Pressure transducer question

CPAP, BiPAP, and other sleep systems

Moderators: SCNVsleep, labman2

Pressure transducer question

Postby slowdavesleep » Wed May 13, 2009 5:38 pm

Aside from the mini piezo based units is the primary technology in sleep DC air pressure transducers potentiometers?

http://www.omega.com/Literature/Transac ... ssure.html

Been doing some reading on the site above and I was just curios.
User avatar
slowdavesleep
Senior Member
 
Posts: 400
Joined: Mon Mar 03, 2008 7:41 pm
Location: RPSGT NY

Postby linuxgeek » Wed May 13, 2009 6:23 pm

I don't know, but I'm thinking they are basically printed circuits that respond to pressure in some fashion.
User avatar
linuxgeek
Senior Member
 
Posts: 1604
Joined: Wed Jan 23, 2008 11:29 pm

Piezo pressure vs Silicon pressure

Postby RBonato » Thu May 14, 2009 9:25 am

Piezo pressure sensors use a piezo disc about the size of a nickel and respond to dynamic changes in pressure. This is the same type of disc used in piezo snore sensors, piezo resp effort bands, and perhaps ankle straps. On the other hand, true high quality pressure transucers use a silicon wafer sensing element and have built-in circuit overload protection. Increased sensitivity and signal quality are huge advantages. Moreover, the silicon technology is responsive to both static and dynamic pressure, as observed during flow limitation and normal breathing. See attached. The electronics used by manufacturers of PT units dictates end signal quality, such as linearity, and for that primary reason product quality varies according to manufacturer. A good quality PT is linear and stable both within a single 8-hour study and between recordings. Usually, there is no need to calibrate a PT, unless you are doing sophisticated recordings such as endoesophageal pressure in addition to typical oronasal cannula recordings. In this case you need to make 100% certain your unit is linear on both negative and positive pressures. Some units are not linear whatsoever and some are linear only on positive pressures. When friends or family ask me to help with snoring or apnea I will not use piezo technology in the tests I perform. Hope this helps.
Attachments
Piezo pressure sensor vs a silicon pressure sensor.pdf
(530.01 KiB) Downloaded 301 times
RBonato
Member
 
Posts: 106
Joined: Fri Dec 29, 2006 8:40 pm
Location: Earth

Postby Rick » Fri May 15, 2009 12:45 am

Below is taken from:
http://www.sensorland.com/HowPage004.html


Pressure sensing using diaphragm technology measures the
difference in pressure of the two sides of the diaphragm.
Depending upon the relevant pressure, we use the terms
ABSOLUTE, where the reference is vacuum (1st picture),
GAUGE, where the reference is atmospheric pressure
(2nd picture), or DIFFERENTIAL, where the sensor has two
ports for the measure of two different pressure.

The piezoresistive pressure sensor, or silicon cell.

This type of pressure sensor consists of a micro-machined
silicon diaphragm with piezoresistive strain gauges diffused
into it, fused to a silicon or glass backplate.

The resistors have a value of approx. 3.5 kOhm. Pressure
induced strain increases the value of the radial resistors (r),
and decreases the value of the resistors (t) transverse to
the radius. This resistance change can be high as 30%.

The resistors are connected as a Wheatstone Bridge, the
output of which is directly proportional to the pressure.


So, now I am confused, the silicon cell pressure transducer uses piezo based resistors, but as resistors, not a power supply (just had an AHA moment) which is what some say is wrong with those mini pressure transducers.

Now, as to linearity, there are articles that state the lack of linearity in pressure transducers does not have an effect on the AHI of a study. To be completely and truly linear a pressure transducer would have the square root transformation of the output signal. I don't think there is a commercially available pressure transducer for use in a sleep facility with this transformation.
User avatar
Rick
Senior Member
 
Posts: 1384
Joined: Wed Mar 19, 2003 8:42 pm
Location: Everett, WA

Postby linuxgeek » Fri May 15, 2009 1:48 am

That's interesting about the piezo based resistors.

I think there are also studies that show the AHI is affected as well.

But, truth be told, I think many labs filter the signal substantially so that it doesn't matter how accurate the device is. Interestingly, we used to have the so-called "nipple breath", the one with the peak in the middle of inspiration. Since I moved us to more quantitative measures years ago, that flow curve type has disappeared from our lab.

The AASM recommends using the square-root transformed signal, but I don't think anyone is doing this commercially. And for one to do the square-root, you need a pressure transducer that is at least linear to pressure. If your null response is not accurate (the zero value) then the hope of square-rooting the signal w/o turning the signal into garbage is out the window.
User avatar
linuxgeek
Senior Member
 
Posts: 1604
Joined: Wed Jan 23, 2008 11:29 pm

Postby DBradley » Fri May 15, 2009 1:59 am

Looks like there is confusion here. So to be very clear I thought some simple explaination is needed.

