Ask HN: What does an electrical engineer actually do at work?

When I said "a lot more standardization", I had things like DRAM or PCIe in mind, which are spec'd by their relevant consortiums (JEDEC and PCI-SIG, respectively). As in, there's clear limits on the way to make it work, and if you meet those limits, it's fine.

I do a lot of POL planning for purposes of thermal management, so yeah. I'm definitely not that specialized.

I used a YouTube playlist[0] to get the basics down, goes from installing KiCad through to soldering a finished board.

[0]https://www.youtube.com/playlist?list=PLy2022BX6Eso532xqrUxD...

I am a board level design engineer and the work is similar in many ways to software. It is very numerical and requires attention to detail.

I spend 25% of the time talking about requirements to other engineers, 50% comparing components to use with datasheets, 12% in Altium Designer and 12% buying components and PCBs.

Huge pain #1: the unbelievably slow process of manufacturing PCBs. Imagine that you were back in the old days of computing where you had to use punch cards with machine code and you had to give your stack of punch cards (your "program”) to the punch card operator. He would run it overnight and you would get your results the next day. If your program failed, then you have to meticulously comb through it and debug in your head. This is modern-day PCBs. Holes in a board that take forever to make. Then you have to pay $80 shipping to get them next day or you can pay your engineers to sit around doing nothing. And the PCB might not even work. Learning feedback loop is very slow.

If you want to do anything remotely interesting like via-in-pad or four-layer boards, first you have to wait at least 24 hours for a custom quote. Half of PCB vendors don’t just give you the formula to make your own quote. Then you have to pay either $1000 and 2 days or $200 and 2 weeks to have 5 copies of your design.

Huge pain #2: reinventing the wheel. When I open a datasheet, I have to read about the device. The pins, the maximum ratings, the application note “gotchas” like “don’t leave this pin floating or else the chip will be unpredictable!” Then I have to make the schematic symbol and footprint by hand. That means manually entering IPC package dimensions into Altium like a braindead zombie. Every package is slightly different. I cannot tell you how many Texas Instruments DC-DC converters I have hooked up. I have no idea why device manufacturers don't just hire someone full-time to make open-source 2D and 3D footprints for the top six CAD tools. SnapEDA and Altium Vault are trying to do this, but the footprints are flawed and they are outright missing a lot of parts. I cannot tolerate mistakes when each board costs hundreds of dollars. The device manufacturers already make footprints to test their parts. Why don’t they share them??

Huge pain #3: High barrier to entry. Very expensive software. In software engineering, professional tools like git, Visual Studio, Eclipse are all free. You can pull code at home from Git and start contributing immediately. The only barrier to entry is the time you need to understand the existing codebase. Even in firmware you can download Code Composer Studio or PSoC Creator for free.

In board design, you need to pay $300 for EAGLE or $5000+500/year for Altium if you’re serious. Sometimes OrCAD goes on sale for $300. Lets say you want to simulate Bitcoin mining ICs frying themselves in their own heat. Or maybe you want to know the radiation pattern of your antenna. You can pay another four digit price tag for simulation software like CST or just copy a previous design like a zombie. Upverter is trying to solve the upfront cost problem with their $125/month SaaS subscription pricing, but I tried their editor 3 months ago and it was 15 FPS with the example board. Not cool. KiCAD is an open source alternative to Altium but as far as I know it is nowhere near comparable.

Huge pain #4: ordering components and PCBs. In my last project I had to order components from China. Ordering from China is not very easy with the language barrier, bad spec sheets, time difference. Alibaba is the place to go for ordering from China, but all the prices are “Contact us”, which means you have to give all your information blah blah blah until you get an email with the price and then pay with wire transfer. Sometimes you get lucky and you can find what you want on AliExpress and pay with credit card.

But the tradeoff to all of this is that if I do it correctly, I can hold something in my hand and give the software engineers a new API to play with. The APIs all stem from the hardware. The work is often more fundamental with equations and physics rather than purposeless corner cases I had to consider when I was in programming. And hardware often has the chance to be featured on the box of a product rather than software which is all just assumed to work. It feels more meaningful.

Lead in the fumes is still solid particulate matter, but that's sill not something you want in you. It's not likely to be a major issue, but it's not completely safe either.

I do hardware engineering because I started my career when it and heavy math were required to get into software courses (and you first computer came as a bag of parts) - the camera itself is spherical build, wedging 26 cameras, 4 beagleboneblacks, a netswitch,IMUs,GPS, and batteries for the souls sucking multiamp drain into a 5" ball (OK, I gave up on the batteries for the moment :-) My primary "actual engineering" has been the control distribution circuits across 8 paired I2C/SPI busses at 4-8MHz and trying not to corrupt the bits in the process (a little bit of whining on Wolfram Alpha with the theoreticals, a lot of being very neat and tidy with the layout - the power distro itself with its various internal taps because power is analog and analog is proof that God hates us, the various passives and actives involved in everybody actually powering up in some more or less controlled fashion, etc - technically for the camera I use the Arducam FPGA against OV2640's and OV5640s, cause that gives me a nice (slower) pure I2C/SPI interface - the real goal is to feed the tensorflow stuff in the cloud, the camera geometry, imu, and gps provide enough correlations for the system to lock onto pretty easily, i.e. stationary spherical image reconstruction and spatial analysis is a no brainer (relatively speaking :-) ), and the back end system is now getting more betterer at using the IMU data stream as a primary assist when the camera is in motion (oh, and then there was the pesky little issue of delivering the GPS clock to 4 linux based computers to establish a common timebase cause we all know that linux is an awesome RTOS)