4 ms·

This is an interesting book. I've been in this space for years now and could easily see myself writing something like this, but I don't think it would look like

This is an interesting book. I've been in this space for years now and could easily see myself writing something like this, but I don't think it would look like this book. I really don't like to see things like any sort of recommendation for use of n-propyl bromide. That shit's neurotoxic. The people who can use it safely will already know about it, know someone who knows about it, or find it on their own. Anyone who finds out about nPB here should not be touching it. Unfortunately, many of the parts of the book that I've scanned are like that. There's a lot of prescription, not a lot of background/theory/underlying details, and no way to tell when the prescriptions are inapplicable or straight-up wrong. Which, often, they are: one of the hallmarks of deep experience is knowing when "the rules" are useful and when they're not, but what we have here is mostly rules. That will get your product out the door, I guess, but it's not going to level you up as an engineer, if that's what you're needing. Another example: stackups are discussed, but there's no mention of slash sheets, which is how you get things done cheaply and correctly. Specifying Rogers material for anything but the nastiest designs is just going to get your pockets drained and your Asian fab annoyed because they have to special order that. If you need it, sure, you need it... but can you get away with something more universal? And then there's things like this: "Most SMPS datasheets will advise you on what bead to use and where to put it". Hahahaha no they won't. And if they will, there's a decent chance they get it wrong. Ferrite beads are so useful and so much trouble that you can't trust an IC datasheet to get it right, even if they want to do the same things you want to do. Which they might not! If you're a more junior engineer trying to level up, give this one a look instead: Analog SEEKrets: https://www.eevblog.com/files/seekPDF.pdf https://www.eevblog.com/files/seekPDF.pdf

Joel_Mckay 2y ago
In general, people have to take a hazing at the certification lab before realizing something that "just works" is just not good enough. There are a lot of different processes for manufacturing electronics, different DFM strategies, and choosing the right path is often a trade secret. Usually made Jr staff read the NASA workmanship standards manual, cables and harnesses guide, and the tin whisker paper. Additionally, they would be expected to get their RF/ham technician license within the year, and practice coding test/boot-loader jigs in C/C++ in their assignments. There is also a tacit discussion about Metrology that lasts on average 3 months if you are smart... Book looks funny (AI slop?), as volume manufacturing is a different skill-set requiring designing to both a standard and factory capabilities. =3
- exmadscientist 2y ago
  > There is also a tacit discussion about Metrology that lasts on average 3 months if you are smart... Man, EEs usually have zero clue about metrology. I worked on a big piece of T&M gear for a while. The looks I got when I said "so, we need to discuss how we plan to calibrate this thing" were... let's call them impressive. I don't think any other person on that project knew what a "traceable calibration" actually was. It will probably not surprise you to learn that that project did not reach the finish line, at least not with my company.
  - Joel_Mckay 2y ago
    Projects can be impossible with some environments, budgets, and teams. Startup success rate is 1:22, and service companies survive 3+ years 6:1 against product companies... Thus, a hardware dependent launch has a 1:66 success rate over 3 years, and if people YOLO production it will go sideways for sure. Best of luck =3
bsder 2y ago
> I really don't like to see things like any sort of recommendation for use of n-propyl bromide. That shit's neurotoxic. I don't even like to see recommendations for IPA (isopropyl alcohol). In 95%+ of the cases ethanol (ethyl alcohol) is just as good and is way less toxic than IPA.
- exmadscientist 2y ago
  There are two problems with ethanol, one practical and one stupid. The stupid problem is that it's taxed heavily and thus is stupidly expensive. When not taxed it is probably denatured, but denatured with what? If you know the denaturant then you can use it with some degree of confidence. Methanol (methylated spirits) is acceptable in almost any technical application but is somewhat toxic, so for stupid reasons it is now rare. Denatonium benzoate is the other common denaturant, but it can leave annoying residues behind. Pick your poison (literally). At one point in my career we had ready access to untaxed, undenatured, 200-proof absolute ethanol. So I used it a lot! Turns out it's significantly more aggressive than isopropanol, and can strip coatings or generally even dissolve things that isopropanol cannot. I don't use ethanol any more, I have better things to do with my life than figure out what it is and isn't going to dissolve.
  - qiqitori 2y ago
    I use 99.9% ethanol for electronics cleaning (not very frequently) and it leaves a white residue on some PCBs (e.g. JLCPCB ones?), but not all. It also seems more aggressive if I accidentally get it on my skin. Doesn't smell as bad as IPA though! I haven't heard that IPA is much worse than ethanol though, beyond what's stated on Wikipedia ("somewhat more toxic"; ethanol is obviously toxic as well). Should I be more concerned about using IPA?
    - alias_neo 2y ago
      > Should I be more concerned about using IPA? Not really. The entire planet was rubbing it on their hands for years during the pandemic. It'll dry your skin out, sure, if you're getting a lot of it on your skin, but you shouldn't be if you're not using it as hand sanitiser, and of course you're not drinking it.
CamperBob2 2y ago
Analog SEEKrets: https://www.eevblog.com/files/seekPDF.pdf https://www.eevblog.com/files/seekPDF.pdf Neat, thanks for the link. Unusual combination of mathematical and practical notions in that one.
penguin_booze 2y ago
> This is advanced material, which assumes that you have fully understood earlier courses. If not, then go back to those earlier texts and find out what you didn’t get. Is this PDF part of a course series? Or the 'earlier texts' mean prerequisites gotten elsewhere?
- exmadscientist 2y ago
  My impression of it is that it's basically perfect to hand to fresh or about-to-be-fresh EE graduates. It fills in some of the massive practical gap between what's taught in college and what you actually need to know to be an effective engineer. So I interpret that as "this builds on your degree". Of course, it's much more applicable than that, but it is not trying to be a from-basics textbook. Which is good, because there are lots of those, and there are very few things that teach what this teaches.