I would refer to a piezo pressure sensor as a piezo electric pressure sensor. This type of sensor does not need a battery. They only generate a signal when the material is deforming (changing). When it no longer deforms, ie upon flow limitation the sensor will generate zero voltage. Basically this type of sensor has a very poor frequency response for use as a sensor to measure breathing signals. They have a small frequency range where the sensor is actually useful.

A silicon pressure sensor is actually a piezo resistive pressure sensor. It is commonly configured in a wheatstone bridge. They do require an external power source. They will maintain their resistance at a constant strain. They have an excellent frequency response for measuring breathing.

I would never use a piezo pressure sensor (non-battery) in measuring breathing.

With respect to the square root transform. This is used to calculate a signal that is linearly proportional to the flow from the pressure signal. The flow is proportional to the square root of the pressure signal. The problem is that the square-root must be applied to the signal with the baseline removed. This is not an easy process. When using this transform you will require a linear (silicon) pressure sensor a piezo pressure sensor cannot generate a proper baseline.

I hope this clears up some issues.

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

Postby Rick » Sat May 16, 2009 12:31 am

Thank you Don.
I was mystified by the piezo based resistors when I first read about them in pressure transducers years ago and really did have a Gestaltic moment as I was posting yesterday when I realized the difference.
Linuxgeek,
The AASM does not recommend using the square root transform, they say you can use a pressure transducer with or without on an adult but must use "a nasal air pressure transducer without square root transformation of the signal" for children; Respiratory Rules for Children 1.B. page 48.
Big thanks to SlowDave for starting this thread.
User avatar
Rick
Senior Member
 
Posts: 1384
Joined: Wed Mar 19, 2003 8:42 pm
Location: Everett, WA

Postby bduce » Sat May 16, 2009 1:12 am

Linux
Embla system provides the square root transform as part of their software packages. Compumedics does if you purchase their Profusion Plus.

My impression is that from a software point of view, it is hard to have the transform work during recording (ie overnight techs can see the transformed signal in real time) but it can work during post-processing.

I am with you though. I believe that we should have it. Those papers that came out by Farre and Ballester showing the similarity of the transformed nasal pressure to the pneumotach was interesting. Unfortunately it never really took off.

For mind, the AASM should have pushed for something like that in their guidelines. Somehow. I'm sure all the companies would have scurried around to get it into their systems. Just my personal view.
User avatar
bduce
Senior Member
 
Posts: 1058
Joined: Fri Oct 20, 2006 10:27 pm
Location: Brisbane, Australia

PT

Postby goudasleep » Sat May 16, 2009 1:42 am

Okay. I've read the links, I've read the threads. I understand the basic physics for both mechanical and electronic strain gauges, or potentiometers. I understand the non-battery powered silicon-diaphragm gauge, but my first question is this: Are the silicon-diaphragm and the piezo-electric PTs one and the same? I get the slight electric current created by the piezo-electric element, is THAT the battery powered one, and the batteries are used to amplify the slight piezo electric current? Or am I completely off track and I need to read the links and the thread again to get it? I've re-read my thread, my first question is circular but y'all understand what I'm asking, I hope.
User avatar
goudasleep
Senior Member
 
Posts: 766
Joined: Mon May 16, 2005 8:37 am
Location: Tejas

Postby DBradley » Sat May 16, 2009 3:24 am

OK. I will try to keep it simple as people out there use different words for the same things.

A piezo based pressure sensor i.e. piezo-electric generates its own very small voltage. The problem with this type of sensor is that the frequency response, i.e. how the sensor outputs a signal when compared to the same input signal at different frequencies is very poor at low frequencies, i.e. breathing frequencies. That is this type of sensor cannot hold a flat signal i.e. flow limitation. The signal will decrease to zero during flow limitation. This poor frequencies response also causes issues with the baseline.

A silicon based pressure sensor i.e. piezo-resistive sensor will change its resistance depending on the strain applied. Even if the strain is constant the resistance is constant. You will see proper flow limitation with this type of sensor and have a correct baseline.

The piezo based pressure sensor does not require a battery. It is inexpensive compared to the more accurate silicon based pressure sensors.

Also I believe that the AASM does not object to the use of the square root transform on the pressure sensors. Actually the AASM scoring manual requires the use of both a nasal cannula and an oral nasal thermal sensor. This combination will allow correct scoring for apneas on the thermal signal and hypopneas on the pressure sensor.

