have discovered this anomoly and
is it characteristic for this model.

-Marc, 
KE2BP