VENTILATOR OPEN SOURCE MIT FULL DOCUMENTATION

Key Ventilation Specifications

Note changes to specifications with increased BPM (Breaths per minute) and I:E
ratios. Adjusting one of these parameters may influence the other, since at higher
BPM, the tidal volume is usually decreased. Consult a clinician.

The minimum controllable parameters in order to ventilate a patient include:

1. Respiratory Rate (RR) (breaths per minute): between 8 – 40

2. Tidal Volume (TV) (air volume pushed into lung): between 200 – 800
mL based on patient weight
3. I/E Ratio (inspiratory/expiration time ratio): recommended to start around 1:2;
best if adjustable between range of 1:1 – 1:4*
4. Assist Detection pressure. When a patient tries to inspire, they can cause a
dip on the order of 1 – 5 cm H2O, with respect to PEEP pressure (not
necessarily = atmospheric).

Airway pressure must be monitored

1. Maximum pressure should be limited to 40 cm H2O at any time; Plateau

pressure should be limited to max 30 cm H2O
2. The use of a passive mechanical blow-off valve fixed at 40 cm H2O is
strongly recommended
3. Clinician require readings of plateau pressure and PEEP (refer to clinical
documentation tab)
4. PEEP of 5-15 cm H2O required; many patients need 10-15 cmH2O

Failure conditions must permit conversion to manual clinician override, i.e. if automatic
ventilation fails, the conversion to immediate ventilation must be immediate.

Ventilation on room air is better than no ventilation at all. Blending of oxygen and air
gas mixture to adjust FiO2 is not important in an emergency scenario. It is certainly nice to
have that ability and can easily be implemented with an oxygen / air gas blender that some
hospitals already have.
COVID-19 can get aerosolized (airborne), so HEPA filtration on the patient’s
exhalation is required or between the ventilator unit and the patient (at the end of the
endotracheal tube) to protect clinical staff from certain infection. In-line HEPA filters can
usually be purchased alongside manual resuscitator bags.
Heat and moisture exchanger should be used in line with the breathing circuit.
Failure conditions must result in an alarm.
This is a minimum requirement set for emergency use. Equipment designed for more regular
use, even if for emerging markets, will require additional features to be used on a regular
basis.
Power Specification
The estimation of the maximum power required by a motor used to compress an Ambu bag
using a double gripper design, actuated from the bottom. Different designs, with other
actuation methods, will change motor specifications, but the power should stay
approximately the same.
Caution: Many designs circulating on the internet significantly underestimate the pressures
needed to inflate a subject’s lungs. There is a reason why the diaphragm is a large muscle.
However, adding more power without great care is equally dangerous. In addition, COVID-
19 compromised patients may require more aggressive motion profiles, i.e. short, quick
breaths and longer inspiration times. In specific, clinicians are reporting I:E ratios of 1:4 in
COVID-19 patients. (Our estimate uses 1:3.)

Theoretical Power Requirement

Independent of the mechanical design of the gripper, the required power output can
be computed from the worst-case values of the following variables:

o Maximum pressure at airway: Pairway,max = 40 cm H2O (pop off cracking pressure)

o Maximum respiration rate: RRmax = 30 bpm
o Minimum inhale/exhale ratio of 1:3: IEratio, min = 0.25
o Maximum volume output: Vmax = 800 cm3

That is, in the worst case the device needs to squeeze of air at a pressure of 40 cm
H2O, in half a second (60 sec / 30 * 0.25).

The volume flow rate needed in the worst-case (peak) scenario is, then:

The power output (in the form of pressurized volume flow in the airway) is:

However, some of the power used for squeezing the bag is lost (bag deformation,
friction, etc.) and let’s estimate that 50% is converted to pressurized volume flow.
Taking this efficiency into account, the power required at the gripper is:

The actual power needed from the motor will be higher, how much higher depends on the
mechanical design. Assuming half the power output of the motor is lost to mechanical
inefficiencies (gears etc.), the power output required from the motor is given by:
Power requirement for 2-finger design

This is an alternative approach to calculating power.