The problem with the square-root transform is due to the fact that you must take the square-root of the signal minus the baseline (value at the start of inspiration for each breath). The flow signal is approximately proportional to the square of the pressure signal. Actually it is a quadratic equation similar to the form Flow = A x Pressure2 + B x Pressure + C.

I trust that this will help.
Don
DBradley
Member
 
Posts: 42
Joined: Thu Sep 27, 2007 1:35 am

Postby slowdavesleep » Sat May 16, 2009 4:51 am

Great explanations, thanks for all the posts.

The piezo-resistor or silicon based pressure sensor produces no charge of it's own. Instead changing the shape of the crystal (via air pressure in this case) changes the resistance the crystal generates when a voltage source (i.e. battery or wall plug) is applied to it. Is that right? I'm assuming the device then further amplifies that signal in order to bring it into a range recordable on sleep DC systems.
User avatar
slowdavesleep
Senior Member
 
Posts: 400
Joined: Mon Mar 03, 2008 7:41 pm
Location: RPSGT NY

Postby sleepyjim » Sat May 16, 2009 3:15 pm

Is it possible for someone to use ms paint or something and show me where there is no flow limitation in one signal and not the other? Is it just the last breath?
Attachments
flows.JPG
flows.JPG (19.8 KiB) Viewed 4517 times
sleepyjim
Member
 
Posts: 79
Joined: Thu Mar 26, 2009 12:33 pm

Postby slowdavesleep » Sat May 16, 2009 4:17 pm

Honestly Jim I see flow limitation in both traces. The bottom trace (piezoresistive) is a more accurate graphing of the pressure since the top signal (piezoelectric) decays to baseline.

My question is how much of that decay is from the AC LFF and how much from the properties of piezoelectric crystals?
User avatar
slowdavesleep
Senior Member
 
Posts: 400
Joined: Mon Mar 03, 2008 7:41 pm
Location: RPSGT NY

Postby DBradley » Sat May 16, 2009 7:08 pm

When Braebon generates any documentation like the document RBonato posted we do everything possible to ensure that it is accurate and fair i.e. comparing apples to apples. The waveforms displayed are from the same subject wearing one cannula. The cannula is split so the pressure wave goes to both pressure sensors with the same inside and outside diameter tubing and the same length of tubing. Both the piezo pressure sensor and silicon pressure sensor are plugged into DC ports with the exact same settings (sampling rates, filters and gains. There are no hardware or software filters being applied. This gives a fair representation of both signals. The only difference would be in the sensitivity/scaling of each waveform. The piezo pressure sensor will output a signal in the 200uV to 500uV range while the silicon pressure sensor can be setup to output a signal in almost any range. For this comparison the signal on the silicon pressure sensor is in the 100mV to 200mV range. The silicon pressure sensors signal is over 500 times stronger than the piezo pressure sensor. Again if one had selected to generate the low voltage output (0-10mV) signal from the silicon pressure sensor then the signal would have been between 2 and 5 times stronger than the piezo pressure sensor's signal.

If one were to plug in these sensors into an AC amplifier then one would only expect that the flow limitation would diminish even more.

Each breath is a flow limited breath. It is important to see how the slope on the piezo pressure sensor is on a fairly larger decline than the silicon pressure sensor. The piezo pressure sensor does not accurately represent what is going on physiologically with the patient. The piezo pressure sensor shows a dimished case of flow limitation to the point when one could easily consider it part of a normal breath. The piezo pressure sensor does not clear show evidence of flow limitation due to the steady decline right into the baseline. This could be taken as a possible normal breath. The sensor signal shows that the patient clearly increased the breathing pressure during the breath (evidenced on the silicon pressure sensor's waveform) yet the piezo pressure sensor shows that there was flow limitation. The third breath shows no evidence of flow limitation on the piezo pressure sensor yet one can see that there is clear evidence of flow limitation on the silicon pressure sensor's waveform. One also should note the baseline difference between both pressure sensors. Again due to the poor frequency response of the piezo pressure sensor the baseline is artificially elevated.

Clearly the silicon pressure sensor represents a more accurate picture of what is going on with the patient.

I hope that this clears up any confusion.
DBradley
Member
 
Posts: 42
Joined: Thu Sep 27, 2007 1:35 am

Postby goudasleep » Tue May 19, 2009 1:37 am

Excellent explanations, thank you.
User avatar
goudasleep
Senior Member
 
Posts: 766
Joined: Mon May 16, 2005 8:37 am
Location: Tejas

Next

Return to Other Equipment

Who is online

Users browsing this forum: No registered users and 1 guest

cron