# Confusion between Electrodynamics texts

• Classical
• Mr.Husky
In summary: Actually Sadiku doesn't contain all the details which are present in Griffiths (as it is a EE book) and also it doesn't mention the word Fourier in it's index. But Sadiku spends more time explaining PDE's vanhees71 mentioned which I think it is good for me. So after Sadiku which EM book should I consult as it deals only with electrostatics.Principles of Electricity and Magnetism by Matthew Sadiku is a good book to consult.

#### Mr.Husky

Gold Member
Hello everyone,

I recently completed kleppner and kolenkow classical mechanics book. Next I am going to learn Electrodynamics. My brother is a EE major and he gave me his copy of "principles of electromagnetics" Matthew Sadiku 4th edition. But a lot of people here recommend Griffiths. So,

1.) Can I use Sadiku as a substitute for Griffiths?
2.) Does Sadiku book prepare me for more advanced em books like zangwill ?

Edit:- I know calculus 1,2 and linear algebra . Will be going to study ODE.

Last edited:
K&K is one of the more advanced beginner texts.
Sadiku is an intermediate text for EE. It should be able to prepare you for Zangwill. It will use vector calculus, so it might be too difficult for you.

Mr.Husky and vanhees71
caz said:
K&K is one of the more advanced beginner texts.
Sadiku is an intermediate text for EE. It should be able to prepare you for Zangwill. It will use vector calculus, so it might be too difficult for you.
Yes vector analysis is difficult but I have enough time to learn it. One more thing, should I have to go through the chapters on antennas, transmission lines, waveguides?

Well, usually EE texts on electromagnetic field theory are pretty good, but one caveat is that they sometimes have a different convention, particularly with regard to the Fourier transform. They often have the opposite sign in the time Fourier integral than physicists, i.e., they usually make the ansatz with ##\exp(+\mathrm{i} \omega t)## instead of ##\exp(-\mathrm{i} \omega t)## when describing harmonic time dependence. That usually drives me nuts when reading an EE text (of course they also use ##\mathrm{j}## instead of ##\mathrm{i}## for the imaginary unit, but that's a minor nuissance).

sysprog, Delta2, PhDeezNutz and 2 others
Ganesh Mammu said:
Yes vector analysis is difficult but I have enough time to learn it. One more thing, should I have to go through the chapters on antennas, transmission lines, waveguides?
Yes. Either you will need them in your future or it will be good for you to have been exposed to them once. As a requirement for Zangwill, the answer is no.

Mr.Husky and vanhees71
vanhees71 said:
Well, usually EE texts on electromagnetic field theory are pretty good, but one caveat is that they sometimes have a different convention, particularly with regard to the Fourier transform. They often have the opposite sign in the time Fourier integral than physicists, i.e., they usually make the ansatz with ##\exp(+\mathrm{i} \omega t)## instead of ##\exp(-\mathrm{i} \omega t)## when describing harmonic time dependence. That usually drives me nuts when reading an EE text (of course they also use ##\mathrm{j}## instead of ##\mathrm{i}## for the imaginary unit, but that's a minor nuissance).
Oh I got it vanhees71.
Does it creat any harm when studying quantum mechanics?

vanhees71
caz said:
Yes. Either you will need them in your future or it will be good for you to have been exposed to them once. As a requirement for Zangwill, the answer is no.
Thank you caz.

Frabjous
Ganesh Mammu said:
Oh I got it vanhees71.
Does it creat any harm when studying quantum mechanics?
No. If you cannot handle a different convention, you have other problems.

Mr.Husky
I don't think it creates any harm, but I think it's already hard enough to start electrodynamics. For me it was the most difficult subject in the early curriculum at university. In Germany it's taught in the experimental-physics course in the 2nd and in the theory course the 3rd semester. In my opinion that's too early, because of all the math one needs: It starts with vector calculus. Then you also need some of the math of partial differential equations (Laplace and Poisson equations for electrostatics; then of course the wave equation for electrodynamics) and the related ways to solve the corresponding boundary/initial value problems. Among them you need Fourier transformations and series (including generalized ones like the spherical and solid harmonics when it comes to the multipole expansion). For that latter subject the treatment of waveguides and cavities are great, because they provide some intuition for why these orthogonal systems of functions occur.

So it's good to first stick to one convention not to have to struggle with one more unnecessary complication. If you get used to the subject, it's not that difficult anymore to switch from one convention to another.

sysprog, PeroK, Mr.Husky and 1 other person
caz said:
No. If you cannot handle a different convention, you have other problems.
A small question caz:-
Actually Sadiku doesn't contain all the details which are present in Griffiths (as it is a EE book) and also it doesn't mention the word Fourier in it's index. But Sadiku spends more time explaining PDE's vanhees71 mentioned which I think it is good for me. So after Sadiku which EM book should I consult as it deals only with electrostatics.

vanhees71 and Mr.Husky
I would stay away from Zangwill. His book is called "Modern Electrodynamics", however he used the ##x_4 = ict## which is anything but modern. I would suggest D.J. Griffiths' text, even though he uses the ##-+++## metric.

