Projecting Calibration interval for Resistors

[lhiga]

Does anyone out there have a process for projecting a calibration interval for resistors. I just received some resistors back from a upper echelon cal lab and the process they used to determine a 36 mos calibration interval has just got me SNOWED.
From the history data they took the difference from the NOMINAL resistance, did a least squares fit of the differences and projected a interval from this line. Does this sound right to you folks??

 

[Marc]

Anyone have any ideas for this?

 

[Ryan Wilde]

I just got off a much needed vacation, and I don't know how long it will be before I get back into the lab. If anyone has access to NCSL-I RP-1, give it a quick browse to see if the least squares method is listed (if so, it would be an A-x method), otherwise when I get into the lab (I am perma-onsite) I will check. I always use the A-3 binomial method.

But thinking about it, it should work. Standard resistors are very stable, and tend to have a fairly linear regression, unless you cook them with a bit o' current. The regression line should show when the resistor will go out of tolerance, although I personally wouldn't set my interval to any more than half of that time period.

Ryan