What's the deal with standard resistor and cap values?

[Taylor]

This is sort like what Seinfeld would ponder if he did electronics, but why are the standard values what they are?

Instead of having more sensible numbers, like maybe 1k, 2k, 3k, 4k... we have 1.2k, 2.2, 3.3, 3.9, 4.7... Caps seem to follow the same numbering.

I know it's possible to get rounder values, but they're not standardized, so they're more expensive. Why can't we have a 5-anything? Why must it always be 4.7? I'm assuming some governing body decided this a long time ago, but why?

[doitle]

http://www.technick.net/public/code/cp_dpage.php?aiocp_dp=guide_standard_resistors_table

That has a good explanation of how they come up with the numbers at least. The reason they do it that way has to do with the fact that let's say you have a 20% resistor. You have a 1.5K and a 2K as opposed to the 2.2K used normally. Let's say the 2K is 20% out of tolerance to the smaller side and the 1.5K is 20% out of tolerance to the high side. 1.8K and 1.6K. Your 1.5K resistor is now higher than your 2.2K resistor. I believe they are placed to avoid such a situation. As long as a 1.5K Resistor is within tolerance it will always be smaller than a within tolerance 2.2K resistor. This would not be the case with 2K though. As the tolerance gets tighter you get more values and they are closer together because as long as the resistors are within tolerance there will still be no overlap.

Hopefully that shot in the dark will tide you over until someone comes in and corrects me.

[R O Tiree]

If you copy the standard E12 values into a spreadsheet (10, 12, 15, 18, 22, 27, 33, 39, 47, 56, 68, 82, 100, so that's 12 values per decade) and divide each one by the one below it, you'll find the multiplication factor comes out at roughly 1.2.

Do the same for the E24 series and the multiplication factor is pretty close to 1.1

Do the same for the E48 series and it's 1.05 within 2.5%

E96 is very close indeed to 1.025 and E192 is 1.0125.

OK, what's the significance of this? Well, tolerance bands mentioned by doitle is one reason. The other is surely down to the fact that we are generally dealing with frequencies, be it audio or radar (extremes). Every octave doubles the frequency, so we have an exponential relationship. Since the resistor values in each series also have an exponentional relationship, then going up or down an arbitrary number of steps in the series will have a predictable and "linear" sounding response in frequency.

[Taylor]

Hmm, pretty interesting. I'm glad there are learned elders out there who figure this stuff out before manufacturing resistors. Wouldn't have occurred to me.

[Paul Marossy]

Huh, I have also wondered about that from time to time, but not really that much. Very interesting to know why they have those odd values, though!