Here are the comparison values for the five 7000 and the 732B.
We had a calibration by Illya last year, and could confirm his values apart from F7000-1 within a few tenths of ppm.
Now, after 9 months we brought together again all four 7000, and made some quick measurements on 10th September evening using my HP3458A, which was not yet fully stabilized, and one more precise one next day. (Has anybody made a comparison with the KS 3458A on 11th? Please send me your results!)
Latter was done by Wolfgang and me, with my HP3458A, compensating each +/- polarity measurement by comparing and correcting to my LTZ #5, which is temperature compensated.
This accounted for the vastly varying room temperature, as my HP3458A features a T.C. of its LTZ reference of about + 0.25ppm/K.
Andrés measurements were corrected for the calibration protocol (+2.4ppm deviation).
His measurements then were all +2ppm high, compared to mine, but otherwise consistent in difference.
Wulf made another direct difference measurement between all 6 references, which were also consistent to my absolute measurements to less or equal than 0.2ppm.
I observed, or estimate, that my F7000-3 drifted about -0.8ppm compared to my reference group during the last 9 months, so the absolute values from my HP3458A should be closest to the actual values.
Unfortunately, we did not have any better baselining reference this year, so I also do not want to give any uncertainty estimation.
I also provide the environmental parameters for my whole trip from Frankfurt area to Stuttgart/Vaihingen.
You can see, that the room temperature on the MM 2021 Day had been between 23 and 27°C, not the best conditions for precision measurements, but I think we all did quite well despite that.
Then in the pressure data you can nicely see, that I had to move my car from 130m level to about 400 ..500m, when quickly ascending / descending the precursors of the 'Schwäbische Alb'.
As we've heard in one talk, atmospheric pressure variations might also influence such precision measurements.. but I doubt that we really are able to detect such an effect for this set up.
I can recommend these data sticks from UNI-T, UT 330C and UT 330A. They are quite cheap, consume very few energy (battery lasts ~ 1.5 years if you sample once per minute).
The user interface is simply horrible, so I provide the Excel sheet for convenient import and display of the .cvs - data.
Please also take notice of my cool basement lab
.
Frank