|
A team of one electrical engineer and one materials scientist + mech engineer in Dortmund, Germany.
We are working on a system that does pretty much everything a modern ventilator does, CMV, IMV, CSV in Pressure and Volume control modes.
The biggest tasks that remain to be solved:
*installing the flow sensor and having true volume control
*adding an alarm
*putting the whole thing in a cool box
*a million more things probably
Please comment on what you think has the greatest pareto improvement potential!
Here's my latest video plus you can find earlier ones on the same channel.
Awesome resource, thanks Erich:
A great checklist for us ventimakers:
|
thanks for the complement! I suggest having a read of volume 3 and 4
Couple of points - oxygen blending needs to be moderately accurate - I'm yet to see a plausible solution to this that will not rely on an O2 sensors
- pressure in a circuit is nearly always positive - make a simple peep valve by submerging you outflow in a bucket of water 5cm then use a 4cm as your breath trigger
- your software should to seamlessly transition between mandatory and spontaneous breathing
- avoid scrolling on your UI, use tabs, but always have certain values displayed at all time
- its a big job - focus on adding value to a single component (volume 4)
hope that helps
Erich Schulz,?mbbs, mba, fanzca 0410 277 408
A team of one electrical engineer and one materials scientist + mech engineer in Dortmund, Germany.
We are working on a system that does pretty much everything a modern ventilator does, CMV, IMV, CSV in Pressure and Volume control modes.
The biggest tasks that remain to be solved:
*installing the flow sensor and having true volume control
*adding an alarm
*putting the whole thing in a cool box
*a million more things probably
Please comment on what you think has the greatest pareto improvement potential!
Here's my latest video plus you can find earlier ones on the same channel.
Awesome resource, thanks Erich:
A great checklist for us ventimakers:
|
O2 level sensing is great when integrated into the ventilator solution with the appropriate transducer inputs.
Alternatively, I wonder if accurate O2 blending could be accomplished via a "metering valve", similar to the way that automotive fuel injection works to control the fuel-to-air mixture. Perhaps it would even be possible to use an automotive injector (new and never used before...) for this purpose. Most fuel injectors are utilizing piezoelectric actuators to perform the metering action,?through?precise?pulsation. The injector orifice(s) determines the volume of fluid that is injected per pulse. The main problem that I see is that it is much easier to seal fluids (like automotive fuel) in the injector mechanism than it is to seal gases (smaller molecules). Regardless, If anyone has the time and resources to play with this concept, they can share their findings with the rest of us.
--Ron N7TFZ
thanks for the complement! I suggest having a read of volume 3 and 4
Couple of points - oxygen blending needs to be moderately accurate - I'm yet to see a plausible solution to this that will not rely on an O2 sensors
- pressure in a circuit is nearly always positive - make a simple peep valve by submerging you outflow in a bucket of water 5cm then use a 4cm as your breath trigger
- your software should to seamlessly transition between mandatory and spontaneous breathing
- avoid scrolling on your UI, use tabs, but always have certain values displayed at all time
- its a big job - focus on adding value to a single component (volume 4)
hope that helps
Erich Schulz,?mbbs, mba, fanzca 0410 277 408
A team of one electrical engineer and one materials scientist + mech engineer in Dortmund, Germany.
We are working on a system that does pretty much everything a modern ventilator does, CMV, IMV, CSV in Pressure and Volume control modes.
The biggest tasks that remain to be solved:
*installing the flow sensor and having true volume control
*adding an alarm
*putting the whole thing in a cool box
*a million more things probably
Please comment on what you think has the greatest pareto improvement potential!
Here's my latest video plus you can find earlier ones on the same channel.
Awesome resource, thanks Erich:
A great checklist for us ventimakers:
|
Measuring oxygen...
Someone with a better background in this should check this basic premise, but the premise is that a pressure differential, across a constriction, measures volume flow. A mass flowmeter, as the name says, measures the mass of the gas flow. Oxygen is about 17% denser than air, so measuring the difference between a mass flow sensor and a differential pressure flow sensor may be able to indicate the ratio of oxygen in the "air" being supplied by the ventilator.?
