An Intelligent Bag-squeezer for COVID-19 Open Source Ventilators and Why It Matters
Rob Giseburt, with the support of his firm Ten Mile Square, has published two long, informative, and extremely important blog posts (part 1 and part 2.) Public Invention has added “The Pressurizer” at the top of our spreadsheet of open-source ventilators and modules on the tab called “Drives”.
In a previous essay, Dr. Erich Schulz and I stated:
A dumb bag-squeezer simply won’t do the job we need doing in this pandemic.
The problem is not the bags; it is the lack of intelligence. You have to have sensors to support spontaneous breathing to make a ventilator that is logistically useful in supporting doctors treating COVID-19. What Mr. Giseburt has done is build a smart bag squeezer. Furthermore, it is completely transparent, free and open; anyone can build one right now. I have personally ordered the parts to do so.
The Pressurizer uses off-the-shelf parts used by 3D printers and CNC machines, and therefore unlikely to be supply-chain constrained in the pandemic. No need to 3D print or laser cuts parts for this!
But more importantly, it uses a standard, well-tested engineering PID control system to provide exquisite control of the pressure. His blogposts explain this very well; it is sophisticated but not new. The result is that the Pressurizer can maintain tight control over the pressure that it provides. At present, the Pressurizer does not sense flow (and thereby volume.) We are working with Rob to use VentMon, the Public Invention module for testing, monitoring, and alarming ventilators. The two put together can give you all of the control modes used in very top-of-the-line Bi-level Positive Airway Pressure (BPAP) machines and intensive care/critical care ventilators — if programmed correctly. This is what the world needs to fight the pandemic by giving doctors well-tested extra tools in their kit that they can confidently deploy.
This is important not because that team is going to make a ventilator — though they might. It is important because it is a freely-reusable smart module that anybody can build and control within their own machine. They haven’t solved the ventilator problem; they have solved the air production problem.
As Dr. Schulz has written in his Ventilator Design Brief (which is essential reading in this field), that all ventilators have the same basic components or modules represented by this schematic:
The Pressurizer is a module that fits the purposes of the mixing chamber, the pump or flow regulator, and the primary pressure sensor at this top of this diagram. It does so in an easily controllable way. If we, the open-source community, start to think in terms of modules, we can re-use the Pressurizer, and, because it has a clearly defined control interface in terms of pressure, we can even choose not to use it profitably.
For example, Blu3 has produced a different air drive mechanism, the Blu3Vent, based on their previously engineered Nemo diving product. In some ways, it is a better solution than a bag squeezer: it uses a diaphragm pump which is also highly controllable. However, they designed this pump themselves for their own purposes, which happen to match what is needed for a pandemic ventilator, but at present only they can manufacturer it. The Blu3 team has gone out of their way to produce a humanitarian solution, but it is not as open and accessible as what Rob Giseburt has done. Possibly a manufacturer could usefully use either by coordinating with either team.
There may be other drive mechanisms; if you know of one, please inform us to add to our spreadsheet. The point is not to grade one as better than the other, but to do what engineers do: define standards and protocols that allow us to build testable, verifiable machines that give doctors a chance to heal patients. In a global pandemic which has already disrupted supply chains, the Pressurizer is important not as a solution but as a reference implementation of a smart module at the heart of pandemic ventilators.
Public Invention is a 501(c)3 public charity.
