Fluke 8846A weird temperature coeficiance at measuring resistance at summer time

[Kiriakos GR]

I was up to perform resistance VS temperature stability test at one Vishay 500R high precision resistor  and I made the discovery that my 8846A multimeter this is by far more prone to temperature coeficiance changes when it inner case temperature changes from 38C to 30C ambient in the summer.

 Actually I were originally alarmed by spotting differences which did not make logical conclusions at my first log of 30 hours when ambient this changing a single Celsius (29C as Low and 30C as high).
The greater mystery recorded is that when environment temperature was lower than a single degree the resistance measurement this was higher at the multimeter.
Basically such behavior this is against of a how a resistor behaves, we all know that Ohms value get higher when the resistor this gets warmer and not the opposite.
And therefore what it was noted as change it was not the resistor value but the delivered measurement.

Neither the current in use at the resistance measurement and neither the size or type of this high precision calibration class resistor, this can justify such a change and or behavior.

Next performed experiment was of me at using a DC fan blower at the side of the DMM case this simply removing the hot air from the metallic case,  by injecting to it air at 30C room temperature.
The R value difference that is recorded above this is of how the 8846A electronics behave when the metallic case temperature this is at 30C instead of 38C.
Generally speaking by owning the 8846A for over a year, it metallic case temperature this is always 8C above room temperature no matter the season.
Due my new experiment in which the 8846A this were externally cooled down to 30C , this translates that this instrument it came at the operating conditions when room temperature this is close to 22 Celsius.

Test results

R High 499,9637  = Case temperature cooled down to 30C

R Low 499,9502 =  ambient 30C   , Case temperature hovering at 38C 

R Difference = 0,0135  ( Ohm )

This is part one of this article, I am now collecting graphs and data as proofs. 
Here is another topic relative to the resistor under test.  https://www.ittsb.eu/forum/index.php?topic=1584.0

[Kiriakos GR]

I think that by me delivering my logged data to the world, this is the best approach for the truth to be confirmed.

Therefore feel free to download the Free TS-dmm-Viewer https://www.ittsb.eu/TsDMMViewer-Data-Logger-2020-Japan&Greece.html
Also download my XLS.file from this topic, load it at the software and explore the entire 100 Hours of Log

Bellow they are high resolution pictures, that is best to download at your PC. 
At ITTSB images viewer ( right click ) and select view image,  then save the larger file at your PC.

My own conclusion this is that I can not trust the 8846A at standard deviation measurements and comparisons them logged due PC software when measuring resistance with it.
Ambient temperature changes they do influence the measurement in a negative way,  such weird temperature coeficiance it makes the evaluation of high precision resistance  DUT this looking as impossible. 
 

[Kiriakos GR]

Get the XLS.file from here.