Possibility constructing a small chamber with a fan moving the "air" at a known flow rate across a heated thermistor, could be used to measure the oxygen percentage of the "air". Thermistors, I've been told, are not fast enough to work as an instantaneous flow sensor for accumulating total flow volume, but speed may not be an issue in this use of a heated thermistor.?
This measurement system would only be good for measuring the input air/oxygen ratio. Because you should be able to know the only two gasses being mixed are air and oxygen. On the output side you would not be able to correctly differentiate between the densities of air, oxygen and CO2.?
Tom, wb6b
|
I believe?that a differential scheme using the fact that O2 (and O3) absorb significant amount?of Far IR energy in the 9 to 10 um (a relatively narrow range) is used in some gas sensors. However, I doubt that we have the time frame necessary to develop such devices. That is also the reason why I am thinking?that O2 metered?injection could conceivably?be implemented in a relatively short time frame, as long as off the shelf automotive injectors can be re-purposed successfully. In principle adding atomized fuel to air to form combustible mixture, is not much different?than mixing O2 with "normal" air to produce oxygen enriched mixture.
--Ron
On Tue, Apr 14, 2020 at 9:12 PM Tom, wb6b < wb6b@...> wrote: Measuring oxygen...
Someone with a better background in this should check this basic premise, but the premise is that a pressure differential, across a constriction, measures volume flow. A mass flowmeter, as the name says, measures the mass of the gas flow. Oxygen is about 17% denser than air, so measuring the difference between a mass flow sensor and a differential pressure flow sensor may be able to indicate the ratio of oxygen in the "air" being supplied by the ventilator.?
Possibility constructing a small chamber with a fan moving the "air" at a known flow rate across a heated thermistor, could be used to measure the oxygen percentage of the "air". Thermistors, I've been told, are not fast enough to work as an instantaneous flow sensor for accumulating total flow volume, but speed may not be an issue in this use of a heated thermistor.?
This measurement system would only be good for measuring the input air/oxygen ratio. Because you should be able to know the only two gasses being mixed are air and oxygen. On the output side you would not be able to correctly differentiate between the densities of air, oxygen and CO2.?
Tom, wb6b
|
Potential electronically controlled metering valve. See?attached image of a typical automotive injector.
On Tue, Apr 14, 2020 at 9:53 PM Dr. Flywheel via <Dr.Flywheel= [email protected]> wrote: I believe?that a differential scheme using the fact that O2 (and O3) absorb significant amount?of Far IR energy in the 9 to 10 um (a relatively narrow range) is used in some gas sensors. However, I doubt that we have the time frame necessary to develop such devices. That is also the reason why I am thinking?that O2 metered?injection could conceivably?be implemented in a relatively short time frame, as long as off the shelf automotive injectors can be re-purposed successfully. In principle adding atomized fuel to air to form combustible mixture, is not much different?than mixing O2 with "normal" air to produce oxygen enriched mixture.
--Ron
On Tue, Apr 14, 2020 at 9:12 PM Tom, wb6b < wb6b@...> wrote: Measuring oxygen...
Someone with a better background in this should check this basic premise, but the premise is that a pressure differential, across a constriction, measures volume flow. A mass flowmeter, as the name says, measures the mass of the gas flow. Oxygen is about 17% denser than air, so measuring the difference between a mass flow sensor and a differential pressure flow sensor may be able to indicate the ratio of oxygen in the "air" being supplied by the ventilator.?
Possibility constructing a small chamber with a fan moving the "air" at a known flow rate across a heated thermistor, could be used to measure the oxygen percentage of the "air". Thermistors, I've been told, are not fast enough to work as an instantaneous flow sensor for accumulating total flow volume, but speed may not be an issue in this use of a heated thermistor.?
This measurement system would only be good for measuring the input air/oxygen ratio. Because you should be able to know the only two gasses being mixed are air and oxygen. On the output side you would not be able to correctly differentiate between the densities of air, oxygen and CO2.?
