[Elecraft] strange K2 160m power problem

[R. Torsten Clay]

In the process of calibrating my LP100 wattmeter, I was hopping back and forth 
between bands and transmitting. I discovered a strange problem with my K2/100 
#3134. This is what happens:

First, I have the K2 hooked up to a dummy load. I also confirmed that this 
happens with a different wattmeter in line. I am keying the K2 with a straight 
key.

1. I transmit on 28 MHz. I get about 100w out.
     I press "band -" several times until I get to 1.8 MHz. Putting the
     key down, I get about 110 w out as expected.

2. I transmit on 28 MHz. I get about 100w out.
     I press "band +" once to get to 1.8 MHz. Putting the key down, I get
     170 w out !!!

In other words, it makes a difference whether I go from 10m to 160m either by 
pressing Band - several times, or Band + once. Pressing Band + once somehow 
screws up the tx agc. The problem does not seem to be in the KPA100, because if 
I turn the power down to 10W I get similar response (10W one way, 15W the other 
way).

My other K2 (similar SN) does not do this, or if there is a difference in the 
160m power it is too small to notice. The only thing different about #3134 is 
that is has somewhat lower 10m tx power, about 7W max without the KPA100. But 
it can still make 100w on 10m with the KPA/100.

Ideas?? If I twiddle the power control AFTER I start transmitting on 160, the 
power returns to the correct level from 170w.

I think this caused me to smoke the matching network on my 160m shunt-fed 
vertical during the IARU contest. I was driving an Ameritron AL1500 amp, which 
only needs about 50w drive to make 1500w out. When I first went to 160 in the 
contest, the swr jumped like crazy, and I confirmed by going outside that a 
doorknob cap in the shunt feed was smoking! I can't remember if I was on 10m 
before I qsyed to 160, but it is possible since I made a bunch of contacts 
there. I suspect that the increased drive put me way over 1500w for a short 
time.

Torsten
N4OGW/5