[Elecraft] K2: PLL Temperature OVERcompensation?

[Bill Coleman]

OK, I have a rather weird problem with my K2 #2548.

This K2 has the A->B mods, the SSB, NB, 160m and KAF2. Back in March, I 
used it in the WPX SSB contest, and it was fine.

In May, I ordered the Temperature Compensated PLL Reference upgrade. I 
built and installed it per the instructions, aligned the K2, and it 
seemed fine.

To be honest (and this will sound like heresy to K2 owners), I haven't 
had much time to actually USE the K2 since then, until late July. I 
dropped into the shack, turned it on 20m. What I discovered can only be 
described as "weird".

The SSB signals sounded sort of watery, and some of them where hard to 
impossible to tune in properly. 40m seemed to be OK, but 20m (and 
higher?) frequencies, the K2 just wasn't working correctly. 

I recently had a chance to diagnose this a bit, and what I uncovered is 
strange:

1) I checked the PLL Temp Compensation board, to see if all the resistors 
were in the right place. They were. (Although I haven't checked their 
values with an ohmmeter)

2) Thinking that the PLL calibration might be off, I ran CAL PLL on the 
cold K2. When it finished, the weird sounding, hard to impossible to tune 
signals on 20m were gone. Ah, problem fixed.

3) I put the covers back on, and left the K2 on overnight, to allow the 
internal temperatures to stabilize. Then I ran CAL PLL again. Again, 20m 
sounds good. I turned the K2 off.

4) Came back later with the K2 cold. Turned it on and the weird 20m 
tuning problem exists again. Ran CAL PLL on the cold K2, and the problem 
is fixed. 

5) Checked the K2 later (thinking the CAL PLL might have warmed the K2 
up), and the problem is still fixed. I assume if I leave the K2 on 
overnight again, the weird 20m signals will return.

Since CAL PLL fixes the problem, it is reasonable to suspect something in 
the PLL -- something introduced with the temperature compensation 
upgrade. It's almost as if the compensation is going too far, or 
something else is changing with temperature.

The really odd thing is that 40m sounds nearly normal, but 20m is 
definitely weird. Perhaps it has something to do with the higher PLL 
frequencies used on 20m. 

I'd really like to use the K2 in the NAQP SSB this weekend, so I'd 
appreciate any suggestions or ideas in diagnosing this. Heck, I almost 
wish I had a replacement RP3 -- my K2 wasn't THAT drifty before I put the 
PLL temp comp mod in.

Bill Coleman, AA4LR, PP-ASEL        Mail: aa4lr@arrl.net
Quote: "Not within a thousand years will man ever fly!"
            -- Wilbur Wright, 1901