This is the corrected exercise set for the trigonometric integrals of MATH 141, Calculus 2, the integral calculus course taken at McGill University in first year.It follows section 7.2 of Stewart: powers of sinx and cosx, powers of tanx and secx, the integrals of secx and sec3x, and products sinmxcosnx.
What the set covers, the traps it targets
It follows section 7.2 of Stewart: powers of sinx and cosx, powers of tanx and secx, the integrals of secx and sec3x, and products sinmxcosnx.
Every answer is exact, as on the calculator-free midterm and final, and every antiderivative in the solutions has been checked by differentiating it back.
The thread running through the whole set: a trigonometric integral is a substitution in disguise. Save ONE factor to be the differential (cosxdx, sinxdx, sec2xdx or secxtanxdx), take it from the ODD power, convert everything else with sin2x+cos2x=1 or tan2x=sec2x−1, and integrate a polynomial. When every power is even, no factor can be saved, and the tool changes: the half-angle formulas LOWER THE DEGREE. When two frequencies are multiplied, the product becomes a sum first.
The traps named explicitly in the solutions: saving the factor from the even power and creating a square root, the wrong sign in cos2x=21+cos2x, ∫sin2xdx written as 3sin3x, the lower bound sec0=1 taken as 0, the lost 21 in ∫sec3xdx, the sign of sin(A−B) when A<B, 1−sin2x replaced by cosx where the cosine is negative, and the csc formula copied from the sec formula without its minus sign.
•On [−π,π], for positive integers: ∫sinmxsinnxdx and ∫cosmxcosnxdx are 0 if m=n and π if m=n; ∫sinmxcosnxdx=0 always.
Part A: the basics (/50)
Exercise 1: An odd power of sine or cosine: save one factor, convert the rest
5 answers to check
When ∫sinmxcosnxdx has an ODD power, the whole integral is a substitution in disguise. Detach ONE factor of the odd power to be the differential (cosxdx=du with u=sinx, or sinxdx=−du with u=cosx), convert every remaining factor of that function with sin2x+cos2x=1, and integrate a polynomial in u.
The figure shows the region whose area part d) computes.
a) Compute ∫sin3xcosxdx. Say which factor you detach and why the square root causes no trouble.
b) Compute ∫cos5xsin4xdx.
c) Compute ∫sin3xcos3xdx twice, once with u=sinx and once with u=cosx. Show that the two answers differ by a constant, and find it.
d) Evaluate ∫0π/2sin2xcos3xdx by changing the bounds with the substitution.
e) A student attacks the integral of d) with u=cosx. Write the first line of his computation and explain exactly where it breaks.
Show the solution
Answers
a)−32cos3/2x+72cos7/2x+C
b)5sin5x−72sin7x+9sin9x+C
c)4sin4x−6sin6x+C and −4cos4x+6cos6x+C; they differ by 121.
d)152
e)Detaching sinxdx from sin2x leaves one sinx=±1−u2: a square root. The factor must come from the ODD power.
a)
The power of sinx is 3, odd, so detach one sinx for the differential: u=cosx, du=−sinxdx. The two factors left, sin2x=1−cos2x=1−u2, convert cleanly. Then ∫sin3xcosxdx=∫sin2xcosxsinxdx=−∫(1−u2)u1/2du=−∫(u1/2−u5/2)du=−32u3/2+72u7/2+C=−32cos3/2x+72cos7/2x+C. The square root is harmless because it sits on the function we SUBSTITUTE: cosx becomes u1/2, a power like any other. The method never needed the other power to be an integer, only the saved power to be odd. Check by differentiating: sinxcosx−sinxcos5/2x=sinxcosx(1−cos2x)=sin3xcosx.
b)
Here the ODD power is on the cosine, 5, so the roles swap: detach cosxdx=du with u=sinx, and write cos4x=(cos2x)2=(1−u2)2. Then ∫cos5xsin4xdx=∫(1−u2)2u4du=∫(u4−2u6+u8)du=5u5−72u7+9u9+C=5sin5x−72sin7x+9sin9x+C. The even power 4 of the sine is simply carried along as u4. Trying to save a factor of sinx instead would leave sin3x, an odd power of the function you are not substituting, and the method stalls.
c)
Both powers are odd, so both routes work. With u=sinx (save cosxdx, convert cos2x=1−u2): ∫u3(1−u2)du=4sin4x−6sin6x+C1. With u=cosx (save sinxdx=−du, convert sin2x=1−u2): −∫(1−u2)u3du=−4cos4x+6cos6x+C2. They LOOK different, and a marker accepts either. To compare, put s=sin2x, so cos2x=1−s: the difference is 4s2−6s3+4(1−s)2−6(1−s)3. Expanding, every term in s cancels and what remains is 41−61=121. Two antiderivatives of the same function on an interval differ by a constant, here 121: neither student is wrong. When the two powers are odd, choose the SMALLER one to save, the polynomial is shorter.
d)
The power of cosx is 3, odd: save cosxdx=du with u=sinx, and convert cos2x=1−u2. Bounds: x=0 gives u=0 and x=2π gives u=1. So ∫0π/2sin2xcos3xdx=∫01u2(1−u2)du=31−51=152. Once the bounds are changed, x never comes back. Sanity check against the figure: the curve stays below 0.19 on an interval of length 2π≈1.57, so the area is below 0.3, and 152≈0.13 fits.
e)
He writes ∫sin2xcos3xdx=∫sinxcos3x(sinxdx) and sets u=cosx, sinxdx=−du. What is left to convert is ONE factor sinx, and sinx=±1−u2, with a sign that depends on the interval: −∫u31−u2du is a harder integral than the one he started with. That is the whole rule of the chapter: an EVEN number of factors converts through sin2x=1−cos2x, an odd number leaves a square root. So the factor you detach must come from the odd power, and the even power is the one you convert.
