ZoidTechnology
commented
1 month ago Thanks for the resources! The pass/fail stuff looks particularly useful. I'll have to give them a proper read. I've got a couple of questions:
- In what ways would an unknown reference resistor distribution impact this? I assume resistors would follow a normal distribution after manufacturing, but binning might affect this? Perhaps leading to a bimodal or rectangular distribution?
- What would a results UI look like using this method? Is there a probability commonly used to determine a result, or should it be user-selectable?
mendip-defender
First off, great work. Don't see my comments as being mean.
You might want to look up Uncertainty Calculations as that would be a more accurate way to go about it. This way you can root sum square the spec/res. See M3003, https://www.ukas.com/wp-content/uploads/2023/05/M3003-The-expression-of-uncertainty-and-confidence-in-measurement.pdf
Resolution can be looked at as half of the digit. So if it's a 0.1 resolution it is common to divide that by two and it gives you 0.05.
Also, the Pass / Fail stuff is covered under, LAB 48 Decision Rules https://www.ukas.com/wp-content/uploads/schedule_uploads/759162/LAB-48-Decision-Rules-and-Statements-of-Conformity.pdf
If you like I can supply you with an example of what the maths looks like for calculating uncertainty.