[Elecraft] K3S KPA3A No Power over 12W

[Mike Walkington]

Hi Mike

Wow! That is a very thoughtful piece of work, and I imagine it will be useful for helping people solve future KPA3A problems.

I spent a fair bit of time last weekend reading about other KP3A issues. As a result, I had a close look at the plug/socket pins. I have a mixed of gold and tin. I cannot see any discolouration that may indicate excessive current and I cannot recall seeing the 12V ERR message; but I operate the radio remotely from the other end of my house so, I cannot rule that out. I used some contact cleaner to clean the pins and tested installed performance again but did not see any meaningful change. But there may be deterioration inside the sockets. Replacing the tin plugs/sockets may be something I should do.

I did review the signal loss behaviour by noting what happens on the P3. I could not reproduce the signal loss I had seen before. I may not have looked closely enough the first time round. Or maybe the contact cleaning changed something. I note someone used their XG3 to examine signal loss behaviour. I can look at doing the same. For now, I am assuming there is no drop in receiver performance above 12W. This may revert thinking about the diodes as the problem.

I did measure using K3S meters the current consumption, and output power using LP-100A, with the following results:
Set 10W (measured 9.8W),current 4.32A
Set 15W (measured 0W),current 3.08A
Set 25W (measured 0W),current 3.7A
Set 30W (measured 0W),current 4.13A
Set 45W (measured 0W),current 5.5A
Receive current at all power levels was 0.98A

So, where to from here. I read that someone used ribbon cables to allow them to extend the KP3A out of the radio chassis. I could look at that but am worried about high current.
I am looking to next measure changes in installed voltages 7R, 7T below and above 12W, ie to see if they are changing appropriately.

I read that someone had noticed that the problem goes away if the K3S RF Board, KPAIO3A, KPA3A "sandwich"is compressed. I will look to try that in case there is a socket/plug seating issue but I cannot see one at the moment. I note there is an Elecraft modkit E740267 that adds a bracket to hold the KPAIO3A down. I do not have that mod. I assume it was devised in response to unseating issues, but I could not find an explanation of why it was introduced.

Thanks for your ongoing interest in helping me solve this one.

Cheers
Mike, VK1OO


Message: 1
Date: Thu, 8 May 2025 16:55:46 +0000
From: Michael Carter<Mike.Carter at unh.edu>
To: Elecraft Reflector Reflector<elecraft at mailman.qth.net>
Subject: Re: [Elecraft] K3S KPA3A No Power over 12W
Message-ID:
	<BL3P223MB03850F4F1E54394696D69C84E488A at BL3P223MB0385.NAMP223.PROD.OUTLOOK.COM>
	
Content-Type: text/plain; charset="iso-8859-1"

Hi Mike,

Sorry for the extended delay in responding - I have not given up on you!

I've done some more thinking about your measurements, and I did some experiments of my own, although not on my KPA3A module. I was curious about the difference in using 'diode test' and 'resistance' measurement methods in-circuit.  I use a Fluke 115 DMM for most of my DC circuit testing, but I suspect most DMMs have similar methods of executing these two test modes. Bear with me here as I explain my experiments.

I tested a known good 1N4004 PN diode in 'diode test': 0.54V forward, 'OL' overload in reverse.  I then measured its resistance: about 1.4 M_ohms forward, 'OL' in reverse.  The resistance result for forward bias was puzzling, but after reading a bit more I realized that the auto-ranging DMM applies a constant current source to the DUT starting with a low value and then steps up to progressively higher current until the DUT resistance is within range.  A more indicative resistance test for a diode would use a higher constant current source value to bring the diode into full conduction.  I may try that again this morning and force a manual resistance range selection.

I then paralleled the diode with a 10 k_ohm, ? watt resistor.  This mimics, although without the fixed 100 ohm series resistors and DC inductor resistances, the R11 parallel path to diodes D5 and D8 in the KPA3A T/R switch.  In 'diode test' mode, the DMM reads 0.54 V forward, 1.94 volts reverse.  In resistance test, it measures approximately 10k_ohms with both probe polarities.

Out of curiosity I measured the 10k resistor alone in 'diode test' mode, something I've never considered doing until now: 1.94V.  The Fluke DMM steps up the applied voltage in 'diode test' mode until sufficient current is registered, then stops ramping up the voltage beyond 2.0 V and just indicates 'OL'.  Since I knew that the diode was good, the result for the paralleled diode and resistor in 'diode test' mode makes sense - the diode is not seen by the DMM with reversed probe polarity, but the resistor allows enough current to flow to 'trick' the DMM and register the same voltage in 'diode test' mode as the resistor in isolation presents.

So, what does this suggest about your T/R switch diodes D5 and D8?  If either diode failed shorted, that would have been evident in both the 'diode test' mode and the resistance measurements you made. If either diode failed open, that would not be evident in the in-circuit resistance tests, but it would have been evident in the 'diode test' mode, and your diodes had forward voltage drops in the correct range instead of 'OL' indications.  My sense is that both D5 and D8 are OK.

So, for me it's back to the beginning of your posts when you checked the two diodes in the +12V input bus.  The 1N5408 is a reversed supply polarity protection diode, and the BZW50-15 is a TVS diode, effectively a Zener that protects against a supply voltage in excess of +15 V DC.  The diode test voltages you measured for those diodes are affected by other circuit elements attached to the +12V bus, but I haven't yet looked at those details.

73,
Mike, K8CN