[volt-nuts] Tilt effect on Keithley 2000
Marvin E. Gozum
marvin.gozum at jefferson.edu
Sun Sep 5 17:58:00 UTC 2010
Orientation has altered the point contact of my miniclips and the
test point. Its nearly eliminated by using alligators instead, this
is in measuring 10uV level changes.
With the Geller SVR board, the orientation of the board changes the
output voltage in the 10uV range too, but output stabilizes after
20-30 min to a new level.
As Charles already suggests, its noticeable that output is fast,
seconds, and over a wide range of uV when mechanical, and slow drift,
minutes, and eventually stabilized when it appears thermally related.
Joe Geller suggests the current draw of the LM399 heater is sensitive
to very small changes in ambient temperature, he exploits this in his
microgust anemometer circuit. When orienting an exposed chip, very
small temperature gradients exists between open air and a table,
particularly with fan or window making small air currents.
At 03:21 AM 9/5/2010, Andreas Jahn wrote:
>>What is the time constant? Slowish (i.e., could be thermal) or
>>rapid (presumably gravitational/mechanical stress)?
>The 20-30uV on my LM399s should be thermal. Time constant
>is about 10 seconds. Additionally the current consumption of the
>reference cirquit changes by 3-5%. So I think that the
>heater within LM399 has not a good thermal layout with respect
>to the temperature sensor on the same chip.
>For the Keithley differences I cannot tell the real time constant
>since I had slow reading with averaging 10 or 50 measurements.
>But since the effect is significantly larger than the reference alone
>there might be a mixture of mechanical and thermal effects.
More information about the volt-nuts