OpenVent-Bristol V2.0 COVID-19 Rapid Manufacture Ventilator BVM Ambubag

13,864

93

32

Introduction: OpenVent-Bristol V2.0 COVID-19 Rapid Manufacture Ventilator BVM Ambubag

We at OpenVent-Bristol are a team of volunteer engineers designing a simple 'low-tech' ventilator for COVID-19 treatment, on a not-for-profit basis.The ventilator is made from readily available parts and manufacturing processes and can be produced in high volumes in most countries in the world at low cost. The design is based on a hand pumped ventilator called an ambu-bag (or BVM), these low cost devices are readily available in most country's health care systems and crucially already have medical clearance.


Please take a moment to share our work and consider helping fund our design if you can: donate here

Video below shows what the V2.0 device does and how to use it:

This prototype was tested on a test lung at the National Physical Laboratory with promising results. A flavor of these results is shown in the video below:

Problem statement:
Ventilation is the only known available treatment for sufferers of COVID-19. Existing ventilation machines in hospitals are complex general purpose machines costing £10s of thousands. The availability of the existing ventilators falls short of the need and many countries will struggle to afford the expensive conventional ventilators. Critically ill patients left without ventilation treatment are in danger of death. One effect of the virus produces a very sticky mucus in the lungs which causes the lungs to collapse and makes it very difficult for the patent to breath of their own accord. Ventilation can be delivered either via intubation or a well sealed mask and delivers positive air pressure into the lungs to keep them inflated at all times. Conscious less ill patients can be aided using a CPAP (Continuous Positive Airway Pressure) ventilation device like a sleep apnoea device, this delivers a constant air of the same flow rate and pressure. The more critically ill patients need a machine to breath for them or assist with breathing which varies the pressure and flow according to inhale and exhale and relies on maintaining a good seal and a PEEP valve to keep the lungs continually inflated. An analogy: Imagine inflating a balloon, letting it deflate completely, then re-inflating it, that is like CPAP. If you inflated the balloon, deflated it half way, then re-inflated it, that's more like the treatment needed for the worse sufferers.

The aim is:

  • For a very simple low cost design
  • made using readily available components
  • that can be manufactured quickly and easily in small quantities or on mass at low cost
  • to work reliably and with the lowest risk to the patient (partly to help speed up the medical product approval process, if this project gets that far)

The USP of this design is how easy and fast it can scale for mass production for the following reasons:

  • Automated production: The production of each part is fully automated (laser cutting)
  • Minimal off-the-shelf components: Very few different components are needed and no exotic parts are needed (not even a motor coupling)
  • Simple mechanism: It uses a super simple mechanism with just an arm mounted to a motor - no complicated mechanisms to go wrong, just one moving part
  • Adjustable settings: Has full adjustment of breath frequency, pressure & tidal volume alarms (currently running PCV mode)
  • Low cost: £100 - £200 total cost is likely possible with volumes of 100s - 1000s (one off cost is around £360)

This device (running software version 5) was tested at the National Physical Laboratory on 04/05/2020 against the MHRA testing plan with good results. Once received I will upload their test report.

An image is attached showing a comparison between a waveform measured from OpenVent-Bristol V2.0 compared to an existing more sophisticated ventilator. They produce very similar output under the lung conditions used.

Operational features of this device:

  1. Pressure control ventilation (PCV). With a software change it is capable of VCV (Volume Control Ventilation), PRVC (Pressure Regulated Volume Control) and likely adaptive mode (where each breath is triggered by the patient)
  2. A user interface consisting of LCD screen and buttons (details are described in the "performance section")
  3. A battery back up (approx. 45min run time)
  4. Back up manual ventilation is possible in seconds - opening the lid of the device exposes the BVM and prevents motorised arm from pressing on the bag
  5. Alarm states including; upper and lower tidal volume alarm, main power disconnect (details are described in the "performance section")

Sensing capability:

  • Airway pressure
  • Airway flow rate (and breath volume)
  • Motor current (used to sense mechanical end stops)
  • Main DC supply voltage sense to detect supply failure (this is not enabled in software V5)

Maximum output:

For safety the device is limited to the the values listed below as these have been advised as the safe limits by our clinical advisor, anything beyond this will risk barotrauma (lung damage). These numbers are lower than those which would be necessary to meet the MHRA requirements V4, more on this described in the video below.

