I personally wouldn't, as you taking a major shortcut in going straight to the answer and showing very little working or skill in doing the integration; I think you're best off doing it the normal way with A/(u+1) + B/(u-1) = 1/(u2-1)_ShiFTy_ said:hehe forgot that u^2 - 1 can be factorised
Will you also be able to use the formula
1/2a. (ln |(a+x)/(a-x)|)
When integrating 1/(a^2 - x^2)
If you get it wrong, you won't get any marks, as you haven't given the marker any oppurtunity to award you some marks "for showing working". Also, if you come across a hard integral, your shortcuts usually won't work or will be too hard to apply to the integral.bboyelement said:our teacher taught us to do the shortcuts ... is that bad and also in the first exercise of cambridge (integration) it also shows you the shortcuts ... i still comtemplating whether to use the substitution or not ...
icycloud said:Letting e^x = sec gives an interesting (but perhaps longer) way of solving this integral without partial fractions .
I=∫e^x/(e^(2x)-1)dx
Let e^x = sec
e^x dx = sectan du
dx = tan du
I=∫sectan / tan^2 du
= ∫sec/tan du
= ∫cosecdu
= ∫(cosec^2+coseccot)/(cosec+cot)du
= -ln[cosec+cot] + c
= ln[sqrt(e^(2x)-1) / (e^x+1)] +c
(=ln[sqrt[(e^x-1)/(e^x+1)]] + c <-- which is the answer you get using partial fractions)
Of course._shifty said:Does that actually work??
Because i tried letting e^x = sinu, e^x dx = cosudu
I got up to:
I= -∫secu du
= -ln (secu + tanu) + c
Then i got stuck...
thx manicycloud said:Of course.
- ln (secu+tanu) + c = -ln(1/cosu + sinu/cosu)+ c
= -ln( [sinu+1] / cosu)+ c
= -ln( [e^x+1] / sqrt(1-sin2u))+ c
= -ln([e^x+1]/sqrt[1-e^(2x)]) + c
= ln(sqrt[1-e^(2x)]/[e^x+1]) + c
Oops, sorry the denominator was supposed to say 9 + e^2xRiviet said:I'm not sure how you got the 1/3's in the answer, but I got tan-1ex + C, by using a substition of u=ex.