The following is an illustration of a 2-finger gripper design:

more direct approach can be used for this design provided the following quantities
can be measured:

o Finger-bag contact area

o Finger lever arm length
o Sweep angle

For one particular prototype, we have:

o Finger-bag max contact area: Abag = 80 mm2

o Finger lever arm length: lfinger = 12 cm
o Sweep angle: αsweep = 30°

The maximum force of the bag on one finger (when fully squeezed) is, using the
same 50% pressure transmission efficiency as before:

The maximum torque needed on each finger is then:

Now we can compute the power required on each finger using the sweep angular
rate ( in half a second):

The total power for the motor (assuming a single motor) is twice that:

which is similar to the result from the calculation based on volume flow rate. Applying
the same 50% motor and gearbox efficiency, we arrive at a similar estimate of 25.2
W.

Recommendation

Therefore, a power supply at 12 V should be specified with a minimum of a 2.5 A

capability for safety. Allowing another safety factor of 2 we recommend a 5 A supply.

Benchtop Validation

During testing of a 2-finger gripper design, E-Vent Unit 002, equipped with an Andy
Mark motor am-3656 (188:1 gearbox) we observed a peak current of under 3 A at
12V or 36 W.

Motor Selection

Motor & Encoder

The mechanical system should be driven with a motor under closed loop control. For
feedback measurement, we are using a DC gear motor with integrated quadrature
feedback.

Prototype component: Andy Mark AM 3656 188:1 gearmotor with encoder. This was
scavenged from a FIRST Robotics kit, it is suitable for testing, and we are working in
validating it’s feasibility for more widespread use under the desired operating
conditions. This is provided as an example only, builders must use their supply chain
to identify correct motors for their application.

Motor Options: Brushed DC motor with gearbox and position feedback. Any
sufficiently high-torque, back-driveable motor with angle sensing, integrated or
separated, should work. We are currently investigating NEMA stepper motors.

Note: It is not ideal to back drive high gear ratio motors, but this is a limited
operation.

Caution: If a stepper motor is used, position should still be taken from the angle
sensor so that missed steps do not cause position drift
Caution: The motor and mechanism, together, must be back drivable in order to
move the mechanism by hand, remove the bag and immediately convert to manual
bagging.

Operating Parameters

Understanding the best way to care for COVID-19 patients is changing daily.

These are our best, current working specifications. Be sure to apply your own safety
factors, recommended 2x.

Assumed nominal operating parameters: Referencing the most recent clinical

documentation of max 40 breath-per-minute (bpm), up to a 1:4 I:E ratio and a pop off
set to 40 H2O.

o Pinion pitch diameter: 45 mm (approx)

o Mating gear pitch diameter: 76 mm (bottom of arms)
o Radial load: 690 N
o Torque: 15 N-m
o Motor oscillation: Approximately 30° back and forth
o Two week operation minimum: 1 million cycles, 100% duty
o Back-drivable
o 188:1 gearbox
o Quadrature encoder integrated (ours provides 7 pulses / rev of motor shaft)

Caution: In deployed use, the motor must be able to operate continuously for several
days, 100% duty cycle. This may require larger motors than expected or increased
motor cooling to prevent overheating.

Mechanical Design

This is a prototype. We are in process of testing and refining this concept design to
increase robustness. The basic concept consists of two arms that gently close in
sync to compress the bag. This must be coupled with a closed loop control system.
Major mechanical design requirements included:

o Be nice to your bag and its hoses – Up to 2x7x24X60x30 = 604,800 cycles will be
needed for 7 day usage. Any design must secure the bag and gently grasp and
squeeze it from both sides to reduce the risk of material fatigue.
o Fail-Safe operation – If the machine fails, a clinician must be able to immediately shut
down, remove the bag and convert to manual bagging.
o Keep It Simple – Empower and support others to fabricate. Hence we are focusing
on the lowest specification system and open-souring our design information for
adaptation to local supply chains.
o Multiple drive motor and sensing possibilities! Enable multiple configurations to meet
local avaialbilty.

The Units version 002 shown here were developed to maximize flexibility research
and testing, specifically so that the components could be changed and the bag and
motor position adjusted. None of this adjustability is needed once a particular motor
and bag size is selected. These units continue to serve bench testing needs.
The basic dimensions are now set and any skilled mechanical designer will be able
to execute this design and adjust it to suit locally available materials and fabrication
technologies. We have ready access to waterjet and laser cutters and 80/20
components, however we are now focusing on designs that can be CNC milled,
stamped, molded, welded and bolted as per your supply chain and capability.

Key Design Notes

o The motor bridge consists of two pieces and structurally supports the two side pieces
o The plastic supports see minimal load and can be replaced with metal.
o The grippers must be rounded and smoothed to prevent long term bag abrasion.
o The gripper should fully grasp the bag when closed to maximise the air expelled.

