LTC6655B long term drift

[BravoV]

Watch out for humidity change too.

[alex-sh]

The MSOP8 package is worse than the LS8 package.

Very much true.
In fact I did not realise that MSOP8 would be that bad in the beginning.
I am testing now LT1236LS8 with a proper pcb cut-outs.

[alex-sh]

Watch out for humidity change too.

This I can hardly control.
I do not even have a lab with const temp / hum.

[BravoV]

Watch out for humidity change too.

This I can hardly control.
I do not even have a lab with const temp / hum.

Although I haven't try it my self, may be one day, as curious what if the chip is sealed inside a glass tube. Of course you need to extend the pins longer like through hole components.

Just an example ampoule glass sealing, google for your self, random youtube video to give your the idea.

https://youtu.be/GEmINGq8HDQ

[Gyro]

The problem might be getting sufficiently hermetic sealing of the leadout wires to the glass - wetting, different thermal coefficients of expansion etc.

You'd be better with a small soldered copper sheet box construction with ready made hermetic feedthroughs, eg. https://www.ebay.co.uk/sch/i.html?_odkw=hermetic+feedthrough&_osacat=0&_from=R40&_trksid=m570.l1313&_nkw=glass+hermetic+feedthrough&_sacat=0

You can also then pass normal copper wires through them to minimise thermocouple effects.