Pressure transducer question

CPAP, BiPAP, and other sleep systems

Moderators: SCNVsleep, labman2

Postby bduce » Tue May 19, 2009 9:11 am

DBradley & RBonato
Do you guys have any references for the phenomena describing the qualities of the silicon and piezo based pressure sensors? I really including them in the chapters of my tech training manuals.
User avatar
bduce
Senior Member
 
Posts: 1058
Joined: Fri Oct 20, 2006 10:27 pm
Location: Brisbane, Australia

Postby DBradley » Wed May 20, 2009 3:31 pm

dbuce: Information on input amplitude response and frequency response of materials is not easy to find. Most suppliers of these types of materials keep the manufactures under strict confidentiality agreements. Thus information for the public is not readily available. With that said some careful digging on the WWW will yield some information that you could use.

I gave a talk just over a month ago on respiratory/airflow sensor technology (thermal airflow sensors(thermocouples/thermistors), pressure sensors(piezo/silicon) and respiratory effort sensors(piezo/RIP)), to over 350 people at the Canadian Sleep Society meeting. Several people came to me after the talk and said that they were not aware of the differences in pressure sensors and that they now look at sensors in a different light. I have not seen this type of information provided in any form to the sleep community. Users of sensors need to understand their limitations and how to chose which is the best sensor for the type of application.

The information contained in my talk came from our own in-house testing as well as our experience in the industry. I would be willing to share some of the slides from my talk. This may give you the information that you require. PM me on what exactly you are looking for and I will see what we can supply.

Regards,
Don
DBradley
Member
 
Posts: 42
Joined: Thu Sep 27, 2007 1:35 am

Postby slowdavesleep » Mon Sep 27, 2010 4:34 pm

Reviving this thread! So I was reading this:
http://hyperphysics.phy-astr.gsu.edu/hbase/pfric.html

If the formula for deriving flow rate from pressure across resistance is as stated in this article where does the need for s square root transform in order to linearalize the flow signal come from?

I know this is more directly related to pneumotachs, however in a nasal pressure transducer the resistor is the upper airway / nose and the two pressure measurements are 1. the cannula 2. atmospheric pressure?

Is that the case?
User avatar
slowdavesleep
Senior Member
 
Posts: 400
Joined: Mon Mar 03, 2008 7:41 pm
Location: RPSGT NY

Postby linuxgeek » Mon Sep 27, 2010 7:05 pm

My understanding is that pressure has a square relationship with respect to flow, unless the flow can be made laminar.

This is also my experience with fiddling around with flow circuits.

As you point out with pneumotachs, there is a method of measuring flow that involves measuring differential pressure across a resistance. In other words, you measure pressure immediately prior to the resistance, and immediately following the resistance. You need to take the square root of that differential pressure to get a linear relationship to flow. If you simultaneously make the flow laminar (typically induces some resistance) you don't need to do the square-root correction.

You need to encompass all the flow to take those measurements. In a cannula system, you aren't encompassing all the flow at both pressure
points. But that's not why it has a square-relationship, it's that the flow is not laminar.

That's how I understand it, but I could be wrong.

To confuse the issue a little more, a cannula system is essentially measuring stagnation pressure.

http://www.princeton.edu/~asmits/Bicycl ... oulli.html
User avatar
linuxgeek
Senior Member
 
Posts: 1604
Joined: Wed Jan 23, 2008 11:29 pm

Postby Rick » Mon Sep 27, 2010 9:44 pm

linux, read just a little farther on and you will see that for exhalation it is a pitot tube, so for half the breathing, no worries about that Bernoulli or Venturi stuff.

I could be wrong but I believe the reasons for the square root transform are relational to the pressure transducer itself and that it does not produce truly linear signals and the square root transform is one way folks tried to linearize those signals. Note: "not truly linear" is in reference to the method of using the bulging of the transducers pressure chamber that changes the values of the piezo resistors in the Wheatstone bridge etched onto the pressure chamber itself to produce the signals of pressure flow.
User avatar
Rick
Senior Member
 
Posts: 1384
Joined: Wed Mar 19, 2003 8:42 pm
Location: Everett, WA

Postby linuxgeek » Mon Sep 27, 2010 11:23 pm

I think that you are probably right about exhalation being different than inhalation and closer to a pitot tube. But, I don't think it's the same as a pitot tube as your reference pressure isn't perpendicular to the direction of flow. Instead it's somewhere far away only sensing room pressure.

I'm pretty sure that the square relationship is due to physics and not to the nature of sensors. It's a square relationship on every pressure sensor I've ever tried, no matter the type. And I've verified that the response to pressure is linear with a manometer, but not with flow.

In fact, I think that you still have to do a square-root correction for pitot tubes, solving for flow. Unless you've managed to produce laminar flow.
User avatar
linuxgeek
Senior Member
 
Posts: 1604
Joined: Wed Jan 23, 2008 11:29 pm

Postby slowdavesleep » Tue Sep 28, 2010 11:20 pm

Do the millar catheters have holes on the end or is the sensor flush with the tip? How would that change the dynamic of a pressure transducer if there was no tube involved?
User avatar
slowdavesleep
Senior Member
 
Posts: 400
Joined: Mon Mar 03, 2008 7:41 pm
Location: RPSGT NY

Postby DBradley » Wed Sep 29, 2010 1:19 am

I have been watching this thread and thought that I would put my team's 2-cents in.

I took slowdavesleep's question and ran it past our Ph.D aeronautical engineer. This was his response.