Points are earned before the solution; after it, you can keep practising.
For you:
Tick the exercises you have done or want to review: a free account, no password, keeps your ticks and your right answers from one visit to the next, tells you which chapter to tackle next and lets you ask for the exercise you are missing. Create your space, an email is enough.
Exercise 2: Only even powers: lower the degree with the half-angle formulas
6 answers to check
When every power of sinx and cosx is EVEN, no factor can be saved: detaching cosxdx from cos2x leaves an odd power behind. The tool changes: sin2x=21−cos2x and cos2x=21+cos2x, which halve the power and double the frequency. Repeat until every term integrates at sight.
The figure shows y=sin2x on [0,π] and the dashed line y=21.
a) Predict ∫0πsin2xdx from the figure, then compute it.
b) Compute ∫cos4xdx. How many times do you use a half-angle formula?
c) Compute ∫sin4xcos2xdx by first writing sin4xcos2x=sin2x(sinxcosx)2.
d) Evaluate ∫0π/4cos2xdx exactly.
e) A student writes ∫sin2xdx=3sin3x+C. Refute it in two independent ways.
Show the solution
Answers
a)2π
b)83x+4sin2x+32sin4x+C; twice.
c)16x−64sin4x−48sin32x+C
d)8π+41
e)Its derivative is sin2xcosx, not sin2x; and it gives 0 on [0,π] for a positive integrand.
a)
The curve oscillates between 0 and 1 around the dashed line y=21, and the part above the line is the mirror image of the gap below it: the region has the same area as the rectangle of height 21 on [0,π], that is 2π. Computation: ∫0πsin2xdx=∫0π21−cos2xdx=[2x−4sin2x]0π=2π−0=2π. The figure IS the half-angle formula: sin2x=21−2cos2x is a constant 21 plus a wave that integrates to zero over whole periods.
b)
cos4x=(cos2x)2=(21+cos2x)2=41(1+2cos2x+cos22x). The square cos22x is again an even power, so use the formula a SECOND time, with angle 2x: cos22x=21+cos4x. Then cos4x=41+2cos2x+81+cos4x=83+2cos2x+8cos4x, and ∫cos4xdx=83x+4sin2x+32sin4x+C. Two uses of the formula for a fourth power, and each one doubles the frequency. The usual slip is forgetting the inner chain rule factor: ∫cos4xdx=4sin4x, not sin4x.
c)
Pairing is the shortcut: sinxcosx=2sin2x, so (sinxcosx)2=4sin22x, and sin2x=21−cos2x. The integrand becomes 21−cos2x⋅4sin22x=81(sin22x−sin22xcos2x). The first piece is an even power again: ∫sin22xdx=∫21−cos4xdx=2x−8sin4x. The second piece has an ODD power of cos2x, so it is a substitution: w=sin2x, dw=2cos2xdx, and ∫sin22xcos2xdx=6sin32x. Altogether ∫sin4xcos2xdx=81(2x−8sin4x−6sin32x)+C=16x−64sin4x−48sin32x+C. Note that halving the degree produced a term with an odd power: the two gestures of the chapter often alternate inside one integral.
d)
∫0π/4cos2xdx=∫0π/421+cos2xdx=[2x+4sin2x]0π/4=8π+4sin(π/2)=8π+41. Check the sign in the formula before using it: at x=0, cos20=1 and 21+cos0=1, while the wrong version 21−cos2x gives 0. Order of magnitude: cos2x runs from 1 down to 21 on an interval of length 4π≈0.79, so the answer lies between 0.39 and 0.79; 8π+41≈0.64 fits.
e)
First, differentiate: dxd3sin3x=sin2xcosx, which is not sin2x. The power rule ∫u2du=3u3 needs du=cosxdx in the integrand, and there is no cosx to play that role. Second, use part a): [3sin3x]0π=0, whereas sin2x≥0 and is not identically zero, so its integral on [0,π] must be positive, and it is 2π. An antiderivative guessed by analogy is always checked by one differentiation before it is used.
Points are earned before the solution; after it, you can keep practising.
For you:
Exercise 3: Tangent and secant: save sec squared, or save sec times tan
5 answers to check
The same idea works for ∫tanmxsecnxdx, with two possible differentials: sec2xdx=du for u=tanx, and secxtanxdx=du for u=secx. The identity that converts the rest is tan2x=sec2x−1.
