This is the rapidly-evolving VentMon by Public Invention. This is for version T0.1 -- "Tester v. 0.1", which is aimed at engineering teams needing to test rapidly manufactured pandemic ventilators.
The VentMon is a physical device designed to help the open-source community develop confidence in rapidly manufactured ventilators systems. At present it serves a tester to be used on the workbench of teams developing ventilators. Eventually, we hope it will serve the important function of providing a monitoring/alarming system as a module within other designs, or as part of an emergent open design for ventilations systems.
The VentMon plugs into an airway and measures flow and absolute and differential pressure within the airway approximately 25 times a second. It measures oxygen, humidity, and temperature within the airway and in the ambient air approximately once a minute.
The basic test function is to use the VentMon to log a data file over a period of time. This log file may be viewed live (with imperceptible delay) via our software tools, allowing an engineer to immediately observe pressures and flows. This same software computes clinically important parameters such as tidal volume, minute volume, and respiration rate, and I:E ratio. If produced over a long period of time, this would be evidence that the ventilator is performing as expected.
Due to a world-wide supply-chain problem, we have had to use a neo-natal flow sensor instead of an adult flow sensor. This limits the instantaneous measurement of flow to 33 standard liters per minute. In many circumstances, this will not matter; but in some test scenarios with highly compliant lungs and low resistance, the flow sensor will "clip".
In our software, this will be visually apparent as the flow curve simply looks like it has been clipped off. We also report this in the PIRDS standard that we publish. However, this error does not currently flow all the they way through our software.
Additionally, the flow sensor in use adds restriction to the flow. All physical lung systems have both resistance and compliance. The fact that the VentMon T0.1 adds restriction due to the use of the SFM3400 may have to be taken account of in your testing.
The VentMon has up to 4 physical connections:
- The air input, which is a female 22mm inner diameter airway tube. This is an international standard. The input to the VentMon comes from the ventilator or air-drive you are testing.
- The air output, which is a standard male 22mm outer diameter port. This should be connected to the test lung/patient.
- A USB connection to provide 5V power to the VentMon. Additionally, the VentMon streams PIRDS JSON data on the serial port, which is useful for testing and checks on verification. It requires additional software, however, to produce an understandable picture of ventilator performance.
- An ethernet connection which is used by the VentMon to stream PIRDS data via UDP.
You are welcome to change the firmware, but our basic firmware streams data to our
public data lake on port
6111and ventmon.coslabs.com.
Your VentMon comes with a small OLED "badge" display for us to support with future firmware improvements. At present, it will display a changing (live) display of the differential pressure. This is useful mainly as an indication that the system is functioning properly. In the future, we hope to evolve this display into a clinically useful display.
To familiarize yourself with the VentMon and "smoke test" basic operation, we recommend the following steps:
- Before connecting to a ventilator or a test lung, simply connect the VentMon to an Arduino IDE and start the Serial Monitor. Do not change the firmware by uploading new code unless you intend to do so. The current firmward outputs 115200 baud; set your Arduino IDE to listen at that baud rate.
- Observe that the VentMon is streaming a series of JSON objects conformant to the PIRDS data standard.
If you cannot observe PIRDS data streaming, please contact us for support.
- Now check that flow is measured by simpling blowing into the ventmon airway tube or using a sanitary dry source of air such as a small fan.
- The PIRDS type code for flow is "F" and the flow sensor the sensor of type "F" named "A" and numbered "0". A PIRDS flow event looks like:
{ "event" : "M", "type" : "F", "ms" : 12103, "loc" : "A", "num" : 0, "val" : 0 }.
- The units of flow according to the PIRDS standard are standard liters per minute times 1000 (or milliliters pers minute.) You should be able to create flow of a few milliliters per minute just by waving your hand quickly at the VentMon input; a fan will produce a much higher flow.
- Observe that you can create a measurable flow.
If you cannot observer a measurable flow, please contact us for support.
Finally, observe that the pressure differential is approximately correct.
