hello friends,
has anybody measured the tempco of the auxilliary burried zener in LT1034,
is it in the region of ca +2mv / degC ?
regards.
40ppm/deg.C = 0.28mV/deg.C (7V) @ 100uA !!, no sign on the 40ppm so can we expect that to be positive ppm's ?
But what TC to expect at 1mA or 5mA ?? 
No noise specs !, 20ppm/khr !
It will have to be Real cheap if its to be worth using your sparetime on 
I would rather bet on a well spec'ed buried zener in a hermetic ceramic house or metal-can and use my time on building a well controlled oven and stable current source 
thanks.
your data matches well with diligentminds.com(ken) statement of around 17% of -2.1mv/degC of Vbe tempco.
the next step should be noise characterization vs current on this (probably *only* accessible burried zener besides LTZ1000)
and if the noise levels are interesting (say sub 1ppm) then compensation strategy with a scaled down Vbe of a thermally
coupled transistor.
as far as the cost: LT1034CZ-2.5 USD 2.5/Qty=1 and USD 2.15/Qty=100
(you are interested *only* in the *auxilliary* zener)
composite:
couple LT1034 (TO-92) with flat side matching with 2N3904 (TO-92) leads opposite,
wrap kapton tape over them, make a small nichrome wire heater over this assembly,
wrap with kapton tape and thermally insulate, followed by heat-shrink tubing, and voila!
*maybe* a synthetic LTZ1000(or LTFLU/SZA263) look-alike.
as far as hermetic sealing goes, maybe dip the whole shebang into a molten pool of solder-glass.
(parylene maybe for the RICH MAN

)
still waiting for my samples

regards.
-zia