15
$\begingroup$

I was recently doing a problem, and apparently got it wrong.

The question asked: what is the polar form of the sum: $(5+3i)$ + $(-10-3i)$.

From my understanding, the sum would equal: $-5+0i$

Now,

the magnitude would equal: $\sqrt{-5^2 + 0^2} = 5$

And,

the angle would equal: $\tan^{-1}(\frac{0}{-5}) = 0$

Instead, the answer is actually $5 (\pi)$, instead of $5 (0)$.

I check on this website: 2pif.com (under the "Complex Numbers" tool), and it also says that 5 (0) is the answer.

So I'm wondering, how in the world does $\tan^{-1}(\frac{0}{-5}) = \pi$

|cite|improve this question
$\endgroup$
  • 4
    $\begingroup$ theta = arctan 0 mean tan theta = 0 but tan theta = 0 does not mean theta = arctan 0. It means theta = arctan 0 + k*pi. We want to solve i) cos theta = -1 ii) sin theta = 0 and iii) 0 <= theta < 2 pi. The answer to that is not arctan 0. It is pi. and indeed tan pi = 0. $\endgroup$ – fleablood Dec 27 '16 at 0:46
  • 7
    $\begingroup$ Of course, rather than use any formulas, it should be obvious, geometrically, that the polar angle of the number $-5+0i$ is $\pi$, since it corresponds to the point $(-5,0)$ in the Cartesian plane, and since we always measure polar angles from the positive $x$ axis. $\endgroup$ – symplectomorphic Dec 27 '16 at 1:58
  • 2
    $\begingroup$ And by the way, the "$-5^2$" part under the root is wrong -- it must be $(-5)^2$. $\endgroup$ – zipirovich Dec 27 '16 at 2:01
  • 2
    $\begingroup$ $\arctan$ like many other trigonometric inverse function, is multi-valued. You are just selecting a branch. $\endgroup$ – Henricus V. Dec 27 '16 at 6:40
32
$\begingroup$

This is a common error among new people into the complex numbers. The argument of a complex number $z = a + ib$ is not $\arctan (b/a).$ It is given by $$\arg z = \begin{cases} \arctan (b/a) &\mbox{if } a>0 &\\ \arctan (b/a)+\pi &\mbox{if } a<0,b\geq 0 \\ \arctan (b/a)-\pi &\mbox{if } a<0,b< 0 \\ \pi/2 &\mbox{if } a=0,b> 0 \\ -\pi/2 &\mbox{if } a=0,b< 0 \\ \mathrm{undefined} &\mbox{if } a=0,b= 0 \\ \end{cases} $$

|cite|improve this answer
$\endgroup$
  • 23
    $\begingroup$ Also known as atan2 $\endgroup$ – 小太郎 Dec 27 '16 at 3:46
  • 6
    $\begingroup$ By the way, this is basically the same error as deducing that $2 = -2$ from the fact that $\sqrt{2^2} =^? \sqrt{(-2)^2}$. $\endgroup$ – Patrick Stevens Dec 27 '16 at 10:30
  • 3
    $\begingroup$ Unfortunately, it seems to also be a mistake in some textbooks. :( $\endgroup$ – egreg Dec 27 '16 at 10:48
10
$\begingroup$

It is an oversimplification to say the argument of $x+iy$ is given by $\theta =\arctan \dfrac yx$. Actually you have to solve the system \begin{cases}\cos\theta=\dfrac x{\sqrt{x^2+y^2}}\\ \sin\theta=\dfrac y{\sqrt{x^2+y^2}}\end{cases} which implies $\tan\theta=\dfrac yx$, but is not equivalent to $\theta=\arctan\dfrac yx$.

The general solution is $\theta\equiv\arctan\dfrac yx\mod \pi$, and you have to choose the value of $\theta$ (mod $2\pi$) in function of the signs of $\cos\theta$ and $\sin\theta$, i.e. of $x$ and $y$.

|cite|improve this answer
$\endgroup$
5
$\begingroup$

To add to the other answers here, the argument is the angle the corresponding vector on the Complex Plane would make with that of $1+0i$

In this particular case, the angle is $180^{\circ}$ or $\pi$ radians as in the picture.

|cite|improve this answer
$\endgroup$
  • $\begingroup$ +1 for the only answer that points out the underlying geometry (with a picture). This is the one the OP should remember. $\endgroup$ – Ethan Bolker Dec 27 '16 at 19:04
3
$\begingroup$

$\tan(x)=0$ has the solutions $x=n\pi$, so $\pi$ is also a solution. Which solution you choose, depends on the signs of $\sin(x)$ and $\cos(x)$. This is why computers have a special function atan2, which calculates the solutions of $\tan(x)=\frac{u}{v}$ and returns the value in the proper quadrant, by accounting for the signs of $u$ and $v$ (as well as for the values where $\tan(x)=\frac{u}{v}$ might be ambiguous/undefined).

|cite|improve this answer
$\endgroup$
1
$\begingroup$

because the angle of a complex number is not given by the ordinary artan, is necessary make an adjust of Pi if the real part is negative, add pi and ist done

|cite|improve this answer
$\endgroup$
1
$\begingroup$

$\arctan(0)=0,$ not $\pi$, but $\tan(\pi)=\tan(0)=0$ As $x$ ranges from $-\frac \pi 2$ to $\frac \pi 2$, $\tan x$ ranges from $-\infty$ to $+\infty$. As we want $\arctan$ to be a function, we have to choose a range of output values. When you use the $\arctan$ to get the polar angle, there is an uncertainty of $\pi$ because $\tan (x+\pi)=\tan x$. You have to use the real and imaginary parts to decide the multiple of $\pi$ to add to the $\arctan$ output to get the proper polar angle you want. You might even need to add $2\pi$. For example, if you want the polar form of $1-i$, the $\arctan$ will give you $-\frac \pi 4$ but you probably want to report the polar angle as $\frac {7 \pi}4$

|cite|improve this answer
$\endgroup$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.