If the power of secx is EVEN, save sec2x and convert the other secants into tangents. If the power of tanx is ODD, save secxtanx and convert the other tangents into secants.
a) Compute ∫tan6xsec4xdx.
b) Compute ∫tan5xsec3xdx. Why is the route of a) closed here?
c) Compute ∫tan3xdx, where there is no secant at all.
d) Evaluate ∫0π/3tanxsec3xdx.
e) Explain why neither rule applies to ∫tan2xsecxdx, and rewrite it as a combination of powers of secx alone.
Show the solution
Answers
a)7tan7x+9tan9x+C
b)7sec7x−52sec5x+3sec3x+C; the power of secx is odd.
c)2tan2x−ln∣secx∣+C
d)37
e)Even power of tanx and odd power of secx: ∫tan2xsecxdx=∫sec3xdx−∫secxdx.
a)
The power of secx is 4, even: save sec2xdx=du with u=tanx, and convert the remaining sec2x=1+tan2x=1+u2. Then ∫tan6xsec4xdx=∫tan6xsec2x(sec2xdx)=∫u6(1+u2)du=7u7+9u9+C=7tan7x+9tan9x+C. The power of the tangent, even here, did not matter at all: it is carried along as u6.
b)
Now the power of secx is 3, odd, so saving sec2x would leave one secx=1+u2: the route of a) is closed. But the power of tanx is 5, odd: save secxtanxdx=du with u=secx, and convert the four remaining tangents, tan4x=(sec2x−1)2=(u2−1)2. What is left of the secants is sec2x=u2. So ∫tan5xsec3xdx=∫tan4xsec2x(secxtanxdx)=∫(u2−1)2u2du=∫(u6−2u4+u2)du=7sec7x−52sec5x+3sec3x+C. Count the factors before writing anything: one secx and one tanx are spent on du, so the powers left are tan4 and sec2, not tan5 and sec3.
c)
With no secant, save nothing yet: split one tan2x and convert it, tan3x=tanx(sec2x−1)=tanxsec2x−tanx. The first piece is the pattern of a) with u=tanx: ∫tanxsec2xdx=2tan2x. The second is the known ∫tanxdx=ln∣secx∣. So ∫tan3xdx=2tan2x−ln∣secx∣+C. A classmate who used u=secx on the first piece gets 2sec2x−ln∣secx∣+C: the same family, since sec2x=1+tan2x shifts the constant by 21.
d)
The power of tanx is 1, odd: save secxtanxdx=du with u=secx, and the rest is sec2x=u2. Bounds: sec0=1 and sec3π=cos(π/3)1=2. So ∫0π/3tanxsec3xdx=∫12u2du=38−1=37. The new lower bound is 1, not 0: sec0=1 is the most common slip on this kind of question, and it would give 38.
e)
The power of secx is 1, odd, so saving sec2x is impossible without a square root; the power of tanx is 2, even, so saving secxtanx leaves one tanx=u2−1. Neither substitution closes. What remains is the identity used as a pure REWRITING: tan2xsecx=(sec2x−1)secx=sec3x−secx, so ∫tan2xsecxdx=∫sec3xdx−∫secxdx. Both integrals on the right are the two antiderivatives nobody guesses, and they are the subject of the next exercise.
Points are earned before the solution; after it, you can keep practising.
For you:
Exercise 4: The secant and its cube: the two antiderivatives nobody guesses
3 answers to check
Two antiderivatives of the chapter are not found by saving a factor: ∫secxdx=ln∣secx+tanx∣+C and ∫sec3xdx, which comes from integration by parts. The first is memorized, the second is rebuilt in four lines.
The figure shows y=sec3x on [0,4π], the dashed horizontal at height 1, and the dashed chord joining the two ends of the arc.
a) Differentiate F(x)=ln∣secx+tanx∣ and conclude.
b) Compute ∫sec3xdx by parts with u=secx and dv=sec2xdx. Explain why the integral you started with reappears, and why that is good news.
c) Finish the integral of Exercise 3 e): compute ∫tan2xsecxdx.
d) Evaluate ∫0π/4sec3xdx exactly.
e) Using the figure and the values π≈3.14, 2≈1.41, ln2≈0.69, check that your answer to d) is plausible, without a calculator.
Show the solution
Answers
a)F′(x)=secx, so ∫secxdx=ln∣secx+tanx∣+C.
b)21(secxtanx+ln∣secx+tanx∣)+C
c)21(secxtanx−ln∣secx+tanx∣)+C
d)22+ln(1+2)
e)The value lies between 1.05 and 1.21, inside the bounds 4π≈0.79 and 4π⋅21+22≈1.50.
a)
By the chain rule, F′(x)=secx+tanxsecxtanx+sec2x=secx+tanxsecx(tanx+secx)=secx, on every interval where cosx=0 and secx+tanx=0. The absolute value is there because dxdln∣g∣=gg′ holds for g of either sign. So ∫secxdx=ln∣secx+tanx∣+C. The formula is found historically by multiplying secx by secx+tanxsecx+tanx, which makes the numerator the derivative of the denominator; on an exam it is quoted, and checked by this differentiation if in doubt.
