Hacker News Reader

The one problem I see is "The Art of Electronics".

I Horowitz and Hill is a LOUSY textbook. And, personally, I find its usefulness as a cookbook overrated.

Going through the Forrest Mims notebooks/cookbooks/etc. for a solution to your problem is generally a way better idea than Horowitz and Hill. The Mims stuff does a really good job of pointing out the pitfalls you're likely to hit as well as the basic pedagogy.

Totally agree it’s a lousy textbook, but it’s not a cookbook either. It’s more like the Cliff’s Notes version of many many many application notes. One needs to be somewhat experienced to take advantage of it. For example, you might be decent at amplifiers but now you need to do a low noise measurement. Read the low noise chapter on AoA first. Now you know the lay of the land and you can understand the vocabulary of the area and navigate the actual datasheets and application notes.

Videos of Anant Agarwal's lectures that closely follow his "Foundations of Analog and Digital Electronic Circuits" book are online: https://ocw.mit.edu/courses/electrical-engineering-and-compu...

Yes, it is too broad. You should look at a school like Berkeley and look at their concentrations. Maybe RF isn’t important if you’re doing systems (unless you’re interested in RF).

https://m.box.com/shared_item/https%3A%2F%2Fberkeley.box.com...

I don't think you can call yourself an EE if you don't know anything about RF.

RF is not a required course in most universities. Some don't even have it as an elective. Were you thinking of electromagnetics? RF is a very specific subfield.

That's true maybe because it became such a big and deeply complex field that it has no space in the basic curriculum?

One might think that it's historically closely tied with the development of the radio and the telephone, see Tesla, Eddison, Bell Labs.

The common wisdom about electromagnetics with regard to EMI and such things seems to be that it's pretty much black magic. Similar for inductor selection in SMPS design.

It's pretty much physics, so a good deal about radiation can be found in medical physics for example, which is simply a different course of study with electronics as a side show, and maybe better as consecutive program for sophomores.

These are awesome recommendations - thanks!

Also for Digital Control Systems, check out this book by Tim Wescott

Applied control theory for embedded systems

It is less math intensive and more intuitive and aims at folks with a software background like yourself.

Why is it to broad to be accomplished? It seems to cover your typical BSEE, and you need an exposure to all of it.

Razavis RFIC is a good one too, but that’s really getting too specialized. Pozar is good for undergrad microwave.

I think it's a bit of a pessimistic take, to be sure, but I do agree that most people will not make it through this in a self-directed manner. Sure, some may, and I wish the OP the best and hope it all works out, but it's hard for most people to tackle much simpler, straightforward topics in a self-learning environment.

In this case, though, I think OP's approach to this shows that they're serious about keeping with it, which is really cool to see.

As someone who survied a BSEE, the range is too huge. I work with digital design as my day job, I only intuitively use EE101, digital logic and computer architecture and occasionally analog when dealing with post silicon issues.

The OP wants to study EE because he has a specific goal. My suggestion was that instead of trying to study everything EE focus only on those subjects that are relevant.

For example: If I was interested in robotics , I would not bother with digital,RF or Analog or eveen communication systems. I woud primarily focus on Control Systems and Embedded Systems.

> The OP wants to study EE because he has a specific goal. My suggestion was that instead of trying to study everything EE focus only on those subjects that are relevant.

This is something that I've seen often in self study plans for software development - the "learn everything and then try to use it" rather than "learn what you need to start solving the problem... and start solving it."

In software development this often takes the form of a self-learner learnings Java, JavaScript, Python, C, and C++. Once in an interview it becomes apparent that they don't know enough about any one language to solve a problem in that language.

This is where a university class (and degree) have an advantage - they've got a set of problems for the student to solve (homework and labs) and then take the student through learning specific knowledge to solve those problems.

This also shows what self teaching often lacks - those small problems that can be accomplished as part of learning how to solve the big problems.

OP has a formal background in applied mathematics and is an experienced software engineer, I don't think he will have any problems with generalization.

I did a basic project and I don't see how you can do robotics without analog systems unless you don't intend to build custom actuators and just buy very expensive off the shelf parts. The digital portion is absolutely trivial.

Analog systems and Analog Integrated Circuits, Analog Systems deals with using Analog chips, analog integrated circuits chips deals with designing such chips, a robotics engineers needs the first one not the latter. And the first one requires EE101 knowledge and signals and systems knowledge for filters etc which the OP is already covering in other courses

> It seems to cover your typical BSEE, and you need an exposure to all of it.

Hard disagree. Much of the page involves what normally would be electives. You need exposure to some subset, but not all of it.

To give you an idea, my undergrad in EE did not require a course on materials (although it was an elective).

Everything in "Phase 2" was an elective - none was required (although many universities do require the "electronic devices" course).

Needless to say, if everything in Phase 2 was an elective, so was everything in Phase 3.

Also, when I look at pretty much any job requiring EE, and intersect it with the courses I took as an EE undergrad, I find that most courses are not needed. EE (and an EE curriculum) is often quite broad. For any given course, there are plenty of jobs that will need that course, but most EE jobs will not. If the submitter has some specific goal in mind, he won't lose much by skipping courses not related to that goal.

To give you an idea, when I worked as an EE, I had to use basic circuit theory 2 or 3 times, digital logic only once, and the physics of electronic devices a lot. The level of EM I needed was satisfied by high school physics, so I won't even count EM. Everything else I took: Control theory, communications, electronic circuits, power/machines: Never used it.