Location via proxy:   [ UP ]  
[Report a bug]   [Manage cookies]                

Ph304 Problem Set 3 Electrodynamics: Princeton University

Download as pdf or txt
Download as pdf or txt
You are on page 1of 3

Princeton University

Ph304 Problem Set 3 Electrodynamics


(Due 5 pm, Thursday Feb. 27, 2003 in Matt Sullivans mailbox, Jadwin atrium)

Instructor: Kirk T. McDonald, Jadwin 309/361, x6608/4398 kirkmcd@princeton.edu http://puhep1.princeton.edu/mcdonald/examples/ AI: Matthew Sullivan, 303 Bowen Hall, x8-2123 mtsulliv@princeton.edu Problem sessions: Sundays, 7 pm, Jadwin 303 Text: Introduction to Electrodynamics, 3rd ed. by D.J. Griths (Prentice Hall, ISBN 0-13-805326-X, now in 6th printing) Errata at http://academic.reed.edu/physics/faculty/griths.html

Princeton University 2003

Ph304 Problem Set 3

Reading: Griths secs. 3.3-3.4, 4.1-4.2. 1. Griths prob. 3.15. 2. Extended version of Griths prob. 3.23. Show that the general form of solutions to Laplaces equation, 2 V = 0, in cylindrical coordinates (r, , z) when there is no z dependence is V (r, ) = (a0 + b0 ln r)(c0 + d0 ) +
k

ak r k +

bk dk cos k + ck rk + k sin k . rk r

If the region of interest includes 0 2, show that the indices k are the positive integers. Use this expansion to motivate an image method for the problem of a line charge per meter at position (r, ) = (b, 0) in the presence of a grounded conducting cylinder of radius a < b, whose axis is the z axis. We hope that a solution exists in which the eect of the conducting cylinder on the region r > a is the same as that of an imaginary line charge per meter placed at some suitable radius c < a. While you are welcome to carry out a solution in which all the Fourier coecients are explicitly evaluated, this is not necessary to achieve the goal of deducing the image method. First, write down the form of the potential for the wire alone, for both r < b and r > b, noting that some Fourier coecients must by zero by symmetry arguments. It is helpful to write the radial dependence in terms of the scaled variable r/b rather than r alone. Likewise, write down the form of the potential due to the conducting cylinder for r > a, which will have the same symmetries at the potential of the wire, since the potential of the cylinder is induced by that of the wire. Then, since the total potential vanishes at r = a, one nds a simple relation between the potential due to the cylinder and that of the wire namely, an image prescription. The potential of a 2-dimensional charge distribution not including the z axis must be nite at this axis, but can have a logarithmic divergence as r . With care you can include terms in ln r in appropriate places in your expansions of the potentials of the wire and the cylinder, to nd that the coecients of these terms obey the same transformation as do the other terms of the expansions. Not for credit: you might wish to construct the image method directly, in the spirit of Griths Ex. 3.2. Experts will also note that 2-D solutions to Laplaces equation are expressible in terms of a function of a complex variable, which have useful application to problems of line charges and cylinders. See, for example, Chap. 4 of Static and Dynamic Electricity by W.R. Smythe (McGraw-Hill, 1968). 3. Griths prob. 3.37. Without loss of generality, set the potential V0 of the conducting sphere to zero. Griths recommends this problem be solved by separation of variables in spherical coordinates. Since the free charge distribution varies as P1 (cos ), the only terms in the Legendre polynomial expansion will be those in P1 . Further, if you write

Princeton University 2003

Ph304 Problem Set 3

the potential with normalized radial coordinates, then continuity of the potential at r = b tells us that there are only two unknown expansion coecients, A1 and B1 :

r b V (a < r < b) = A1 + B1 b r V (r > b) = C1 b r


2

P1 (cos ),
2

b P1 (cos ) = (A1 + B1 ) r

P1 (cos ).

Then, since V (a) = 0, there is really only one unknown.... However, this problem is also susceptible to treatment by the image method! Carefully deduce the image charge density on the appropriate image sphere. As both charge densities and vary as cos = P1 (cos ), both of these charge distributions have only dipole moments. Hence you should be able to quickly write down the potential both for r > b, and for a < r < b, recalling from Ex. 3.9 that the eld inside a cos charge distribution is uniform. This last remark oers a third way to solve this problem. Since the electric eld must vanish inside the conductor for r < a, a surface charge must be induced at r = a whose interior eld exactly cancels the interior eld from the given charge distribution at r = b. Show that deduced this way is the same as that obtained from the potential that you have previously calculated. 4. Griths prob. 3.40. 5. Griths prob. 3.41. 6. Griths prob. 3.44. Since conductors are equipotentials, the integrals mentioned in the statement of Greens reciprocity theorem in prob. 3.43 lead to the total charges on the conductors. Hence, this theorem is more often stated as: if a set {i} of conductors is at potentials Vi when charges Qi are placed on them, and instead charges Qi would result in potentials Vi , then Qi Vi = Qi Vi .
i i

A useful trick is to consider a point charge as residing on a tiny conductor, so that in a second scenario with the same conductors, the amount of the charge on this tiny conductor can be set to zero, and this conductor will take on the potential that would exist at the position of the absent charge. Greens reciprocity theorem can be used to deduce charge distributions on conductors by imagining those conductors to be suitably subdivided. For an example of such a procedure, see prob. 4, Ph501 Set 4, which is presented separately at http://puhep1.princeton.edu/mcdonald/examples/straw.pdf

You might also like