1
$\begingroup$

Could anyone kindly help me evaluate the period of the following function:

$$f(x)=\sin\left(2\left(\tan^{-1}\left(\frac{\tan(x)}k\right)\right)\right)$$

where $k$ is some positive real constant.

Had it been only $\sin2(\tan^{-1}(\tan(x)))$, I could have found its period but this constant $k$ is proving to be a challenge for me.

I guess, for certain values of $k$, f may not be periodic at all, can we find under what condition this function will be periodic?

Thanks for your time.

PS: I am in electrical engineering and I used to find the period of function ten years back, but now I have really forgotten the concepts and formulas and I don't have any book at my disposal either.

$\endgroup$
  • 2
    $\begingroup$ It is of the form $f(x)=g(\tan(x))$. Since $\tan$ is periodic $\pi$, so is $f$. $\endgroup$ – Simply Beautiful Art Aug 22 '17 at 23:50
  • $\begingroup$ @SimplyBeautifulArt $0$ is of the form $g(\tan x)$. It has no (fundamental) period at all, though. $\endgroup$ – Cauchy Aug 23 '17 at 0:29
  • $\begingroup$ @Cauchy Well, its no answer, but its certainly close $\endgroup$ – Simply Beautiful Art Aug 23 '17 at 0:33
  • $\begingroup$ @SimplyBeautifulArt: Thanks for editing the question and for your comment. But, assuming a fundamental period of $\pi$ somehow doesn't support analysis result of my circuit that I am getting from my simulator. And why the period will not be dependent on $k$. What if $k=\pi$, for example? $\endgroup$ – MoM Aug 23 '17 at 0:39
0
$\begingroup$

The period of sin is 2π. Therefore $$\tan^{-1}\left(\frac{\tan(x)}k\right)$$ will need to change in increments of π - {0,π,0,π}; $$\left(\frac{\tan(x)}k\right)$$ will be {tan(0), tan(π)} (both 0). tan(x)/k is only 0 when x is a multiple of π. Thus the period of $$f(x)=\sin\left(2\left(\tan^{-1}\left(\frac{\tan(x)}k\right)\right)\right)$$ is π.

$\endgroup$
0
$\begingroup$

As written the function $$f(x):=\sin\left(2\arctan{\tan x\over k}\right)$$ is undefined at odd multiples of ${\pi\over2}$. But it turns out that these singularities are removable, and that $f$ is real analytic on all of ${\mathbb R}$ when $k>0$. Put $$\arctan{\tan x\over k}=:\alpha\in\left]{-{\pi\over2}},{\pi\over2}\right[\ .$$ Then $${\tan x\over k}=\tan\alpha$$ whenever $\tan x $ is defined, and we 0btain $$f(x)=\sin(2\alpha)={2\tan\alpha\over1+\tan^2\alpha}={2k\tan x\over k^2+\tan^2 x}={k\sin(2x)\over k^2\cos^2 x+\sin^2 x}\ ,$$ whereby now the RHS makes sense for all $x\in{\mathbb R}$. This RHS is of period $\pi$ by inspection, and it is easy to see that no number of the form ${\pi\over n}$, $n\geq2$, can be a period, whatever the value of $k>0$.

$\endgroup$
  • $\begingroup$ Thank you, Sir, now I understand. I will recheck my circuit simulation. $\endgroup$ – MoM Aug 23 '17 at 21:21

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.