Demystifier and Mr.Husky
Dr Transport said:
After I learn vector analysis I decided to get my copy of mathematical methods book by mentor Oorudin.

dextercioby said:
I would stay away from Zangwill. His book is called "Modern Electrodynamics", however he used the ##x_4 = ict## which is anything but modern. I would suggest D.J. Griffiths' text, even though he uses the ##-+++## metric.
Thanks Dextercioby. I am now looking for a book some advanced than Griffiths but still explaining things clearly.

Another question here :-
Can anyone explain more about the notation used in special relativity.

Ganesh Mammu said:
A small question caz:-
Actually Sadiku doesn't contain all the details which are present in Griffiths (as it is a EE book) and also it doesn't mention the word Fourier in it's index. But Sadiku spends more time explaining PDE's vanhees71 mentioned which I think it is good for me. So after Sadiku which EM book should I consult as it deals only with electrostatics.
Like I said earlier, Sadiku is an EE text. Zangwill is physics. So they emphasize different things. What advanced book you use depends on the direction you are heading; i.e., do not worry about it now.

Mr.Husky
caz said:
Like I said earlier, Sadiku is an EE text. Zangwill is physics. So they emphasize different things. What advanced book you use depends on the direction you are heading; i.e., do not worry about it now.
Ok caz. I think it is waste of time now thinking what to study in future. Rather I am now interested in knowing more about special relativity. So can you answer my previous question in #15?

bob012345
The sign convention deals with how you do inner products for 4-vectors in relativity. In modern approaches, the time portion has a different sign than the spatial part so you can have -+++ or +---. An older approach uses i which allows you to define the inner product ”normally” because i2 changes the sign. The modern approach is preferable and eventually becomes required in general relativity. People get really worked up over the topic. Personally, it does not bother me in EM texts.

Last edited:
bob012345 and Mr.Husky
caz said:
The sign convention deals with how you do inner products for 4 vectors in relativity. In modern approaches, the time portion has a different sign than the spatial part so you can have
-+++ or +—-. An older approach, uses i which allows you to define the inner product
”normally” because i2 changes the sign. The modern approach is preferable and eventually becomes required in general relativity. People get really worked up over the topic. Personally, it does not bother me in EM texts.
I think it takes time to understand this stuff. Thank you caz for replying!

Frabjous
dextercioby said:
I would stay away from Zangwill. His book is called "Modern Electrodynamics", however he used the ##x_4 = ict## which is anything but modern. I would suggest D.J. Griffiths' text, even though he uses the ##-+++## metric.
Well, one has to get used to the mess with the signature in relativity anyway. The HEP people usually use the west-coast convention ##\eta_{\mu \nu}=\mathrm{diag}(1,-1,-1,-1)## (though not all as e.g., Weinberg) and the GR people mostly use the east-coast convention ##\eta_{\mu \nu}=\mathrm{diag}(-1,1,1,1)## (though not all like e.g. the newer editions of Landau and Lifshitz and Adler). That's all fine, but the use of the old ##\mathrm{i} c t## convention is a sin. If only Sommerfeld wouldn't have done this, I'd say don't read any book using it, but Sommerfeld is so good (imho the best theory books on classical physics ever written)... His vol. 3 is also a very good intro to electricity and magnetism, while the (also highly recommended) treatment of vector calculus is in vol. 2 on hydrodynamics.

I don't know, what's modern on Zangwill's book. It's as old-fashioned as Jackson and his is very comprehensive and much better than Zangwill (at least the 2nd edition where he sticks consistently with Gaussian units, which are the 2nd-best choice for theoretical electrodynamics; the best being Heaviside-Lorentz units). A truly modern approach and still accessible for beginning graduate students is Landau and Lifshitz vol. 2. Another even more modern book is Lechner, making also use of modern theory of generalized functions/distributions, solving many of the age-old troubles with point charges (even massless ones), as far as this unphysical classical point-particle model makes sense at all (it doesn't really of course, but Lechner pushes it as far as one can, at least to my knowledge).

