[Elecraft] KPA100 failure

[Don Wilhelm]

Paul,

I have three possible guesses.
First is that the KPA100 was developing the full power output that it is 
capable of (often 180 watts) because the wattmeter diodes D16 and D17 
have been damaged by static (wind, rain, snow, or nearby lightning).
If those diodes are damaged the MCU thinks there is no power output and 
drives to it full power capability.  If left to continue, the PA 
transistors in the base K2 will eventually destroy themselves.  The PA 
transistors in the KPA100 are quite hardy and seldom need to be replaced 
(I have only replaced 1 set in the last 14 years of repairing K2s).

The 2nd possibility is that there is some oscillation.  That can happen 
more frequently on 40 meters, and the KPA100 tries to transmit at about 
8 MHz.  Be certain your KPA100 is at the latest level.  If you have blue 
toroid cores at RFC1 and L16 it is the latest.  If not the latest, 
upgrade it with the KPA100UPKT.

Thirdly, were the PA transistors tight?  If not, the PA transistors will 
heat the board around the transistors and not be thermally  coupled to 
the heatsink.  Are the thermal pads properly in place?
After initial assembly (and anytime afterwards if the thermal pads are 
changed) you must re-tighten the screws after about 10 hours of 
operation at high power.

A 4th possibility is that you have a problem in the T/R switch.  After 
you get it up and running, do the Diode Voltage checks in the manual. 
If you have an older manual but have the latest upgrade installed, use 
the chart in the KPA100UPKT instructions rather than what is in your manual.

Note well, the KPA100 is rated for only 10 seconds of keydown power at 
100 watts, so you did exceed its maximum rating.

If you melted the fuseholder without blowing the fuse, you may have had 
bad contact from the wire to the fuse.  That would also result in a low 
voltage to the KPA100 and cause it to draw greater than the normal 
amount of current because the K2 controls try to keep the power constant 
- less voltage requires more current for the same power.

Hopefully some of that information will answer your question.
Do all initial testing into a dummy load and not an antenna and bypass 
the KAT100 by connecting the dummy load (and perhaps a wattmeter) 
directly to the KPA100 - remove the coax between the KPA100 RF out and 
the KAT100 RF IN jacks.

73,,
Don W3FPR



On 1/5/2019 2:34 PM, Paul Wilton wrote:
> I was using my KPA100/KAT100 combination today to check out some changes to my OCFD antenna.  Previously I have had problems with the balun heating up and the SWR increasing over an extended transmit on 7MHz (eg 50w on an FT8 over).   I have made an improvement to my OCFD antenna by adding another choke to reduce the common mode current.  So in order to test it out, I planned to run 100w CW for 15-30seconds watching the SWR meter on the KAT100 to see if I was going to get the SWR increase.  This was done with a “cold” PA.
> 
> So I started the test and unlikely previously, the SWR remained good.  However at about the 30s mark, I saw some smoke coming from the KPA100 and then it went dead.  It also melted my fuseholder in the PA supply (although didn’t actually blow the fuse).
> 
> I have opened up the KPA100 only to find that the two 22R power resistors around the two final transistors (R38 and R39) have both unsoldered themselves from the board and were lying in the bottom of the case!
> 
> Clearly it is an easy job just to put them in again but before doing so I want to try and understand what is happening
> 
> So a few questions:
> a) I am hoping that my two finals are still OK.  Is there a simple test I can do perform applying power?  Using a DMM in resistance mode, I measure approx 2M between collector and emitter and 11.2R between base and emitter (other than the missing 22R resistors I haven’t disconnected anything else).
> b) Any idea what could have caused that much heat that the two resistors unsoldered themselves?  It can’t be a DC failure since there is a capacitor blocking the path in the event of a PA transistor shorting to ground.  I am concerned that there is some sort of oscillation going on ie the PA is starting to hoot.
> c) Any advice on how to proceed?
> 
>