[Elecraft] Power hunting

[Jim Brown]

Exactly right, Joe.

AND, the crystal filters in the K3 are not ideal. Soon after I started 
running RTTY with my K3s in 2008, I saw incidental AM resulting from the 
ripple in the stock 2.7 kHz 5-pole filter. When I replaced it with a 2.8 
kHz 8-pole filter, which is flatter, but still not ideal, that 
measurably reduced the incidental AM.  And as Joe observes, real filters 
have both magnitude and phase response.

So while Joe and the WSJT team have done their best to minimize the 
extent to which they can excite IMD, real radios are not theoretical, 
and WSJT signals DO produce some IMD in real radios.

73, Jim K9YC

On 7/23/2023 10:08 AM, Joe Subich, W4TV wrote:
> This is true for any "N of M tone" FSK mode where N is 1.  However,
> the DSP filtering in both the K3 and K4 has a finite delay which
> causes some "spreading" of energy from one bit to the next which
> means there is *some* level of both tones present during the transition.
> 
> This two tone condition *can* result in IMD.  Since both the K3 and K4
> are "100W, 12V" final amplifiers, the finals will begin to go into gain
> compression (the early stages of saturation) at around 80W (look at the
> data sheets on the final devices) and will produce IMD if run with the
> PWR control above the onset of compression - *EVEN WITH A "CONSTANT*
> *CARRIER, CONSTANT ENVELOPE" MODE*