The Student Room Group

STEP Maths I, II, III 1989 solutions

Scroll to see replies

Reply 60

II/7:

t = x^{\alpha} \Rightarrow \dfrac{\text{d}t}{\text{d}y} = \alpha x^{\alpha - 1}\dfrac{\text{d}x}{\text{d}y} \Rightarrow \dfrac{\text{d}y}{\text{d}x} = \alpha x^{\alpha - 1} \dfrac{\text{d}y}{\text{d}t}

\Rightarrow \dfrac{\text{d}^2y}{\text{d}x^2} = \alpha (\alpha - 1)x^{\alpha - 2} \dfrac{\text{d}y}{\text{d}t} + \alpha x^{\alpha - 1} \dfrac{\text{d}}{\text{d}x} \left( \dfrac{\text{d}y}{\text{d}t} \right)

Note that \dfrac{\text{d}}{\text{d}x} \left( \dfrac{\text{d}y}{\text{d}t} \right) = \dfrac{\text{d}}{\text{d}t} \left( \dfrac{\text{d}y}{\text{d}t} \right) \dfrac{\text{d}t}{\text{d}x} = \dfrac{\text{d}^2y}{\text{d}t^2} \dfrac{\text{d}t}{\text{d}x}.

\Rightarrow \dfrac{\text{d}^2y}{\text{d}x^2} = \alpha (\alpha - 1)x^{\alpha - 2} \dfrac{\text{d}y}{\text{d}t} + \alpha^2 x^{2 \alpha - 2} \dfrac{\text{d}^2y}{\text{d}t^2}.

Substituting this horrendous crap leads to: \alpha^2 x^{2 \alpha} \dfrac{\text{d}^2y}{\text{d}t^2} + (\alpha(\alpha - 1) -b \alpha)x^{\alpha} \dfrac{\text{d}y}{\text{d}t} + x^{2b + 2}y = 0.

Now choosing \alpha = b + 1 and substituting will smash up dy/dt one time bro', leaving

(b+1)^2 t^2 \dfrac{\text{d}^2y}{\text{d}t^2} + t^2y = 0 \Rightarrow (b+1)^2 \dfrac{\text{d}^2y}{\text{d}t^2} + y = 0 \Rightarrow y = A \cos \left( \dfrac{t}{b+1} \right) + B \sin \left( \dfrac{t}{b + 1} \right)

= A \cos \left( \dfrac{x^{b+1}}{b + 1} \right) + B \sin \left( \dfrac{x^{b+1}}{b + 1} \right).

If A = 1, then as x -> 0 then y -> and dy/dx -> 0, but as B can be anything, this is not a unique solution.

Reply 61

II/4:

\text{f}(x) = \dfrac{(x-a)(x-b)}{(x-c)(x-d)} \Rightarrow \ln \text{f}(x) = \ln(x - a) + \ln(x - b) - \ln (x-c) - \ln(x - d)
\Rightarrow \dfrac{\text{f'}(x)}{\text{f}(x)} = \dfrac{1}{x - a} + \dfrac{1}{x - b} - \dfrac{1}{x - c} - \dfrac{1}{x - d}
\Rightarrow \dfrac{x \text{f'}(x)}{\text{f}(x)} = \left( \dfrac{x-a}{x} \right)^{-1} + \left( \dfrac{x - b}{x} \right)^{-1} - \left( \dfrac{x - c}{x} \right)^{-1} - \left( \dfrac{x - d}{x} \right)^{-1}
= \left( 1 - \dfrac{a}{x} \right)^{-1} + \left( 1 - \dfrac{b}{x} \right)^{-1} - \left( 1 - \dfrac{c}{x} \right)^{-1} - \left( 1 - \dfrac{d}{x} \right)^{-1}

If |x| is much greater than |a|, |b|, |c| and |d| we can use the first two terms of the binomial expansion with sufficient* accuracy.

\dfrac{x \text{f'}(x)}{\text{f}(x)} \approx 1 + \dfrac{a}{x} + 1 + \dfrac{b}{x} - 1 - \dfrac{c}{x} - 1 - \dfrac{d}{x}

\Rightarrow \dfrac{x^2 \text{f'}(x)}{\text{f}(x)} \approx a + b - c - d.

