Where Can Physicists Begin Learning Advanced Engineering Principles?

[pierce15]

Hello,

I have a very strong background in math and physics (completed most of undergraduate degrees). I want to learn mechanical and electrical engineering principles and believe that I can start learning at an advanced level. Can you recommend some sources?

 

[jasonRF]

I am an electrical engineer and can say that the field is quite broad - I'm pretty sure mechanical is the same in that regard.

If there is a "core" to undergrad electrical engineering curriculum, it would be: analog and digital electronics, signals and system, electomagnetics, probability and random processes, and probably feedback-control systems (which I never took). Beyond this core you would have a whole host of specialties from information theory to quantum electronics to power systems to signal processing to semiconductor devices to control systems to computer engineering to electromagnetics (antennas, etc.) to ... I think you get the picture.

In my opinion, If there was one topic that could help a physicist think like an electrical engineer it would be signals and systems (assuming you have already taken electronics from the physics department). I have worked with a number of physicists we have hired at my company, and the lack of signals and systems (and probability theory - who knew one could get a physics degree without it?) is a noticeable gap. Standard books are Oppenheim and Willsky (1st edition is fine):
https://www.amazon.com/dp/0138097313/?tag=pfamazon01-20
and Lathi (1st edition is fine),
https://www.amazon.com/dp/0941413349/?tag=pfamazon01-20
The two books are fairly different but either would be fine - Oppenheim is probably the most widely used. MITs videos for the course the Oppenheim was written for can be found at
https://ocw.mit.edu/courses/electri...-science/6-003-signals-and-systems-fall-2011/
There you will also find homework assignments to work on. I would follow the MIT course, if I were you.

On the surface, signals and systems will look like a utilitarian course on continuous and discrete Fourier analysis, Laplace and z transforms, and simple ODEs; but beyond the math is the way it helps you think about time and frequency domains simultaneously, which is tremendously useful. Learning how to use block diagrams to represent both continuous and discrete time linear systems, including those with feedback, is crucial. This is not a tremendously hard topic, so you should be able to work through it at a reasonable pace.. This subject will have none of the rigor you are used to in theoretical math classes. But you should be used to this from your physics classes, where insisting on mathematical rigor would grind all progress to a halt in many cases. You can supplement (but not replace!) a typical book with ones that fill in the rigor, if you want. To fully do all of this rigorously requires real analysis (my math friends claim Lebesgue integration is really needed), distribution theory, and complex analysis.

Just my 2 cents!

Jason

 

[deskswirl]

I agree with Jason. I used Lahti's book to prepare my notes when I TA'ed Signals and Systems. The ideas you need from Signals and Systems is that every linear, time-invariant system can be represented by an impulse function (a.k.a. green's function) or by its frequency response. This very powerful method is used over and over again in engineering problems to represent circuits, devices and their (bio)mechanical, fluid (simple), thermal and acoustic equivalents.

The other predominate way of thinking which he did not mention but which is fully equivalent and useful (especially in control theory) is the state-space approach to linear systems which you may have seen in your differential equations course.

Both the time-frequency and state-space methods are often used in other fields where the linear system approximation can be valid over a restricted range . Personally I've used them to develop control models in epidemiology, parametric gain processes, and characterize a homemade erbium doped fiber amplifier among others.

The Oppenheim and Willsky book can be a bit difficult to get into so I'd also recommend you look at Continuous and Discrete Time Signals and Systems by Mandel and Asif which has really nice worked examples: https://www.amazon.com/dp/0521854555/?tag=pfamazon01-20

In terms of the other foundational EE topics you need electric circuits perhaps beyond what you may have seen in you introductory physics courses. For that I'd recommend Circuits, Devices, and Systems: A First Course in Electrical Engineering, 5th Edition by Smith and Dorf: https://www.amazon.com/dp/0471839442/?tag=pfamazon01-20

This is a long lived (for a very good reason) classic which includes solving linear circuits in time/phasor/s-domain, networks/frequency response, three phase power, electric machines, digital logic, microprocessors and a bit more depending on which edition you get. I would not say the book is superficial but it will show you a great portion of the general EE forest without focusing too much on anyone tree.

Really the reason I like this book is because it will teach you how EEs think.

 

[Aufbauwerk 2045]

I can't advise anyone else, but I can share my personal experience which is relevant to this topic.

I added a couple of years of electrical and computer engineering to my studies, worked on some projects, and got lots of practical experience with circuits, microprocessors, digital signal processing, and robotics. I also did lots of programming. Eventually I went back to physics, which suits me more.

I realized that I could not just skip over basic engineering material. Fortunately I had learned to not take my physics teachers seriously when they said "that's just engineering" and claimed it would be trivial for their brilliant minds to jump in and run circles around the engineers if they wanted to. They were arrogant.

The problem is that knowing Maxwell's equations and basic circuit theory does not mean you can just skip over learning the electrical engineering basics, which are not taught in physics classes. If you have not studied and used basic data structures and algorithms, you can't just jump in and be an advanced programming student. There's actually a lot to learn if you want to understand digital circuits and work on chip or computer design. The principles of mechanics are one thing, learning all the techniques to apply them to real world structures is something else. Of course we can learn these things, but it takes time and I think there are no shortcuts.

Of course if you want to learn signal processing, and you already know the math that's used in signal processing, that is a big advantage. Although again I think there is a difference between knowing the math, and having experience applying it to the type of problems they give in engineering courses.

Therefore I can't advise on any specific advanced material. But I hope my comments are helpful anyway.

 

[Logical Dog]

believe that I can start learning at an advanced level.

You can't IMO. You will have to start from the basics of electrical engineering like introductary circuit analysis etc

 

[Logical Dog]

the only difference is you would be able to pick up some concepts much easier and faster perhaps.