[Assuming you are referring to the “Flow Resistance for a Tube”? P1-P2 is the static pressure drop between points 1 and 2 along the tube, so you need measurements at two different locations. With a cannula, resistance doesn’t enter into it. You have (essentially) two measurements at a single location -- the stagnation pressure of the flow, and the static pressure (which is the local atmospheric pressure). The difference between them (which is what is actually measured directly because it is a differential transducer referenced to atmosphere) is the dynamic pressure, ½ρv2. Because the velocity and flow rate are proportional, the flow rate is proportional to the square root of the measured pressure (the density can be considered constant for flows less than ~100 m/s).]

Just too clear things up I would add that there are a couple of important points here that need to be made to remove any confusion. Firstly you need to be very careful in taking an example case and applying it in our field. These equations require a laminar flow to be accurate. This requires a fairly long length of tube before you take the first pressure measurements and another long length after the second measurement for any accuracy.

Secondly, using a pitot tube scenario you are in fact converting the kinetic energy in the flow to potential energy (the pressure measurement). This occurs at the stagnation point and is transmitted down to the differential pressure sensor. Notice that equation given above is very similar to the actual equation for kinetic energy of a mass (Ek = 1/2mV2). To determine the actual velocity you must take the square root. Since in our field we do not require a quantitative value for flow we can just take the square root of the measured differential pressure with respect to atmosphere and take the square root. (note: this does require the use of a differential pressure sensor that has a linear output(pressure is linearly proportional to the voltage), a piezo based pressure sensor cannot be used with any hope of accuracy.)

I hope this clears up some of the confusion.

Be careful on taking an example and applying it our field. Make sure you fully understand all the physics involved.
DBradley
Member
 
Posts: 42
Joined: Thu Sep 27, 2007 1:35 am

Postby somnonaut » Wed Sep 29, 2010 2:29 am

You have a Ph.D aeronautical engineer on staff?
User avatar
somnonaut
Senior Member
 
Posts: 14154
Joined: Thu Dec 11, 2003 6:03 pm
Location: http://tinyurl.com/righteousdude

Postby linuxgeek » Wed Sep 29, 2010 4:20 am

Ok, I see that a cannula is essentially measuring dynamic pressure (stagnation pressure - static pressure).

In my testing, if I measure differential pressure across a resistance in a tube, I get a signal that is the square of flow. If I put in a screen (like the Hans Rudolph pneumotach apparatus) and I take the same measure in the section where flow is laminar, I get a signal that is linear to flow. BTW, that apparatus is not very long, it seems to do it with fine screens, but maybe it produces more resistance that way. But, I'm guessing you probably mean "long" for the ratio of the length to the cross-sectional area. And in that case it the diameter would be of each divided section of the screen, which is rather small.

What I don't understand, is why flow is not already laminar as I thought it had to do with the Reynolds number and I think the tube dimensions and flow rate satisfy that already. This is where I'm still kinda foggy on what's going on. But, this is why I test things, so I don't have to rely on my limited understanding.

Anyways, I know this is probably getting a bit esoteric but it was good to hear the explanations.

I'd like to add that a cannula has to be positioned just right (parallel to flow) to be a measure of stagnation pressure. Probably still pretty close, but I've seen some weird things happen when the cannula bends so that it is slanted to the direction of flow.
User avatar
linuxgeek
Senior Member
 
Posts: 1604
Joined: Wed Jan 23, 2008 11:29 pm

Postby slowdavesleep » Wed Sep 29, 2010 2:03 pm

Still trying to muddle my way to some acceptable level of understanding of all these posts. Extremely interesting. Related article:

http://jap.physiology.org/cgi/reprint/71/6/2317.pdf
I need a lot more background reading to really get a god grasp on this.
User avatar
slowdavesleep
Senior Member
 
Posts: 400
Joined: Mon Mar 03, 2008 7:41 pm
Location: RPSGT NY

Postby DBradley » Wed Sep 29, 2010 2:52 pm

For an obstruction flow meter to work accurately you typically require a length of tubing with constant diameter of 5 times the diameter before the first pressure measurement and a length of tubing with constant diameter of 3 times the diameter. You may also want to look at obstruction meters. They typically create a larger pressure range. This is the same concept as putting the screen in the path of airflow. However you normally want to measure the pressure as close as possible to the obstruction.

In the case shown on the webpage you could take a ¼” diameter tube, make sure that you sealed the tube in the nostril so there were no leaks and have two pressure port taps. The tube would need to be at least 150 meters in length. So spend thousands of dollars on extremely sensitive pressure sensors to shorten the length a bit but you will still end up with an unusable device.

Linuxgeek make a very good point on that the opening of the nasal prongs should be perpendicular to the flow of air to accurately measure the stagnation pressure. Also the normal to the opening of the nasal prongs should be parallel to the flow of air.

I typically inform techs to do the following three things for optimum pressure measurement.
1. Make sure the patient has cleared their nose and that nose hairs have been trimmed back. (i.e. remove as many obstructions as you can) This will also improve comfort for the patient.
2. Trim the nasal prongs so that their opening is perpendicular to the flow of air. This will also improve comfort.
3. Tape the cannula as close to the nares as possible. Also when doing this try to center the nasal prongs within the nasal openings.

As to laminar flow, I typically think of laminar flow as all air moving in the same direction. No turbulence or eddies in the flow. A low Reynolds number will normally indicate a laminar flow. However the Reynolds number is a ratio of forces acting within the medium. This has many factors. So you must stabilize the factors to produce high viscous forces and low internal forces to create a laminar flow. Increasing the length of the tube will reduce internal forces as the flow of air stabilizes over the length of travel.

I hope this has helped.

Don
DBradley
Member
 
Posts: 42
Joined: Thu Sep 27, 2007 1:35 am

Previous

Return to Other Equipment

Who is online

Users browsing this forum: No registered users and 1 guest

cron