As |x| gets very large, f(x) tends to 1 (can be seen by writing f(x) as \dfrac{(1 - a/x)(1 - b/x)}{(1 - c/x)(1 - d/x)}, and x^2 is always positive, so x^2/f(x) is positive. So f'(x) (the gradient of f(x)) has the same sign as a + b - c - d.

Let z = \dfrac{(x -a)(x -b)}{(x-c)(x-d)} \Rightarrow z(x-c)(x-d) = (x-a)(x-b). Now, the value of z much be such that there is a root of this equation (i.e. a value of x). As it's a quadratic in x, the discriminant of this quadratic, the expression shown, will be greater than or equal to zero.

Reply 63

1989 Paper 1 number 16

Reply 65

1989 Paper 2 nos. 12 & 13

Reply 66

1989 Paper 2 nos. 15 & 16

Reply 68

1989 Paper 3 nos. 13-16
Now can ANYONE do numbers 11 and 12?

Reply 69

1989 Paper III
Sorry, there is an error in my solutoion for number 13 so here is an amended version.

Reply 70

brianeverit
1989 Paper 3 nos. 13-16
Now can ANYONE do numbers 11 and 12?


I have quite a different answer to the end of 15.

I get E(Z) = 0 as z(c^2 - z^2)^n is an odd function.

For Var(Z):

Note that \displaystyle \int_{-c}^c (c^2 - z^2)^n \, \text{d}z = \frac{(n!)^2 (2c)^{2n+1}}{(2n+1)!}. Now, let z = c \sin \theta and we get \displaystyle \int_{-\pi/2}^{\pi/2} c^{2n+1} \cos^{2n+1} \theta \, \text{d} \theta = \frac{(n!)^2 (2c)^{2n+1}}{(2n+1)!}

For Var(Z), we want to evaluate \displaystyle \frac{(2n+1)!}{(n!)^2 (2c)^{2n+1}} \int z^2(c^2 - z^2)^n \, \text{d}z. Letting z = c \sin \theta as before transforms this to

\displaystyle c^2 \frac{(2n+1)!}{(n!)^2 (2c)^{2n+1}} \int_{-\pi/2}^{\pi/2} c^{2n+1} \cos^{2n+1} \theta \, \text{d} \theta - \frac{(2n+1)!}{(n!)^2 (2c)^{2n+1}} \int_{-\pi/2}^{\pi/2} c^{2n+3} \cos^{2n+3} \theta \, \text{d}\theta. Using what we know for the first integral, and replacing n by n + 1 for the second, we get

c^2 - \dfrac{(2n+1)!}{(n!)^2 (2c)^{2n+1}} \cdot \dfrac{((n+1)!)^2 (2c)^{2n+3}}{(2n+3)!} = \dfrac{c^2}{2n+3}.
Quite a nice answer, makes me wonder if there's an easier way.

Reply 71

Having looked at number Paper III 15 again I agree with your solution.

Reply 72

Dystopia
STEP 1, Q3

\textbf{x} = x\textbf{a}, \; \textbf{y} = (1-x)\textbf{b}

\textbf{c} = \frac{1}{3}(2\textbf{a} + \textbf{b}), \textbf{d} = \frac{1}{3}(\textbf{a} + 2\textbf{b})

(By the ratio theorem.)

\overrightarrow{CY} = (\frac{2}{3} - x)\textbf{b} - \frac{2}{3}\textbf{a}
\overrightarrow{DX} = (x - \frac{1}{3})\textbf{a} - \frac{2}{3}\textbf{b}

These vectors are perpendicular, so the scalar product is zero.

(\frac{2}{9} + x - x^{2})\textbf{a}.\textbf{b} - \frac{2}{9} = 0

(Note that the scalar product is commutative and distributive.)

\textbf{a}.\textbf{b} = \frac{-2}{9x^{2} - 9x - 2}

Let f(x) = \frac{-2}{9x^{2} - 9x - 2}, \; 0 \leq x \leq 1

f'(x) = \frac{2(18x - 9)}{(9x^{2}-9x-2)^{2}}

Minimum and maximum values occur at the endpoint of a range or at a turning point. There is a turning point when x = 1/2.

f(0) = 1, \; f(\frac{1}{2}) = \frac{8}{17}, \; f(1) = 1

So \frac{8}{17} \leq \textbf{a}.\textbf{b} \leq 1

Note that \textbf{a}.\textbf{b} = |a||b|\cos\theta = \cos\theta

Where \theta = \angle AOB, \; 0 < \theta < \pi

The maximum value is \theta occurs when \cos\theta = \frac{8}{17}

Suppose that \lambda \overrightarrow{CY} = \overrightarrow{CE},\; \mu \overrightarrow{DX} = \overrightarrow{DE}