Files

CAD files for the Unit 002 design and a pre-production model can be found on
our Downloads page.

This prototype is adjustable for fast bench testing. Note – these images do not show
the limit switch described below.

Notice that the bag support plates are adjustable both horizontally and vertically.
This allows different bags to be supported between the arms. Additionally, the bag is
entirely exposed from the top. This allows a clinician to immediately and intuitively
grab and remove the bag from the device and squeeze manually, in case of machine
failure. This safety feature is supported by a back-drivable gear train. This ensures
that even if the device fails with the arms closed, a clinician can pull the arms apart
by hand to remove the bag.

A limit switch enables the machine to set starting position when it is powered on

Here is a close-up of the gear-train for actuating the arms.he arms are geared
together directly which forces them to move together and apart in sync. The geared
motor drives one of the arms. On the right you can see a limit switch that tells the
machine when the arms are fully open. This helps the machine locate arm position
when it turns on. The switch is mounted on channels so that the position is
adjustable.

Note: Other homing and sensing is possible as a function of the position feedback
method chosen.

Plumbing

This document describes critical design requirements of the patient breathing circuit. This
details a key dead space issue, which if not addressed, will result in a patient breathing in
expelled CO2 and deoxygenation fast with immediate adverse result. For a detailed primer
on Breathing Circuits read Mapleson’s Breathing Systems.

Normally, self-inflating manual resuscitators are directly connected to the patient’s

endotracheal tube adapter. Manual resuscitators have a “patient valve” that directs oxygen /
air gas mixture into the patient and shunts the exhaled gas out to the environment.
(See Wikipedia as a good primer on self-resuscitating manual resuscitators.) Integrated
into the end bag valve mask (BVM) are a number of critical features:

o Oxygen connection and reservoir

o Pop-off valve for safety (location not important)
o One-way valve that guides air to the patient
 o Exhalation valve (this stays closed while there is any pressure on the bag)
o PEEP valve that is installed post the exhalation valve and maintains backpressure
o Sensing port for manometer connection (we use this for our pressure sensor
connection)

Caution: Manual resuscitator bags are in no way FDA approved for use as long-term
ventilation solutions.

This page described the means by which a patient should be connected to a manual
resuscitator-based ventilator. Some considerations include:

o The ventilator must be placed as close to the patient as possible.

o Bag should be secured to ventilator to prevent an awake patient from pulling on it or
otherwise disengaging the bag from the mechanism. This is a fault condition that
should be detected by pressure sensing.
o Care must be taken to prevent rebreathing of CO2 due to long hoses. A fundamental
challenge is the location of the one way and expiratory valves, which are typically
directly integrated into the bag.
When a manual resuscitator is placed into an MIT E-Vent, or similar design, the
system cannot be placed right up against the patient’s head. In addition, patients
need to be turned intermittently for routine care and patients can thrash and move in
their beds. Even when patients are paralyzed, the paralytic may wear off at times
and we must consider how to keep the patient safe from inadvertent breathing circuit
disconnection or extubation. Therefore, a safe method to extend the “reach” and
flexibility of the manual resuscitator to a patient lying on a hospital bed is needed. If a
simple tube is used to do so, it creates a critical safety concern of “dead space.”

Note: In a 1 m long tube of nominal 2 cm diameter, there is an unacceptable 314 mL

dead space that the patient will breath in and out and not be oxygenated.

Dead space simply means volume in the respiratory circuit that does not participate
in gas exchange in the lungs. Our natural anatomy has dead space as well.
Considering gas exchange occurs at the alveoli in our lungs, every anatomical
structure above it can be considered “dead space”: nasal/oral passages, pharynx,
larynx, trachea, and primary/secondary/tertiary bronchi. Extending the tubing
through which bidirectional flow of inhaled / exhaled gas mixture occurs only
increases dead space and the patient, see Wikipedia.

A way to move the patient valve of the manual resuscitator closer to the patient is
critical in solving this issue. Standard ventilator circuits have two limbs, one for
inspiration and one for expiration, so that gases can be recaptured by the ventilator.
Single limb ventilator circuits with a patient valve located distally already exist on the
market, but are not necessarily optimized for use with a manual resuscitator.

