Here's the result of a one year measurement of typical drifts one might expect in a basic home setup. Before you guys get upset, I fully
understand that this is NOT metrology-grade.
I took monthly readings of a Malone VREF10-003 (30 ppm) 10V reference against an HP 3456 6-1/2 digit and Advantest R6871E 7-1/2 digit
DMM. The equipment was left off except for a two-hour warmup once a month when I made the measurement. The measurements were
done by averaging the results of 5 pairs of readings, each pair yielded (x - y)/2 where x = leads in normal position, and y = leads reversed. The
same test leads were used throughout.
The graph shows the ppm error off of the nominal 10V along with the approximate temperature in the lab. The temperature was offset by
60 degrees so it would plot on the same axes (i.e, a plotted value of 4 means 60+4 = 64 deg F).
Of course, since the Malone reference could be drifting in addition to the two DMMs, it's hard to reach a definitive conclusion. But, feel free to
weigh in with your conclusions (other than I need to get a calibrated 1 ppm-class reference)!