WO2013156439A1

WO2013156439A1 - Method and device for controlling the operating temperature of a loudspeaker

Info

Publication number: WO2013156439A1
Application number: PCT/EP2013/057801
Authority: WO
Inventors: Ivan Bourmeyster
Original assignee: Arkamys
Priority date: 2012-04-20
Filing date: 2013-04-15
Publication date: 2013-10-24
Also published as: EP2839676B1; FR2989858A3; FR2989859A1; EP2839676A1; CN104429099A; FR2989859B1

Abstract

The invention relates to a method for controlling the temperature of an electrodynamic loudspeaker comprising a membrane (140) excited by an electric motor fed by an excitation signal (221), characterised in that it comprises the following steps: a) obtaining the spectral distribution of the electrical impedance of the motor at a moment t; b) identifying an impedance peak in a pre-defined frequency range of the spectral distribution obtained in step a); c) determining the difference in frequency between the central frequency Fc₁ of the impedance peak identified in step b) and the central frequency Fc₀ of an impedance peak identified in the same pre-defined frequency range in a spectral distribution of electrical impedance of the motor, obtained at a moment t₀ previous to t₁ ; d) obtaining a correlation relation between the difference in frequency (Fc₀-Fc₁ ) and the temperature of said loudspeaker; and e) determining the temperature of the loudspeaker according to the result of step c) and the correlation of step d). The invention also relates to a device for implementing said method.

Description

METHOD AND DEVICE FOR CONTROLLING THE OPERATING TEMPERATURE OF A SPEAKER

The invention relates to a method and a device for controlling the operating temperature of a loudspeaker. The invention is more particularly, but not exclusively, adapted to the field of electrodynamic loudspeakers for mobile phones.

FIG. 1, relating to the prior art, schematically represents the components of an electrodynamic loudspeaker. Such a loudspeaker (100) of substantially conical shape, comprises:

a chassis (1 10) supporting all the components of said loudspeaker;

a permanent magnet (120) fixed to said frame;

a coil (130) consisting of a wire wound in turns surrounding the permanent magnet (120) without contact;

- a membrane (140) of substantially conical shape attached to the frame (1 10) on the periphery of its larger diameter and to the coil on the periphery of its smaller diameter.

The coil (130) and magnet (120) assembly constitutes a motor. When a periodic electric signal, called the excitation signal, is applied to the terminals (135, 136) supplying the coil (130), this motor transforms the electrical energy into mechanical energy, in the form of vibrations, which vibrations are transmitted to the membrane (140), which vibrates, produces fluctuations in air pressure, that is to say a sound. The efficiency of this engine is not perfect, a portion of the electrical energy of the excitation signal is converted into heat, especially Joule effect. This heat increases the temperature of the loudspeaker and leads, when said temperature reaches critical levels, to a distortion of the sound emitted, by the modification of the electromagnetic characteristics of the engine, the thermal expansion of the materials and the loss of linearity in the electromechanical transformation. excitation signal. This heating may even result in the irreversible degradation of the loudspeaker, for example, by the breaking of the constituent wire the turns of the coil.

According to the prior art, various techniques can limit the heating of the speaker and protect it from overheating. These techniques are, for example, passive in nature, consisting in producing the loudspeaker by combining components whose characteristics make it possible to limit heating, in particular by increasing the number of turns of the coil, by reducing the mass of the membrane, by ensuring good ventilation and increasing the gap between the coil and the magnet. These technical solutions are difficult to implement in the case where the speaker is reduced in size and is housed in a small space. Technical solutions of an active nature for regulating the temperature of said loudspeaker are also used. Thus, document US-A-2005/0039465 discloses a device using a temperature probe and a Peltier cooling module connected to the loudspeaker. An electronic circuit triggers the operation of the Peltier effect module so as to cool the loudspeaker, which module is powered by the excitation signal, when the temperature measured by the probe reaches a critical value.

This example of thermal regulation of the prior art is also difficult to apply when the loudspeaker is integrated in a reduced size system such as a mobile phone, and moreover, produces a significant attenuation of the sound volume when it is turned on. . However, the user of a mobile phone wishes to use said phone in meeting mode and for this purpose, said mobile phone includes a mode of amplification allowing from the speaker of this phone to produce an audible sound several meters, for a period of up to several tens of minutes.

