[Elecraft] RF in the K2 on RTTY

[Ken Beals]

OK you Elecraft gurus, I've got one for you:

K2/100 #3664 with all mods to date, all options installed, using KAT-100 
tuner.  Normal assembly, amplifier in the radio box, tuner in the skinny 
package underneath the radio.  Antenna is a random vertical loop, about 
125 feet or so around, peak at about 50 feet.  Feeding with coax, and 
have a ferrite choke at the feedpoint.  I'm not sure that has anything 
to do with it, since my previous antenna was a 100 foot wire fed against 
radials, and problem was also occurring.  At first I suspected RF into 
the power supply or radio, but now I'm not sure.  The main reason I 
suspect RF is that the problem goes away if I use a dummy load.  It also 
occurs on two different power supplies.

Occurs most notably on 40 meters. Power set to max will make it occur 
every time, reducing the power will reduce the occurrences until at 
about 25 watts it does not occur.  On RTTY, any transmission longer than 
about 7 or 8 seconds causes the K2 to demand an approximately 35 amp 
current spike from the power supply just after the K2 returns to receive 
mode.  This causes the supply voltage to drop to about 8 volts, so the 
K2 does a sort of power up reset (Elecraft in the display, frequency 
changes to the remembered power up frequency, etc.).   I'm monitoring 
the supply current with a sense resistor (0.03667 ohms) using a scope in 
differential mode to monitor the voltage across the resistor.  If the 
transmission is less than the 7 or 8 seconds, no problem.  The current 
spike starts about 1.5 to 2 ms after the K2 returns to receive mode, and 
the high current lasts about 2 ms, tapering off a bit before it returns 
to normal receive current.  The power supply recovers pretty quickly, 
the whole glitch is over in 3 ms.  I think the only reason the glitch 
stops is that the K2 reset puts it back in receive mode and shuts off 
whatever is pulling the current.  This supply is an Astron RS-35A with 
all the recommended RF bypassing done.  I have an MFJ 25 amp switching 
supply, and the problem also occurs, although the current spike is 
obviously big enough to drive that supply into total shutdown.  Regular 
key down transmit current is about 15 amps.

This will naturally also occur with the radio in SSB mode putting in the 
RTTY tones.  I also duplicated it using just a plain tone into the 
microphone.  This problem does not occur in CW mode.  I can leave a full 
power carrier on as long as I want and no problem on return to receive. 

Some other observations:  if I disconnect the KAT-100 and use an 
external tuner on the antenna, the problem goes away.  When the radio 
returns to receive mode, the bargraph seems to take about a half second 
to go down to zero.  On CW I notice that the bargraph instantly falls to 
zero when you come out of tune mode.  Also, sometimes when I activate 
the RTTY tones, it seems that the power control has lost control, since 
I get about 170 watts out (read on external wattmeter).  This doesn't 
happen every time, and I haven't figured out exactly what triggers it.  
Sometimes it settles out at 110 watts just fine.

This really smells like RF getting in somewhere, but I can't figure what 
it's doing to the radio to begin to try and choke it off.  All the audio 
cables have chokes/shielding/beads/bypasses.  Looking at the KPA-100 
schematic, it would seem that the only devices capable of that kind of a 
current spike are the finals or the TR switch diodes, and I can't figure 
how they would be turning on on the return-to-receive cycle.  It seems 
like some sort of stored energy being dissipated as the transmitter is 
shut off.  I'm leaning toward the SSB module since there is no problem 
in CW mode ( I know, just stay on CW!!), but I can't make the connection 
between the SSB module and the PA..

Any ideas you folks can toss my way I'd appreciate.  Since I didn't have 
any problems in the initial construction, I didn't get to dig into the 
schematics the way some of you have.  I'm hoping someone can take all 
these pieces and complete the puzzle. 

tnx,

Ken K6MR