- The VentMon measures pressure in the airway in absolute terms, but more useful is the differential pressure, which looks like this:
{ "event" : "M", "type" : "D", "ms" : 12037, "loc" : "A", "num" : 0, "val" : -11 }
The VentMon uses two independent BME680 sensors to compute the pressure differential. The ambient pressure is kept as a running average over a period time. The instantaneous pressure in the airway is published as an event about 25 times a second, as is the difference with the ambient air.
- Observe that the pressure in the Airway is of type "P" and numbered "A" and "0". The amBient pressure is "B" and "0". Following medical practice, these absolute pressures are measured in cm H2O. The VentMon PIRDS standard defines the units reported for pressure to be hundredths of cm of H2O.
- Therefore, at close to sea level, the "val" field for these absolute pressures should be close to 10000. The amBient pressure is reported very rarely compared to the Airway pressure; it may be challenging to find the "B" field in the serial output.
- Differential pressure (when not in an active breathing circuit) should be at most tens (less than one cm H2O). This represents miscalibration in our two sensors; we do not believe this inaccuracy is clinically significant as long as it is small.
If the pressure sensors do not seem to be working, contact us for support.
Although you are free to make any changes you wish to the VentMon firmware and to process its output via any means you want, we have created what we hope is a convenient means of reviewing the live peformance of a ventilator and test lung combination.
The simplest way to use the VentMon is to use our public data lake, which logs data for you on a server hosted by Public Invention and its volunteers (Geoff Mulligan.) This produces a live, real-time display modeled on what clinicians like to see. This approach currently offers no privacy; anyone can look at any of the data in the data lake. If you don't want your data to be visible, don't attach an ethernet cable to the VentMon.
The data lake records data indefinitely in a log file named after your IP address. This, and any other data logs, can be accessed by going to the VentMon Data Lake. From there, you can see your raw PIRDS data. By clicking on one of the links such as that of Robert L. Read's personal Ventmon, you can see live plots of differential pressure, flow, volume, and other parameters that are clinically important. For each parameter, you can set high and low "alarm limits". These will turn red when ventilator behavior is outside of this specified range, and will be blue while you are inside it. (You will want to discover you own IP address, which may change occasionally depending on your internet service provider, to find the Breath Plot of your own device.)
Breath Plot has a toggle switch that let's you switch live data on and off. When switched off, you will be examining whatever is on the screen until you turn the live data back on, at which point it will show you about the last 10 or 20 seconds of data live.
The BreathPlot web page can, however, can be used with any system that produces data in the Public Invention Respiratory Data Standard. You can create and import an array of JSON objects in the PIRDS standard and display it.
- Make sure the VentMon is powered up via USB.
- Connect a CAT-5 ethernet cable from the port to directly to your router (NOT your computer.)
- Attempt to make sure you router allows outgoing IP packets to ventmon.coslabs.com on port
6111. You may or may not have to manipulate your router firewall configuration. - Determine your IP address as it appears to the outside world.
- Using any modern browser, browse to http://ventmon.coslabs.com/.
- Hopefully there you will see your IP address in a list that looks like this:
- Identify your IP address in the list, and click on the Breath Plot link.
- Hopefully you will see a live breath plot, changing constantly based on the actual flow
within the VentMon. If you ventilator is rhythmically forcing air through the VentMon, it
should look something like this:
- Observe the features on the right may or may not be red if they are within default limits. These features change slowly as the samples change. You will see about 12 seconds of breath data by default.
If you wish to see the software in use without having a live VentMon, you can paste a sample PIRDS data file into the text area and hit the "Import Trace" data file. These files were created by the "Export Trace" function.
The Import and Export Trace capabilities use the JSON binding of the PIRDS data standard. Although offering little control at present, this provides a crude way to store, transmit, and communicate breath data.
The Breath Plot data page has four areas:
- An upper most meta-control area. This allows you to use a different data server, if you wish, as longs as it serves PIRDS data in the JSON format. It also allows you to control a few other parameters.
