Archive: 8 posts

---

Is there an easy way to take the reciprocal of any given number? I can't think of a quick way off the top of my head without tons of manual bits and pieces that will result in it only being applicable to one number.

2014-08-26 22:40:00 Author: Tynz21 Posts: 544

---

If you take a reciprocal of a number that's less than one, i.e. an analog percentage, you will end up with a value greater than one. If you want the reciprocal of a value so that you can use it to do division (using multiplication) then I think it would be easiest to do the division by using a negative feedback loop and not bother trying to get the reciprocal.

2014-08-27 09:01:00 Author: fluxlasers Posts: 182

---

If you take a reciprocal of a number that's less than one, i.e. an analog percentage, you will end up with a value greater than one. If you want the reciprocal of a value so that you can use it to do division (using multiplication) then I think it would be easiest to do the division by using a negative feedback loop and not bother trying to get the reciprocal. How would using a negative feedback loop help? The whole poi t it to be able to take a number greater than 1 and use it to divide (I.e. multiple by the reciprocal) without having to preprogram any values.

2014-08-27 14:28:00 Author: Tynz21 Posts: 544

---

I think he means if you want to get some output you can't easily calculate directly, but you can verify if a given output matches the input, and you know which inputs are inversely related to the output, you can take any output, calculate if it matches the input, and feed the difference back into the output. In this case you can verify division by multiplying the output by the denominator to get numerator, and only the denominator is inversely related to the output. So compare the output of a multiplying AND of your output and your denominator input to the real numerator, and feed that difference back into your output, and your output should automatically steer towards the correct output. You may need to dampen the feedback somewhat by multiplying with a fractional constant (battery) to prevent it from wildly overshooting the target.

2014-08-27 15:30:00 Author: Rogar Posts: 2284

---

I think he means if you want to get some output you can't easily calculate directly, but you can verify if a given output matches the input, and you know which inputs are inversely related to the output, you can take any output, calculate if it matches the input, and feed the difference back into the output. In this case you can verify division by multiplying the output by the denominator to get numerator, and only the denominator is inversely related to the output. So compare the output of a multiplying AND of your output and your denominator input to the real numerator, and feed that difference back into your output, and your output should automatically steer towards the correct output. You may need to dampen the feedback somewhat by multiplying with a fractional constant (battery) to prevent it from wildly overshooting the target. I get what you're saying, but I would still have to preprogram in all the inverse values that it checks for so this isn't really what I'm looking for. I'm not sure if what I want even exists to be honest.

2014-08-27 17:14:00 Author: Tynz21 Posts: 544

---

No, you don't preprogram any values (except the feedback factor), you let the feedback loop find the correct answer. I've only just started experimenting with this kind of feedback, but I think the principle is sound. By the way, just to make sure, we're talking about the division function here, not the reciprocal function. You just want the reciprocal to make division, right?

2014-08-27 17:26:00 Author: Rogar Posts: 2284

---

No, you don't preprogram any values (except the feedback factor), you let the feedback loop find the correct answer. I've only just started experimenting with this kind of feedback, but I think the principle is sound. By the way, just to make sure, we're talking about the division function here, not the reciprocal function. You just want the reciprocal to make division, right? Id prefer both but can make do with either or. I'll have to take a look when I get home.

2014-08-27 17:55:00 Author: Tynz21 Posts: 544

---

Here is how I would do it First equalize scale with exponents then just do division like this (best in binary) 3/5 = 3 >= 5? F = 3 * 2 6 >= 5? T = (6 - 5) * 2 2 >= 5? F = 2 * 2 4 >= 5? F = 4 * 2 Repeat until accurate enough (so far this was 0.1001 in binary) then just convert to anaIogue

2014-08-28 00:21:00 Author: amiel445566 Posts: 664

Posts: 1077139    Threads: 69970    Members: 9661    Archive-Date: 2019-01-19

Aus dem Archiv wurden alle persönlichen Daten wie Name, Anschrift, Email etc. - aber auch sämtliche Inhalte wie z.B. persönliche Nachrichten - entfernt.
Die Nutzung dieser Webseite erfolgt ohne Speicherung personenbezogener Daten. Es werden keinerlei Cookies, Logs, 3rd-Party-Plugins etc. verwendet.