CPAP machine anyone?
My neighbor got a new CPAP (continuous positive airway pressure) machine and the supplier said to throw this one away. He asked me if I wanted to play around with the old one (he knows me pretty well). Of course I do. It's a ResMed unit and the PCB says Sydney, Australia.
So what it does is give you positive air pressure in your airway all night. It has a control unit which has an SD card. The SD card appeared to have daily files on it giving the prescriber the data on the performance. In the control unit is a blower. The left side connects to the right side which has a water reservoir for humidity control. Then the air comes out a long tube to your mask.
The main unit has a menu system for some simple user settings, but there is no way to control the air pressure in the menu. He said they can do it in the office if needed after a sleep study. There is an interface connector on the back but not one I recognize. I presume that is where they connect an external computer or proprietary unit to set it up.
Some of the parts used, other than the usual suspects such as voltage regulators and the like -
stm32f103zgt6
ARM® Cortex®-M3 STM32F1 Microcontroller IC 32-Bit 72MHz 1MB (1M x FLASH
Sensiron sdp711-01 differential pressure sensor
GE NPA-C02035 Surface-Mount Pressure Sensor
On the top side of the PCB are a whole lot of gold pads, perhaps for testing? In the water reservoir unit is a humidity sensor that sends it's data back through the interconnect between units.
Right now it blows a steady stream of air and doesn't require the humidity unit to run. Might be useful for a project someday.
Surely, that is essentially a ventilator? It would be my guess that those gold pads would for an automatic test machine to test the board after it was populated and soldered to see if it needed to go for reworking or not?
A number of projects for turning CPAP machines into ventilators are in various stages of development. A couple of challenges: CPAP machines are expensive as a starting point unless the machines are donated, which raises the second problem: how do you vet a donated machine to be sure it is a good starting base for a ventilator?
The alternatives, like the MIT bag-to-ventilator project seem to be better solutions: cheaper, easier to test, easier to use, etc. These are fascinating, and workable, alternatives that say a lot about the design, manufacturing, and marketing of medical devices.
But there's a more basic problem, as we have seen in the serological testing area. When the FDA makes the testing and approval cycle irrelevant grifters and frauds flood the market with shit that doesn't work. That's more than a "somebody got taken" monetary issue when the device or test in question may be a contributing cause to a serious illness or death rather than a preventative. That means, along with figuring out how to build inexpensive, easy to produce medical equipment, we have to figure out what sort of testing and certification is required so it doesn't kill a bunch of people.