b)
With u=secx, dv=sec2xdx: du=secxtanxdx and v=tanx. Then I=∫sec3xdx=secxtanx−∫secxtan2xdx. Convert with tan2x=sec2x−1: ∫secxtan2xdx=∫sec3xdx−∫secxdx=I−ln∣secx+tanx∣. So I=secxtanx−I+ln∣secx+tanx∣. The unknown I reappeared with a MINUS sign, which is exactly what makes the equation solvable: 2I=secxtanx+ln∣secx+tanx∣, hence ∫sec3xdx=21(secxtanx+ln∣secx+tanx∣)+C. The constant is added at the end only, once I has been isolated. Had it reappeared with a plus sign, the equation would read 0=… and the choice of u would have to change.
c)
From Exercise 3 e), ∫tan2xsecxdx=∫sec3xdx−∫secxdx=21secxtanx+21ln∣secx+tanx∣−ln∣secx+tanx∣+C=21(secxtanx−ln∣secx+tanx∣)+C. The logarithms combine to −21ln∣…∣: dropping the 21 of part b) at this step is the classic way to lose the mark.
d)
At x=4π: sec4π=2 and tan4π=1. At x=0: sec0=1, tan0=0, and ln1=0. So ∫0π/4sec3xdx=21(2⋅1+ln(2+1))−21(0+ln1)=22+ln(1+2). This is the exact answer expected without a calculator; no decimal is required.
e)
The figure gives two bounds. The curve stays above height 1, so the area exceeds the rectangle 4π⋅1≈0.79. The curve bends upward (it is convex), so it stays below its chord, and the area is less than the trapezoid 4π⋅21+22≈0.79×1.91≈1.50, using sec34π=(2)3=22. Now estimate the answer: 1+2≈2.41 lies between 2 and e≈2.72, so ln(1+2) lies between ln2≈0.69 and 1, and the answer lies between 21.41+0.69=1.05 and 21.41+1≈1.21. It sits well inside (0.79,1.50): plausible. A sign error in part b), giving 22−ln(1+2)<0.37, would fall below the rectangle and be caught at once.
Points are earned before the solution; after it, you can keep practising.
For you:
Exercise 5: Products of different frequencies: turn the product into a sum
5 answers to check
∫sinmxcosnxdx with m=n is neither a power nor a substitution. The product-to-sum formulas turn it into two terms that integrate at sight: sinAcosB=21[sin(A−B)+sin(A+B)], sinAsinB=21[cos(A−B)−cos(A+B)], cosAcosB=21[cos(A−B)+cos(A+B)].
The figure shows y=sin3xsinx on [0,π], with its signed areas.
a) Compute ∫sin5xcos3xdx.
b) Compute ∫sin2xsin6xdx.
c) Compute ∫sin2xcos5xdx. What happens to sin(A−B) when A<B?
d) Evaluate ∫0π/4cos4xcos2xdx.
e) Compute ∫0πsin3xsinxdx and explain the answer on the figure.
Show the solution
Answers
a)−4cos2x−16cos8x+C
b)8sin4x−16sin8x+C
c)6cos3x−14cos7x+C; sin(−3x)=−sin3x.
d)61
e)0: the two positive lobes exactly balance the negative one.
a)
With A=5x, B=3x: sin5xcos3x=21[sin2x+sin8x]. Then ∫sin5xcos3xdx=21(−2cos2x−8cos8x)+C=−4cos2x−16cos8x+C. Why not substitute? u=sin5x gives du=5cos5xdx, and there is a cos3x, not a cos5x: the frequencies do not match, so no factor is the differential of anything. And ∫sin5xdx⋅∫cos3xdx is not an option: an integral of a product is never the product of the integrals.
b)
sin2xsin6x=21[cos(2x−6x)−cos8x]=21[cos4x−cos8x], since cosine is even, cos(−4x)=cos4x. Then ∫sin2xsin6xdx=8sin4x−16sin8x+C. The minus sign belongs to the SUM angle: sinAsinB is half of cos(A−B) minus cos(A+B). Swapping the two is caught by a check at x=4π, not at x=0 where both versions give 0: sin2πsin23π=−1, and 21[cosπ−cos2π]=−1. Correct.
c)
Here A=2x<B=5x: sin2xcos5x=21[sin(−3x)+sin7x]=21[−sin3x+sin7x]. Sine is ODD, so the negative angle brings out a sign; it is the step students skip, writing sin3x for sin(−3x). Then ∫sin2xcos5xdx=21(3cos3x−7cos7x)+C=6cos3x−14cos7x+C. With the sign lost, the answer would be −6cos3x−14cos7x, whose derivative is 2sin3x+sin7x=sin5xcos2x: the integral of a different product.
d)
cos4xcos2x=21[cos2x+cos6x], so ∫0π/4cos4xcos2xdx=[4sin2x+12sin6x]0π/4=4sin(π/2)+12sin(3π/2)=41−121=61. The value sin23π=−1 is where the marks go: at x=4π, the angle 6x is 23π, not 2π.
e)
sin3xsinx=21[cos2x−cos4x], so ∫0πsin3xsinxdx=[4sin2x−8sin4x]0π=0−0=0. On the figure, the curve is positive on (0,3π) and on (32π,π) and negative in between: the integral is a SIGNED area, and the negative lobe weighs exactly as much as the two positive ones together. This is not a coincidence of these two frequencies: for any positive integers m=n, sinmxsinnx is half a difference of two cosines of NON-ZERO integer frequencies, and each of them integrates to 0 over [0,π]. Exercise 10 uses exactly this on [−π,π].
