Re: [Elecraft] Strange K3 & KPA500 Issues fixed with a longer coax cable?

Hank P Wed, 19 Nov 2014 14:08:43 -0800

There are certainly other possibilities.

Making an amplifier unconditionally stable with any combination of sourceimpedance ,load impedance , drive level , and supply voltage and achievingany reasonable gain , efficiency , bandwidth , and able to withstand aboutanything a customer can dream up at the same time is a real challenge . Andwhat these source and load impedances are in band or well out of band have astrong bearing on stability.

Then to add to the fun, if you are not a high quantity buyer of thesemiconductors , and therefore cannot demand there be no changes inprocessing, you can do a design and a couple years later all of a sudden youhave a stability problem that never existed because the vendor decided tomake a process change such that the device still met the typical specs ,buthad a much stronger propensity for half freq oscillations. Some processchanges that raise low frequency gain without much effect on high frequencygain can make an amp have a tendency for low frequency oscillations becausethe decoupling /low frequency circuitry loading is now not sufficient tosuppress that .

Enough of that , what CAN be happening here is with certain cable lengthsbetween the amp and the K3 , in band or out of band impedances presented tothe K3 could well make the K3 PA want to have a spurious output like halffrequency oscillation or low frequency oscillations . I do not recall if theKPA500 has any resistive broadband padding at the input in normal operation, I believe they say they switch in a 3 dB pad when extreme overdrive isapplied. Resistive pads are a great way to limit in band or out of bandimpedance excursions - BUT you give up amp gain dB for dB.

Same is true at the output of the amplifier -(or combinations of in andoutput) certain loads may make the amp more prone to a spurious output athalf the driving frequency - even though the load may look like a perfect 50ohms in band at the fundamental frequency . This is another joyous featureof semiconductor power amp designers get to experience whether bipolar orFET .

The designer has to be careful what the out of band impedances of the outputfilters are - and just because you have a 1:1 SWR in band with your antenna, the out of band impedances can be darn near anything .

As I recall , in this case , he had the same problem even into a dummy loadon the output. So just as an experiment it might be interesting to see of a3 dB resistive pad between the amp and the K3 made the whole thing totallyindependent of cable length. I am m not talking about cable loss , I amtalking about limiting the possible impedance excursions especially out ofband .

Enough of this - I spent 30 years of my life designing and putting intoproduction these beasts , I enjoy the fruits of the Elecraft designersefforts .have no interest whatever in doing that again. I like antennasmuch better they don't half f or oscillate at low frequencies- at least sofar .