[Elecraft] Problems on 40M with 100 watt section

[Ronald Cox]

Tom,
I am having a very similar problem on my KPA100 which was shipped  
right after K2 serial number 5372 last month. Thinking it must be the  
final transistors, I ordered a new set. I would rather not change  
them if there is any possible solution. I just went into my KPA100  
and moved the RFC away from the SO-239 as much as possible. Now I can  
modulate a PSK31 signal at any power setting to a dummy load, but  
when I put the real antenna load on with an SWR of 1.1-1. it goes  
into parasitic oscillation after about 10 seconds of the  
transmission. I must have a more serious problem, or I need to do  
something more (to the RFC ?) to eliminate the parasitic oscillation.  
Here is what I wrote to Elecraft support before I saw your note on  
the reflector:

Gary,
I think I have had some kind of early failure in the KPA100. I made  
about 10  successful PSK 31 QSOs running 25 to 45 watts into an  
antenna with less than 1.5:1 SWR on 7.070 MHz. I Have been improving  
the station throughout this time with the purchase and installation  
of a new high end Creative Labs sound card, and a MicroHam USB  
Interface II. With the addition of each item, the K2/KPA100 seemed to  
be more prone to blow a fuse when PSK31 modulation was applied. I  
always run the modulation so that no ALC bars are showing, and I have  
lowered my SWR from 1.5:1 to 1.0:1 thinking that may be a  
contributing cause. With the antenna in the attic I thought a first  
it was a problem of RF getting back into the rig/power supply, but  
after installing the MicroHam USB Interface II, and putting $30 worth  
of clamp-on RF filters on all the leads in and out of the K2/KPA100,  
I am convinced that the problem is in the KPA100.

Here are other actions/tests I have done to isolate the problem:

1. Installed a 13.8 Volt float battery on the power supply leads to  
eliminate possible power supply influence to problem, no change  
except the fuse blows before I can turn off the rig now.

2 . The K2 has no problems before the 11 watt switch over to KPA100.

3. After switch over to KPA100 there is a slow power variation at the  
11 watt setting. Power varies from 9 to 15 watts at about a 1 or 2 Hz  
rate on a good 50 Ohm dummy load until 10 to 20 sec when the KPA100  
draws over 20 Amps and blows the fuse if PSK31 modulation is present.  
Using tune, the power can be increased to 100 watts with out blowing  
the fuse. The power variation between 11 and 20 watt settings is  
still there using tune.

4. I have disassembled the KPA100 several times and checked solder  
joints re-soldering all that looked suspect. The thermal pads under  
Q3 and Q4 have come apart with the repeated disassembles. I now have  
electrical tape in place of the thermal pads on the last reassembly.  
This is not a root cause factor as I just confirmed that the problem  
is the same as before the last disassembly and have not operated the  
K2/KPA100 on transmit since the one last test.  I will request a new  
set from Scott.

I think the only components that could breakdown, blow the fuse and  
recover like it does (able to put out 100 watts on tune) is Q1 and  
Q2. R38 and R39 have a slight brown discoloration, but I would expect  
that with operation at high power. During the few QSOs I had, I was  
checking the heat sink temperature on longer transmissions. The heat- 
sink did get warm to the touch, but was within the temperature where  
I could keep my hand on it. I have installed a PC muffin fan now on  
top of the heat-sink. I probably did run the KPA100 at 45 watts on  
PSK31 for a few minutes during those QSOs. The KPA100 has never  
displayed any indication that it was exceeding its calibrated temp  
setting or done anything I would call a temperature power reduction.  
Do you think I should replace Q1 and Q2 and all four thermal pads?  
Please advise.

Ron Cox, W9KFB
K2 #5372