The invention aims to solve the disadvantages of the prior art and concerns for this purpose a method for controlling the temperature of an electrodynamic loudspeaker comprising a membrane excited by an electric motor powered by an excitation signal, which method comprises the steps of: a. obtain the spectral distribution of the electrical impedance of the motor at time tj;

b. identify an impedance peak in a predefined frequency range of the spectral distribution obtained in step a);

vs. determining the frequency difference between the center frequency, Fa, of the impedance peak identified in step b) and the center frequency, Fc ₀ , of an impedance peak, identified in the same predefined frequency range in a spectral distribution of electrical impedance of the engine obtained at a time t ₀ prior to ti;

d. obtaining a correlation relation between the frequency deviation (Fc ₀ - Fa) and the temperature of said loudspeaker;

e. determine the temperature of the speaker according to the result of step c) and the correlation of step d).

Thus, the method that is the subject of the invention makes it possible to determine the temperature of the loudspeaker without using a specific probe, and without increasing the size of said loudspeaker.

The invention can be implemented according to the advantageous embodiments described below, which can be considered individually or in any technically operative combination.

Advantageously, steps a) to e) are repeated for each time interval Bt = (ti-t ₀ ) when the loudspeaker is in operation. Thus, the temperature of said speaker can be controlled during its operation.

Advantageously, step e) comprises the steps of:

ei. determining an instantaneous temperature, T, at the time ti by the frequency difference between the impedance peaks; eii. determining the power of the excitation signal and the energy portion of this signal converted into heat;

EIII. depending on the result of step eii), determining the temperature of the loudspeaker at a time (h + z) such that ε> 5t. Thus the method according to the invention allows, according to this embodiment, to provide heating of the speaker when it is in operation.

Advantageously, the method which is the subject of the invention comprises, after step e), a step consisting of:

f. reduce the power of the excitation signal in a band of frequency, called thermal band, if the temperature determined in step e) exceeds a critical threshold.

Thus, according to this embodiment, the method according to the invention makes it possible to protect the loudspeaker against overheating by acting directly on the excitation signal, without integrating additional cooling device.

Advantageously, the method which is the subject of the invention comprises, according to one of its embodiments, the steps of:

boy Wut. obtain the spectral distribution of the power of the excitation signal; h. from the distribution obtained in step g) determining the power of the signal in a frequency band, called the mechanical band, centered on the center frequency, Fa, of the impedance peak obtained in step b); i. obtain a permissible excitation power for holding the membrane of the loudspeaker;

j. if the power obtained in step h) is greater than or equal to the allowable power obtained in step i), reduce the power of the excitation signal in the mechanical band.

Thus, the method that is the subject of the invention makes it possible, by analyzing the signal in the same frequency band as for measuring the temperature, to mechanically protect the membrane of the loudspeaker.

According to an improvement of the embodiment allowing the thermal protection of the loudspeaker, the method of the invention comprises after step c) and before step d) the steps of:

k. obtain a frequency value, F _stab , called stability criterion;

I. comparing the frequency difference obtained in step c) with the stability criterion;

m. perform steps d), e) and f) if the condition (| | Fc ₀ -Fci \ \ <F _stab ) is true and return to step a) if this condition is not satisfied.

Thus, the attenuation of the excitation signal is realized only when necessary in order to avoid "pumping" phenomena, that is to say close fluctuations in the sound level or the audio spectrum of the sound emitted from above. loudspeaker.

Advantageously, the power reductions of the excitation signal of the steps f) and j) are performed by filtering said signal in selective frequency bands and the method according to the invention comprises according to this embodiment the steps of:

o. determining, before step f), the characteristics of a filter capable of selectively reducing the spectral power of the signal during said step f);

p. determining, before step j), the characteristics of a filter capable of selectively reducing the spectral power of the excitation signal during said step j).

Thus, the selective action of the method that is the subject of the invention limits the audible effect of the thermal and mechanical protections on the sound emitted by the loudspeaker. Updating the characteristics of the filters makes it possible to adapt the protection and its effect to the real risk.

Advantageously, the excitation signal is a digital signal converted into an analog signal to excite the loudspeaker motor and the filters of steps o) and p) are digital filters applied to the excitation signal before it is converted into an analog signal. Thus, the characteristics of the filtering can be adapted very precisely to obtain the mechanical and thermal protection of the speaker by limiting the audible effect of these filters on the sound emitted by the speaker.

Advantageously, the predefined frequency range of step c) is between 300 Hz and 1000 Hz. The Applicant has determined that the impedance peaks in this frequency range are particularly correlated with the speaker temperature. .

Advantageously, it is equal to 5.10 ^"3 seconds This delay is at the same time sufficient to carry out the calculations necessary for the analysis of the signals and sufficiently short to realize the thermal and mechanical protections of the loudspeaker without it being necessary to significantly reduce the power of the excitation signal, that is to say, without degrading too much the excitation signal and therefore the sound.