Note: Solving this problem requires creativity – No manual resuscitator manufacturer

makes an approved solution and no manufacturer makes all the parts that will
assemble together correctly.
Industry Notes

In reviewing products available on the market, we have some notes:

o No bag makers supply extension hoses with the appropriate fittings.

o Bags designed for reuse, i.e. autoclavable, are the only bags that can potentially
survive under repeated use. We do not have any information about lifespan.
o Some Ambu bags do not have detachable heads, but they do incorporate pop-off and
easily attached PEEP valves in their designs. They can only be used if extended with
a separate head and extension tube. Ambu Mark 5 and Silicon Oval heads are
detachable and may be available as parts. They have easily combined manometers
and PEEP valves.
o Laerdal bags do have detachable heads, however in the adult sizes these heads do
NOT come with pop-off valves; these are available in the Pediatric model. The
pediatric model head will probably fit the adult bag.
o When a long tube is used, without a dual limb circuit and one way valving to address
the dead space issue, this may affect the volume delivered to the patient; it may be
necessary to increase the inspired volume.
o Addition of the HEPA filter will cause a pressure drop and may affect PEEP settings.
o Tightness of all connections is important.
Caution: In the worst-case scenario, placing the head as close as possible to the
patient will reduce the dead space, but it is not an optimal or safe solution, especially
for patients with reduced inspiratory volume.

Two version of single limb circuits are shown below. The first made use of readily
available components and two printed adaptors to make them fit together, with the
HEPA filter placed between the exhalation port and the PEEP. The second used a
single limb breathing circuit, with most of the necessary features integrated, and a
HEPA filter added inline between the porcine and the breathing circuit. This is a
better position as it filters air heading both in and out of the patient, including any air
that escapes from the pop-off valve. It may also help to moisten air inbound to the
patient.

The best location, if you have only one filter, for it is between the endotracheal tube
and the breathing circuit.
Dead Air – Additional Potential Solution

Our friends at the US Air project, an Oregon based group of fabricators currently
scaling up, have dissembled Ambu bags and developed an adaptor, that
demonstrates the feasibility of DIY circuits with a little feasibility.

Ambu has readily available complete patient valve assemblies, though limited
information is available on their website. Ambu has readily available complete
patient valve assemblies with complete valve assemblies are available for the Mk IV
and Silicone Oval lines:
Mk IV Ped (w/ pop-off and Mport) #299 000 508
Silicon Oval (w pop-off and Mport) #470 000 503
The ideal solution is to order these valves, with a hasty solution being “harvesting”
valves from disposable SPUR II units. Note: Valve is not replaceable from the bag
once removed. This method places the pressure port in the correct position to
monitor patient airway pressure.

Connecting these valves can be accomplished with a adaptor. All Ambu valve
assemblies have a conical taper from injection molding that allows for a reliable
adapter connector to be quickly fabricated. This adaptor can be 3D printed (fast, but
sub-optimal), CNC milled with a sub 1 min cycle (in 6061) or injection molded
(ProtoMold quote of $1.63 in 5000 volumes, in HDPE). A hose can now be used to
connect this valve to an intact manual resuscitator. The extra valve assembly on the
intact manual resuscitator should be inactive for practical purposes.

Caution: Care must be taken to ensure that connections are secure. Medical grade
adhesive should be used.
Electrical Hardware
This page provides a description of the minimum hardware set required to control the
ventilator as described in the other documentation.

Motor & Encoder

The mechanical system should be driven with a motor under closed loop control. For
feedback measurement, we are using a DC gear motor with integrated quadrature
feedback. This is the safest solution.

The beginning point for motor selection should be the Power Calculation based on
the force needed to compress a bag.

Motor selection will then vary highly as a function of your geometry and we provide
information on Motor Selection, based on our design.

Power Supply

Nominally, a power supply that can deliver 12 V and 5 A is expected to work. Rapid
deceleration of the motor causes supply-voltage spikes and must be avoided via
correct motor motion profile design. Capacitors should be added across the H-Bridge
power leads for extra protection.

An alternate power supply would be to use a car battery with a 5 A car battery
charger connected. This will allow for very high instantaneous power draws and
voltage spike absorption. The battery will double as a ~2-3 hour backup in case of
building power loss.
Controller

Microcontroller for timing, measuring, and actuating: Arduino Uno – readily

available and easy to program, with extensive online support and documentation. It
provides 6 A/D pins (for potentiometers and pressure sensors) and 13 digital I/O
pins, including dedicated hardware interrupts (for encoders) and PWM pins for H-
bridge driving. Other industry validated controllers can be used, and we may
implement them in the future.

