[volt-nuts] Traveling Standards

[Andreas Jahn]

Hello Bob,

yes you are right. The absolute ageing drift gets lower the more working 
hours are on the reference.

But the ageing rate is only a statistically value (averaged over a large 
quantity of individual references)
And: the environmental condititons have to be constant during the ageing.
Each mechanical stress (eg due to soldering) to the package may issue a new 
level of ageing.

For well aged plastic references the ageing effects may be invisible against 
humidity drift.
See also:
http://cds.linear.com/docs/Design%20Note/dn229f.pdf


At the moment I am doing some longterm drift measurements on some 
references.
The first was a hermetically sealed AD586LQ with stress reduced mounting in 
my ADC13 voltage converter.
On the attached picture you can see that the drift against a LTZ1000 #2 
(dark blue) was initially
around 3.4 ppm/sqrt(kHr) (light blue extrapolated line) from "day of birth" 
= day 200 on the x-axis.

Around day 410 I tried some "accellerated ageing" by loading the reference 
with 15 mA load during night.
The result is extrapolated by the pink line with 2.4ppm/sqrt(kHr) from day 
410 on.

Suddenly around day 470 the ageing stopped nearly. Except of some noise it 
seems to be stable.
I dont know the reason for this.

The last 5 days on the diagram is now without the 15 mA loading during 
night.

Best regards

Andreas


----- Original Message ----- 
Yep - and I have a question.  What does the spec of Xppm/sqrt(1,000 hrs) 
mean?
If we expect Xppm over the first 1,000 hours, does the stability improve 
with
long term operation?  What would be expected over the second 1000 hours, 
etc?

Would we expect a drift of 1.4 * X ppm over 2000 hours, sqrt(10) * X ppm 
over
10,000 hours, and so on?

Thanks,

-- 
=======================================================================
Bob Smither                                      Circuit Concepts, Inc.

-------------- next part --------------
A non-text attachment was scrubbed...
Name: ADC13_longterm.PNG
Type: image/png
Size: 48875 bytes
Desc: not available
URL: <http://www.febo.com/pipermail/volt-nuts/attachments/20121020/fff65d5d/attachment-0001.png>