Search This Blog

Follow by Email

Saturday, September 26, 2009

A High Frequency Oscillatory Ventilator Design For Use in Pandemics

This is a continuation of some of the ideas I expressed in a previous post on HFOV design. You can see it here:

Some of the news items I have been reading say that there will be a particular shortage of high frequency oscillatory ventilators. Many ICU units do not have any or may only have one. This type of ventilator is required to care for patients with the most damaged lungs. Here is a little more information on a design for an HFOV. This is very preliminary. It surely needs more work. Someone would have to to build and test a prototype to determine if it is feasable.

What is HFOV?

An HFOV (High Frequency Oscillatory Ventilator) is an advanced ventilator design that is sometimes used in ARDS (Acute Respiratory Distress Syndrome) patients when a conventional ventilator will no longer provide adequate ventilation. Using a HFOV is considered a “lung sparing” technique.

When using conventional ventilators, the ventilation levels can be increased by increasing the percentage of oxygen fed to the ventilator, increasing the stroke volume, or increasing the rate or frequency. Other measures to improve ventilation can be increased PEEP (Positive Expiratory End Pressure) levels, reversed I/E (Inspiration/Expiration) ratios, methods to increase the average airway pressure, or even PLV (Partial Liquid Ventilation) has been tried, where a perflourocarbon solution (perfluorooctyl bromide) is put into the lungs to reduce lung damage without reducing oxygen transfer. ECMO (Extra Corporeal Membrane Oxygenation) is another method that can prevent damage to lungs by the use of high pressures and oxygen levels in conventional ventilation.

The HFOV can maintain a fairly high mean airway pressure, resulting in better ventilation without causing as much lung damage. With the high rates required for HFOV, each breath is less than the dead space in the lungs. There are various mechanisms that explain how it works (, but it does work quite effectively. The breathing mechanism is similar to a dog panting at a high rate.

How HFOV Works

To make a HFOV work, you need to have a system that maintains a set average airway pressure and then have another device that oscillates this column of air at a desired rate, amplitude and I/E ratio. That’s the fundamentals of what it does.

Now to optimize this design you need to have a gas management system that controls the oxygen level, temperature, humidity and inlet pressure and flows to the HFOV device. We do not have to worry about designing these parts. They are all standardized respiratory equipment that is also used on conventional ventilators.

Other design considerations are that the air flow goes through the tubing in such a way as to optimize gas exchange, and we will also need other alarms to warn us of low or high pressures, improper rates and loss of supply gas. We may also want additional alarm systems that warn us of equipment failure modes. The tubing should be relatively non-compliant and the system should have minimal dead space.

Fig 1. HFOV design from IEEE Transactions on Biomedical Engineering
this link.

This design would be based mostly on the diagram in Fig 1. The controls and operator interface could be modeled to be similar to the 3100B from Sensormedics. The 3100B is the most commonly used HFOV for adults. Many RTs are already trained in its operation. By making the controls and alarms similar to the 3100B, it could be more easily deployed in a pandemic situation.

This is how the device in Fig 1 works:

  • A filtered, humidified air/oxygen mixture is fed into the feed tube near the ET (Endo Tracheal) tube. The flow rate is monitored and controlled by the mass flow meter
  • It travels down the tube towards the oscillator unit and exits via the servo controlled restriction valve.
  • The pressure sensor is that thing on the tube between the inlet and outlet ports. The electronics control system will receive this pressure signal and adjust the servo controlled restriction valve so that the average airway pressure is equal to the desired set point.
  • The pneumotach is not really required for operation. They have to do measurements for their study. Vacuum is not really required either, as average airway pressures will always be positive.
  • The oscillator is that plunger looking thing on the right hand side. It looks and works like a speaker in the 3100B but is really a special purpose built device. They call it the driver.
  • The plunger moves in and out at the desired rate, wave shape and amplitude as determined by the driver circuitry and the operator settings.
  • Now, you can see, that column of air is going to push and pull air in and out of the ET, which goes into the lungs. When the air comes out of the ET, the fresh bias flow gas will flush it away and out toward the servo controlled restriction valve. Fresh bias air is pushed into the lungs when the plunger moves toward the ET tube.
  • The oscillations of the plunger will change the instantaneous pressure in the tube positive and negative with respect to the average pressure.
  • The position feedback device improves the performance of the oscillator circuit and can also be used as a part of a safety system

Parts And Controls

I will group parts into 3 general categories.