Tom, wb6b
|
I did some research and found out that automotive suppliers actually produce special injectors for engines fueled by natural?gas (CNG) or liquid petroleum?gas (LPG). Though O2 molecules are smaller than either CNG or LPG, these injectors may still work well with a source of compressed O2 gas to support the computer controlled mixing system that we are seeking. Since the injector produces a very precise volume of gas on each injection pulse, the total volume delivered depends on the number of pulses driving the injector per unit of time. These parts operate from 12VDC without further?modifications.
Here is product info from Bosch:?
Example part on Amazon:
--Ron
On Tue, Apr 14, 2020 at 10:00 PM Dr. Flywheel via <Dr.Flywheel= [email protected]> wrote: Potential electronically controlled metering valve. See?attached image of a typical automotive injector.
On Tue, Apr 14, 2020 at 9:53 PM Dr. Flywheel via <Dr.Flywheel= [email protected]> wrote: I believe?that a differential scheme using the fact that O2 (and O3) absorb significant amount?of Far IR energy in the 9 to 10 um (a relatively narrow range) is used in some gas sensors. However, I doubt that we have the time frame necessary to develop such devices. That is also the reason why I am thinking?that O2 metered?injection could conceivably?be implemented in a relatively short time frame, as long as off the shelf automotive injectors can be re-purposed successfully. In principle adding atomized fuel to air to form combustible mixture, is not much different?than mixing O2 with "normal" air to produce oxygen enriched mixture.
--Ron
On Tue, Apr 14, 2020 at 9:12 PM Tom, wb6b < wb6b@...> wrote: Measuring oxygen...
Someone with a better background in this should check this basic premise, but the premise is that a pressure differential, across a constriction, measures volume flow. A mass flowmeter, as the name says, measures the mass of the gas flow. Oxygen is about 17% denser than air, so measuring the difference between a mass flow sensor and a differential pressure flow sensor may be able to indicate the ratio of oxygen in the "air" being supplied by the ventilator.?
Possibility constructing a small chamber with a fan moving the "air" at a known flow rate across a heated thermistor, could be used to measure the oxygen percentage of the "air". Thermistors, I've been told, are not fast enough to work as an instantaneous flow sensor for accumulating total flow volume, but speed may not be an issue in this use of a heated thermistor.?
This measurement system would only be good for measuring the input air/oxygen ratio. Because you should be able to know the only two gasses being mixed are air and oxygen. On the output side you would not be able to correctly differentiate between the densities of air, oxygen and CO2.?
Tom, wb6b
|
On Tue, Apr 14, 2020 at 09:53 PM, Dr. Flywheel wrote:
thinking?that O2 metered?injection could conceivably?be implemented
I've seen designs where the O2 and Air valves were each activated for controlled time to mix the gasses. The advantage of the fuel injector is it could be pulsed many times in an inhalation cycle so the air/oxygen bursts should be fully mixed by the time it gets down the inhalation circuit tubing. If there is a pinpoint nozzle in an injector, maybe it could be enlarged, should that actually be an issue.? Possibility only be providing the oxygen enriched "air" while the lungs are filling and if the later part of the cycle leaks gas while maintaining PEEP and such, providing only air, could help reduce oxygen usage. Especially with "low pressure" ventilators where leakage may be part of the design to ease flow control and for the system to be less stiff and more tolerant to a patient?coughing. Tom, wb6b
|
On Tue, Apr 14, 2020 at 10:31 PM, Dr. Flywheel wrote:
injectors for engines fueled by natural?gas
Cool, that's even better...
|
The oxygen sensors based on LED are reportedly unreliable. Dr.KP Prasad was narrating an incident yesterday when an expired patients oxygen was still showing 97% saturation! I will have the flow measurement sorted out in the next few days. It is maths just beyond my current understanding. - f
On Wed 15 Apr, 2020, 11:06 AM Tom, wb6b, < wb6b@...> wrote: On Tue, Apr 14, 2020 at 10:31 PM, Dr. Flywheel wrote:
injectors for engines fueled by natural?gas
Cool, that's even better...
|
On Tue, Apr 14, 2020 at 11:07 PM, Ashhar Farhan wrote:
I will have the flow measurement sorted out in the next few days
You probably already have this information. But, just in case it is of help, a link is in this message. /g/VentilatorDevelopers/message/1598Good wishes on all your (and everyone's) hard work. Tom, wb6b
|
To Culyma group
Keep up the good work ?