Points are earned before the solution; after it, you can keep practising.
For you:
Part B: problems and reasoning (/50)
Exercise 6: Substitute or rewrite first, then recognize the trigonometric integral
6 answers to check
On a final, a trigonometric integral rarely arrives in the textbook form. It hides behind an inner function (x, x2, 3x), behind a double angle, or behind a quotient that becomes a product of powers once everything is written in sin and cos. The first line of the answer names that preliminary step; the rest is the method of Part A.
a) Compute ∫xcos3xdx.
b) Evaluate ∫0πxsin2(x2)dx.
c) Compute ∫tan3xsec43xdx by two different routes, and reconcile the answers.
d) Compute ∫sin2xcos3xdx. Why is the product-to-sum formula a poor idea here?
e) Compute ∫cos2xtan3xdx.
Show the solution
Answers
a)2sinx−32sin3x+C
b)4π
c)12sec43x+C, or 6tan23x+12tan43x+C; they differ by 121.
d)−52cos5x+C; cos3x is a power, not a single frequency.
e)2cos2x−ln∣cosx∣+C
a)
The inner function comes first: w=x, dw=2xdx, so xdx=2dw and the integral becomes 2∫cos3wdw. Now an odd power of cosine: save coswdw=dv with v=sinw, convert cos2w=1−v2: 2∫(1−v2)dv=2v−32v3. Back to x: ∫xcos3xdx=2sinx−32sin3x+C. The factor x1 is the signal: it is (twice) the derivative of the inner function, which is why the first substitution closes.
b)
w=x2, dw=2xdx, bounds 0 and (π)2=π: ∫0πxsin2(x2)dx=21∫0πsin2wdw. The power is even, so lower it: 21[2w−4sin2w]0π=21⋅2π=4π. The answer reuses Exercise 2 a) through one substitution. Keeping the bounds 0 and π after changing the variable would give 21(2π−4sin2π), a number that cannot be simplified without a calculator: a strong hint that the bounds were not changed.
c)
Put w=3x, dw=3dx: the integral is 31∫tanwsec4wdw, which has BOTH an odd power of tan and an even power of sec. Route 1, save secwtanwdw=du with u=secw: 31∫sec3w(secwtanwdw)=31⋅4u4=12sec43x+C. Route 2, save sec2wdw=du with u=tanw and convert sec2w=1+u2: 31∫u(1+u2)du=31(2u2+4u4)=6tan23x+12tan43x+C. Reconcile with sec2=1+tan2: 12sec43x=12(1+tan23x)2=121+6tan23x+12tan43x. They differ by the constant 121: both are correct. Route 1 is shorter; forgetting the factor 31 of the inner substitution costs the mark in either.
d)
Rewrite the double angle first: sin2x=2sinxcosx, so sin2xcos3x=2sinxcos4x. One factor sinx, odd power: u=cosx, du=−sinxdx, and ∫2sinxcos4xdx=−2∫u4du=−52cos5x+C. The product-to-sum formulas multiply two SINGLE frequencies, sinmxcosnx; here cos3x is a cube, so the formula would first need cos3x=43cosx+cos3x, then two more products: correct in the end, and four times longer. The form of the integrand, a power of cosx times its derivative up to a constant, decides.
e)
Write everything in sines and cosines: cos2xtan3x=cos2x⋅cos3xsin3x=cosxsin3x. The power of sinx is odd: save sinxdx=−du with u=cosx, convert sin2x=1−u2. Then ∫cosxsin3xdx=−∫u1−u2du=−∫(u1−u)du=−ln∣u∣+2u2+C=2cos2x−ln∣cosx∣+C. The power of cosx is negative, −1, and that changes nothing: only the saved power has to be odd. Remember that −ln∣cosx∣=ln∣secx∣ if you want to compare with a classmate's answer.
Points are earned before the solution; after it, you can keep practising.
For you:
Exercise 7: Read the answer before computing: symmetry, signs and a hidden square root
6 answers to check
A definite trigonometric integral can often be predicted, or at least bounded, before any antiderivative is written: a symmetry of the graph, the sign of the integrand, a mirror between sin and cos. Those predictions are also the best way to catch an error in a substitution that LOOKS legitimate.
The figure shows y=cos3x on [0,π].
a) Using the substitution x=2π−t, show that ∫0π/2sin2xdx=∫0π/2cos2xdx. Deduce the value of each without any half-angle formula.
b) Compute ∫0πcos3xdx and ∫0πsin3xdx. Explain the first answer on the figure.
c) Evaluate ∫−π/4π/4(tan3xsec2x+sec4x)dx with as little computation as possible.
d) A student computes ∫0πcos2xdx with u=sinx: he writes cosxdx=du and cosx=1−u2, so the integral becomes ∫001−u2du=0. Find the error and give the correct value.
e) Yet ∫0πsin2xcos3xdx IS correctly computed with u=sinx and bounds 0 and 0. Compute it, and explain why this substitution is legitimate when the one in d) was not.
