Millions of Ventilators

Worst case projections place the quantity of ventilators required at around 1 million in the USA.  World-wide the situation is generally much worse.  It is likely that the potential number of ventilators needed in countries with less developed healthcare systems will be in the many millions, and the supply is in the low thousands.

Existing designs rely on specialty parts with little existing stock and long lead times, and are mostly limited to being made in quantities of a few thousand over the next few months. If the curve doesn’t flatten and stay flat wordwide – we won’t have months.

Most community and open source designs of electro-mechanical ventilators have similar supply problems to the commercial designs; they require specific parts which are hard to source in +10k quantities.  This has led us to look at very old designs, which are purely pneumatic (air-powered), and also extremely simple to manufacture (can be made from raw materials) and operate (no moving parts; no settings)

We introduce the ARMEE Ventilator – it provides the most difficult piece of ventilation, can be built on the scale of hundreds of thousands per week, starting immediately, in a single US State alone.

The original device it’s based on, created by the US Army in 1965, has been tested on animals and humans successfully. We have made small modifications to make it suitable for COVID-19 care (adding positive PEEP pressure), but otherwise it is good to go.

We are in the final testing and iterating phase. When it’s ready for full release, we expect it to make big waves. We hope we’ll have your assistance spreading the word.

I want to help!

I am a doctor or in charge of a hospital and I would like to help you validating Project Armeevent. Send us your contact info at:

I am an engineer/maker and want to get started right away, these are the links to designs and prints (you need to set up your Test Lung and Sensor Environment too!)

The 1965 Army Ventilator

Our design is sourced from the design by Harry Diamond in a 1965 US Army research patent which uses two plastic plates bonded together and 3 calibration screws to set PIP, PEEP, and Inhale:Exhale ratio. Once calibrated the device is static and can only be controlled by changing the incoming pressure. It allows the patient Assisted Controlled breathing (patient can instigate an inhale via suction), and otherwise controls switching via Pressure (will change from inhale to exhale when PIP is reached, and vice-versa when PEEP is).

No moving parts: the main design runs entirely off basic properties of fluid dynamics and has no moving parts to wear down or deteriorate. Once calibrated, there is only one control which can be tampered with, and otherwise the device runs autonomously.

No electronics/motors needed: this device – and its planned accessories making up the entire ventilator system – is designed to be run entirely off a pressurized air intake.  No electronics/motors needed.

Quickly-Manufacturable: The speed of manufacturing comes down to a lot of factors, including lead-in (setup) time, but we estimate that a medium-size injection-molding facility could produce devices at a rate of 10-100ka day, with millions of units possible in a week or two (note though: lead-in time could delay this by a week or more).  Milling/machining could also produce similar numbers (a factor of 2-3 times slower perhaps), without a significant lead-in time

What does it do?

This device is just a pressure-controlled switch which directs a high-flow stream of air down one of two channels with a particular timing and min/max pressures.  To illustrate, let’s say the pressure stream starts by going to the left in the below diagram, pointing towards the patient (or test lung). Air flows down that channel, building up pressure in the lungs. As it does so, that pressure propagates back to this device and through the left control channel, flowing through it and pushing at the power nozzle. Once enough flow goes through, it’s enough to push the main jet of pressurized air and direct it from the left side (patient side) to the right side (exhalation port).  We can control the pressure this happens at by narrowing the channel width – setting the Positive Inspiratory Pressure (or “PIP”). That’s all we need to switch from an inhale to exhale.

Now that the jet is exiting the right channel, the lungs can naturally deflate – and they also get a bit of assistance from entrainment (suction) from all the air rushing out the exhalation port. This continues until the lung pressure is low enough – and thus the pressure on the left control channel is low enough – that there’s a flow from the right to the left across the power nozzle strong enough to push the jet stream back from the right channel to the left – and back to the patients lungs in an inhale. We can control when this exhale ends by narrowing the right channel width via another calibration screw – giving us all we need to make full cycle of inhalation and exhalation at well-defined pressures.

We can change the relative timing of inhalation/exhalation and set the Respiratory Rate by adding resistance (via another screw) on the exhalation port, but that’s all there is to the original 1965 US Army design.

We have subsequently explored many improvements on this design, aiming first and foremost for Positive-End Expiratory Pressure (“PEEP”), as this original design uses a negative pressure. We have found geometries and settings which give comfortable ranges necessary for COVID-19 treatments (PIP 25 cmH2O, PEEP 10 cmH2O, RR 10-20 cpm, TV 350-450 ml) with a stable cycling rate for models of the average patient, and have produced prototypes proving the design works. We are currently working on finalizing our release model which will narrow down the properties into the most usable design we can, which we will then be releasing for testing, regulatory approval, and manufacturing. We ask anyone interested in taking part in this process to reach out to us via the links on this page.

Our final offering will likely be a set of different solid-state single-use disposable models, each designed to treat one particular type of patient (lung volume, compliance, resistance), with clear estimations of tolerance ranges, so medical staff can make informed decisions before prescribing.


The device works. It provides Pressure Control ventilation with Assisted Control, capable of treating the average patient independently. It provides the basic functions of ventilation in an elegant way with stable PIP and positive PEEP – just what’s needed for COVID-19.

A Ventilator that takes 5 minutes to manufacture. Sam Passenger do it in a minute 25

Posted by Christopher Jung on Monday, April 20, 2020

Hundreds of these devices could be made per day by one CNC milling or injection-molding machine, with hundreds of thousands of total production possible in a single US State in a week. These machines are found commonly worldwide.