[Elecraft] KPA100 spurs

[Geoffrey Mackenzie-Kennedy]

Matti TF3MA wrote:

> I just installed a new X1 XTAL 18.432, as a part of upgrading
> the KPA100. I used components obtained locally.
> Stability of KPA100 is improved, at least it stopped oscillating
> on 17m band when I use SSB.
>
> But Now I find "lots" of spurs.
>
> Here is a list:
>
> 18.080, 21.415, 24.903, 24.984, 28.015, 28.260, 28.671, 29.314
>
> these are definetly from X1 as the freq shifts when PA is turned on
> by setting the power >10w.

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

The birdies which involve the HV Bias oscillator are created because either 
(1) The fundamental or a harmonic of the Bias oscillator is at a spurious 
response frequency of the receiver or (2) The fundamental or a harmonic of 
the Bias oscillator mixes with the fundamental or a harmonic of one or more 
than one of the other oscillators, or with a product generated by other 
oscillators (multiple mixing), to produce a signal at the frequency to which 
the receiver is tuned or a spurious response frequency. An example of (1) is 
the 28.671 MHz birdie resulting from a spurious response 4Fs - 6Fo = IF 
where Fs equals 36.864 MHz  i.e. the second harmonic of the Bias oscillator. 
Fs = spurious response frequency, Fo = LO frequency.

In my K2/100 the 16MHz Bias oscillator's fundamental and its harmonics were 
getting into the K2 receiver via the ribbon cable and loudspeaker cable. As 
a temporary fix I put a "clip on" ferrite designed for ribbon cable onto the 
ribbon cable towards the KPA100 end, and re-routed the speaker cable 
removing it from the KPA100 board, using RG-174 coax. While this worked for 
me, it did not in another K2/100 for reasons probably connected with RF 
grounds.

The best fix IMHO is to remove the existing Bias oscillator / rectifiers and 
repackage them in a small shielded box made from single sided PCB, copper 
facing inwards, to sit in the space occupied by the original circuitry. All 
leads entering or leaving the box should be well filtered, with a single 
ground connection between the repackaged circuit's groundplane and the 
KPA100's groundplane. Although not yet fully completed, I have gone down 
this route and am satisfied with the results so far. I will still keep the 
speaker's circuit away from the KPA100's board for "non birdie" reasons.

Please contact me off list if I can be of further help.

73,
Geoff
GM4ESD