Search Support Database

The effects of cannula with and without filtering



This knowledge base article describes the effects of cannula with and without filtering on measures of snore pressure.


One of the parameters measured in the Nox-T3 system is the snore pressure derived from the nasal pressure signal.  The T3 recorder has a built-in pressure sensor and by connecting a nasal cannula to the pressure input of the T3, the nasal pressure signal is measured.  The general assumption is that the pressure inside the cannula is the same all over the cannula and the same as measured in the nose.  This is, however, a simplification and the cannula does affect the pressure signal being recorded.

The cannula is basically an assembly of tubes of different shapes, diameter and length and in some cases filters.  Those items together form air containers and flow resistance that affect the dynamic response of the system.  The pressure that appears at the pressure sensor in the T3 recorder is therefore not exactly the same as the pressure that appears at the nose.  The effect is different for different frequencies of the input pressure and the question is therefore, how the cannula affects the measured pressure at different frequencies and is the cannula suppressing the snore pressure signal?                              


To answer those questions, a pressure generator was built that is capable of sweeping a sine-wave formed pressure signal over the range between 10Hz and 90Hz.

Three different sweeps where performed on the same T3 recorder. 

The first sweep was performed by connecting the T3´s pressure port without tubing directly to the test device pressure chamber.  This sweep, including no cannula was used as a reference for the amplitude of the pressure signal at each frequency.

The second sweep used the exactly same settings as the first sweep, but in this case, the pressure chamber was connected to the T3 recorder by using the standard Nox cannula of 70 cm length (CANAF-0100) with the filter removed.

The third and last sweep used again the same settings but this time, the filter was left on the CANAF-0100 cannula.


The figure below shows the calculated effect of the cannula on different frequencies. 

Figure 1. Frequency response of the cannula with and without filter measured for the frequency range between 10Hz and 90Hz. The reference value is the applied pressure of 100% gain.

As can be seen from the figure above, the cannula is indeed affecting the amplitude being measured.  At 10 Hz, both cannulas do have the gain of 100%, that is, the pressure in has the same amplitude as the reference value.  From this point however, the measured pressure rises and is actually higher than the applied pressure. This is due to resonance in the cannula that reaches its maximum around 50Hz. Resonance occurs due to the storing of the vibrational energy within the cannula system, so even small periodic driving forces can cause large oscillations in amplitude at the system’s resonant frequencies (in this case maximum at 50Hz). The resonance effect of the filtered cannula is substantially lower than for the cannula without filter.  After 50Hz, the cannula gain declines and the filtered cannula reaches its unit gain at 60Hz while the non-filtered has the same slope, but reaches the unit gain first at 80Hz.  At 90 Hz both cannulas attenuate the signal, the non-filtered by 20% and the filtered by 40%.


From the measure above it is clear that the cannula itself has considerable effect on the measured amplitude of the snore signal, depending on the frequency of the signal.  For example there is a two-fold difference in measured amplitude between 50Hz and 90Hz, with the signal at 90Hz being twice as weak as the signal at 50Hz. Due to the resonance gain, around 50Hz, the attenuation effect is, however, not significantly affecting the capability of the cannula to pick up snores and both cannulas do have bandwidth close to covering the band of interest. These results show that the overall effect of having the cannula with and without filter is negligible.  However, the cannula itself can never be considered to be anything more than a qualitative measure of snoring, as the amplitude of the snore frequencies is distorted in the cannula itself in addition to the effects of different nose shapes, the application and movement of the cannula during the night.

It should also be noted that the Nox cannulas used for this comparison are relatively short in length (70cm).  It is expected that the resonance would move to a lower frequency if a longer cannula would be used and that would basically have the effect that the system would not cover the snore bandwidth between 10-90Hz; making the snore measures even less reliable.


For questions or further information please contact



Ver 1.0

*Please note that not all features or all products are available in all markets. Our products are medical devices and subject to registration/approval in each market area. For more information on product availability please contact

Print Friendly and PDF
Was this article helpful?
2 out of 2 found this helpful

Have more questions? Submit a request