TiN

This whole discussion does not really apply to the LTZ1000 circuit, that is the bare 7.15..V output, w/o step up amplifier, as all drift parameters were attenuated by orders of magnitude, depending on the resistors, R1-R5. In the eevblog thread, several volt-nuts measured the real dependency, in contrast to the datsheet, which simply claims 1:100 attenuation for each of these resistors: R1 R2 R3 R4 R5 author 120 70000 70000 12-15k 1k -1/770 -1/250 -1/1400 +1/100 janaf -1/600 -1/238 -1/1000 +1/87 lars -1/656 TiN 1/714 1/250 1/3333 1/105 bbs38hot -1/670 -1/238 -1/1184 +1/74 -1/75 Andreas Therefore, even humidity effects at most affect R4/R5, with 1/75 impact. I also think, that it's possible for volt-nuts to measure such residual effects, by assembling several references, keeping the parameter of question (e.g. humidity or temperature) constant for the master reference, varying this paramater for a DUT, and evaluating the drift by monitoring the difference between the references. Same goes for longterm drift, by using a group of many references (A man with one, two, three clocks..) In the end, maintaining ultra precise references at first requires to care for controlled/constant ambient parameters, than to put too much effort to design references, which are immune to all possible drift factors. Frank