Then \lambda (\frac{1}{6} \textbf{b} - \frac{2}{3} \textbf{a}) = \frac{1}{3} ( \textbf{b} - \textbf{a} ) + \mu ( \frac{1}{6} \textbf{a} - \frac{2}{3} \textbf{b} )

So \frac{1}{6} \lambda = \frac{1}{3} - \frac{2}{3}\mu

-\frac{2}{3}\lambda = -\frac{1}{3} + \frac{1}{6}\mu

\lambda = \mu = \frac{2}{5}

\textbf{e} = \textbf{c} + \frac{2}{5}\overrightarrow{CY} = \frac{2}{5}\textbf{a} + \frac{2}{5}\textbf{b}

By the cosine rule, AB^{2} = \frac{18}{17}
Let F be the midpoint of AB. By Pythagoras, OF^{2} = 1 - \frac{9}{34} = \frac{25}{34}, \; OF = \frac{5}{\sqrt{34}}

OE = \frac{3}{5} \times \frac{5}{\sqrt{34}} = \frac{3}{\sqrt{34}}


OF=1/2(a+b) and OE=2/5 (a+b) so surely OE=4/5(a+b)=4/5OF not 3/5?

Reply 73

I made a sign error at one point (accidently multiplied by A sinh t instead of -A sinh t) so there's a bit of a mess on the last little bit (I'd have wrote it out again in a real exam) but it should be understandable - STEP II 1989 Q9 is attatched (It'd probably take a very long time to latex)

Reply 74

brianeverit
1989 Paper 3 nos. 13-16
Now can ANYONE do numbers 11 and 12?


I did 12 and did the first part of 11 but have no idea about the last part

Reply 75

brianeverit
1989 Paper III
Sorry, there is an error in my solutoion for number 13 so here is an amended version.


holy **** how did you think of that?:eek3:

Reply 76

Haha 1989 ... I bet the STEP papers were monstrous back then! :eek:

Reply 77

Original post
by squeezebox
...

'SimonM'
I wasn't sure if the OP was still active.

STEP I - Q1

solution

Reply 78

Original post
by brianeverit
1989 Paper 1 numbers 10,12-15


Hi, I´ve been trying to do Q10 and I just saw your solution which is different to my method.

Couldn´t you just calculate the loss in kinetic energy and then equate that to the work done by the force? i.e. :
\Delta K.E = \int^d_0 \frac{M\omega (v^2 +V^2)}{v}dx
where x is the distance from the start of the field and v=\dot{x}

I was having a bit of trouble with the integration so I was wondering if anyone could give me a hand:colondollar:

It might just all be wrong but any help would be appreciated.

Reply 79

Original post
by ben-smith
Hi, I´ve been trying to do Q10 and I just saw your solution which is different to my method.

Couldn´t you just calculate the loss in kinetic energy and then equate that to the work done by the force? i.e. :
\Delta K.E = \int^d_0 \frac{M\omega (v^2 +V^2)}{v}dx
where x is the distance from the start of the field and v=\dot{x}

I was having a bit of trouble with the integration so I was wondering if anyone could give me a hand:colondollar:Well, you can't really integrate that without ending up doing what Brian did.

Let's write v(x) for v, to emphasise that v is a function of x.

Then you have

\frac{1}{2}v(0)^2 - \frac{1}{2}v(d)^2 = \int_0^d \frac{\omega (v(x)^2 +V^2)}{v(x)}dx

So, here's the problem - at this point you have absolutely no idea what v(x) is, which is going to make finding the integral pretty tricky.

So, let's relabel some of the variables: I'm going to write 'x' instead of d, and 't' instead of 'x'.

\frac{1}{2}v(0)^2 - \frac{1}{2}v(x)^2 = \int_0^d \frac{\omega (v(t)^2 +V^2)}{v(t)}dt

Now diff both sides w.r.t. x:

v(x)\frac{dv}{dx} =   \frac{\omega (v(x)^2 +V^2)}{v(x)}

Which is Brian's first line. So from here you can solve for v as a function of x, which will give you the desired result.

Obviously it's quicker to just use the v dv/dx form of acceleration to start with.
(edited 15 years ago)

Quick Reply