Electronics > Metrology

Tales from the mouth of a voltage reference ageing box

(1/3) > >>

Andreas:
Hello,

as some other volt-nuts I operate some ageing boxes to pre age mainly 5V and 7V voltage reference ICs.
From time to time I am harvesting those with the least drift to build a stable ADC or voltage reference cirquit.

After a 9 day power outage in my lab I had some (partly) unexpected excursions in my daily measurements.

The references in the ageing box are operated 24/7 so the temperature within the box is ~10 deg C warmer than the environment.
Every morning the box is heated for ~1.5 hours to 50 deg C (+/- 0.1 deg C) during the measurement of the references.
Most of the references are 5V buried zeners in hermetically tight packages (AD586LQ/LT1236AILS8-5) but also one 10V and some 5V devices in plastic packages.

A further description of the DIP8 ageing box can be found here:
https://www.eevblog.com/forum/metrology/small-oven-controler-for-voltage-reference/msg1655015/#msg1655015

with best regards

Andreas

Andreas:
Chapter 1: REF102AP

This reference (10V) in a plastic DIP8 is a lower grade version of the REF102CP which is used in the infamous 10V reference.
Even in the daily heated state (where I would think the references are "dry") I get around 6 ppm seasonal changes which correspond to humidity which is ~40% rH during winter and ~60-70% rH during summer in my area.
After power outage (in summer) I can see additional 6 ppm initial drift which needs about 1 week to stabilize again to the value of daily heating.
As comparison some AD586LQ in cerdip package which are not affected by seasonal changes and the power outage.
So the humidity coefficient (unheated) of my sample of REF102AP is around 0.5 ppm / % rH

with best regards

Andreas

Andreas:
Chapter 2: LT1019CCN8-5

other manufacturer this time a 5V bandgap reference with similar story after power outage in plastic DIP8 package.

This reference I have choosen to be evaluated because it has a (undocumented in newer datasheets) internal heater which
could be used to do a oven together with the internal temperature sensor.
Unfortunately the heater is connected with common ground and creates large offset shifts when operated.

Seasonal changes are already somewhat higher ~23 ppm in dayly heated state.
Excursion after power outage additional ~30 ppm
And again you can see that it takes nearly 2 weeks to get to the old ppm value before power outage.
So the humidity coefficient (unheated) is around 2.5 ppm / % rH

Some older drift data of the LT1019CCN8-5 is here:
https://www.eevblog.com/forum/metrology/ltc6655b-long-term-drift/msg2360088/#msg2360088

with best regards

Andreas

Andreas:
Chapter 3: MAX6250A + AD586MNZ

Again plastic DIP8 packages with 3 samples of MAX6250A and one AD586MNZ

The MAX6250A show still a relative high average drift of ~5 ppm/year after 7 years of ageing. (initially ~10 ppm / year)
Overlaid by some seasonal changes of ~5ppm due to humidity.
Excursion after power outage additional ~5-6 ppm with time constant ~1 week to settle after power outage.
So the humidity coefficient (unheated) is around ~0.5 ppm / % rH

The AD586MNZ (DIP8) shows directly the difference to the AD586LQ (CERDIP8)
With daily heating the AD586MNZ has settled to <1 ppm/year ageing drift and < 2 ppm seasonal changes.
But during power outage up to 5 ppm drift due to humidity.
So the humidity coefficient (unheated) is around <0.3 ppm / % rH

with best regards

Andreas

Andreas:
Chapter 4: MAX6350

Maxims "best" reference with 1 ppm/K tempco.
But again the plastic DIP8 package spoils the usage as precision device.
The MAX6350 show still a relative high average drift of ~5 ppm/year after 7 years of ageing. (initially 10 ppm/year)
Overlaid by some seasonal changes of ~5ppm due to humidity.

Excursion after power outage additional ~3 ppm with time constant 3 days to settle after power outage.
So the humidity coefficient (unheated) is around 0.2 ppm / % rH

with best regards

Andreas

Navigation

[0] Message Index

[#] Next page

There was an error while thanking
Thanking...
Go to full version