100 years in a 100% controlled environment is easy.

Meaning: you are missing a lot of variables in your game that will mess with your plans.

So spend some times and do requirements engineering on an abstract level. Then select the techniques to solve them.

OK, say constant operation at room temperature indoors.

No matter how well temperature and moisture controlled your environment is, I highly doubt that 100 years is easy. 100 years? 876,000 continued hours of operation? Large companies regularly have expensive recalls because of premature failures for all kinds of reasons. Components fail for myriad reasons and I think to get 100 years you’ll need careful component selection, de-rating, plating, thoughtful solder selection, conformal coating, etc.

>OK, say constant operation at room temperature indoors.

Suddenly a wild humidity change appeared.

You must spend a lot of time defining your environmental parameters.

Yes, obviously humidity changes with temperature. I don’t see how that’s important. 99% of dead boards out there are not dead because of indoor, room temperature level humidity fluctuations, and the plenty of PCBs still working from the 70s don’t seem much affected by them either. There’s thermal cycling concerns, but that’s dominated by the board’s own heat generation. Do you have some data that would indicate that modern boards are sensitive to this?

API for central management of APs. And a server instance to do this.

Absolutely.

That's one thing that ubiquiti and the like do very well.

Combine that with the openwrt wiki/databases and you can support a LOT of different hardware with different capabilities relatively easily.

A smart move from China would then to sanction delivery of all R/L/C/D components to US that cost less than a cent.

The US tech industry would starve within a week by running out of components.

Every PCB is having hundreds if not thousands of them. And no industry in the west is able to build then in the required amount and quality.

> Or when they built their own GPS system and it ended up being far more accurate than any other system in use?

It actually is. But yes - that project was a shitshow for a long time.

Galileo HAS now offers 30cm accuracy with less than 100s convergence time not needing additional correction servers/stations.

Also spoofing resistant thanks to cryptographic signing (Galileo Open Service Navigation Message Authentication).

Both free for use. Forever. Classic GPS doesn't offer this.

My impression is that the US system does, but the higher accuracy is still reserved for military use?

The magic of the military GPS codes are AFAIR the extremely broad signals and use quasi non-repetitive codes. It's very difficult to jam and spoof.

Galileo HAS service is free for anyone to use.

Didn't Israeli students show that you can recover audio from the vibrations of bulb filament with a fast photo diode?

I'd test that with a CCD line sensor plus a wide aperture lens and reading it out with 8kHz. Then you have 128 audio pixels that can cover an entire city.

Line of sight might be an issue there. I'm thinking more high-end clandestine eavesdropping. Fun fact: curtains are a pretty good defeat for laser microphones, but if the building is really old and made of solid stone, you can point at the rock instead!

The rock?! That’s incredible. I would have guessed it was too dense to pick up normal speaking volume. Then again, even the window glass vibration seems pretty magical to me.

Beamforming is standard in modern conference room gear. It's being used for making a video focus on the active speaker and optimizing his audio.

Have a look at the "Meeting Owl" for example.

It works great up to a limit (around 5m) then you will need additional microphones closer to the speaker.

XMOS targets this space specifically.

https://www.xmos.com/ a descendant of the Transputer.

If somebody wants to play around with Zynq 7010's - have a look at the EBAZ4205 board. They can be bought from Aliexpress (20-30€). These are former Bitcoin Mining controllers.

Some people reverse engineered the entire thing. It can be found in GitHub. And there's an adapter plate available for getting to the GPIOs.

For a less complex entry there are also Chinese FPGAs ("Sipeed" boards which use a GoWin FPGA. They are quite capable and the IDE is free.

Xilinx tool chain is also no-cost.

For some/smaller parts. Once you start going higher than Artix or the token Kintex parts you need to pay up.

A similar technique is very popular in industrial automation to spot leaks in compressed air pipes and their connections from far away. These leaks are extremely loud in the ultrasonic range. It's overlayed with a camera picture.

That's ultra expensive gear.

I’ll guess «ultra expensive» partly due to very low volumes made for a customer that can afford a lot.

For servers there are passive cooled n100 variants on Aliexpress with multiple Intel i226 2.5GE network interfaces. We benchmarked them - they are able to serve line rate on all interfaces with a few routing/firewall routes. We bunkered a few of them in case we have a serve emergency situation - this allows us to setup infrastructure in zero time.

The next bigger variant (n305) is getting too hot - but at that point I'd start looking into variants which use an AMD notebook CPU.