Advantageously, the frequency range framing the thermal band is centered on a frequency of 3.10 ³ Hz. This frequency of 3 kHz is greater than the majority spectral content of the human voice, the attenuation of the signal power in this frequency band thus effectively limits the heating of the speaker without degrading the sound volume felt by the listener especially in the case of an application on mobile phone in conference mode.

Advantageously, the spectral distribution of the impedance is obtained in step a) of the method that is the subject of the invention, by measuring a transfer function between a signal proportional to the current of the analog excitation signal of the loudspeaker. and the delayed digital excitation signal of said loudspeaker. Thus, the method can be implemented easily on any device comprising a digital audio source, by adapting the delay time, without significant branch on the speaker side.

The invention also relates to a device adapted to implement the method of the invention on a loudspeaker whose motor is powered by an excitation signal, which device comprises:

i. signal processing means;

ii. a computer, able to modify the characteristics of the processing performed by the signal processing means;

iii. means for measuring the spectral distribution of the impedance of the loudspeaker, characterized in that said means comprises a measuring resistance, impedance of less than ^1/50 of the impedance of the loudspeaker, connected in series on the power supply of the loudspeaker motor and traversed by the excitation signal, and means adapted to deliver a signal proportional to the voltage across said resistor.

Thus the device of the invention can be easily adapted to a device without significant modification.

The invention also relates to a sound diffusion apparatus, in particular a mobile telephone, comprising a loudspeaker, which apparatus comprises a device according to one of the embodiments of the invention.

Advantageously, the apparatus forming the subject of the invention comprises means able to implement a mode of over-amplification of the sound diffusion and means able to implement a method of attenuation of the excitation signal according to one of the embodiments of the method which is the subject of the invention, when said mode of boosting is selected. Thus, the apparatus of the invention can be used at a high volume without risk of overheating.

The invention is explained below according to its preferred embodiments, in no way limiting, and with reference to FIGS. 1 to 6 in which:

- Figure 1 relating to the prior art shows in a sectional view a schematic diagram of an electrodynamic loudspeaker;

FIG. 2 is a block diagram of an exemplary embodiment of the device that is the subject of the invention;

FIG. 3 shows a functional diagram of the means for measuring the spectral distribution of the impedance of the loudspeaker according to an exemplary embodiment of the device that is the subject of the invention;

FIG. 4 shows, in FIG. 4A, a functional diagram of the computer implementing spectral analysis and filtering according to an exemplary embodiment of the device which is the subject of the invention and represents, FIG. 4B, an example of comparison of spectra of FIG. impedances measured at different temperatures of the loudspeaker;

FIG. 5 shows, FIG. 5A, an example of a Bode diagram of a low-pass filter and FIG. 5B, an example of a Bode diagram of a band rejection filter;

and FIG. 6 represents a flowchart of a method for measuring and monitoring the temperature of a loudspeaker according to an exemplary embodiment of the method that is the subject of the invention.

Figure 2, according to an exemplary embodiment of the device (200) object of the invention, it is inserted into the audio processing chain of an apparatus such as a telephone. The device according to the invention receives as input a digital signal (21 1) originating from an audio source, and outputs a digital signal (212) that has been processed in one of the modes. embodiment of the method which is the subject of the invention. This digital signal (212) at the output of the device (200) which is the subject of the invention is amplified and converted into a signal (221) for excitation of the loudspeaker (100) by a module (220) for converting and amplification according to a method known from the prior art. The device according to the invention comprises means (230) for measuring the spectral distribution of the impedance of the loudspeaker (100), a computer (240), and a means (250) for processing the signal. The computer (240) is able to modify the characteristics of the processing performed by the signal processing means (250) on the signal (21 1) introduced into the device that is the subject of the invention.

FIG. 3, according to an exemplary embodiment, the means (230) for measuring the spectral distribution of the impedance comprises a measurement resistor (331) connected in series with the loudspeaker (100) at the output of the module ( 220) of conversion and amplification. Said measurement resistor (331) is thus traversed by the excitation signal (221). The resistance value of this resistance (331) measurement is low, of the order of 50 to 100 times lower than the speaker impedance (100). Thus for a loudspeaker (100) whose electrical impedance is of the order of 8 ohms, the impedance of the resistor (331) measurement is chosen to be of the order of 0.1 Ohm. Said spectral distribution measuring means (230) comprises a module (332) voltmeter, connected in shunt across the resistor (331) measurement, and delivering a signal (340) proportional to the voltage across said resistor ( 331), that is to say proportional to the intensity of the loudspeaker excitation signal (221), which intensity is a function of the electrical impedance of said loudspeaker (100). The signal (340) from the module (332) voltmeter is amplified by a low noise amplifier (333) and converted by an analog / digital converter (334) into a digital signal (341).