A team of one electrical engineer and one materials scientist + mech engineer in Dortmund, Germany.
We are working on a system that does pretty much everything a modern ventilator does, CMV, IMV, CSV in Pressure and Volume control modes.
The biggest tasks that remain to be solved:
*installing the flow sensor and having true volume control
*adding an alarm
*putting the whole thing in a cool box
*a million more things probably
Please comment on what you think has the greatest pareto improvement potential!
Here's my latest video plus you can find earlier ones on the same channel.
Awesome resource, thanks Erich:
A great checklist for us ventimakers:
|
Update from Eddy, the Emergency Ventilator out of Dortmund, Germany.
In a video (1:33):
https://youtu.be/gBZ2cvdbsgg
To your points, Erich,
1. Eddy does precise, timed FiO2 control. The O2 and air valves are open after one another to achieve the required O2 concentration. For exemple, at FiO2 = 60.5, Both O2 and air valves are open for the exact same length of time.
The oxygen concentration will be controlled by 2 flow sensors: one at
2. As soon as I'll have a realistic lung model, I'll add my PEEP valve and an option for 'sofrware PEEP', ie. closing the exhaust valve when the pressure falls to the predefined minimum at exhalation. (The rubber balloon's pressure curve is very abrupt, almost binary.)
3. Intermittent Mandatory Ventilation implemented and trigger pressure oprtion added.
4. New, bigger screen, no scrolling needed.
In progress:
Adding a muffler for shockwave attenuation at the beginning of inspiration
Adding the flow sensor to have true volume control.
Adding alarms.
|
On Sat, Apr 18, 2020 at 11:53 AM, <culyma@...> wrote:
Eddy, the Emergency Ventilator out of Dortmund, Germany.
I'm new to groups.io, so not sure if I'm replying correctly to the post, but in reference to the posted video about Eddy, I have some feedback. I hope I'm not speaking out of turn here, just want to offer some feedback from medical/respiratory physiology standpoint. (I'm a pulmonary and critical care physician) Based on the video, EDDY uses a single hose for inspiration and expiatory, which looks like a standard 25mm id X 6 ft long CPAP hose. This will not work for a ventilator. CPAP systems have built in air leak so most of the exhaled air escapes through holes in the mask. This differs from ventilator systems where there is no leak. The problem with the single hose is that the volume in the hose (referred to as dead-space) is larger than the tidal volume. The hose has a volume of nearly 900 ml (this is an edit, my earlier math was wrong) while a typical ventilator tidal volume for a 6ft male is between about 450 ml and 650 ml This means NO FRESH AIR will reach the patient, rather, they will just be cycling the same air in and out of the hose.? You can test this at home by breathing through the tube without having it connected to a vent (one end in your mouth and one end hanging free). After a while you will feel like you're suffocating even though you are moving plenty of air in and out of your lungs. I'm haven't been able to completely determine everyone's expertise here, but I'm happy to lend a hand with any questions from a pulmonary or respiratory physiology standpoint. EDIT to add info per Gordon's response Nearly all true ventilator tubing sets use separate inspiratory and expiratory hoses with a Y connector such that air really only flows in one direction. From machine to patient on the inspiratory limb, then from patient to machine on expiratory limb. This way the only dead-space is within the endotracheal tube and conducting airways of the lungs (distal to the common tube of the Y connector).? Here's an image: Some specialized systems will use a single hose from the machine to the patient, then have valves at the patient so exhaled air is blown out into the room. However, this isn't a good option for COVID patients as it results in significant aerosolization of the virus which increases the risk to HCWs (health care workers). Steve S
|
Thank you for pointing out the dead space issue. ? You may want to explain how valves are used to get around that
toggle quoted message
Show quoted text
On Apr 18, 2020, at 18:37, Steve S <skol0016@...> wrote:
?On Sat, Apr 18, 2020 at 11:53 AM, <culyma@...> wrote:
Eddy, the Emergency Ventilator out of Dortmund, Germany.
