Low cost, locally sourceable, ventilator for emergency cases - designed during COVID-19

This project is an attempt at designing open-source easy to scale solutions to attend to ventilator shortages around the world. My inspiration came from an article in the economic times about ventilator shortages in India during COVID-19. 

Why is this needed?

India is predicted to have a total of 30,000 ventilators, many of which are already in use. If the pandemic grew, whilst other existing diseases and conditions continue to last, we would soon run out of ventilators. Many Indian health-tech companies rely on international companies to design their machines, some of which are produced locally. 

Since the design process is not locally carried out, ability to scale these systems relies on imported parts, which are hard to get with aviation bans in place  (https://economictimes.indiatimes.com/industry/healthcare/biotech/healthcare/aviation-ban-may-cause-shortage-of-ventilators-in-india/articleshow/74726896.cms?from=mdr). Hence locally made SMART ventilators are the need of the hour. 

Here is my solution for one such ventilator. Additional modules for air temperature control, air humidity control and data analytics with ML for performance enhancement can be added. Happy to help you engineer those modules if you are pursuing this project.  

Updated learnings after working alongside Healthcare professionals in San Francisco and Mumbai

1. The ventilator needs to be non-invasive 

2. The ventilator needs to be closed loop to prevent microbes from spreading in the room and infecting other patients. 

3. The MHRA specifications highlighted below are necessary to comply with for a COVID-19 ventilator to be ready for use.

Steps to solve this problem

I've attached a minimum specifications requirement published by the MHRA below - please use that as a base reference while working alongside the appropriate health professionals to make your ventilators. 

1. Identify performance metrics required for ventilators to be considered use-able by industry - air velocity, air pressure, alternating frequency, pipe diameters, pipe length, etc. 

2. Identify the appropriate modes necessary for patients with COVID-19. 

3. Identify required safety metrics for device to be considered reliable and safe to use. 

4. Identify local supply chain in individual countries to support the scale-able production and execution of this idea.

5. Work alongside doctors to validate the functioning of such a device - when approved, conduct patient trials on permission, under medical supervision. 

Disclaimer and notes

- DO NOT implement this system without design inputs from appropriate medical professionals. 

- DO NOT implement this system without the permission and involvement of appropriate medical professionals. 

- ENSURE hygiene and consistency of performance before providing support to a patient.

- This design has not been tested aggressively on patients and is my DIY thought approach to building low cost ventilators for situations of emergency.

- I do not take responsibility for any mishaps/casualties that may occur due to change in design/inappropriate installation.

- I recommend using low level C register by register programming approach as opposed to using off-the-shelf prototyping platforms like Arduino, Raspberry Pi, mBed, etc. for driving the logic in this machine. The mentioned prototyping platforms use bootloaders that are notoriously known to stall causing the system to break. This can be fatal in the application of a ventilator. Writing a low level execution program helps reduce bloating in your code ensuring clean, lean, execution on a low power microcontroller.

- For people attempting to 3D print parts, especially valves and internals that will interface with the pipes, it's imperative that one ensures all broken, warped, unfinished, bulged layers are eliminated and cleaned. ABS curing using acetone may cause constant release of VOCs during device operation, and can be toxic, hence must be eliminated. Methods alternate to FDM process are recommended. Metal fabrication or injection molding would be optimal for such a project. 


Device Specifications (Defined by the Government of Maharashtra, India)


Adult and Pediatric

• Volume controlled (VCV), assist/control 

• Pressure controlled (PCV), assist/control 

• Pressure support (PSV) 

• Continuous positive airway pressure (CPAP) 

Combined Ventilation modes



• APRV, BIPAP mode + PSV 

• Non invasive ventilation (NIV) + PSV 

• Spontaneous Ventilation (SPONT / CPAP) + PSV 

Parameter Selection

• Tidal Volume: 2ml to 2500ml 

• Inspiratory Time: 0.2 to 9 seconds 

• I:E Ratio: 6:1 to 1:199 

• Respiratory Rate: 1 to 120 rpm 

• FIO2: 21% - 100% 

• O2 100%: Starts oxygenation sequence for inhalation 

• Inspiratory Sensitivity: Triggering by flow: 0.5 to 15 L/Min. 

• Triggering by Pressure: 0.5 to 10cm H2O below peep 

• Expiratory Sensitivity for PSV: 5% to 80% of the initial flow with passages of 5% 

• PEEP/CPAP: 0 to 50cm H2O 

• Pressure Control Ventilation (PCV): 5 to 90 cm H2O 

• Pressure Support Ventilation (PSV): 0 to 90 cm of H2O 

• Inspiratory Pause (Programmable in VCV): 0 to 4 sec. 

• Manual Inspiration: One Inspiration


They have luminous and audible signals and have messages on display 

• High Inspiratory Pressure 

• Low Inspiratory Pressure 

• Oxygen and air low pressure or lack of pressure 

• Low pressure of one of the gases (Oxygen or air) 

• Lack of main electric power 

• Battery low 

• Battery exhausted 

• High Continuous Airway pressure 

• Technical Failure 

• Mask disconnection during NIV 

• Oxygen not adequate 

• High Tidal Volume 

• Low Tidal Volume 

• High FiO2 Percentage 

• Low FiO2 Percentage 

• Apnea 

• Leak in NIV 

• Circuit disconnected 

• Low PEEP 

• High Respiratory Rate 

• High Exhaled Minute Volume 

• Low Exhaled Minute Volume 

Other Features and Controls

• Trends 

• Regulation of Alarm Sound Volume 

• Inbuilt Nebulizer 

• Manual Inspiration 

• Inspiratory / Expiratory Pause (Instantaneous) 

• Inhaled Oxygen Sensor 

My design below is a basic, general overview on how this system can work - each feature mentioned above must be tested comprehensively before allowing its use in true medical situations.


I've written some pseudo code/english to help write the logic and feedback loop for this system. Please adapt this to suit your microcontroller and IDE - happy to help with projects using C/C++. 

There are additional data collection, IoT, and ML modules that can be added to these low cost systems to make them more useful even beyond simply helping the patient in time of need. This would help the systems learn and enhance their performance based on patient type,and condition. 

I am uploading this work on the 21st of March, 2020 in the spirit of encouraging people to solve this problem. This work is open source for learning and prototyping purposes, not for commercial use without appropriate permissions and conversations.