I'd, however, not bother with these advanced books for beginning. I like the book by Griffiths very much, including a very nice approach to what's infamously dubbed "hidden momentum" (although it's just relativistic momentum). Another somewhat older very good book is Abraham and Becker (available in a nice Dover edition). Particularly, it has a very good introductory chapter on vector calculus (of about 40 pages).
A more accessible "relativity first approach" is by Schwartz. Stay away from Berkeley physics course vol. 2 (Purcell) which at least I found confusing when I first studied the subject. We also liked the volume on electrodynamics in the theory series by Greiner. Last but not least another gem is of course vol. 2 of the Feynman lectures.

sysprog and Mr.Husky
vanhees71 said:
Well, one has to get used to the mess with the signature in relativity anyway. The HEP people usually use the west-coast convention ##\eta_{\mu \nu}=\mathrm{diag}(1,-1,-1,-1)## (though not all as e.g., Weinberg) and the GR people mostly use the east-coast convention ##\eta_{\mu \nu}=\mathrm{diag}(-1,1,1,1)## (though not all like e.g. the newer editions of Landau and Lifshitz and Adler). That's all fine, but the use of the old ##\mathrm{i} c t## convention is a sin. If only Sommerfeld wouldn't have done this, I'd say don't read any book using it, but Sommerfeld is so good (imho the best theory books on classical physics ever written)... His vol. 3 is also a very good intro to electricity and magnetism, while the (also highly recommended) treatment of vector calculus is in vol. 2 on hydrodynamics.

I don't know, what's modern on Zangwill's book. It's as old-fashioned as Jackson and his is very comprehensive and much better than Zangwill (at least the 2nd edition where he sticks consistently with Gaussian units, which are the 2nd-best choice for theoretical electrodynamics; the best being Heaviside-Lorentz units). A truly modern approach and still accessible for beginning graduate students is Landau and Lifshitz vol. 2. Another even more modern book is Lechner, making also use of modern theory of generalized functions/distributions, solving many of the age-old troubles with point charges (even massless ones), as far as this unphysical classical point-particle model makes sense at all (it doesn't really of course, but Lechner pushes it as far as one can, at least to my knowledge).

I'd, however, not bother with these advanced books for beginning. I like the book by Griffiths very much, including a very nice approach to what's infamously dubbed "hidden momentum" (although it's just relativistic momentum). Another somewhat older very good book is Abraham and Becker (available in a nice Dover edition). Particularly, it has a very good introductory chapter on vector calculus (of about 40 pages).
A more accessible "relativity first approach" is by Schwartz. Stay away from Berkeley physics course vol. 2 (Purcell) which at least I found confusing when I first studied the subject. We also liked the volume on electrodynamics in the theory series by Greiner. Last but not least another gem is of course vol. 2 of the Feynman lectures.
And can Griffiths prepare me for Lechner because its contents look more promising than Zangwill.

Ganesh Mammu said:
And can Griffiths prepare me for Lechner because its contents look more promising than Zangwill.
After looking at couple of pages of Lechner's book, Does mathematical methods at the level of boas prepare me to tackle it?

Ganesh Mammu said:
And can Griffiths prepare me for Lechner because its contents look more promising than Zangwill.
I don't know Sadiku. So it's hard to say.

It's anyway wise to first check several books in the library, which you get along with best. I'd not recommend Lechner to start learning E&M. It's too much advanced. I think Boas ( Mathematical Methods in the Physical Sciences) is a good math for physicists book. Particularly the chapter on vector analysis should be very helpful to work through before starting electromagnetism.

Mr.Husky
vanhees71 said:
I don't know Sadiku. So it's hard to say.

It's anyway wise to first check several books in the library, which you get along with best. I'd not recommend Lechner to start learning E&M. It's too much advanced. I think Boas ( Mathematical Methods in the Physical Sciences) is a good math for physicists book. Particularly the chapter on vector analysis should be very helpful to work through before starting electromagnetism.
Now I got it. After I complete Griffiths and Boas then I will move to Lechner. Thank you vanhees71 for your reply!

Thanks all for your replies. I will be studying Griffiths from now. No more questions.

@vanhees71
I can see why you like Lechner, but do you really think that it would be a good FIRST graduate EM text?

Mr.Husky
I guess one should first have a more physics centered text. The radiation-reaction problem is not that important after all. At the end it boils down to the result that one needs quantum theory to describe matter anyway. On the other hand, the relativity-first approach should be good as a first graduate text. For this I'd recommend Landau Lifshitz vol. 2 and then of course also vol. 8.

Mr.Husky
vanhees71 said:
... That's all fine, but the use of the old ##\mathrm{i} c t## convention is a sin.
There are many sins in the world today but using the old ##\mathrm{i} c t## convention is not on my list.
vanhees71 said:
Stay away from Berkeley physics course vol. 2 (Purcell) which at least I found confusing when I first studied the subject.
I disagree. Purcell is a fine book written by a Nobel prize winner.

