Resistor color code

[Nigel Goodwin]

A LONG time ago (pre-Windows), I wrote a number of programs using Turbo Pascal - one was for working out parallel resistors, you told it just the value you wanted, and it displayed a list of suitable preferred values (in order), and the percentage of error for each combination.

Unfortunately all the programs were on a long dead 386 laptop!

But in any case, it was pretty easy to write.

 

[Dean Huster]

Ron, I had a voltage divider program written in something and then compiled, so have no way to play with it. But it seemed pretty slick right at first. Just give it the input voltage and desired output voltage and it went through and spit out the two standard resistor values that provided an output closest to your desired output. You could even choose whether you wanted to sort through the 5% standard values or the 1% values. I thought it was great. I figured, "Hey, if it's sorting through the entire set of possibilities at blinding computer speed for the best two values, that has to be a lot closer than me selecting my own bleeder resistor and calculating a multiplier. It'd be going through all of the bleeder/multiplier values for the best ratio. Then one day, I designed my own divider and then used the program to make one better. The two resistors it selected yielded an output voltage that had a larger error than my design (and I wasn't adding little values in series to get the "perfect" multiplier value, either). I started trying different input voltages and output voltages and sure enough, I found all sorts of anomalies where the program provided a divider ratio that was worse than what I'd come up with. I quit using that one.

What in the heck is so hard about a reiterative program that begins at the bottom of the resistor chart for the multiplier and progressively selects one of 144 other resistors, calculating the output voltage each time, comparing with perfection and saving the values the provide the lowest error, then advances to the next value up for the bleeder and tries again. In all cases, you can eliminate starting at the values you know will be wrong in order to quicken the process. A program like that HAS to provide the best possible ratio of standard values.

Dean