  • Oscillator driver and driver circuit.
  • Sensor, actuator and associated circuitry.
  • Control and display system.

Oscillator Driver and Driver Circuit

For the oscillator driver we would want to use a big, high power subwoofer type speaker. It should be tough and able to handle high duty cycles and long periods of operation. It should have a metal cone to make it inflexible. We might have to glue a metal plate to the cone to make it more rigid. In order to reduce the dead space, we could make a mold of the front of the cone surface in resin or silicone with an air access hole drilled in the center to mate with the speaker cone assembly.

I am not certain how well a speaker will work though. There must be technical reasons why the designers of the 3100 use that design. Speakers, even subwoofers have a certain compliance and harmonic resonance built into them that is at a higher frequency than the rates we would need to use. It may require a large speaker using only a small portion of it's maximum designed excursion in order to minimize the effects on the output airflow pattern caused by the damping effects of the speaker cone suspension.

I am not sure if a position sensor is absolutely required, but something can be attached to the back of the cone if it is. Cooling may also be required. We could use lots of air and fans, or perhaps an active system using peltier devices.

The drive circuit would be a high output audio amplifier. I think it is best to use one designed for automotive use. These are generally more rugged, modular and can easily run on a 12 volt battery for electrical backup purposes.

Sensor, Actuator and Associated Circuitry

The sensor and circuitry would be similar to the one my son Jeff used in his ventilator design (Norman). It would convert the pressure pulse to a digital value encoded and sent on an RS232 port. We may wish it use more than one pressure sensor in order to provide redundancy for safety reasons. The pressure controller and alarm board would be a servo controlled valve and driver circuitry that operates by RS232. It could also house the audio alarm. This alarm would also engage and cause the valve to open if communications were lost. These circuits would have to be hand built unless there is a commercially available alternative.

Control and Display System

The control and display system would be a computer. It would probably be a PC and probably a laptop. A laptop has its own integral battery backup system. A program such as Labview can be run to show a display that looks similar to the control interface from the 3100B.

The instantaneous pressure readings received from the sensors could be integrated over time for display. The minimum and maximum pressures would be the peak recurring pressure extremes integrated over a short time interval. The average pressure would be integrated over a longer time period.

Operating parameters could be entered by selecting the appropriate box on the screen and entering the parameter via the keyboard. Alarms could also be displayed and color coded.

The computer would also output an audio signal to the speaker amplifier. This wave shape is normally a square wave pulse produced by a pulse circuit with variable duty cycle, frequency and amplitude in the 3100B. In our machine, we would have full control of the waveform via software. We could add pre-emphasis and custom wave shaping to the output to compensate for physical design shortcomings in the speaker and driver circuit or shortcomings in the housing and tube.

Here are some resources if you want to learn more about HFOV.

Guidlines for the use of HFOV

HFOV guidelines from Stanford Hospital and Clinics

The use of HFOV in surgical patients.

Slide show of HFOV in the adult patient.

University of Virginia experience with HFOV.

ARDS and HFOV from Express Healthcare.

Ventilation article from

Wickipedia Article about HFOV (please improve this)

Spec sheet for the 3100B

Picture of a 3100 HFOV

Competency exam for 3100B operators.

Video showing operation of the 3100B

Video showing initial operator calibration of the tubing set for the 3100B


  1. Awesome... I really love the way you write your article here... And also it shows facts which I like. Thanks for sharing this.

  2. Really impressive blog you got here. Thanks for your opinion...
    Heating and Cooling Mississauga

  3. This is getting looked at by who are trying to work on an open source ventilation system to help Covid-19. Please come help out.

    Makers of linear electromagnetic motor piston pumps.


Commenting has been limited for this post

Note: Only a member of this blog may post a comment.