Christian Schreiber+49 (0)345 / 133 17-829
Signal processing in the SONAPHONE outputs different sound levels, which are significant depending on the application.
As a basis, raw level values are obtained from the filtered raw data with a sample rate of 1 ms. The derived level values are calculated and stored in much larger time intervals. The sample rate can be varied in the menu "Measurement settings" between 4 ms and 128 ms. It should be noted that the set sample rate has an influence on the representation of data in the spectrogram.
Level values are:
The level values Lmin, Lmax, Lpk, Leq are calculated from the data from the beginning of each measurement and shown on the display. Pressing the "Reset" button in the Levelmeter App will start a new measurement with newly calculated level values. During data recording, the values apply to the recording period.
Through signal processing, ultrasound can be converted into a secondary "downmixed" signal in the audible range.
In the heterodyne method, a narrow frequency in the ultrasonic range is selected (e.g. carrier frequency of 40 kHz +/- 2 kHz) and transformed into the audible range via difference frequencies. The method is used in the traditional analogue test equipment, which work in the narrow band around 40 kHz. The broadband digital SONPAHONE integrates this approach. By means of a shift of the carrier frequency (movable line in the spectrogram in the broadband 20 ... 100 kHz), the corresponding narrow-band audio signal is made available.
In many cases a qualitative evaluation is already possible with the heterodyne method via the audible impression, for example in leak detection and in the basic evaluation of bearing condition.
In most cases, the ultrasound signals are distributed over a broad frequency range. This means that the information in the acoustic signal cannot be completely detected in a narrow frequency range. In the case of the SONAPHONE, which operates in broadband, it is possible to convert the signals using the phase vocoder method. Doing this, the entire frequency range of 20 ... 100 kHz is compressed by a factor of 32. Despite little loss of information, the original acoustic situation in the broadband can be reproduced, the audible impression is available for the entire ultrasound range.
Both methods "sound" differently. The user can listen to the signal via built-in speakers or headphones. The sampling rates of the audio signals are 8 kHz. The volume of the signal changes in the same way as the intensity of the original high-frequency signals. In addition, the variation of the original signal over time is preserved so that the dynamics of events in the audible signal are reflected.
In addition, the SONAPHONE records the audible signals. They are available in WAV format and are thus also available for further data processing.
The term "watertight" describes colloquially a tightness requirement that is specified for a component and its application. The precise description of the requirement is made on the leakage rate. For the term "watertight" this would be 10⁻² mbar l/s (using the test medium air).
To define a leak requirement, information such as test pressure, the medium for testing and the maximum allowed leakage rate must be known. For example, if the test is set to the limit of 10⁻² mbar l/s using the medium air, there will be no water leakage through the leak in dense components - but permeability to gases or less liquid media (such as gasoline or oil) may still be possible. The tightness requirement of 10⁻² mbar l/s or "watertight" cannot be reliably and reproducibly fulfilled with commercial ultrasonic technology in normal industrial environments. For further explanation see FAQ L3 (Minimum detectable leak rate using air-borne sensors).
When testing with air-borne sensors, the sound pressure is usually indicated by a level value in dB or dBμV. Using the level in dB the measured values are converted to a logarithmic ratio based on a reference value. Thus, measured values can be represented in a large dynamic range. The level output is determined by the type of instrument and the sensor used.
The air-born sensors BS10 and BS30 of the SONAPHONE use a broadband-sensitive microphone capsule that converts the sound pressure into an AC voltage. The sound pressure level is based on the acoustic physical comparison variable of 20 μPa. 20 μPa correspond to a sound pressure of 0 dB, further examples are:
20 µPa = 0 dB
1 Pa = 94 dB
2 Pa = 100 dB
The calibration and adjustment of the SONAPHONE microphone capsule is done by comparing it with a reference microphone (calibration of the microphone voltages) using a 40 kHz reference signal (*1). Based on the reference signal the characteristic curve of the broadband-sensitive microphone capsule is shifted in the entire measuring range of 20 ... 100 kHz.
In the SONAPHONE, the level values are finally calculated from the broadband sound pressure between 20 ... 100 kHz. The meaning of the different level values is explained in a seperate FAQ.
SONAPHONE Pocket (and comparable analog instruments)
For analog ultrasonic instruments that are available in the market (for example, SONAPHONE Pocket) such a signal calibration is not meaningful. The reason for this is the use of an ultrasonic capsule (Murata capsule), which is sensitive to ultrasonic frequencies in a relatively narrow frequency range around 40 kHz. The level values displayed on these devices must be referenced differently. The reference value is 0 dBμV at a voltage level of 1 μV (rms value). The value varies depending on the characteristics of the individual ultrasonic capsule. On the device display the microphone voltage is displayed in dBμV, for negative values the voltage is below 1 μV.
