Ultimate Electronics (2021)
ultimateelectronicsbook.com
ultimateelectronicsbook.com
Spend $125 on Horowitz and Hill’s The Art of Electronics, 3rd Edition. It’s the last electronics text you’ll ever need.
I use AoE as a reference, but I’ve read full chapters from time to time when I’ve worked on relevant projects. I’ve never done an end to end read. Maybe I will when I retire lol.
I want to set aside time to read chapter 5 at some point. The precision electronics chapter is supposed to be bomb. Culminates in a design review of the autoranging circuit of one of the common handheld DMMs.
I shudder to think what will happen if youtube closes or radically changes operation. It will be even worse than burning the library of alexandria
You’re right: AoE is a rough sled as a brand new electronics learner.
I read Electronics for Inventors, and have even pushed myself through respectable chunks of AoE, but Platt’s book really helped to drive home some concepts and spans a mere 350 pages.
“Make: Electronics” by Charles Platt and "Practical Electronics for Inventors" are both good practical recommendations. What to prefer really depends on your background. I'd add "Electronic Devices" by Thomas L. Floyd if you prefer more academic/systematic approach.
1) At the level you can understand it, you don't need it. There are better, more in-depth, shorter books.
1) At the level you can't, it's incomprehensible. You'll run into a brick wall.
It's hard to recommend something without knowing background, but to do electronics well, you need to know basic linear algebra and differential equations, and understand (at least on a cursory level) the Laplace domain. You also need to be comfortable with poles, zeros, and rational functions.
That's a high bar. The most common mistake I see is trying to get further without being over that bar. Below that, you have popular books, many cartoonish, which explain what voltage and current are and the basic components. That's enough to do some basic Arduino projects and robotics, but not really get what's going on more deeply.
Above that bar, there are much more in-depth and theoretical approaches one is ready for the Art of Electronics.
Horowitz and Hill tries to compress that stuff down, toss in a bunch of oddball topics, and it doesn't really work.
Someone in the thread mentioned it was designed for physicists. It might be okay there. That's a kind of unique background where someone might understand the theory without the application. However, most of the time I've seen it used, it's a recipe for frustration.
brilliant.org has decent courses on linear algebra, including nodal analysis ($100/year or so). You don't need to know a lot; you do need to know how to express a linear equation as a matrix and solve with row reduction, and eventually, how to apply that to nodal analysis.
alcumus and AoPS textbooks are pretty adequate for rational functions, complex numbers, roots-of-unity, etc. You should know everything in their intermediate algebra and precalculus textbook (and ideally, their title textbooks: AoPS Volume 1 and 2). These aren't too long, are much more advanced than normal high school curricula, and are worth doing in either case, skipping the parts you know. Next step is something like MIT 6.302 OCW, which is brilliant.
For differential equations and Laplace, any differential equations textbook will be adequate. I would not do a deep dive here. A shallow one will more than suffice. If you kind-of-get-it, that's good enough.
If you lack calculus background for this dive, OSU MOOCulus or AoPS Calculus are good places to go. OSU will be easy, step-by-step, and AoPS will be challenging, so it depends on your background.
For most people, this takes 2-4 years, and you can't really cram it (which is both good and bad news; a little time each day is almost as good as learning full-time here; you can read about the spacing effect).
It's probably only good for a beginner with like 180 IQ. You can cover weeks of material in 5 pages, and it's well over 1000 pages total.
I'm an electrical engineer and I own a dozen or so textbooks for referencing. I don't even use AoE for referencing because it's so dense.
There are known caveats to the general idea (Nash equilibrium, Hotellings's law) but I'm just trying to point out that even despite the general correctness, there are people who are very heavily vertically integrated. And you would in fact expect those people to exist, because local minima are a fact of reality and you need people to understand larger parts in order to see past local phenomena.
'Ars Longa, Vita Brevis'[2] is a well-liked post that is in line with what you're describing. It's a bit funny, because the weak form argument (specialization is generally good) is totally obvious and there's basically no way to attack it. Specialization is the thing that has made society possible, and if anything it's more effective now than it was at the invention of simple tools. But the strong form requires some pretty specific assumptions.
[1] https://en.wikipedia.org/wiki/I,_Pencil [2] https://slatestarcodex.com/2017/11/09/ars-longa-vita-brevis/
That said, I bought “Learning the Art of Electronics: A Hands-On Lab Course” and really like it. There’s a good bit of practical, accessible knowledge there. Imho worth checking out for anyone who likes the idea of TAoE but finds it just a bit deep.
FWIW, my formal education is CS with a dose of digital electronics thrown in. For anything analog, I’ll try and find existing circuits close to what I need.
https://hackaday.com/2017/01/18/forrest-mims-radio-shack-and...
