Some thoughts on the Ventilator project.
Motors, I would avoid brushless model motors, they are built down to a price, not up to a spec. I doubt they could run continuously for long without wearing out. I wonder if Dyson is going to use a standard hoover motor in his ?, I would worry about reliability. UK Gov spec says run for 14 days plus, that's 336 hrs. The average hoover is used once a week for 20mins, and fails after 3 years, that's 52 hrs.
I would stick with standard, well engineered, Industrial grade blower fans. They last for a long time, some are still going after 10 years plus.
Shame we couldn't use a computer fan, they generally run for years without problems, modify a standard cooling fan ?.
Battery, I would choose lead acid, its a mature technology been around for 30+ years so no surprises. easily available in quantity if you stick to burglar alarm sizes. Li-ion potentially better, however charging requirements are complicated, and they still go off with a bang if not treated properly. Also sourcing good quality ones in the 1000's would be difficult.
Standard compressor - no they use oil to lubricate the pump, and put oil mist into the air stream. It tastes crap (! ) and would need a lot of filtering to clear it. However maybe those used to recharge aqualungs would do ?, could they run 24/7, or use a reservoir system and feed 10 people at once ?.
Reliability and notifying faults. Well start with a hardware watchdog system, a retriggerable monostable that has to be reset every sec or so else it causes a hardware reset and sounds the alarm, will ensure the processor doesn't fall asleep. Next every so often the program checks that all sensors are giving allowable values. I remember one 3D printer trying to heat the bed of an unattended printer, nearly caused a fire cause the sensor cable had broken. You could even use a 2nd processor to keep a check on things and if it dissagreed with the current machine state it could also sound the alarm.
I think the Medtronics documents are very useful. Not as a set of DIY build it yourself plans, everyone builds one in their garage and dashes down the hospital with it, but all the clever stuff, the thinking about how it has to work is all done for us. Basically al we got to do is engineer something to meet their spec, and we would be home free, no need to spend years verifying performance is what is needed.
Thoughts ?