2

Referring to my previous query Sine function approximation circuit. How does this work?. Can anybody suggest method to transform the circuit to work on all 4 quadrants?

gari
  • 103
  • 4
  • The circuit you found is not a sine function approximation circuit. So you could not expand it to all four quadrants. – Uwe May 30 '22 at 13:22
  • 2
    What are you trying to do? Why do you need calculate a sine function as analog circuit? This can be done digitally, and depending on your needs it might be a better solution for you. – Eugene Sh. May 30 '22 at 14:32
  • Yep, what Eugene says: you'd need some kind of detector in which quadrant you are. Building that detector is already a quantization (namely, four quadrants: two bits.) At that point, this analog circuit just adds a bunch of new, hard problems (e.g. balancedness, smoothness of derivative at the borders) that you didn't need solve in the first place, to do something easy (a sine). From the top of my head, extending this to four quadrants is thus very hard, whilst building a sine generator differently is easier. – Marcus Müller May 30 '22 at 14:45
  • so, it's very hard to advise how to modify that circuit without understanding why precisely you want to use this (pretty esoteric!) method of generating a sine. What's the reason? What do you do with the sine? Without, it's really hard to help you properly. – Marcus Müller May 30 '22 at 14:48
  • want to avoid look up tables ROM in the circuit – gari May 30 '22 at 14:51
  • Thank you, @gari! Aha! but now we only know what you want to avoid, but that doesn't tell us what you want to achieve! – Marcus Müller May 30 '22 at 14:52
  • want to achieve all analog solution. It is purely academic – gari May 30 '22 at 16:22
  • 1
    Sorry, if this is purely academic, then you still need to describe requirements. We really can't guess what a suitable approach would be from "here's a circuit whose operation I don't even explain, but I want to extend it to four quadrants": This, plainly spoken, is a pretty terrible linear-digital to sine-domain analog converter, and extending it to four quadrants would be done on the digital side, quite clearly, but you say you don't want to do that, so there's no solution. – Marcus Müller May 30 '22 at 16:26
  • you see, $y=\sin(x) $ can be pretty well approximated by $y=0$. That does, admittedly, not look like a sine at all, but it doesn't have that much more error than the circuit you show over there, works in all quadrants, is arbitrarily fast and very low in power consumption. You probably will still not want that solution! But that's because you do have system requirements in your head that you don't tell us. But we need these to actually help you. – Marcus Müller May 30 '22 at 16:30

0 Answers0