Cyclic error in original K3 synthesizer

Bob Benedict, KD8CGH

I found a few old (09) references in the mail list archives to " cyclical error problem with the old synthesizer".
Does anyone have more info on this? I ask because I have an older K3 (#70) with a second receiver so I haven't upgraded to two new synths. I did just order an installed a K3EXREF and I'm wondering how much it will help.

Examples (

"Back in the day, Elecraft informed us about that cyclical error problem with the old synthesizer. It would now be good to have a definitive statement about the frequency setting accuracy in the new synthesizer.
The same applies to the TCXO error correction by the KREF3 module. Since corrections are applied by rewriting the declared value of the 48MHz reference frequency, presumably this can still only be changed in 1Hz increments.
A related question (raised by Don's post, just in) is the magnitude of any difference between the displayed sidetone pitch and the actual value within the synthesizer?
73 from Ian GM3SEK"
"All true.  Note that the standard oscillator takes two hours to stop
drifting instead of four.  Also the TCXO isn't.  There is no temperature feedback implemented.  One is essentially seeing the stability of the native oscillator.  Too bad feedback isn't implemented, I'd bet a lot of drift noted would disappear.
Not discussed is the "last digit" syndrome.  Only believe the last digit displayed if you can check that it is accurate.
With the standard synthesizers, as one tunes up the band the true frequency deviates from the displayed by up to +/-4 Hz.  There are times when one tunes up and the RX actually goes to a lower frequency.  Same problem when tuning downwards. In other words, there is electronic "backlash".  One has no way of knowing which point of this sawtooth error curve you are when choosing to calibrate to WWV.  This is essentially fixed with the new synthesizers.

73 de Brian/K3KO"



   Bob   KD8CGH