  • Pressure limited to 45 cmH2O
  • Tidal volume limited to 800 ml (with tolerance +100 ml)

This is our website for more project information: https://openventbristol.co.uk/

We are working with Helpful Engineering (a US based volunteer engineering) who are working to develop this design for the US market and FDA clearance.

I'm not selling this product, just releasing the design as open source to help others. This design doesn't currently meet all MHRA requirements and does not have medical product approval which will be needed for treatment. A list of improvements for the next version V3.0 include:

  • PRVC ventilation mode
  • Adaptive ventilation mode (sensing patients breath)
  • Enclosure chassis to be made from 1.2mm stainless steel grade 304 using laser cutting and CNC bending processes
  • Enclosure lid to be made from a medical approved transparent plastic that can be laser cut
  • Membrane switches for hygiene control benefit
  • Wipeable cover for LCD screen for hygiene control
  • Flow sensor to be made from medical safe plastic and integrated into chassis

Version 5 GitHub code link: https://github.com/RealFreshRate/OpenVent/tree/mas...

MHRA Rapidly manufactured ventilator system specification: https://www.gov.uk/government/publications/specifi...

Disclaimer:

All users of this design and device shall be deemed notified of the warnings stated herein. This device is a simply designed, fast produced ventilator. This device should not be used in place of an existing hospital ventilator. These should only be used as a last resort where a patient has no other alternative due to the lack of availability of existing ventilators. This is not a fully medically certified device and should not be relied upon as such. The device is designed for use by trained medical professionals and should only be used by trained medical professionals; it is not intended for home use. The designers and manufacturers of this device shall not be held liable for any death or injury that may result from this device. The designers and manufacturers of this device give no guarantees or warranties as to the efficacy and/or safety of this device. This design isn't connected with dyson or any other ventilator projects including Dyson CoVent design.

Credits:

Credit to the people below for helping to make this all possible:

  • Sam Reilly - Software Certification Engineer
  • Sadie - ITU nurse
  • Tom Breddal
  • Dr Emilio Garcia - Consultant Anesthetist in UK ITU
  • Ross Goodwin - mechanical engineer
  • P3 Medical
  • all people who have kindly donated to help fund this project

Step 1: Performance

This section provides a more detailed explanation of the performance of the system. Please see attached images and video.

Step 2: Buy Parts

The bill of materials is attached to this step as a .csv file which can be opened in excel.

The total comes to around £360 for a one off unit, this cost can be brought much lower with scale, perhaps to around £150- £200 with volumes of 100s-1000s. The next design V3.0 will be made with more DFM (Design For Manufacture) methods

Step 3: Lasercut Parts and Assemble Chassis

Everything of the chassis is laser cut from 8mm acrylic, but laser ply or MDF can be used instead.

Construction method is similar to the previous design (captive nuts and bolts).

Once yours is assembled it should look something like the attached images.

DXF files and STEP file of the assembly are stored on GrabCAD in this link: https://grabcad.com/library/openvent-bristol-v2-0-...

Quantities and materials are:

  • Q1 "Motor plate V2.0 rig 2" - 8mm acrylic
  • Q10 "arm contact ring" - 8mm acrylic
  • Q1 "Left side big hole 2" - 8 acrylic
  • Q1 "Left side big hole hinged part V2.0 rig 2" - 8mm acrylic
  • Q1 "Right side small hole V2.0 rig 2" - 8mm acrylic
  • Q1 "Right side small hole hinged V2.0 rig 2" - 8mm acrylic
  • Q1 "Motor plate V2.0 rig 2_bk" - 8mm acrylic
  • Q1 "Base plate V2.0 2" - 8mm acrylic
  • Q1 "V2.0 front plate 2" - 8mm acrylic
  • Q1 "Buttons for buttons rig V2.0" - 8mm acrylic
  • Q1 "back lower part V2.0" - 8mm acrylic
  • Q3 "Base fins" - 8mm acrylic
  • Q1 "Top panel rig V2.0" - 8mm acrylic
  • Q1 "Lid latch rig V2.0" - 8mm acrylic
  • Q1 "Back panel rig V2.0" - 8mm acrylic

