[Elecraft] KPA100 power on 20m/30m, update

[Geoffrey Mackenzie-Kennedy]

The "Trap theory" makes a lot of sense, and Don's comment about low
tolerance C80 and C81 capacitors would seem to support it. But if the
resistors in parallel with C80 and C81 are the only "Inductors" that form
the L part of the traps, I would have thought that by simply changing R19
and R20 to so called Non Inductive types would cure the 30/20 meter problem.
But this might just move the problem to a higher band because of the
inductance of leads AND PCB traces between C80/R19 and between C81/R20. But
are these the only "traps"? Consider the low Q notch C80/R34 and C81/R35.

If the "traps" are formed only by C80/R19 and C81/R20 plus lead inductance
and the L of interconnecting PCB traces, yes increasing the values of C80
and C81 will shift the trap frequency down, but what happens if the value of
the new caps is at the high end of their tolerance range? Does the problem
start to affect 40m? (Slide rule warming up).Probably not. The same thought
about the notch circuits. So I am not totally convinced that the C80/R19 and
C81/R20 is all that there is to it, I hope that I am wrong, but I do wonder
about the stack of bypass capacitors at the centre tap of T1 as well, and
the other 2W resistors for starters.

Increasing the value of C80 and C81 will also affect the amount of feedback
around the finals Q1 and Q2 to some extent, but without doing the sums for
the parallel tuned trap situation, I think only by a small amount IF T1 and
the bypass caps at T1's centre tap are well behaved. The effect of a series
resonant notch would be different.  The behaviour of C58, C59 and C83 also
need to be checked. A measurement of the gain of the PA between 3.0MHz and
30MHz when driven from a pure resistive source, with everything to the left
of R34 and R35 disconnected except bias, would tell much. Dummy load at T2's
output. No doubt Elecraft already have this information.

So I can understand why Elecraft have not yet issued an official statement.
There are several "spurious" resonant circuits to be considered IMHO.

Some thoughts.

73,

Geoff
GM4ESD

P.S. Welcome home Vic!

----------------------------------------------------------------------------
-----

Vic Rosenthal wrote:

> I wish we could get an official statement from Elecraft on this issue.
>
> <begin MY OPINION>
>
> Apparently the metal-film resistors used in this application vary widely
> in inductance.  Some resistors have an  inductance that seem to resonate
> with the capacitors near the 30/20 meter bands, basically putting a
> parallel trap in series with the drive, and causing the driver stage to
> have to work harder.  This shows up as high current and/or reduced
> output on 30/20 meters.  Hence whether or not you need to change the
> capacitors and if so to what value depends on the resistors that you you
> have.  Note that this depends on the physical construction of the
> resistor and it's (unwanted) inductance, not its resistance.
>
> <end MY OPINION>