Motor Driver

Use any motor driver with sufficient voltage and current ratings to meet the motor
power specifications. Closed loop servo controllers can also be employed. Our
control strategy is PWM with a H Bridge. For fastest implementation using cheap,
off-the-shelf parts, we recommend an Arduino compatible motor shield. More
industrial versions will be safer. Main recommended motor controller specifications:

o Speed plus position command input that is robust

o Accepts quadrature encoder input for closed loop feedback
o Has integaretd safety features, including but not limited to, temperature sensing and
current limiting that can be communicated to the controller. (What action a fault
condition should initiate is not yet defined, ideally this will throw an alarm and
message.)
o Capable of handling up to 15 A. (This is to handle random spikes; you must conduct
testing to determine your requirement.)

Caution: Motor controllers can create significant heat inside an electronics enclosure,
be sure to consider cooling. Solutions include active cooling with a fan, connection to
a protruding heat sink or connection to a large metal chassis.

For reference, we are using a RoboClaw Solo motor controller to control a

brushed DC gear motor. The RoboClaw firmware uses a velocity PID controller and
a position PID controller to command the motor to a desired position at a desired
velocity. The PID values must be tuned in advance. Other motor controllers with
similar functions will work, we do not recommend any specific controller.

Inputs

Pressure sensor – Receives a voltage proportional to the pressure in the patient’s

lungs. Used to determine max pressure reached during inspiration, and to trigger
when the patient is attempting to breathe in during assist mode. Minimum pressure
sensor selection:

o Differential (to sense negative pressures)

o Range of up to 100 cm H2O. This is a 2x safety factor.
o Accuracy on the order of 0.5 cm H2O

Note on Plumbing: The pressure sensor must be connected to the Ambu bag’s
sensing port or somewhere in the airflow, as close as possible to the patient, past
any valves.
Control knob potentiometers (POTs) should all be single turn, 10 KΩ. Single turn is
to allow for specific settings to be marked on the face plate.

POT 1 – Varies inspired volume, sets angular oscillation of the arms. During
operation, each arm varies by a maximum of approximately 20°, corresponding to
fully squeezing a large bag. This dial varies position from 0% (fully open) to 100%
(fully compressed).

POT 2 – Varies the BPM. This sets the rate from 0 to the maximum BPM given in the
clinical document.

POT 3 – Varies the I:E ratio. Range as given in the clinical document.

Note: It is not essential that this be settable, one value greater than 1:1 can be
selected and the POT repurposed as a threshold for over pressure. (Multiple
clinicians have indicated that varying I:E is not critical.)

POT 4 – Sets the pressure threshold for detecting assist control. This varies as
described in the clinical document.

Switch – Power on / off

E-stop – Instantly deactivates the system. This can be the main power switch, but a
single pish must fully depower the system. This will allow the bag to be removed and
immediate conversion to manual bagging in the case of any major failure.

Toggle Switch – Mode selection from volume to assist control.

Momentary Button 1 – Used to temporarily silence alarms. This must be debounced.

Momentary Button 1 – Used to confirm a change to one of the POTs. This is a

necessary safety feature. This must be debounced.

Limit Switch – Used for homing the arms positions.

Output

LCD screen displays airway pressure in cm H2O. Other functions can be

incorporated later. We are using a 20×4 character LCD display as this will display the
minimum information, described in under interface. Any display better than this will
be sufficient.

Audible alert buzzer will identify multiple fault conditions.

Controls
Notice: Any control strategy must fulfill the requirements described in Key Ventilation
Specifications.
The goal of the high-level controller is to provide a controlled volume of air to the patient in a
set amount of time. There are two control phases: the inspiratory phase and the expiratory
phase (see Figure 1). There are three input parameters, referencing the Key Ventilation
Specifications:

Parameters

Tidal Volume (TV): The total volume of air to be delivered to the patient.

BPM: Breaths per minute, also called respiratory rate (RR). Typically varies between 8-30
BPM.

IE Ratio (IE): The ratio of the duration of the inhale to the duration of the exhale. For example,
a 1:3 ratio means that the exhale phase lasts three times longer than the inhale phase.
Typically varies between 1:1 to 1:3, with a maximum of 1:4 currently being observed in COVID-
19 patients.