4A, according to an exemplary embodiment of the computer (240) of the device according to the invention, the latter comprises a processor, memory means for storing data and instruction programs of the processor to perform calculations and only input and output ports and a clock. The digital signal (341) from the analog / digital converter (334) is introduced into said computer (240), which computer comprises a spectral analysis program (441). According to an exemplary embodiment, the digital excitation signal (212) of the loudspeaker is delayed by a delay module (440) and the delayed signal (412) is also inputted to the spectral analysis program (441). . The delay applied by the delay module (440) to the digital excitation signal (212) of the loudspeaker makes it possible to compensate the processing time of the excitation signal (221) until the delivery of the digital signal (341) from the measurement module of the spectral distribution. The spectral analysis program (441) determines the transfer function of the loudspeaker from the ratio of the two signals (341, 412) at the input of said program and delivers information corresponding to the spectral distribution of the electrical impedance of the loudspeaker. speaker. The time delay applied by the delay module (440) is adjusted to the application. Thus, the device of the invention makes it possible, by a simple method, a limited number of components and without significant intervention on the audio circuit, to obtain, in real time, the spectral distribution of the electrical impedance of the loudspeaker. . A program (442) included in the computer memory determines, from the spectral distribution provided by the spectral analysis program (441), the position of the center frequency Fa and the half-height width, f, an impedance peak, in a frequency range included, according to an exemplary embodiment, between 300 Hz and 1000 Hz. If such an impedance peak is detected, the information relating to its central frequency, Fa, its width half-height, /;, and the measurement date are stored in a memory (443) live accessible in read and write, called historical RAM, according to the acronym of "Random Access Memory". A comparison program (444) compares the position of the central frequency Fa of the last impedance peak recorded at time t with the position of the center frequency Fc ₀ of the impedance peak in the previous record, at the time t ₀ , in the same frequency range. Said program (444) delivers information, for example the quantity (Fci-Fc ₀ ), relative to the displacement of the impedance peak. A correlation program (445) determines the temperature, T, of the loudspeaker at time t; of the recording, by reading a correlation table connecting the impedance peak displacement and the speaker temperature. Said correlation table is read by said program (445) in a memory (446), called a characteristics memory, accessible in read-only mode, preferably of type ROM (according to the acronym for "Read Only Memory") or EPROM ( according to the English acronym of "Erasable Programmable Read Only Memory") and having a record of the correlation table for the speaker model considered. 4B, according to an exemplary behavior of the loudspeaker, when the temperature of said loudspeaker increases, the spectral distribution of the impedance at time ti present according to a plot (471) in a diagram having on the abscissa the frequency (480) and on the ordinate the impedance (490), a peak centered on a frequency Fa, which frequency Fa is greater than the center frequency Fc ₀ of the peak detected in the plot (470) of the impedance at time t ₀ prior to tu À For example, for a loudspeaker with an electrical power of 1 Watt, the center frequency of the impedance peak changes from 800 Hz to 400 Hz when the speaker temperature goes from 70 ° C to 120 ° C.

Returning to FIG. 4A, according to an exemplary embodiment of the device that is the subject of the invention, it comprises a program (447) for calculating the spectral power of the digital signal (21 1) originating from the audio source. Said power calculation program (447) uses information contained in the memory (446) of characteristics relating to the loudspeaker and its location in the receiving apparatus, which characteristics make it possible to calculate the heating, ΔΓ, of said loudspeaker. speaker after a lapse of time 5t, after the time t according to the spectral power of the signal (21 1) audio. By way of example, without limitation, the program (447) of calculation performs the operations of:

calculating the excitation energy Ee of the loudspeaker as a function of the power, Pe, of the excitation signal of the loudspeaker, estimated from the spectral power of the digital audio signal 21 1 and the characteristics of the amplification of this signal: Ee = Pe. ε with ε> Bt;

calculate the part Qe of the excitation energy Ee, transformed into heat;

- calculate ΔΤ: ΔΤ = {Qe - Qd) / (M,

where Qd is the amount of heat dissipated without heating the loudspeaker, M the mass of the loudspeaker and C the mass heat of said loudspeaker. These quantities are determined by simulation or by experimentation and the corresponding empirical laws are stored in the memory (446) of characteristics, for example in the form of coefficient tables. A summation and comparison program (448) performs the sum T + AT and compares the result to a critical threshold Stores you in the memory (446) of features. Based on the results of the summation and comparison program (448), a parameter program (449) calculates the parameters of a digital filter to be applied to the audio signal (21 1), which parameters are transmitted to the medium (250). signal processing.

