How were standard resistor values chosen?

[alank2]

So one of the things I got from my local radio shack clearance was a 500 resistor pack with various standard ones.  220, 330, 470, 560, 1.5k, 2.2k, 3.3k, 4.7k, etc.  Curiously how were these values selected originally instead of say even numbers?

[KJDS]

It's a matter of ratios

so E6 is the simplest set still in use, which means that parts with a 20% tolerance just about overlap
100, 150, 220, 330, 470, 680
Then E12, for 10% tolerance parts
100, 120, 150, 180, 220, 270, ....

Then E24 for 5% parts
100, 110, 120, 130, 150, 160, 180, ....

Then E48 for 2% parts and so on up to E192
100, 101, 102, 104, 105, 106, 107, 109, 110, ...

[eejake52]

Here is a good explanation of standard resistor values:
http://www.logwell.com/tech/components/resistor_values.html

[alank2]

Thanks guys - that makes perfect sense!

[cosmicray]

Out of curiosity I pulled down a 1946 copy of Reference Data for Radio Engineers to see what it listed for resistor values.

5% - 51 ohm to 10 meg
10% - 56 ohm to 10 meg
20% - 68 ohm to 10 meg

There is also a column for "old standard resistance values"
50, 75, 100, 150, 200, 250, 300, 350, 400, etc up to 10 meg
If the old standard values are from the 1938 standard, there appear to be color band values for tolerances from 1 to 9, and 5, 10 and 20 %. The actual resistor value band colors appear to be standardized at least from 1938, possibly earlier.

[Rene]

You may find this link useful:

http://www.resistorguide.com/resistor-values/

[SArepairman]

i heard it had to do how sailors in the british navy organized rope diameters. (tollerance)

[T3sl4co1l]

The aggravating part is, within rounding error... by choosing the most regular increments (equally spaced on a log plot), they necessarily chose values that are the least likely to satisfy certain worst-case ratios.  That is, to realize some given ratio in a voltage divider or whatever, for any combination of any two values picked from the list, you will have a worst case error equal to half the increment between values.

Whereas if they hadn't played with them, and stuck with the "rounder", more organic, but mathematically more oddball values that were in general use before the standard, there would be a higher likelihood of being able to pick a lucky pair of resistors that get very close to the desired ratio.  The base multiplier might not always work out (you might have to do it with 1k or 10k resistors instead of the 3k that would fit your circuit more appropriately), but it's almost always easier to scale resistances than ratios, so this is often preferred.

Indeed, what it does achieve, is relative independence from resistance value.  A 1k and 2.2k resistor have almost the same ratio as 3.3k and 6.8k, or 4.7k and 10k.  You might not be able to pick a good ratio, but you can pick that suboptimal ratio from practically any scale of resistance you wish.

Maybe resistance really was more important back then.  E-series codes date from the days of tubes, where pots were cheap relative to the overall construction, especially with regards to active devices (a tube might cost $1, not like the micro-cents a transistor-on-die costs today).  And everything drifted with temperature and age, so things like TV sets were in constant need of adjustment.

I suppose it all washes out today, as 1% (and better) resistors have never been cheaper.  Sucks if you want to keep a thousand different parts in stock, but...

Tim

[gilbenl]

http://en.m.wikipedia.org/wiki/Preferred_number