Step 4: Motor Mount

As shown in the pictures;

  • attach 4 cable ties (the beefier the better, I used 8mm wide ties) through the 6 holes in the back plate such that 4 of them sit parallel to each other in the horizontal axis
  • then pass 2 jubilee clips (I used 25-40mm clips) through the loops in the cable ties and around the motor
  • Then tighten the clips until it grips the motor nicely, but don't tighten it too much!!! Because it can crush the casing and damage the motor

Step 5: Mount Motor Arm

Laser cut the motor arm from 6mm mild steel. DXF is in the following link: https://grabcad.com/library/openvent-bristol-v2-0-...

Attach the motor arm to the motor shaft using the M8 nylock nut and bolt. Hold the nut in its 'captive' slot and tighten the bolt up against the flat of the motor shaft with a hex key.

Attach some of the small laser cut oval pieces to the end of the arm with the long M6 bolt (I used a 100mm) and nut to distribute the pressing force on the bag. I didn't cut enough ovals so I spaced them out with nuts and washers (as you see in the pictures). It is best to use a nylock M6 nut if you have one here. You may also want to add some bubble wrap or something around the end of the arm for extra cushioning as it may wear into the bag over extended use.

Step 6: Build Flow Sensor

Instructions for flow sensor build are in a separate Instructables post here: https://www.instructables.com/id/How-to-Make-an-A...

Step 7: Pressure Sensor

Step 8: Build Electronics

Follow the electronic system diagram in the attached image. Steps are as follows:

  1. Buttons on the LCD shield are connected to pin A0, problem is A0 is also used for current sensing on the motor driver shield so we need to de-solder the A0 male header pin from the LCD shield and attach a jumper wire connecting A0 to A3 on the LCD shield (shown in images)
  2. Wire up your gauge pressure sensor and differential pressure sensor (for flow meter). I wired these on floating wires (shown in pics) but a better way would be to use an Arduino proto shield
  3. Add your potential divider circuit to your proto shield
  4. Wire up your 2.1mm DC power socket with parallel connection to 12V lead acid battery and power connection to both motor shields
  5. Make an Arduino shield sandwich of:
    1. LCD shield (on top)
    2. Motor driver shield 1st
    3. Motor driver shield 2nd (we need 2 because 1 alone can't provide enough current, a better motor driver will be used in V3.0)
    4. Proto shield
    5. Arduino Uno (bottom)
  6. Wire up motor to the motor driver shield

Circuit testing:

You will need to install the following libraries for full functionality of the test codes:

  1. Do a test upload with the blink program
  2. Test the gauge pressure sensor on pin A5 with code from here: https://www.instructables.com/id/How-to-Read-MPX50...
  3. test the differential pressure sensor on pin A4 with code from here: https://www.instructables.com/id/How-to-Make-an-Ai...
  4. test the LCD screen using code from here: https://www.velleman.eu/downloads/29/vma203_a4v03.... Note; the default analogue ranges didn't work with my buttons, maybe my resistors were out of tolerance or something, so you may have to change these too

Step 9: Software

Download the software from GitHub in the link below:

Version 5 code: https://github.com/RealFreshRate/OpenVent/tree/mas...

Make sure the main file (e.g. OpenVent_Bristol_rig_V2.0_PCV_V5.ino) is in a folder with the same name and the "Analogue_filtering.ino" file is in that same folder, that way when you open the main file the analog_filtering file should open in the second tap in the Arduino IDE.

You will need to make sure you have the following Arduino libraries installed:

Step 10: Test Whole System

Test your whole system using the Arduino plotter to create some dynamic graphs like in this video:

Arduino Contest 2020

Second Prize in the
Arduino Contest 2020

Be the First to Share

    Recommendations

    • STEM Contest

      STEM Contest
    • Role Playing Game Challenge

      Role Playing Game Challenge
    • CNC Contest

      CNC Contest

    32 Comments

    1
    520mianabdullah
    520mianabdullah

    16 days ago

    I have l293N i am confusing now how can i control my direction as there are two pins for direction control in l293N and in L298p only one pin used for direction please help sirI have l293N i am confusing now how can i control my direction as there are two pins for direction control in l293N and in L298p only one pin used for direction please help sir...

