[volt-nuts] Homebrew LTZ1000 reference?

[Mike S]

At 02:21 PM 11/29/2009, Alan Scrimgeour wrote...
>I'm interested in the possibility of developing 
>a reasonably priced reference that has 
>significantly better long term drift than the 
>LTZ1000 - long term drift seems to be the 
>biggest problem with voltage references.

The older Motorola buried zener references, as 
used in the Fluke 732a have "drift rates [which] 
are normally distributed with a mean of 0.47·10-6 
/ yr and a standard deviation of 0.5·10-6 / yr." 
- 
http://assets.fluke.com/appnotes/Calibration/deaver_msc01.pdf 
Since those rates are actually for the entire 
system, the reference itself may be even better 
(or worse, if there are offsetting drifts in the 
surrounding circuitry, I suppose).

That's better than the LTZ1000, in spec and 
practice (ibid). NLA, except by scavenging 
existing equipment (which should also mean 
they've "settled down"). The same ones appear to 
have been used in the Fluke 332* Voltage 
Standards, which can sometimes be found on eBay for ~$100.