I'm new to groups.io, so not sure if I'm replying correctly to the post, but in reference to the posted video about Eddy, I have some feedback. I hope I'm not speaking out of turn here, just want to offer some feedback from medical/respiratory physiology standpoint. (I'm a pulmonary and critical care physician) Based on the video, EDDY uses a single hose for inspiration and expiatory, which looks like a standard 25mm id X 6 ft long CPAP hose. This will not work for a ventilator. CPAP systems have built in air leak so most of the exhaled air escapes through holes in the mask. This differs from ventilator systems where there is no leak. The problem with the single hose is that the volume in the hose (referred to as dead-space) is larger than the tidal volume. The hose has a volume of over 3.5 L while a typical ventilator tidal volume for a 6ft male is between about 450 ml and 650 ml This means NO FRESH AIR will reach the patient, rather, they will just be cycling the same air in and out of the hose.? I'm haven't been able to completely determine everyone's expertise here, but I'm happy to lend a hand with any questions from a pulmonary or respiratory physiology standpoint. Steve S?
|
On Sat, Apr 18, 2020 at 11:53 AM, <culyma@...> wrote:
Eddy does precise, timed FiO2 control.
Hi, Are there any issues with supping the unmixed O2 and air during separate times in the inhalation phase; does the lung care? Or could there even be an advantage to sending air or full 100% O2 first, then following up by the other? Tom, wb6b
|
Just to add a few points
Agree with Steve and Tom's comments. Screen looks good!
This design needs an "expiratory valve" (aka "insipatory valve"), and also a blending chamber (some of the UQ students were looking at hacking a fire extinguisher for this purpose). Timing based blending needs to consider that line/cylinder pressure (especially dynamic pressures at high flow) may vary between the O2 and air supplies - its a tricky problem)
Both Steve and Culyma (and everyone who isn't flat out already on the FDA approval of this project) - I encourage you both to check out the - comments, corrections and clarifications are very welcome. It's in 4 volumes - all four are important!
I wrote these words for a Medium article this morning they seem relevant: The good news is that we are closer to the goal of making smart safe ventilators than you may realise. We have all the components we need and many clever designs. Our challenge is to synthesise the efforts of multiple teams in a way that extracts the best from each. Ventilators break down neatly into various components: electronic controllers and interfaces, oxygen-air blenders, pressure regulators (or pumps or bellows), inspiratory and expiratory manifolds, scavenging units. Each component has multiple subcomponents such as valves and sensors that can be addressed individually.? The quickest pathway to usable solutions is to start identifying the best candidates for each particular component. Hopefully this is an effort that we can address collectively. The starting point is identifying each component and characterising its attributes. Such component level reviews, if accompanied with links to complete open designs, are the building blocks of the solution we need. Ultimately I really need help with Volume 4, which aims to point the global efforts in a way that they can build on each others.
Culyma - it would be great to see some the component level analysis that you did.
Erich Schulz,?mbbs, mba, fanzca 0410 277 408
On Sun, 19 Apr 2020 at 10:20, Tom, wb6b < wb6b@...> wrote: On Sat, Apr 18, 2020 at 11:53 AM, <culyma@...> wrote:
Eddy does precise, timed FiO2 control.
Hi,
Are there any issues with supping the unmixed O2 and air during separate times in the inhalation phase; does the lung care? Or could there even be an advantage to sending air or full 100% O2 first, then following up by the other?
Tom, wb6b
|
Are there any issues with supping the unmixed O2 and air during separate times in the inhalation phase; does the lung care? Or could there even be an advantage to sending air or full 100% O2 first, then following up by the other?
