[Elecraft] K2 w/KSB CW reverse BFO

[Ron D'Eau Claire]

Yes... As you know, during CAL PLL the LO oscillator frequency is
measured and the control voltage required to put it on the desired
frequency is measured and stored in memory. 

During CAL PLL is the only time the actual LO frequency is measured. In
all subsequent operation, the logic puts the voltage level that was
stored during CAL PLL on the VCO and ASSUMES that will put the LO back
on the original frequency. If anything changes the frequency the LO will
generate when supplied that control voltage, it will no longer be on the
desired frequency. Things such as replacing covers, temperature changes,
etc., all will introduce an error. Also any error in producing the exact
voltage needed will produce a frequency error. That's where the
resolution of the digital-to-analog converters (DACs) affects the
frequency accuracy..

The same sort of errors occur with CAL FIL when the BFO frequency is
measured and the corresponding control voltage is stored in memory. 

The whole CAL PLL and CAL FIL process is just a fancy way of calibrating
the VFO and BFO "dials". In an old analog radio we'd have dials with
pointers and paper scales that we could mark up to "calibrate" the LO
and BFO. We'd measure each oscillator with a good frequency counter at
several points over their ranges and put marks on the paper scale so we
could reset the pointer to that place in the future. How close to the
desired frequency would be when we turned the dial back to that mark at
some future time would be subject to several things. 

First, we would NOT measure every possible dial setting. We'd take a
bunch of samples, and then estimate the frequency in between. Halfway
between two "marks' would be halfway between those two frequencies, etc.
That's pretty good, but it's not perfect. The algorithm in the K2 logic
does something similar. It defines certain points at which the frequency
was measured and then calculates ("interpolates" to use the $5 word) the
settings in between those points that will produce the frequencies in
between those points. 

Secondly, when we marked the dial on the exact frequencies we did
measure, the mark itself would be wide enough to see. That  means that
when we put the dial back on that frequency we have to squint and look
carefully to get the pointer set exactly on the mark. We can't be
absolutely exact. There will always be some 'resettability" error.
That's what happens in the K2 because the DAC's don't have perfect
resolution. 

Next, the original calibration was only as good as the accuracy of the
frequency counter used to measure the frequency of the oscillator. Any
error there will produce a permanent error in the calibration. That's
why it's important to have the oscillator right on 4.00000 MHz so the
internal frequency counter in the K2 will be as accurate as possible
when running CAL PLL or CAL FIL. 

And FINALLY, and drift in the oscillators due to temperature or other
environmental conditions - or anything else that will change the
oscillator frequency without moving the "dial" such as replacing a
cover, etc. - will add errors to the system. 

Once calibrated, the K2 is only as good as it's LO and BFO frequency
stability over time and under changing environmental conditions.

Ron AC7AC
K2 # 1289

I've noticed that putting the lid on affects the frequency a bit, so I
find myself struggling to adjust the oscillator so it will be right when
I put the lid back on!  

Does anyone disagree with this?

Vic K2VCO