Hello, I am playing around with some reference voltages and found out that my LM399 references have 20-30 uV change in output voltage depending on the orientation of the reference. So when calibrating my voltage references with a Keithley 2000 I found out that the reading changes on the Keithley too depending of the orientation of the instrument. For example a 5 V reference is showing 4.99998 V with Keithley flat on the table. On the right edge of the instrument the reading is 4.99991 V On the left edge I get 5.00004 V. All in all the difference is 130 uV or 26 ppm!! A second Keithley shows similar effect with 120uV difference. Has anyone observed similar effect on other precision instruments like Solartron, Agilent or Flukes. Is this effect restricted to LM399-based instruments like the Keitley 2000 or are other references like LTZ1000 affected too? best regards Andreas

Andreas wrote: >I am playing around with some reference voltages and found out that my LM399 >references have 20-30 uV change in output voltage depending on the >orientation of the reference. > >So when calibrating my voltage references with a Keithley 2000 >I found out that the reading changes on the Keithley too >depending of the orientation of the instrument. What is the time constant? Slowish (i.e., could be thermal) or rapid (presumably gravitational/mechanical stress)? I have operated Fluke 732As and 5440Bs in various orientations and have not noticed any changes. With the 732s I was specifically looking for changes with orientation, and didn't see any to the limit of my methods (around 0.1 ppm). With the 5440B I wasn't specifically looking for an orientation effect, but I think I'd certainly have noticed 1 ppm. Of course, neither instrument uses an LM399 or LTZ1000. Best regards, Charles