    16174011454634216144307563905149.jpg
    0
    maovx
    maovx

    3 months ago

    I'm trying to get Volume control ventilation but I'm not sure how to change your code to achieve that. Where do I start?

    0
    maovx
    maovx

    Reply 3 months ago

    Oh I see, thank you! The code is really easy to read, well done!

    0
    2017mc311
    2017mc311

    4 months ago

    I am facing a error in programming.
    Arduino motor shield rev r3 not is not supporting can u help please?

    0
    Darren Lewis
    Darren Lewis

    Reply 4 months ago

    Have you installed the library for the motor shield?

    0
    Captain_Nemo
    Captain_Nemo

    11 months ago

    This thing will kill people. People who need to be put on a ventilator are too fragile to use a BVM on for more than a few minutes. This will put too much stress on their lungs, and they will die.

    0
    krisferrari
    krisferrari

    Reply 5 months ago

    wow, what a respect to author

    0
    ingfrancom
    ingfrancom

    10 months ago

    Your project is a failure.
    It is absurd to use the manual respirator, when you can use an axial fan, or centrifuge.
    Have you seen a car with legs?

    0
    ingfrancom
    ingfrancom

    Reply 10 months ago

    Use esto en lugar de la goma

    unnamed (10).jpg
    0
    maovx
    maovx

    Reply 7 months ago

    Como se llama eso?

    0
    pepemech
    pepemech

    Question 8 months ago

    Hi very nice work!!!
    I have a question, why don't appear "Q1 "Motor plate V2.0 rig 2_bk" - 8mm acrylic" in the general assemble?

    hola.jpg
    0
    johan.pitty.jp
    johan.pitty.jp

    Question 10 months ago on Step 9

    Hi Darren
    a question when loading the program is just those two files that are together plus the libraries you mention.

    0
    khoile78
    khoile78

    10 months ago

    Dear Darren, can i use MPX53DP 50KPA as pressure and flowrate sensor in stead of MPX5010DP 10KPA

    2
    technovative
    technovative

    11 months ago

    Impressive and admirable work. Strong evidence that complex exorbitantly priced ventilators aren't the only solution. Thank you for participating in the effort to help sustain lives.

    0
    Captain_Nemo
    Captain_Nemo

    Reply 11 months ago

    Those "complex exorbitantly priced ventilators" are that way because they've had exponentially more R&D put into them. A company recently published "plans" for their ventilator design, it didn't have the actual plans in it, it was just all the engineering specifications, and it was massive, hundreds of pages long.

    0
    Darren Lewis
    Darren Lewis

    Reply 10 months ago

    Yes you’re right, all of their paper work is needed to help prove that the device will be safe. I think that was MedTech?

    0
    Karmudjun
    Karmudjun

    Reply 11 months ago

    Here in the United States our litigious society has truly made medical grade ventilators both complex and costly. The complexity is required, ventilation is not a one setting therapy, you can do more harm than good if you approach ventilation that way. For our national certification, the durability of the device must be proven with many 'worst case scenarios' attempted - not to cause the device to fail because we can, but to prove that the device WON'T fail under the worst of normal circumstances. When you add in product liability - that drives up the cost of US ventilators. Since they are made with proven and durable parts (not off the shelf parts), most everyone can trust the quality control behind each of the parts used. Off the shelf doesn't have that quality control factor.

    I'd use this ventilator - but only due to the fact I know I can trust the build. How long the sensors would last in real world conditions? That is the only issue I have and I trust all my respiratory staff to be able to manage a patient on a device like this if it did fail.

    0
    Darren Lewis
    Darren Lewis

    Reply 10 months ago

    Agreed it’s not an easy task but we’re working on it

    1
    mtaustin
    mtaustin

    11 months ago

    I can't quite see your whole hook up but it looks like you are delivering patient air out of the Ambu and through that long white tube coming from the Ambu out? Assuming that your exhaled air from the patient is also traveling through that long white tube and exiting the system through the original Ambu valve? If this is true, how are you dealing with the dead space created by the long white tube to the patient and the build up of CO2? Each new patient breath will contain a tremendous amount of CO2 from his last exhaled breath and certainly mess with your desired FI02 delivery. Just wondering.