Example: http://www.zpag.net/Electroniques/Kit/Getting_Started_in_Ele...
Compulsion and nostalgia have kept me from tossing this one (and also its neighbor on the shelf): Commodore 64 Programmer's Reference Guide.
"The Art of Electronics", by Horowitz and Hill, is far more useful. It's not an easy book, but it teaches you how to actually design and build stuff. The original target audience was physics grad students who need to build some piece of hardware for their research. Their explanations of how to think about what transistors do are very useful. Horowitz and Hill get across some basic useful concepts, such as:
* Components are not ideal, but today, many of them, operated within their limits, are pretty close to ideal. So, if you're not doing something that requires very high speed or very high power or very low noise, you can often use off the shelf parts in a straightforward way, without getting too clever. Arduino is this concept taken to the max.
* Learn to read data sheets. That tells you what you can buy and what problems you can get pre-solved.
* Design and build well-behaved sub-assemblies and connect them up. Don't try to do the whole thing at once.
* Learn to design PC boards. It's not that hard, and the tools are quite good.
Sure, but the bigger insight in the first place would be that all measurements have error tolerance. Which is a slightly different topic, so I'm not sure that it's the last book I'll ever need.
In particular reading on switch mode powersupplies it sounds like inductors, resonant circuits and electro magnetic interference are black magic. And this field developed a lot since the first book.
1. https://web.archive.org/web/20230516064619/https://ultimatee...
2. https://web.archive.org/web/20200214154059/https://ultimatee...
Reading the comments here, it seems like the summary of feedback is simply "please write more sections". It's true that I haven't had much time to devote to this project lately due to other work commitments, my apologies.
Adding any new section involves three major intertwined components:
1. the conceptual explanatory text,
2. the equations and algebraically-solved examples, and
3. the schematics and interactive simulations.
Right now my focus is on building out all three (probably roughly equal in time demand) and I am wondering if that is a style that is working? Any other feedback on the content or style itself? Thanks -- your feedback is encouraging :)
Also - a way to download it as a epub and / or PDF would be excellent.
I guess there is probably an ideal middle ground where it can be presented as a “you can think of it kind of like this analogy, but then here’s the actual concept you should learn instead of learning the analogy” kind of way.
Which is to say not very well beyond a certain level of detail. But in neither case is there a more useful first order approximation to use.
My goal in the use of analogies is to give them a basic "if there is more of that, there must be less of this"-kind of intuition. This is a really, really useful skill that helps you in practise, because you will automatically do a double-check if you miscalculate or if the result doesn't fit your intuition (and thus sharpening it).
After understanding the basic systemic relationships one can go and have a more accurate look, but depending on who we are talking about and what they want to achieve that might not be needed.
Many electronic calculations could also be done on a deeper physical level, which would be more accurate, but also much more expensive to calculate. This is not practical for the typical electrical engineer, which is why they go for the simplified calculations.
Similarily there are amateur tinkerers which will be able to do most of what they want by relying mostly on analogies and that is okay as well. The math is there if you wanna go deeper as an amateur, just as physics is there if you wanna go even deeper as an electrical engineer.
That being said the analogy also serves a purpose for the electrical engineer. Everything is quicker if you intuitively know that voltage has to rise at point x and less current goes through point y etc. Sometimes you don't even need to know how much
If you have hydraulic explanations for these, please do share.
At this point you'll be halfway through typing "but electrons are real, not an analogy" - but what do you think of when you think of an "electron"? If it's a ball, or even a point, then that itself is an analogy which will break down in all sorts of annoying ways. Electrons are slippery little non-localized bastards which can tunnel through solid objects. And don't get me started on holes. People like to think of electrons as "real" and holes as "not real", simply because they're defined by their absence, but in semiconductor physics they are both equally real and important.
If you use electrons, you have to face all sorts of basic problems like "why does electricity travel much faster than electrons move?" and "Why is the sign wrong on all of my calculations?" and "Which direction is current actually flowing?" These confuse people with an intermediate understanding.
It's much better to start with the field as primary. People understand gravitational fields and magnetic fields reasonably well. From these you can build an understanding in which it's the field carrying all the energy and the electrons are just along for the ride. Each carrying its own tiny electric field.
* the "standard NPN" (2N5172) is mostly discontinued; a few places have them for ~$1 apiece
* the "standard FET" (2N4302) is very discontinued; SynthCube has a few NOS parts for $5(!) apiece
* the CA3080 OTA is so discontinued that the first reissue by Rochester has also been discontinued; Alfa has cloned it as the AS3030E but compatibility is not 100% guaranteed.
Some of these are overlapping - for instance, many of the FETs are used as buffers because the CA3080 has a weak output - but resolving them will practically be redesigning the circuit from scratch...