I would like to add features to the resistor portion of the PCB calculator.
The reason for this is my primary need from a resistor calculator is simply to tell me the correct/nearest e-series value for a value I have just calculated. I use 1% resistors almost all the time so E96 is what I need. But the calculator doesn't do E96. I prototyped adding this and so I see inside why it doesn't (computational complexity). I have taken this into account for my proposal. However, there is also one existing piece of functionality I would have to change. I will mention that first. That is, the functionality of automatically excluding the entered value from the list of possible results works against you when you are looking for the closest match. For example, if I were to type in '100' (100K) the closest match is 100K and I'd like that as a match. The proposal assumes this functionality has been removed and I will discuss at the end what to do about replacing that functionality. I would propose adding e96 (and possibly e48) to the selections in the calculator. When you enter a value using e96 mode it will calculate the nearest single resistor to that value. It also will calculate the nearest you can get with two resistors. Did you know that with at most two e96 resistors in series you can create any value to 3 significant figures? For example, say you want 164K, that's 105K + 59K. If you want 241 Ohms, that's 182 + 59. Oh, you actually wanted 165K? That's an E96 value by itself, a single 165K resistor. This property isn't an accident, it is a feature of the E96 series. It is a feature (to lesser precision) of the other series too. Given E96 resistors are 1% accurate that really means you never need more than 2 resistors, you are already within 1% of the requested value (often much better). But if it is desired I could calculate 3 resistors or 4 to approximate the value even more closely. However I do not see a lot of value in this. These 3 and 4 resistor values would not be guaranteed to be the best possible as the search would not be exhaustive and not all topologies would be considered. This algorithm is not well suited for the available option of two excluded values. However I think I can make it work. I use values below 10 Ohm and (rarely) above 1M, so I would extend the range it will attempt to match down to 1 Ohm and up to 10M for E96. I could even go further. This is not a problem for the E96 algorithm. And due to how the algorithm works it would not cause the algorithm to use resistors outside the 10 to 1M range for the 2 resistor solutions unless you request a value at the ends of that range. This is the case as the resistors returned for the 2 value answers are always in the same decade as the desired value or the top portion of the one below. So if you ask for 97.6K it's never going to suggest you use a 97.5999K and 0.1 Ohm resistor in series. Now about the issue of automatic exclusion of the entered value. My preferred solution would be to remove this functionality. I just don't think it is a common use. As the 'exclude values' remain the user could just enter the initial value into the exclude box if she wants another approximation. The target value could even be automatically entered into the exclude box when you enter a value but I do not recommend as the common case is to find the nearest e-series value and it would make that a two step process. I also propose that optionally an option to exclude higher or lower results could be added. The reason for this would be if you want a resistance which is *at least* the entered value or *at most* the entered value. I do feel this would make the window busy and is not needed in the most common case. However if you did need a value that was *at least* the entered value and a lower one is returned then you are in a situation where you either need to consult E-series tables yourself, do a bunch of math on your own, or start entering iterative values into the target resistance until you get a value that meets your criteria. So there is an argument for it. -- You received this message because you are subscribed to the Google Groups "KiCad Developers" group. To unsubscribe from this group and stop receiving emails from it, send an email to devlist+unsubscr...@kicad.org. To view this discussion on the web visit https://groups.google.com/a/kicad.org/d/msgid/devlist/CAEZguRJ6iLBm8VWX_1ibobGq565gY-xAHkSWA0iVt119Fy_hqA%40mail.gmail.com.