Joel_Mckay 2y ago

In general, people have to take a hazing at the certification lab before realizing something that "just works" is just not good enough. There are a lot of different processes for manufacturing electronics, different DFM strategies, and choosing the right path is often a trade secret. Usually made Jr staff read the NASA workmanship standards manual, cables and harnesses guide, and the tin whisker paper. Additionally, they would be expected to get their RF/ham technician license within the year, and practice coding test/boot-loader jigs in C/C++ in their assignments. There is also a tacit discussion about Metrology that lasts on average 3 months if you are smart... Book looks funny (AI slop?), as volume manufacturing is a different skill-set requiring designing to both a standard and factory capabilities. =3

exmadscientist 2y ago
> There is also a tacit discussion about Metrology that lasts on average 3 months if you are smart... Man, EEs usually have zero clue about metrology. I worked on a big piece of T&M gear for a while. The looks I got when I said "so, we need to discuss how we plan to calibrate this thing" were... let's call them impressive. I don't think any other person on that project knew what a "traceable calibration" actually was. It will probably not surprise you to learn that that project did not reach the finish line, at least not with my company.
- Joel_Mckay 2y ago
  Projects can be impossible with some environments, budgets, and teams. Startup success rate is 1:22, and service companies survive 3+ years 6:1 against product companies... Thus, a hardware dependent launch has a 1:66 success rate over 3 years, and if people YOLO production it will go sideways for sure. Best of luck =3

bsder 2y ago

> I really don't like to see things like any sort of recommendation for use of n-propyl bromide. That shit's neurotoxic. I don't even like to see recommendations for IPA (isopropyl alcohol). In 95%+ of the cases ethanol (ethyl alcohol) is just as good and is way less toxic than IPA.

exmadscientist 2y ago
There are two problems with ethanol, one practical and one stupid. The stupid problem is that it's taxed heavily and thus is stupidly expensive. When not taxed it is probably denatured, but denatured with what? If you know the denaturant then you can use it with some degree of confidence. Methanol (methylated spirits) is acceptable in almost any technical application but is somewhat toxic, so for stupid reasons it is now rare. Denatonium benzoate is the other common denaturant, but it can leave annoying residues behind. Pick your poison (literally). At one point in my career we had ready access to untaxed, undenatured, 200-proof absolute ethanol. So I used it a lot! Turns out it's significantly more aggressive than isopropanol, and can strip coatings or generally even dissolve things that isopropanol cannot. I don't use ethanol any more, I have better things to do with my life than figure out what it is and isn't going to dissolve.
- qiqitori 2y ago
  I use 99.9% ethanol for electronics cleaning (not very frequently) and it leaves a white residue on some PCBs (e.g. JLCPCB ones?), but not all. It also seems more aggressive if I accidentally get it on my skin. Doesn't smell as bad as IPA though! I haven't heard that IPA is much worse than ethanol though, beyond what's stated on Wikipedia ("somewhat more toxic"; ethanol is obviously toxic as well). Should I be more concerned about using IPA?
  - alias_neo 2y ago
    > Should I be more concerned about using IPA? Not really. The entire planet was rubbing it on their hands for years during the pandemic. It'll dry your skin out, sure, if you're getting a lot of it on your skin, but you shouldn't be if you're not using it as hand sanitiser, and of course you're not drinking it.

CamperBob2 2y ago

Analog SEEKrets: https://www.eevblog.com/files/seekPDF.pdf https://www.eevblog.com/files/seekPDF.pdf Neat, thanks for the link. Unusual combination of mathematical and practical notions in that one.

penguin_booze 2y ago

> This is advanced material, which assumes that you have fully understood earlier courses. If not, then go back to those earlier texts and find out what you didn’t get. Is this PDF part of a course series? Or the 'earlier texts' mean prerequisites gotten elsewhere?

exmadscientist 2y ago
My impression of it is that it's basically perfect to hand to fresh or about-to-be-fresh EE graduates. It fills in some of the massive practical gap between what's taught in college and what you actually need to know to be an effective engineer. So I interpret that as "this builds on your degree". Of course, it's much more applicable than that, but it is not trying to be a from-basics textbook. Which is good, because there are lots of those, and there are very few things that teach what this teaches.