Show the solution
Answers
a)Both equal 4π, since their sum is ∫0π/21dx=2π.
b)∫0πcos3xdx=0 and ∫0πsin3xdx=34.
c)0+38=38
d)1−sin2x=∣cosx∣, not cosx, on (2π,π). The value is 2π.
e)0: the integrand is exactly g(sinx)cosx with g(u)=u2(1−u2), no root involved.
a)
With x=2π−t, dx=−dt, and the bounds x=0, x=2π become t=2π, t=0. Since sin(2π−t)=cost: ∫0π/2sin2xdx=−∫π/20cos2tdt=∫0π/2cos2tdt. Call the common value J. Adding, 2J=∫0π/2(sin2x+cos2x)dx=∫0π/21dx=2π, so J=4π. The two graphs are mirror images about x=4π, and each fills exactly half of the rectangle [0,2π]×[0,1]. This two-line argument replaces the half-angle computation whenever the interval is a quarter period.
b)
cos3x: odd power, u=sinx, cos2x=1−u2, bounds sin0=0 and sinπ=0: ∫0πcos3xdx=∫00(1−u2)du=0. On the figure, cos(π−x)=−cosx, so the graph on (2π,π) is the positive lobe turned upside down: the two signed areas cancel. For sin3x: u=cosx, du=−sinxdx, bounds 1 and −1: ∫0πsin3xdx=−∫1−1(1−u2)du=∫−11(1−u2)du=2−32=34. This time the integrand is positive on (0,π), so a positive answer was expected; a result of 0 or a negative number would signal a bound written in the wrong order.
c)
Split. The function tan3xsec2x is ODD (tan is odd, sec is even, and an odd power of an odd function times an even function is odd), so its integral over the symmetric interval [−4π,4π] is 0 with no computation. The function sec4x is even: save sec2x, u=tanx, sec2x=1+u2, bounds tan(±4π)=±1: ∫−11(1+u2)du=2+32=38. Total: 38. Seeing the odd piece saves half the work and removes half the chances of a sign slip.
d)
The error is cosx=1−u2. From u=sinx one only gets cos2x=1−u2, hence 1−u2=∣cosx∣, which equals cosx on [0,2π] and −cosx on [2π,π]. The substitution silently flips the sign of the integrand on the second half, and the two halves then cancel. The integrand cos2x is non-negative and not identically zero, so an answer of 0 is impossible, which is the check that should have stopped him. The honest route: every power is even, so lower it, ∫0π21+cos2xdx=2π. The rule of Exercise 1 e) is the same rule seen from the other side: with an even power there is no factor to save, and forcing one creates a square root.
e)
Here the integrand is sin2xcos2x⋅cosx=sin2x(1−sin2x)cosx, EXACTLY of the form g(sinx)cosx with the polynomial g(u)=u2(1−u2). The substitution rule ∫abg(sinx)cosxdx=∫sinasinbg(u)du holds with no condition on the monotonicity of sinx, so ∫0πsin2xcos3xdx=∫00u2(1−u2)du=0. In d), the student never had g(sinx)cosx: he had cosx⋅cosx and replaced ONE cosx by a square root, which is where the sign was lost. Symmetry confirms e): sin2(π−x)=sin2x and cos3(π−x)=−cos3x, so the integrand is antisymmetric about 2π, as on the figure.
Points are earned before the solution; after it, you can keep practising.
For you:
Exercise 8: Five statements to correct
5 answers to check
Each statement below was written on a MATH 141 paper. Each is false. For each one, say what is wrong, write the correct statement, and give the quickest check that exposes the error.
a) ∫cos2xdx=2x−4sin2x+C.
b) To integrate sin3xcos2x, set u=sinx, since du=cosxdx appears.
c) ∫tan2xdx=3tan3x+C.
d) Since ∫secxdx=ln∣secx+tanx∣+C, by symmetry ∫cscxdx=ln∣cscx+cotx∣+C.
e) For positive integers, ∫−ππsinmxcosnxdx=0 when m=n, and =π when m=n.
Show the solution
Answers
a)∫cos2xdx=2x+4sin2x+C
b)Set u=cosx (the odd power is on sin): −3cos3x+5cos5x+C.
c)∫tan2xdx=tanx−x+C
d)∫cscxdx=−ln∣cscx+cotx∣+C=ln∣cscx−cotx∣+C
e)∫−ππsinmxcosnxdx=0 for ALL integers m, n, including m=n.
a)
The half-angle formula was taken with the wrong sign: cos2x=21+cos2x, so ∫cos2xdx=2x+4sin2x+C. Quickest check: plug x=0 into the formula used, cos20=1 must equal 21+cos0=1, whereas 21−cos0=0. A second check: differentiating the false answer gives 21−2cos2x=sin2x, so the student integrated the wrong function.
b)
The power of cosx is 2, even; saving cosxdx leaves ONE cosx=±1−u2, and the method stalls. The odd power is on the sine, so save sinxdx=−du with u=cosx and convert sin2x=1−u2: ∫sin3xcos2xdx=−∫(1−u2)u2du=−3cos3x+5cos5x+C. The fact that a cosx appears somewhere is not enough; what matters is what is LEFT after detaching it.