1 µV = 0 dBµV
Due to the use of different capsules and calibration methods, different characteristic curves as well as the consideration of different frequency ranges (narrowband / broadband) the displayed values of the analog instruments in dBμV and the SONAPHONE in dB cannot be easily compared or converted. The reasons again at a glance:
Digital Instruments (SONAPHONE)
Analogue Instruments (SONAPHONE Pocket)
Calibration of the level via physical reference / sound pressure
Calibration of the level via voltage signal of individual sensor
|Characteristic curve of used microphone capsule (broadband)||Characteristic curve of used ultrasonic capsule (around 40 kHz)C|
Specific calculation of level values
Specific calculation of level values (depending on supplier)
Display of broadband levels (20 … 100 kHz)
Display of narrowband levels (around 40 kHz)
*1: In the acoustics a 1 kHz reference signal is used
There is a variety of applications and suitable methods to find and assess leakages. Acoustic testing with air-borne sensors is used when leaks are to be found relatively quickly and without great technical effort. Pressure differences cause air to leak from the system (overpressure) or into a vacuum system (underpressure). Starting at certain leak sizes turbulent flow is created, which in turn is responsible for the generation of ultrasound.
However, the detectable leak rate using air-borne sensors is limited by physics and the performance of available electronic components. On the one hand, several factors determine if and to what extent a turbulent flow is created. As example, due to physical characteristics a turbulent flow only arises above a leak rate of 10⁻² mbar l/s. Furthermore, a pressure difference of approx. 500 mbar between the system and the environment is needed. On the other hand, the electronic components used in ultrasound instruments constrain the lower measuring scale. Economically reasonable sensor technology limits the signal to noise ratio and thus the system’s sensitivity.
The above-mentioned reasons are mainly responsible that tightness requirements of 10⁻² mbar l/s or "watertight" cannot be reliably and reproducibly met in industrial environments using standard ultrasound technology!
However, there are applications where locating and evaluating leaks with less demanding requirements is relevant, e.g. in the field of energy efficiency in compressed air and steam systems as well as in leak testing using ultrasound transmitters (e.g. in transport and shipbuilding). In the field of quality assurance, requirements need to be defined and approved by the customer in order to get a qualitative assessment of components and subassemblies.
The indication of values of leakage losses are for orientation, as they are influenced by many factors. In ultrasound testing, the acoustic radiation largely depends on the leak size, leak shape, surface condition of the material, pressure difference, outflow velocity and outflow profile as well as the temperature. An important role is also played by the measuring distance and the measuring angle. The multiplicity of factors creates a complexity that is difficult to model through simple analysis.
Traditional, narrow-band instrument technology around 40 kHz is very well suited for the localization of compressed air leaks. However, the validity of the sound pressure level in this narrow frequency range with respect to the determination of the amount of loss (quantification) has to be questioned, since the frequency maximum (amplitude) of the ultrasonic signal stochastically occurs in a wide frequency range. Therefore, the loss of a leakage should be evaluated in a wider frequency range. The range between 20 and 100 kHz has been proven to be practicable. Broadband digital instruments can provess this additional information and evaluate leakages on the basis of their frequency characteristics (integral sound pressure level).
However, the determination of the volumetric flow loss in l/min remains subject to statistical fluctuations as a result of the aforementioned factors. Therefore, it is recommended that a monetary estimate of leakage losses should always be made for the total loss of multiple leakages detected. When performing statistical analysis of many leaks, the measurements will get sufficiently accurate and reliable.
For the same reasons - and in particular the fact that leaks have various shapes - the mathematical derivation of a circular leak size from leak losses in l/min may be less meaningful. Nevertheless, for basic understanding and orientation, some hole sizes with estimated leakage losses should be shown here (basis: Leakage loss calculation by Postberg+Co. GmbH):
Hole size (mm)
Leckage size at 6 bar (l/min)*
Leckage loss at 6 bar (m³/year)**
Monetary loss (€/year)**
If, in a volume of 1 liter, the gas pressure drops by 1 mbar within 1 s, this corresponds to a leakage rate (outflow of gas) of 1 mbar l/s. The unit mbar l/s is relatively difficult to imagine, but is typically used alongside units such as Pa m m³/s, Torr l/s or atm cc/s in the vacuum applications.
The indication becomes more understandable if the loss quantity is indicated in other units. For example, units such as g/a or oz/y are used in air conditioning technology. For the indication of leakage losses in compressed air systems, the unit l/min has prevailed.
Examples of unit conversion (at Δp = 1,013.25 mbar, 273.15K or 0°C, medium air):