Any opinions on the DMMSure LT1021 DMM CALIBRATOR

[iMo]

I saw a much worse soldering art, btw.. I would clean up the flux before a shot, however.

Interestingly they understood the epoxy package of the LT1021 is not good for this application, thus they potted it into something.

The issue with the epoxy package in case of the LT1021 (and any other reference in epoxy package where you claim X ppm stability) is the hysteresis due to humidity (the epoxy is a sponge pumping water in and out changing its volume, and thus pressing at the silicon die inside).
Those epoxy packages contain up to 5% of water. It takes 3 days at 150C to get rid of the water (experts say)..

I saw here 60ppm hysteresis after a couple of months the ref was switched off with the epoxy LT1021.

With a metal LT1021 it would work better (considering up to say +/-4ppm/C TC), with the potting the epoxy package it may somehow help but none potting is hermetic so they just made the water-content-time-constant longer.

You have to run the epoxy LT1021 at say 50C 24x7, afaik, to deal with that humidity issue.

[coromonadalix]

hand crafted for sure,  as description   lollll   

[CalibrationGuy]

iMo, Fluke apparently agrees with you. In their $12,000 754 calibrator, there are 2 of these references in the metal can package. I have clients who've used them for over a decade without needing calibration. Of course, they are not claiming ppm accuracy.

TomG.

PS - They ARE on a pcb though, just saying....

[GigaJoe]

a a while ago analog did plastic package with cavity over crystal as soic8 package,  comparing to regular soic8 they was a thicker
for plastic DIP8  i don't remember such thing. so the problem of plastic it transfer any forces to crystal , and that caused voltage shift...

so i try to mitigate , hook chip on flex wire, dry chip for 30 hours ( 120C) use a glass bottle with oil submerge it,  use a thermal chamber for constant temp.
combine all 3 options after 1 year of settling, let say, i got around 6 PPM for 6 month period voltage shift.  I cant say it may shift more, as my measurement was periodic. as well i have no idea what possible affect this shift as much as i can to try eliminate mechanical,  humidity and temperature shift.  Much possible i think, it a chip itself , non select ,reclaimed , made 1987

[Jay_Diddy_B]

Hi group,

A few observations from a DMMSure listing on eBay.



The datecode on the LT1021 is 0450. This is the 50th. week in 2004. This suggest that the LT1021 are new old stock or recycled.



The grade is C. The C indicates the best initial accuracy, but the worst tempco.

Regards,
Jay_Diddy_B

[GigaJoe]

If I extract gold from a "gold socketed", will I be rich?  I think it solid plan ... what else this thing need for ..

[Doctorandus_P]

Makes me wonder where they find the 10kV version of the LT1021.

[CalibrationGuy]

Makes me wonder where they find the 10kV version of the LT1021.

Just 10K pieces in series, of course. Just a board layout issue.

TomG.

[Alex Nikitin]

Makes me wonder where they find the 10kV version of the LT1021.

Just 10K pieces in series, of course. Just a board layout issue.

TomG.

Only 1000 pieces, it was actually done previously, IIRC.

Cheers

Alex