This isn't ideal for two reasons. 1) You won't know what fraction of oxygen is reaching the patient. Of course a lot of mixing will occur in the tubes just from turbulence, but there might be significant variation if the oxygen and air don't fully mix. It could be like a train with separate cars on a track, one breath is high FiO2, the next breath is low. Averaged over time it might equate to the actual FiO2, but thats not what the lungs are seeing at each breath.?This could give a false impression of how much oxygen the patient is actually requiring which could complicate patient assessment. Probably not a huge issue, but hard to say for sure. 2) High oxygen concentrations can be damaging to cells. We sometimes need to use high concentrations, but we avoid it when we can. If the gases didn't mix and the patient was alternating between high concentrations breath and low concentration breaths then the high concentration breaths will put more oxidative stress on the cells. I don't know what the real world result of this would be, but the cells are already pushed to the limit. Steve S
|
On Sat, Apr 18, 2020 at 05:44 PM, Erich Schulz wrote:
?
This design needs an "expiratory valve" (aka "insipatory valve"), and also a blending chamber (some of the UQ students were looking at hacking a fire extinguisher for this purpose). Timing based blending needs to consider that line/cylinder pressure (especially dynamic pressures at high flow) may vary between the O2 and air supplies - its a tricky problem)
What is the current proposed method of O2 introduction? I'm not a?fluid dynamicist, but I think the gases will both need to be supplied throughout the entire inspiratory phase. I agree that time based blending likely won't work well. I can think of two approaches that could be used. The obvious is an adjustable valve to alter the flowrate of the oxygen, however, these aren't as easy to source as binary (on/off) valves. The other is a binary (on/off) valve (like typical solenoid valves) using a PWM (pulse width modulation) type principle to rapidly open and close the oxygen valve during the inspiratory phase to add the desired amount of oxygen. Altering the rate of opening and closing will alter the FiO2. I don't "think" much of a mixing chamber would be required. If the oxygen is supplied throughout the entire inspiratory phase it should mix just from turbulence. If we have any concepts for the proposed system I have some aerospace colleges who specialize in airflow and turbulence patterns who might be able to throw together a computer model.? What types of valves are currently being prototyped for this vent system? Was oxygen exposure considered during valve selection? Oxygen is very corrosive to many materials and may lead to early failure that won't become apparent during RA (room air) testing. Possibly more concerning is the risk of spontaneous combustion when high levels of oxygen contact hydrocarbons, such as those frequently used to lubricate solenoid valves. I don't know the actual probability of this but it should be considered. Also, I'm new to this project and have had a tough time figuring out where everything stands. Is there a centralized document location? If not, it might be a good idea to set something up. I know this project is being worked on by many different people in different locations, but its tough to get an overview of what going on and what needs to be done.? Perhaps a specific topic thread that lists the current needs and problems to be overcome so contributors can easily identify areas they might be able to help. Steve S
|
On Sat, Apr 18, 2020 at 06:47 PM, Steve S wrote:
What is the current proposed method of O2 introduction?
The primary project on this group is using a compressed O2/air mixer device that is apparently part of the normal things around a hospital. Here is a post that suggest the possibility of using a fuel injector as a valve that can be pulsed at a rapid rate. If a regular solenoid valve was not fast enough to provide small enough bursts of air/O2 for easy mixing. Likely there is a need to mix O2/air where the normal mixing device is not available.? /g/VentilatorDevelopers/message/1609If both the air and O2 were pre-regulated by accurate pressure regulators, the ratio of on/off times could work. Many regulators experience a large output pressure drop when the flow switches form off to on. Don't know if the regulators were of identical design, if that drop in pressure when the regulators can't keep up with the flow, would cancel each other out in the overall mix calculations. (Hmmm.. With widely different on/off times for the air and O2, and reserve air/O2 stored in the internal space of their respective plumbing, especially at high pulse rates, the pressure drops over time/volume may be a messy thing to compensate for. Hopefully a regulator that can stay accurate when switched from off to on is possible.) Tom, wb6b
|
