Is my Fluke 8840A "messed up"

[PwrElectronics]

Hello EEVblog!!!

My first post!  I have been reading here for a while though.

A few years ago I found a Fluke 8840A in my employers recycling barrel.  There was a tag on it from the calibration tech stating "resistance readings off".

I finally got around to checking into this unit so checked it against 2 other reasonably decent meters I have:  A Fluke 45, last calibrated in 2007 and a Agilent U1242A I have owned for at least 8yrs with factory calibration.

I grabbed a pile of resistors and measured them all with all 3 meters.

I got these results:  The final column is the reading of the 8840A divided by the average from the 45 and U1242A.  I guess I am not seeing there is a big problem?



I was using the 2 wire mode on the 8840A, maybe the 4 wire is the problem.  I would need to make a harness to check this out.  Thoughts?

[Kleinstein]

The specs on the Fluke 8840 are likely much tighter than the fluke45 or the HP1242. So the resistor reading may be off by only a small fraction, not visible in this test. So the problem may not be large. Possibly just large enough to not let the meter pass calibration to specs.

For the smaller resistors 4 wire sensing would be kind of mandatory.

[coromonadalix]

Seems fine to me,  the last column seem consistant over a wide range of values, nothing extreme or abnormal at first glance.

If you want top notch precision wayyyyyy under 1% and a few 5% of all your tested values, you could send it for calibration or calibrate it ??


If you push it against a 6.5 to 8.5 digit calibrator  : you may end with the out of ranges values ... and may be another story with 0.1 % precision resistors, if you want to go this far ?

[TiN]

PwrElectronics
As Kleinstein outlines, F45 or 1242A are not suitable instruments for testing 8840A. 
No amount of math will help you to extract data that isn't there (true value of resistors under test) to judge 8840A readings accuracy.

On the positive side, meter like 8840A should be relatively cheap to send in for calibration (would expect $100 or less), if you want to get nice report with actual errors.
Adjustment also should be no problem.

[coromonadalix]

The cal lab in my town  charge me 125$ for an 34401a,   and 150$ for an older keithley 196  loll

50,000 count dmm or less its 75$

[PwrElectronics]

Thanks.

Yes, I know I cannot use my 2 lesser meters to check the calibration on the 8840A.  It was more about trying to figure out why the calibration tech had thrown it out.  Whatever is wrong with it (if there really is?) it does not seem to be a major fault as I would have thought.  I assumed it would be so far off that it would be obvious.  That or 1 or more ranges would stand out as being grossly wrong.

Looks like from the manual that calibration adjustments are done in firmware so there are no trimmers, etc inside.  I am not going to be messing with that in any case.  Just seems odd to me the tech tossed it.  Maybe it was considered too old to bother with?