Those 4 Caddock 1776-10 decade divider resistor arrays I mentioned a few days back turned up. Herein follow the details from our Inbound Inspection Department for the information of anyone else who may be tempted by some.
-snip-
Nominal accuracy +0-0.5%, nominal ratio accuracy ±0.1%. All specimens well within their specifications, particularly good ratio matching.
-snip-
What about the measuring meter's uncertainty? I'm guessing it's small enough that the possible range of values for the resistors remains within the specs, but you know I had to ask...
My spreadsheet has the actual standard deviation of the readings in, but I wasn't going to fuss with posting that. The whole uncertainty, including the meter uncertainty, for the
absolute values comes at about 900 ppm (k=2) or 0.09% for the 10M end of things down to 130 ppm (k=2) or 0.013% at the 10k end ,
ratio values on the other hand just boil down to the 24 hour uncertainties of the meter, which is about 25ppm (k=2) or 0.0025%.
If one was going to be really finicky you'd work out the Johnson noise of the resistances and RMS add that to the existing measurement uncertainties
assuming that HPAK haven't already accounted for that in the range by range uncertainties which, judging from the trends of the figures in the data sheet they already have (0.015%+0.001%) for 10M range, 24 hours, versus (0.002%+0.0005%) for 10k range, 24 hours.