http://elecraft.85.s1.nabble.com/K3-slightly-off-frequency-tp7629167p7629183.html
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
Not discussed is the "last digit" syndrome. Only believe the last digit
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
> Immediate apologies, this is a tad long but hopefully interesting.
>
> My K3 has the new synths (2), K144XV with REF Lock option and all is
> well without doing the KREF3 modification. I guess this is an it either
> works or it don't situation.
>
> With respect to the post about off frequency, not a simple answer.
>
> Firstly accuracy and resolution or precision are not to be confused, one
> doesn't imply the other. The K3 display may read to 1Hz but that doesn't
> mean it is accurate to 1Hz. If folks in a net all set their various
> digital displays to the same frequency it doesn't mean they are all on
> the same frequency.
>
> As has been mentioned there is a perfectly good method of adjusting
> calibration detailed in the handbook using WWV. It doesn't directly tune
> or otherwise adjust the reference oscillator (TCXO) but lets whatever is
> involved with controlling the synths what the reference frequency really
> is so that the error can compensated for mathematically. That is my
> understanding. I have been using this WWV method for the past nine years
> and was convinced that the TXCO was probably drifting with age and have
> often said that the CONFIG: REF CAL is not a one off adjustment. This
> unfortunately turns out not to be the whole story.
>
> Now it has to be said, the TXCO does drift, shock, horror. The high
> stability option is described as being +/-1ppm but actually seems to
> achieve 0.5ppm, at least mine does. How do I know this? I don't have a
> frequency counter, or at least I didn't and in fact you don't need one
> as will be discovered later. What I do have is the K3 external reference
> frequency lock option fitted and recently purchased a Leo Bodnar GPS
> disciplined reference clock which provides the necessary 10MHz drive for
> the ext ref option. With a new toy you have go to play and what I
> discovered follows.
>
> With the GPS source up and running, turning on the K3 and enabling the
> external ref feature the REF CAL measured TXCO frequency will be
> displayed, counted by the ref lock module. This frequency will be seen
> to change with warm up. My TXCO has a positive drift of approx 26Hz+/- a
> couple in the first three minutes, for the next 30 minutes it drifts a
> further 6Hz. Thereafter it slowly drifts until after four hours it tops
> out at +57Hz or so and by this measurement method stops drifting.
>
> However, there is more. At the same time as noting the REF CAL frequency
> every minute, for the first hour, I also recorded the front panel (FP)
> temperature. This started at 23C and over the measurement period
> increased to 38C. It was observed that the rise in FP temp and the drift
> follow a similar trajectory. As would be expected.
>
> I have monitored the REF CAL frequency for a couple of months now and
> note that the TXCO drift on my unit always maxes out at more or less the
> same reading.
>
> I have reached the conclusion that if at least two hours of warm up are
> allowed and a WWV cal operation is then performed you will be in a
> pretty good place with respect to total drift and subsequent overall
> frequency (dial) accuracy.
>
> It is easy to discover the pattern and extent of the drift of the TXCO
> in your K3 without any external measuring equipment, external ref locks
> or such.
>
> 1. From cold tune to a WWV frequency you will be able to receive
> reliably for a few hours. Higher the better.
> 2. Wait 5 minutes or you will be chasing a moving target.
> 3. Perform a REF CAL and note the reading.
> 4. Perform a REF CAL every half hour, less if you feel like it and note
> the readings.
> 5. Eventually the reading will really slow down and more or less stop
> changing.
> 6. Choose a reading that indicates that the bulk of the drift has
> happened, (1-2 hours)
> 7. Set REF CAL to that reading and sleep well. Good enough for HF.
>
> Given the highest and lowest frequency readings a further experiment can
> be performed to discover the consequences of the TXCO drift on dial
> accuracy. It only takes a few minutes.
>
> 1. Set REF CALL to the lowest recorded frequency and tune the K3 to a
> stable test frequency, I used my XG3, note the "dial" reading in fine
> frequency display, 3 digits after the decimal point. I used the CW auto
> spot feature to do the final tune.
> 2. Set REF CAL to the highest recorded frequency and retune the K3 as
> above noting the "dial" frequency.
>
> The lowest to highest REF CAL frequency is the total TXCO drift in Hz.
> The lowest to highest "dial" frequency will indicate the consequence of
> that TXCO drift in Hz. You will be pleased to note that the two do not
> match. The K3 tuned frequency change is rather less than the TXCO drift.
>
> Reset the REF CAL reading to that you previously chose to be your K3's
> sweat spot.
>
> The above shows why I was wrong in relating the REF CAL frequency
> changes to TXCO ageing. In reality it all comes down to warm up time.
>
> I hope this was worth the read.
>
> Regards,
>
> Mike VP8NO
>
>
>
> On 09/04/2017 05:51, Roger via Elecraft wrote:
>> KREF3MDKT This modification increases the output levels from the KREF3
>> Reference Oscillator to provide proper drive levels for K3 transceivers
>> equipped with a KSYN3A synthesizers the KRX3 or KRX3A sub receiver and
>> the
>> K144XV 2-meter transceiver.
>>
>> 73, Roger
> ______________________________________________________________
> Elecraft mailing list
> Home:
http://mailman.qth.net/mailman/listinfo/elecraft> Help:
http://mailman.qth.net/mmfaq.htm> Post: mailto:
[hidden email]
>
> This list hosted by:
http://www.qsl.net> Please help support this email list:
http://www.qsl.net/donate.html> Message delivered to
[hidden email]