Without a calculator, how can I calculate the sine of an angle, for example 32(without drawing a triangle)?
Asked
Active
Viewed 5.8k times
2 Answers
12
You can use first order approximation $\sin(x+h)=\sin(x)+\sin'(x)h=\sin(x)+\cos(x)h$
where $x$ is the point nearest to $x+h$ at which you already know the value of the $\sin$ function and its derivative $\cos$ function too.
Like for $\sin(32^0)=\sin(30^0)+\cos(30^0)*(\frac{\pi}{90})$
Here you need to take $h$ in radians which is $\frac{\pi}{90}$ for $(32^0-30^0)=2^0$
Aang
- 14,920
8
Using Taylor expansion you get an approximation up to the desired precision
Edoardo Lanari
- 3,674
-
-
1I don't know why OP accepts this answer, but apparently MacLaurin series approximation doesn't work well when the angle is large, say $438748327^{\circ}$. Avatar's answer is much better than this one in terms of MSE standards. – Shuhao Cao Jun 19 '13 at 15:55
-
What about $2\pi$-periodicity? The expansion is for radiant value. – Edoardo Lanari Jun 19 '13 at 15:57
-
@Lano My point is, expansion using terms $x - x^3/3! + \cdots$ doesn't work well unless $x\ll 1$. Otherwise you have to wait $(n+1)!$ to majorize $x^{n+1}$ in the error term, which normally results computation of many terms. If you use known sine values like $\sin(\pi/24)$, you could use first order approximation(two terms in Taylor) like in Avatar's answer. – Shuhao Cao Jun 19 '13 at 16:01
-
1@ShuhaoCao I think Taylor's series is actually better answer. We could always mod the angle to 360 so it should not be the problem – Thaina Sep 24 '20 at 04:32
Cdoes it: http://sourceware.org/git/?p=glibc.git;a=blob;f=sysdeps/ieee754/dbl-64/s_sin.c;hb=HEAD involving some iterative process. – Shuhao Cao Jun 19 '13 at 15:49