# Confusion about accuracy vs tolerance of an operating resistor

Let's say I bought a 100 Ohm resistor from a manufacturer with a tolerance of 10%. What I understand from the term tolerance here is that: it is the amount that a resistor may initially vary from its "stated nominal resistance". So in our case when I have this resistor in my hand I can say this can be between 90 to 110 Ohm.

But imagine I have a good ohmmeter and I measured this resistor as 98 Ohm.

If I install this resistor to a circuit, what can I say about the variation or accuracy of this resistor by time? Can it still vary between 90 Ohm and 110 Ohm during its operation or it will be around 98 Ohm always? Im a bit confused between tolerance and uncertainty accuracy/uncertainty here.

• 'Accuracy' doesn't have a quantitative definition. – Chu May 5 '18 at 0:22

Like awjlogan said, the nominal value for that part is fixed after you measured it.

Over time you will get drifts based on that nominal value. Humidity, power, temperature, time, overload and other things will affect the value in the later life of the resistor.

As an example, here is a table out of a Vishay datasheet:

As you can see there are several mechanisms which can have a serious impact on the value. The typical value shows, that the most critical things for these resistors are moisture and time (max. 3% and 1% after 10000h).

Depending on the make of the resistor different mechanisms will have different effects, so check the datasheet of the resistor you are using.

No, it will be 98 $\Omega$ always (not accounting for aging or temperature etc). If you're building lots of devices, you can't measure each individual resistor. Therefore, you must design your circuit to have the desired performance within the tolerance quoted by the resistor's manufacturer. Most components have a Gaussian distribution around their nominal values.

What you have done is usually called hand matching, which is absolutely fine if you are making small quantities of high value devices but is not scalable for mass manufacture. For example, if you were building a one-off precision piece of scientific equipment, you might wish to match the components in a filter to a better tolerance than you would expect by simply grabbing them from the packet.