Wavetek/Fluke 7004 system

[Dr. Frank]

Well, I was the other lucky guy to pick up the 3rd 7000.

Many, many thanks to branadic, for organizing the purchase, the transport by TiN to Germany, and organizing linear PSU and plug!

On 18.12.20, in Stuttgart, we had a brief handover session, i.e. comparison of all 4 units and my LTZ #5 'hot reference', using branadics DMMs.

The differences of voltages between the 4 units which TiN measured before in U.S., see blue stickers, was confirmed to +/- 0.5ppm or less, so that we have confidence, that the calibration made by TiN was still valid after the transport, and conditioning cycles afterwards. My LTZ #5 was in accordance with the mean of the 7004 system to less than 0.4ppm, and relative to my 7000 #3 to about 0.1ppm (!).


I maintain my lab volt meanwhile based on 8 references, (plus 3 additional ones) which I'm monitoring for many years.
By taking baseline point, that is a comparison to calibrated standards, I'm able to transfer S.I. volt to my reference group, and I'm also able to make a prediction about the drift of this group, which I estimate to be accurate to less than 0.5ppm/year. The last baseline points were taken on June 2019 during the Metrology Meeting using a 732A and a 7000 which had less than 0.5ppm uncertainty, and another comparison against a 7000 in October 2019, with the same uncertainty.

TiN's calibrations, being literally transferred across the Atlantic Ocean, were accurate/uncertain to 0.3ppm.

I brought my 7000 #3 and the LTZ #5 'hot' back to my lab, that's another 250km ride by car, no change of latter reference voltage was observed, compared to the others.
(That's different, compared to another LTZ reference, which I recently sent 'cold' via DHL, and which drifted upwards, due to transport temperatures below 5°C.)
 
The voltage of the 7000 in comparison to my reference group (drift extrapolated mean value) agreed within 0.2ppm to TiN's measurement.

I don't want to exaggerate on this result - the complete error budget has to be calculated more carefully, of course, to get a trustworthy uncertainty figure.

But anyhow, I found this whole experiment of a complete ring transfer: MM2019 - TiN / U.S. - Stuttgart - My Lab very interesting and instructive.
Interesting, because transfers / comparisons and apparent correlations can be realized at a level of few tenths of a ppm.
Instructive, because the transport, transfer and comparison methods have a lot of pitfalls, which have to be accounted for and to be overcome.

Maybe this experience is also valuable for the American Cal Club, and the planned EU Cal Club.   

Frank

[branadic]

Just in case you were wondering like I was, we have printed a short piece of the enclosure using FDM printer and PLA material with the CAD files linked above. It does fit the W/F7000 case but also the HB TP 14 220 R1 cases from Fischerelektronik that are similar to the ones used for the 7000 system.
As most printers can't print the full length of the case in one piece, I've tested if two pieces can be glued together with acetone and it seems to work perfectly. Time to have some cases printed in preparation of MM2021.
Only front bezel design is missing, but I think this is less important.

-branadic-

[branadic]

So last night the two parts of the encluse were printed in PLA. You can either use superglue or acetone to stick them together. For those of you with a printer not large enough to print it in one piece I've attached a version with both parts and guides on one of them.
The glued part was sanded, coated with primer and later with filler. Ready to get a finish in whatever color you prefer.

-branadic-

[branadic]

Update on this topic, I've designed a front bezzel too. Data for the transport case and bezzel available here:

https://www.eevblog.com/forum/metrology/wavetekfluke-7004-system/msg3369676/#msg3369676

-branadic-