[Elecraft] K2 a little deaf?

[w2bvh]

I dragged my HP8640B out of the "lab" (basement) to the second floor 
shack to do some antenna testing about a week ago. Before dragging this 
50 lb instrument back down two flights of stairs I decided to do a quick 
MDS measurement on my K2 and got some results I'm a little concerned about.

Yesterday I did an informal test. K2 set for cw, 1.5 KHz bandwidth, 
14.05 MHz, AGC on, preamp off. Attached siggen, tuned the K2 to 550 Hz 
tone and cranked down the siggen amplitude until the signal and noise 
were about equal, "by ear". Siggen output was -130 dBm. This was enough 
to cause me to look further.

Today I took some numbers. Same setup, but with a Fluke dvm instead of 
ears and the siggen phase locked . Started at -90 dBm drive and tuned 
the K2 to peak voltage reading. Then made the following measurements:

_Sigen (dBm)      Meter(V)_

-90                      .38
-100                    .15
-110                    .050
-120                    .017
-130                    .011
(off)                     .010

Then did a rough voltage-ratio-to-dB "calibration" as follows: Set 
siggen to -100 dBm & measured the same .15V ob the K2, then ran the 
siggen output to 103 dBm and measured .204V. So 3dB signal difference is 
a voltage ratio of 1.36.

Finally measured MDS driving the K2 until the measured signal voltage 
was (.011 * 1.36 = .015 V).  I got this value at a drive level of -123 
dBm. Tried a second set of measurements with the preamp on using the 
same technique and got the same MDS of -123 dBm. Tried a third set of 
measurements with preamp off  and AGC off and got a slightly worse MDS 
of -121.

Now I know there are flaws with this measurement, especially in 
measuring a signal near the noise. But I know the provenance of the 
siggen, use it all the time and get good measurements from it. I trust 
it. If it was a matter of 2 or 3 dB I wouldn't be writing to the list, 
but this is 6-11 dB worse than the 2 numbers on Sherwoods site (-129 and 
-135).

FYI, another informal "measurement" I made was on  6M with the XV50. A 
signal was easily audible over the noise with the siggen set at -140 dBm 
drive!

On air results are equivocal: yesterday afternoon I couldn't hear my 
friend Joe N2CX 145 mi away who was qrp portable on 20M (something I've 
done multiple time in the past), but I had 3  consecutive easy q's to 
eastern EU last night on 30M...

On the K2, gain could be missing or noise could be high anywhere: front 
end, I.F, audio stages.

What do you think.  Could / should I send it away for work. It'd have to 
go somewhere with better instruments, since the K2 isn't exactly 
"broken". Or should I don't worry be happy?

73, --Lenny W2BVH

ps -FYI this is K2 1520, I built in 1999 & been in use here ever since.