Brief description of test setup and data sets.

Voltage reference runs were performed with two instruments. HP3457 and HP3456.  In all tests the HP3457 was programmed to read the extra bits register giving one extra digit of precision, or 7.5 digits.  The HP 3457 includes an HP44492 Multiplexer Card allowing programing input source.

Six input sources were used.  

1.  The 5V reference output of a DMM check multimeter accuracy verfification tool
2.  A simple 10V LM399 based reference.  (LM399 Ref 1)
3.  A simple 10V LM399 based reference. (LM399 Ref2 )
4.  An Ebay AD584-M based referenced sourced from vendor Kmoon on Ebay.  Label says 10V output is 10.00276 volts
5.  The USA Cal Club FX reference.
6.  The USA Cal Club PX referenced

Data were collected using a EZGPIB to collect observations as rapidly as possible  which turned out to be approximately every 70 seconds.  In addition to the data from the HP3457 each data line contains a date/time code, temperature deg C, relative humidity and a HP3456 reading of the FX reference.

USA_Cal_Club_Round3_Test 1.csv contains a trial run performed before arrival of the Cal Club gear.

USA Cal_Club_Round3_Run1.csv contains the initial run with all sources connected.  Unfortunately there was a storm which caused power glitches sufficient to reset digital clocks in my home and desktop computers.  The run was controlled by a laptop and appeared at a casual glance to be performing corrrectly, but later examination showed that all HP3457 measurements were incorrect, and there were a few glitches on the temperature and humidity readings.  The data also confirms that the heat pump thermal control for my lab area turned off as is shown by the gradual temperature drop.   Only the first couple of hours of this run is completely valid for all measurements.

USA_Cal_Club_Round3_Run2 is  a fully valid test series.

The equipment was then rearranged to measure the three locally owned references differentially with the FX reference.  Data lines contain time, temp, humidity, LM399-1, LM399-2 and KMoon

USA_Cal_Club_Round3_Differential.csv  showed anomalous behavior on LM399-1 so a second run was performed.  Again funny results, but I decided I could debug this reference against the other two local refernces and moved on.

Resistance checks were performed with HP3457, programmed to used extra digit register and average 10 readings to report each data line.  Time and resistance are reported.  Temp and humidity were not recorded based on lab stability.

Titles for the measurements indicate the content.  The RM box was operated to get at least 8 valid readings per resistance.  In the data set there is a transition reading where switching occurred during the 10 reading average which should be discarded.