zener stability at 50-60 Celsius

[001]

Hi!

I got late 1980th voltage reference
It consist of common use zener at owen  (constant temperature about 50 Celsius) and current source
Is it good idea to use zener at so high temperature?
Is it ok to use simple zener at owen without any buffer etc? I see what in my device zener is directly connected to output plug (a sort of simple parametric regulator)

Thanx!

[Kleinstein]

The zener diode is relatively low impedance (often in the 10 Ohms range). So there is no buffer needed in most cases. The buffer can be outside at least. In the early days it might have been attractive to have the amplifier or similar also at a constant temperature, but this is not such a big issue any more with modern OPs, well better than 1 µV/K. So it is more like keeping the power consumption inside the oven low, to avoid to much self heating.

40-50 C is not yet that hot: the LM399 is running in the 90-100 C range and the LTZ1000 is also usually run above 40 C, maybe 70 C.
The 40-50 C is about the lowest temperature one can keep constant with just a heater if the outside can go to 30-35 C.

[001]

Thanx! 

Is it good idea to use common zener at constant temperature  instead special  unit without owen?

[Kleinstein]

In the old days low noise compensated zener references like 1N829A where a common choice. By choosing the right current one can trim the linear TC, to get a low TC - at least in a small temperature range.
An small extra oven could be easier and better than trying to compensate the second order effects and was also common practice.
With the  linear TC adjusted via the current,  the oven does not need to be super stable.

Non heated references can get a low TC, but usually only over a small temperature range and the compensation circuit also adds to the drift and price. So chances are a TC adjusted zener in an oven can be lower TC than a good non heated reference.

Compared to the LM399, a temperature regulated 1N829 ref diode could have similar stability / drift and lower noise. However the extra oven is quite some effort and the diodes might need screening for really low noise ones.

A suitable LTZ1000 circuit should be more stable and usually lower noise.

[001]

So my unit is probably old russian
the old  zener is marked as 2C190D (russian letters) so it is 0.0002% grade by datasheet
Is it still good?
Can I substitute it with some 1N*** zeners? What?

[Vgkid]

That is a a rather low temperature coeficcient , I would keep it. Plus the other components were matched to it. Since this is ovened , it will drift less than most meters over a temperature shift.