Mr.Husky and weirdoguy
Well, I don't say you shoud throw away old textbooks using the ##\mathrm{i} c t## convention. For me the best general theory textbook (series) ever written are the 6 volumes by Sommerfeld (Lectures on theoretical physics), where this convention is used too. I'd only never recommend to learn special relativity from a textbook using it, because it's utmost confusing. If you are used to the relativistic four-vector formalism it doesn't do much harm anymore.

Only because a textbook is written by a Nobel laureate it's not a guarantee for being a good textbook. I find Purcell's E&M utmost confusing. As a much better book, also written by a Nobel laureate and with the same good intention for providing a "relativity first" approach, I'd recommend rather M. Schwartz, Principles of Electrodynamics, Dover Pub. (1972).

It's of course only my personal opinion and reflects my personal experience when trying to learn from Purcell's textbook. For me the revelation about E&M were the Feynman Lect. vol. 2 and after a while A. Sommerfeld, Lects. on Theoretical Physics, vol. 3. It may well be that other students find Purcell's book helpful.

hutchphd, Mr.Husky, Demystifier and 1 other person
vanhees71 said:
Well, I don't say you should throw away old textbooks using the ##\mathrm{i} c t## convention.
I'm glad. Reading an old book is like shopping in a supermarket where you know where everything is.
vanhees71 said:
Only because a textbook is written by a Nobel laureate it's not a guarantee for being a good textbook. I find Purcell's E&M utmost confusing. As a much better book, also written by a Nobel laureate and with the same good intention for providing a "relativity first" approach, I'd recommend rather M. Schwartz, Principles of Electrodynamics, Dover Pub. (1972).
I'll check it out. Thanks.

hutchphd, Mr.Husky and vanhees71
Mr.Husky said:
Yes vector analysis is difficult but I have enough time to learn it. One more thing, should I have to go through the chapters on antennas, transmission lines, waveguides?
Vector analysis cannot be put off. It is required if you want to understand electromagnetism.
The problem is that most texts treat them as separate subjects.
You could try "Classical Electromagnetism" by Franklin which coordinates Vector Analysis with the electromagnetism. It is more advanced than Griffiths, but the first three chapters, which introduce vector anaylsis, are fairly simple.

Sure, it's nonsense to think you could learn physics without using the only adequate language to discuss it, and that's almost all in terms of analytical/differential geometry.

You get quite a long way with Euclidean vector calculus. In classical mechanics and (in the non-covariant "1+3 formalism" of relativity) electrodynamics you only rarely need 2nd-rank tensors (in point-particle mechanics the tensor of inertia for rigid bodies, in continuum mechanics and electrodynamics the stress/Maxwell stress tensor and, if you cover crystal optics, the dielectric tensor).

For the fully covariant relativistic electrodynamics, which has great advantages concerning consistency and in some sense even implification from a more advanced point of view, you need tensor analysis for the pseudo-Euclidean affine Minkowski manifold. That's not too difficult to learn with some experience in the Euclidean vector calculus formalism.

I'm not so sure, whether it makes sense to use more advanced and modern techniques like the Cartan calculus of differential forms or the coordinate- and frame-independent formulations although they have some calculational advantages compared to the more traditional style using the Ricci calculus, which sometimes tends to be quite an "index battle". I'd think that's not so necessary for the introductory undergraduate topics than rather for advanced topics in General Relativity.

What also seems to be underrepresented in the traditional curriculum is (Lie-)group theory. If there is one unifying concept of all physics it's symmetry principles and thus the extended view on geometry (in the wider sense) introduced by Klein's Erlanger program somewhat earlier than it became the utmost important theoretical tool of modern physics in the 20th century. I think at least some glances on these topics already in the earlier theory lectures (e.g., Poisson brackets and their Lie-algebra structure in the analytica-mechanics lecture) can help greatly to understand the necessarily more abstract quantum mechanics and (relativistic) quantum field theory.

sysprog and Mr.Husky
I found electrodynamics to be the hardest topic in my undergraduate curriculum.
In my second course on undergraduate intermediate electrodynamics (time-dependent stuff) my professor gave us copies of Sadiku, and I hated it. Probably a bias against EE texts.
Griffiths is good enough for most topics, but if I remember correctly, the treatment of the Lieanrd-Wierchert potentials is utterly confusing. In that regard I was pleasantly surprised by Vanderlinde's book.

By the way, talking about older books (that I have not read), what about Panofsky and Phillips? seems quite interesting.

Mr.Husky and vanhees71
We are talking about a first book in electromagnetism. You know, what's taught to freshmen and sophomores. Not upper division, and certainly not graduate level.

Well, my first course ever on electromagnetism used Alonso & Finn and Feynman lectures. Nowadays I would complement them with "Student's Guide to Maxwell's Equations" by Fleisch.

But it is weird since Griffiths was mentioned. Do people really use and recommend Griffiths for a first book ever on electromagnetism?