c)
The power rule needs the derivative of the inside: dxd3tan3x=tan2xsec2x, not tan2x. Convert instead: tan2x=sec2x−1, so ∫tan2xdx=tanx−x+C. Check: dxd(tanx−x)=sec2x−1=tan2x.
d)
Symmetry between sec and csc always brings a SIGN, because dxdcscx=−cscxcotx and dxdcotx=−csc2x. Differentiate the proposal: cscx+cotx−cscxcotx−csc2x=−cscx. It is off by a sign. The correct formula is ∫cscxdx=−ln∣cscx+cotx∣+C, which is the same as ln∣cscx−cotx∣+C because (cscx+cotx)(cscx−cotx)=csc2x−cot2x=1.
e)
sinmxcosnx is an ODD function (odd times even) for ALL m and n, so its integral over the symmetric interval [−π,π] is 0 in every case, m=n included. For m=n, directly: sinmxcosmx=2sin2mx, whose integral over [−π,π] is 0. The value π belongs to the other two families: ∫−ππsin2mxdx=∫−ππcos2mxdx=π for m≥1, by the half-angle formula. The complete table, 0 for every mixed product and π only for the square of one function, is what Exercise 10 uses.
Points are earned before the solution; after it, you can keep practising.
For you:
Exercise 9: The Mercator map: where the integral of the secant comes from
5 answers to check
On a Mercator map of a globe of radius R, the circle of latitude φ (in radians, 0≤φ<2π) has true length 2πRcosφ but is drawn with the width 2πR of the equator: the map stretches it horizontally by the factor secφ. To keep angles true, Mercator stretched the map VERTICALLY by the same factor at every latitude, which places the parallel of latitude φ at the height y(φ)=R∫0φsectdt above the equator.
The figure compares the parallels every 15∘ on the globe (spaced Rφ, evenly) and on the map.
a) Using the Fundamental Theorem, compute y′(φ) and explain why the parallels spread out on the map, as in the figure.
b) Compute y(φ) in closed form, and explain why no absolute value is needed.
c) Give the exact heights of the parallels of 45∘ and 60∘. On the globe, 60∘ is 34 as far from the equator as 45∘. Show, without a calculator, that on the map the ratio is larger than 34.
d) Mercator's navigators used tables of lntan(4π+2φ). Show that secφ+tanφ=tan(4π+2φ).
e) A wall map is cut at the height y=Rln3. Find the exact latitude of its top edge. Then explain why no Mercator map, however tall, can show the pole.
Show the solution
Answers
a)y′(φ)=Rsecφ, which is at least R and grows with φ.
b)y(φ)=Rln(secφ+tanφ)
c)Rln(1+2) and Rln(2+3); ratio above 34 since (2+3)3=26+153>17+122=(1+2)4.
d)Both sides equal cosφ1+sinφ.
e)φ=arcsin54=arctan34; secφ+tanφ→∞ as φ→2π−.
a)
y(φ)=R∫0φsectdt with sec continuous on [0,2π), so by the first part of the Fundamental Theorem y′(φ)=Rsecφ. This is exactly the vertical stretch Mercator wanted, equal to the horizontal one. Since secφ≥1, the map rises at least as fast as the globe (dφd(Rφ)=R), and since secφ INCREASES on [0,2π), each band of 15∘ is drawn taller than the one below it: that is the spreading visible on the figure, slight near the equator (sec0=1) and dramatic above 60∘ (sec3π=2).
b)
y(φ)=R[ln∣sect+tant∣]0φ=Rln∣secφ+tanφ∣−Rln∣1+0∣=Rln∣secφ+tanφ∣. On [0,2π), both secφ≥1 and tanφ≥0, so secφ+tanφ≥1>0 and the absolute value can be dropped: y(φ)=Rln(secφ+tanφ). Dropping it WITHOUT that sentence costs the justification mark; it is the kind of line a marker looks for.
c)
sec4π=2 and tan4π=1, so y(4π)=Rln(1+2). sec3π=2 and tan3π=3, so y(3π)=Rln(2+3). To compare the ratio with 34 without decimals, compare 3ln(2+3) with 4ln(1+2), that is (2+3)3 with (1+2)4, since ln is increasing. First (2+3)2=7+43, then (7+43)(2+3)=14+73+83+12=26+153. Next (1+2)2=3+22, then (3+22)2=9+122+8=17+122. Since 26>17 and 153>122, the first is larger, so y(π/4)y(π/3)>34: the map exaggerates the distance between the two parallels, which is the price of true angles.
d)
Write secφ+tanφ=cosφ1+sinφ and put θ=2φ. The double-angle formulas give 1+sinφ=cos2θ+sin2θ+2sinθcosθ=(cosθ+sinθ)2 and cosφ=cos2θ−sin2θ=(cosθ−sinθ)(cosθ+sinθ). Dividing, secφ+tanφ=cosθ−sinθcosθ+sinθ=1−tanθ1+tanθ, after dividing top and bottom by cosθ>0. This is the addition formula tan(4π+θ)=1−tan4πtanθtan4π+tanθ with tan4π=1. So y(φ)=Rlntan(4π+2φ): the same antiderivative in another form, the one the navigators tabulated.
e)
Solve ln(secφ+tanφ)=ln3, that is cosφ1+sinφ=3. With s=sinφ and cosφ=1−s2>0 on this range: 1+s=31−s2; squaring, (1+s)2=9(1−s)(1+s), and dividing by 1+s>0, 1+s=9−9s, so s=54. Check, since squaring can add a root: cosφ=53, secφ+tanφ=35+34=3. The top edge is at φ=arcsin54=arctan34, a little above 53∘. For the pole: as φ→2π−, secφ→+∞ and tanφ→+∞, so y(φ)=Rln(secφ+tanφ)→+∞. Every latitude below 2π has a finite height, but the heights are unbounded: no finite sheet reaches the pole.
