LM399 drift tendency with age (which direction)?

[Alex Nikitin]

Here is another attempt to observe this drift - 20 min run, 10 min before ACAL and 10 min just after. There is about 3 min gap in the data when ACAL was run and the measurement re-set. The log interval is 5sec and the HP3458A is set for NPLC50. The source is the same "not very stable  " LTZ1000 reference. Vertical scale is 0.5ppm/div.

Cheers

Alex

[HighVoltage]

I have now two 34470A instruments and both have this behavior after an ACAL.

My older instrument has a significant jump of sometimes more than 20 uV and then finds it way down to the almost original value after 20 to 30 min. This instrument was just sent to Keysight for inspection / repair / calibration and hopefully will be working perfectly afterwards.

The newer instrument has a similar jump as shown in the last graph by Alex, less than 1 ppm and I am also not too much worried. But somehow the implementation of the ACAL procedure needs to be addressed by Keysight in a future FW update, it seems.

@Alex,
What software are you using to log the data from the 3458A and 34465A at the same time?
 

[Alex Nikitin]

@Alex,
What software are you using to log the data from the 3458A and 34465A at the same time?

I log the data to a USB stick directly on the 34465A and use the Keysight meter to trigger the HP3458A, which is in turn logged over GPIB with EZGPIB program. Then I combine both logs in Excel. Not that sophisticated, I know  .

Cheers

Alex

[kj7e]

@HighVoltage and @Alex Nikitin,

Quick opinion, does the 34470a perform significantly better (subjective) than the 34465a in terms of drift, enough so to justify almost twice the cost?  No need for a multi paragraph answer, just a quick opinion.  I may have an opportunity to upgrade my 34465 to a 34470.  I have not seen any conclusive evidence to show the 34470 performs significantly better.  I think Frank did a comparison once, using one of HV's meters IIRC, not sure where I read that now.

[Dr. Frank]

I have now two 34470A instruments and both have this behavior after an ACAL.

My older instrument has a significant jump of sometimes more than 20 uV and then finds it way down to the almost original value after 20 to 30 min. This instrument was just sent to Keysight for inspection / repair / calibration and hopefully will be working perfectly afterwards.

The newer instrument has a similar jump as shown in the last graph by Alex, less than 1 ppm and I am also not too much worried. But somehow the implementation of the ACAL procedure needs to be addressed by Keysight in a future FW update, it seems.

Hello,
what was your measurement setup, to identify this 20uV jump?
This jump is measured in which DCV range, and relative to which input voltage?

I never recognized such a behavior  on my 34465A, but would like to reproduce that, if you would kindly guide me, where / how to look at that problem, please.

Frank

[Alex Nikitin]

Here is the same measurement done with the reversed leads from the LTZ1000 voltage reference (so the voltage readings are negative on both meters). Again, an ACAL is followed by about 0.7ppm drift over few minutes on the Keysight 34465A.

Cheers

Alex

[HighVoltage]

Hello,
what was your measurement setup, to identify this 20uV jump?
This jump is measured in which DCV range, and relative to which input voltage?

I never recognized such a behavior  on my 34465A, but would like to reproduce that, if you would kindly guide me, where / how to look at that problem, please.

Frank

In my case:
- Geller Labs SVRT 10V reference that was turned ON for at least 24 h
- 34470A set to 10V DC manual range, 100 NPLC

In trend chart mode one can easily observe this behavior. After a few hours of collecting data, just press ACAL and after the ACAL is finished, the measurement value will jump up. In my case sometimes more than 20 uV. And if you let it sit for a while, it will start go down again to the original value as it was, before ACAL was pressed.

For my second 34470A, this jump is much smaller but also visible.