In addition to these inputs set by the clinician, the high-level controller uses two more inputs:
the motor encoder position and the system pressure. Its job is then to translate all these inputs
into the motor commands the low-level controller needs: desired motor speed and position.

Note that because our device does not directly measure volume, the tidal volume (VT) input
of our controller is specified as a percent of a full compression of the Ambu bag instead of
Liters. The percent (%) of bag compression from 0 – 100% maps to the encoder pulses that
correspond to how far the fingers of the device move towards or away from each other and
this determines the volume of air delivered.

Figure 1 – Flow, Pressure, and Volume profiles for volume-control ventilation over 2 breath
cycles; PEEP is illustrated on the Pressure plot. Image courtesy AK.

Our controller converts the three inputs into four parameters:

Period (T): The length of time (in seconds) of an inhale/exhale cycle.

T = 60 / BPM

Tin: The length of time (in seconds) of the inspiratory phase.

Tin = T / (1 + IE)

Tex: The length of time (in seconds) of the expiratory phase.

Tex = T – Tin

Vin: The rotation rate of the inspiratory phase (in pulses/second).

Vin = VT / Tin

In addition to these four parameters, there are three user-set parameters:

Th: The amount of time (in seconds) to hold the compression at the end of the inhale for
plateau pressure.

Ve: The velocity of the fingers in the expiratory phase (in pulses/second). Note that during
exhalation, our device does not control flow rate out of the patient. This velocity is simply the
velocity of the fingers opening and is not related to expiratory flow rate.

Pmax: The maximum allowable pressure (set to 40 cmH20).

These six parameters are then used to control a state machine that switches between
phases in the control loop (see Figure 2).

Time t is the amount of time spent in the current state.

In addition to these three inputs, there are three measurable pressure parameters that must
be taken into account (see Figure 2):

Pip: Maximum pressure during inhale. We consider 40 cmH20 to be the upper pressure limit
for safety. This also corresponds to the over-pressure release valve limit on some Ambu
bags.
Pplat: The plateau pressure of the inhale. An important diagnostic number for clinicians.

PEEP: The residual pressure in the system after exhale. We do not directly control this
value, but it is typically controlled manually via a PEEP valve on the Ambu Bag.

State Machine Summary

Figure 2 – State machine for controlling the breathing cycle.

During the setup phase, we initialize the program, start serial communication with the
motor controller, and home the encoder.

Figure 3 – We home the encoder using a limit switch.

In the inspiratory phase, we command the motor to go to position VT at velocity Vin.

After Tin seconds, we switch to the pause state.

In the pause state, we hold for time Th and measure the plateau pressure. We then
switch to the expiratory phase.
In the expiratory phase, we command the motor to go to position 0 at velocity Ve. After time
Tex, we switch back to the inspiratory phase.

Plateau Pressure

Each time the arms close, we implement a 0.15 s pause before they open. This does not affect
the I/E ratio, but it is necessary to hold the air into the patient. During this phase the airway
pressure is measured and displayed. This indicates “plateau pressure” and will guide clinical
decision making. This pressure will be displayed until the next cycle and update. Other
pressures are less important but can be addressed in a more complicated control strategy.

Alarms Functions

Caution – Not yet fully implemented.

All alarms must simply, concisely, and clearly alert the clinician of the type of fault, so that the
clinician can decide how to proceed. For example, a mechanical fault requires a different
clinical response than a patient who stops breathing on assist mode.

There are several different failure modes in volume control, which require different responses.
For example, a compression may cause a lung overpressure fault, or the position setpoint
may not be able to be reached. There may be a leak in the system, e.g. a failed bag or
disconnected hose.

Alarms specifically related to pressure sensing are describe in Pressure Measuring.

Assist Control

Figure 3 – Assist Control

Assist Control differs from regular volume control in that the Exhale state is split into 3 states.
In the first Exhale state, the fingers move to their home position at the edge of the bag. In the
second state, Exhale Pause, the fingers pause for a short time and measure the PEEP. In the
third state, Listen, we wait either for the patient’s own inhalation to trigger the Inhale state, or
we wait for a set amount of time (like in normal Volume Control) and then trigger the Inhale
state automatically. Read the Clinical page for more information.

In the Assist Control mode, a notification (not necessarily a full alarm) should also sound
whenever a breath is NOT activated by the patient and the system’s timer kicks in to command
a breath. Other faults must be detected, such as mechanical failures to reach desired
positions, etc.