Points are earned before the solution; after it, you can keep practising.
For you:
Exercise 10: Two tones on one wire: the energy of a signal and its hidden coefficients
5 answers to check
An electrical signal carries two tones, f(t)=3sint+2sin2t (volts, with t scaled so that one cycle is [−π,π]). Up to a constant factor, its energy over one cycle is E=∫−ππf(t)2dt. The figure shows f(t)2, whose area is E, and a dashed line at height 213.
This is a typical final exam problem: the computations are short, provided the right identity is chosen at each step.
a) Show that ∫−ππsin2ntdt=π for every positive integer n.
b) Show that ∫−ππsinmtsinntdt=0 for positive integers m=n.
c) Deduce E without expanding anything but the square, and interpret the dashed line of the figure.
d) The signal h(t)=sin3t is known to be of the form asint+csin3t. Find a and c by multiplying by sint, then by sin3t, and integrating over [−π,π].
e) Use d) to evaluate ∫0πsin3tdt and the energy ∫−ππsin6tdt of h, and compare with the direct routes.
Show the solution
Answers
a)∫−ππ21−cos2ntdt=π
b)21∫−ππ[cos(m−n)t−cos(m+n)t]dt=0
c)E=9π+4π=13π; the rectangle of height 213 on [−π,π] has the same area.
d)a=43, c=−41: sin3t=43sint−41sin3t.
e)∫0πsin3tdt=34 and ∫−ππsin6tdt=85π.
a)
Even power, so lower it: sin2nt=21−cos2nt, and ∫−ππsin2ntdt=[2t−4nsin2nt]−ππ=π−0=π, because sin(±2nπ)=0 for every integer n. The value does not depend on n: over a whole number of periods, sin2 averages 21 whatever the frequency.
b)
Product to sum: sinmtsinnt=21[cos(m−n)t−cos(m+n)t]. Since m=n, both m−n and m+n are NON-ZERO integers, and for a non-zero integer k, ∫−ππcosktdt=[ksinkt]−ππ=0. Hence ∫−ππsinmtsinntdt=0. The hypothesis m=n is used exactly once: for m=n, the term cos0=1 appears and integrates to 2π, giving back part a).
c)
f(t)2=9sin2t+12sintsin2t+4sin22t. By a), the first and last terms integrate to 9π and 4π; by b), the cross term integrates to 0. So E=13π: the energies of the two tones simply ADD, and E is the energy of the first tone (9π) plus that of the second (4π). On the figure, the area under f(t)2 equals that of the rectangle of height 213 on an interval of length 2π, since 213⋅2π=13π: the bumps above the dashed line exactly fill the gaps below it. Without a) and b), the same result needs three antiderivatives, one of them a product to sum, and three evaluations at ±π.
d)
Multiply sin3t=asint+csin3t by sint and integrate over [−π,π]: by a) and b), ∫−ππsin4tdt=aπ+0. The left side lowers the degree twice: sin4t=83−2cos2t+8cos4t, whose integral over [−π,π] is 83⋅2π=43π. So a=43. Multiply by sin3t instead: ∫−ππsin3tsin3tdt=0+cπ. On the left, sin3tsint=2cos2t−cos4t and sin2t=21−cos2t, so the integrand is 41(cos2t−cos4t−cos22t+cos2tcos4t); over [−π,π] every term gives 0 except −cos22t, which gives −π. So cπ=−4π and c=−41. Check with the triple-angle identity sin3t=3sint−4sin3t: it gives sin3t=43sint−sin3t, the same decomposition.
e)
∫0πsin3tdt=43∫0πsintdt−41∫0πsin3tdt=43⋅2−41⋅32=23−61=34, the value found in Exercise 7 b) by saving a factor. The energy of h uses c) again: ∫−ππsin6tdt=∫−ππ(43sint−41sin3t)2dt=169π+161π=1610π=85π, the cross term vanishing by b). The direct route lowers sin6t=(21−cos2t)3, expands a cube and lowers cos22t and cos32t in turn, to find the same constant term 165 and the same 165⋅2π=85π. Both routes are valid; the second is where sign errors live.
Points are earned before the solution; after it, you can keep practising.
Something missing from this set?A type of exercise your teacher gives, a problem you are stuck on, a mistake you spotted: tell me, I read every request.
Sending goes through your free space: that is how I can answer you and tell you when the exercise is added. An email is enough, no password. Create your space.
Rather not create an account? Write to me directly.