- The plotting area shows three plots aligned in the horizontal dimension of time. The first plot is differential pressure (in the airway against ambient pressure.) The second is flow (change in volume over time.) The final plot the "Event plot". The Event Plot uses an abstract scale. For convenience it draws ths flow plot. However, its main purpose is to show how the trace has been broken into breaths for the purpose of computing clinically significant measures. Ocasionally measurements such as temperature and humidity are rendered as markers on this plot.
- On the right are the clinical measures, such as PIP, PEEP, Tidal Volume, Minute Volume, and Respiration Rate. This are has default high and low values. When the calculated values are outside this range, they number is drawn in red to draw attention to the fact that one parameter is outside the specifications.
- Finally at the bottom is a simple user interface for importing and exporting traces.
The VentMon is not a test lung. The purpose of a test lung is to provide realistic and compliance and resistance simulating a healthy or diseased human lung system. It is not possible to predict the clinical performance of a ventilator unless it has been extensively tested with a physical test lung. VentMon, can, however, measure some aspects of ventilator without a test lung attached as a preliminary test. Alex Izvorski of the A.R.M.E.E. and Robert L. Read of Public Invention have discussed this extensively in How To Make Your Own Accurate Test Lungs for Testing Emergency Ventilators.
The VentMon is both open hardware and open software. The firmware that runs on the board and makes sensor measurements is released under the MIT License. Other libraries and files may be released under different open source licenses if they are derived works of starting points released under other licenses.
Public Invention does not apply for patents. All of the hardware designs for the VentMon are open. The particular design documents and drawings are released under Creative Commons licenses as indicated on the document, generally either CC0 or CC-BY-SA.
The VentMon is not a commercial product; it is given away free of charge by Public Invention to teams addressing the COVID-19 pandemic with the generous support of the Mozilla Opens Source Support Foundation (MOSS) and Protocol Labs.
However, the parts in the VentMon are expensive, and we may not be able to meet demand. To request a VentMon please contact us with a description of your need and a link to your project. A small donation helps make this work possible.
For technical support, please contact [email protected] by email. Some common problems and solutions are listed below.
If you have just unboxed the device you may find that the airway pressure does not change in relationship to the flow. If this is the case, it is possible that the I2C addresses of the two BME680 pressure sensors are reversed. Please contact [email protected] for support or change the addresses in FW yourself.
The two addresses should be set as follows:
AIRWAY pressure sensor address - 0x77
AMBIENT pressure sensor address - 0x76
As of May 1st, 2020, the VentMon is a rapidly changing system which can be expected to have evolving firmware. Please check the GitHub repo often for firmware updates.
The VentMon is offered free-of-charge; however, if you decide to stop using it, please contact us about shipping it back to us. There may be others who want one, and the COVID-19 pandemic has created supply chain shortages.
You may build your own VentMon based on our instructions or make modifications and improvements to this design. If you make improvements, we hope you will submit them back to use under and open-source license so that we may publish them to the whole community.
Before the VentMon is used as a monitor with real patients, sanitization procedures will have to be provided. At present, the VentMon is intended as a tester only, and is not to be used with real patients. However, it should still be sanitized periodically. We recommend a quick wipe with isopropyl alcohol. If you have breathed into the VentMon, do not allow another person to breath it into, or any other part of your breathing circuit, without a full sanitization. The VentMon is not autoclavable at present. We are not entirely sure how to fully sanitize it.
The Sensirion Flow Sensor that we use does not need further calibration. We have checked the neonatal flow sensor against an adult flow sensor with a 500ml syringe; they are both extraordinarily accurate. Additional double-checks are always valuable, but we do not believe a user of the VentMon will have to worry about calibrating the flow measurement.
Thanks to the Public Invention volunteers Lauria Clarke and Geoff Mulligan.
A special thanks to Protocol Labs and the Mozilla Open Source Support Foundation (MOSS) COVID-19 Solutions Fund which made generous grants to support this project.