Circuit Diagram

Notice: Any circuit must fulfill the requirements described in Key Ventilation Specifications.

Arduino-Based Prototype Circuit

Caution: This circuit does not contain the alarms necessary for alerting clinicians of failures!
This circuit diagram does not contain the 12V supply, the motor, the encoder, or the buzzers
necessary. Additionally, mass-manufactured designs should be made without development
boards such as Arduinos and motor shields. This diagram is for reference only.
In this design, the logic is fed from a 5V linear regulator from the 12V supply. This 5V
feeds both the Arduino, LCD Screen, and Encoder power. The key functions that this
circuit must fulfill are described in our Key Ventilation Specs page.

Operation

Note: While the mechanical hardware design is changing rapidly and some of the
underlying electronics are being upgraded for prototype purposes. the operating
principles are relatively design frozen.

Figure 1. The two main components: the control box and the ventilator at a testing facility.

Scenario

The system will initially be setup by a clinician explicitly skilled in respiratory management.
The parameters must be tuned and the patient monitored carefully. Once stable, they can be
left in the care of a clinician not skilled in respiratory management. In the case of any change
in patient condition or an alarm, the system must again be tuned by a skilled clinician. In the
case of any serious concern, the bag can be manually removed from the machine and
squeezed by hand until the fault is addressed.
Figure 2. Control the ventilator and read pressure information from the top of the control box.
Figure 3. There are 6 pieces of information displayed on the LCD screen which update after
each breath cycle. Alarm messages will display on the top bar.

Figure 4. Ensure that all wires are connected and that the power switch is in the off position.

Figure 5. The ventilator fingers in ward motion controls the inflation and deflation of the bag.
The clear sheets on either side of the fingers are for mounting the bag. These are replaced
with adjustable metal mounts.
Figure 6. Correct bag placement is centered between the fingers vertically and axially.
Confirm placement by manually moving driving the fingers to lightly touch the bag. Supports
should be adjusted vertically and laterally to support the bag in this operating position.

Operation

1. Gently open the fingers by hand if not already open. This takes some force, but it is
safe.
2. Position the Ambu Bag in the cradle between the fingers, following the directions
given in Figure 6.
3. Ensure that all dials shown in Figure 2 are turned counter clockwise to the lowest
position.
4. Check that nothing is in the way of the fingers.
5. Power up the system. The system will start moving immediately and home to the fully
open position. Then it will slowly move to the edge of the bag. (This position is hard
coded; it could be settable in a more sophisticated version.)
 6. Set the desired Respiratory Rate, Tidal Volume and I:E knobs, as shown in Figure 2,
and confirm the values on the display, shown in Figure 3. Press the Set button to
apply.
7. Increase the Tidal Volume to a low setting (as determined by clinician) and press
the Set button to apply. The system will start pumping. Confirm correct operation.
8. Once the machine is confirmed pumping, connect to patient and increase the Tidal
Volume to the desired larger value and press the Set button.
9. Monitor the peak and plateau pressures and adjust parameters as per clinical
guidance. PEEP pressure should be observed to match the setting on the PEEP
valve.
10. Do not leave the patient unattended.
11. Monitor vital signs, listen for alarms and respond.
12. If Assist Control mode is desired, increase the Threshold knob and press
the Set button to apply. This will increase the setpoint with respect to the PEEP.
13. Respiratory Rate should beset to less than the expected patient respiratory rate, i.e.
the machine waits longer than the patient would.

Note: Spinning potentiometers (POTs) will cause the display valued to change, but
will not change machine behavior until the Set button is pressed. If the Set button is
not pressed within 5 seconds to apply the new settings system will alarm.

Caution: If a POT is changed, but Set is not pressed, a potentially confusing situation
will arise whereby the machine is different from knob settings. Potential solutions
better than the alarm include: 1. Shielding the potentiometers to prevent accidental
turning. 2. Using a potentiometer that must be pulled out to set. 3. A rotary encoder
with a pushbutton; this will allow the button to be activated, set and deactivated.
More sophisticated hardware and software solutions can be implemented.

Alarms

In the case of an alarm, the system should attempt to maintain operation. The
Silence button will pause the alarm for 1 minute. Alarm conditions are checked for
once per cycle.
Caution: Not all of these alarms are fully implemented. The logic is still being
developed. This is not an exhaustive lists of failure conditions.