According to an exemplary embodiment of the device according to the invention, the signal processing means (250) is a program implemented by the computer (240) and which performs a digital filtering of the audio signal (21 1). Thus, several filters may be cascaded to said audio signal (21 1). By way of nonlimiting example, the filtering comprises a low-pass filter, a band-cut filter or a single-band or multi-band band rejection filter.

5A, according to an exemplary embodiment, the filter applied to the audio signal is a low-pass filter whose Bode diagram (570) in a logarithmic reference with the abscissa frequency (580) and the ordinate (590) the gain, comprises a bandwidth (571) and a rejected band (572) beyond a cut-off frequency (575) corresponding to an attenuation of -3 dB relative to the gain of the pass-through portion. According to an exemplary embodiment of the filtering performed by the device according to the invention, said cut-off frequency (575) is chosen in the upper part of the spectrum of the speech signal, which is greater than or equal to 3.10 ³ Hz for a high - speaker used to broadcast mainly conversations. Beyond the cut-off frequency (575), the attenuation of the filter is -6 dB / octave. According to an exemplary embodiment of the method which is the subject of the invention, when such a low-pass filter is applied to the audio signal, the parameter adjusted as a function of the temperature of the loudspeaker is, on the one hand, the application of the filter to the signal, cutting frequencies above 3.10 ³ Hz essentially affecting the tone of the voice without affecting the volume felt by the listener, while reducing the signal excitation power, then the gain in the bandwidth (571) of the filter, this gain then being adjusted to a value less than 0 dB.

5B, according to another embodiment of the method according to the invention, the filter applied is a band rejection filter. The plot (577) of the Bode diagram of such a filter has attenuation centered on a frequency (576) characterized by a bandwidth l _f about this center frequency (576) at -3 dB relative to the gain in the bandwidth. Thus, the adjustable parameters of such a filter are the gain in the pass-through portion, the center frequency (576) of the rejected band, and the bandwidth 1 _f around this center frequency. The digital filtering of the signal makes it possible to combine the different types of filters according to the desired result in order to minimize the audible degradation of the sound emitted by the loudspeaker. The appropriate filter combination is determined experimentally or by simulation.

FIG. 6, according to an exemplary embodiment of the method that is the subject of the invention, implemented by a device as described above, said method comprises a first measurement step (610) consisting of obtaining the impedance spectral distribution of the loudspeaker, preferably in a frequency range between 300 Hz and 1000 Hz. According to an identification step (620), an impedance peak, centered on a frequency Fa, is identified in the spectral distribution obtained. during step (610) of measurement. During a comparison step (630), the position of the central frequency Fcj of said peak is compared with the position of the center frequency Fc ₀ of the impedance peak in the spectral distribution obtained at an earlier date t ₀ . According to a stability control step (635), the difference | | Fa - Fc ₀ | | is compared with a stability criterion F _stab . As the heating process of the loudspeaker is progressive, a too large difference between the central frequencies of the peaks, when the measurement dates t ₀ and t i are close together, is a measurement artifact. Thus, the subsequent stages of signal processing are implemented only if the loudspeaker heating is proven. According to a correlation step (640), the difference (Fci-Fc ₀ ) is correlated with the temperature T of the loudspeaker, thus determining the temperature of said loudspeaker at time t _} . According to a signal analysis step (650), the spectral distribution of the power of the excitation signal is obtained. According to a prediction step (660), said spectral distribution of the power of the excitation signal is used to predict heating, ΔΓ, of the loudspeaker at time (ti + z) with ε> ht. According to a thermal risk analysis step (665), the quantity (T + AT) is compared with a critical temperature Te. If (T + AT) ≥Te then a filter to reduce the power of the excitation signal according to a selected frequency spectrum is calculated during a parameterization step (670). In parallel, according to a step (680) of mechanical analysis, the power of the excitation signal in a frequency band centered on Fa according to a bandwidth /; is analyzed. During a step (685) of mechanical comparison, the power determined during the mechanical analysis step is compared to a critical value defining a risk of displacement of the speaker diaphragm out of its allowable mechanical travel . If the power determined during the step (680) of mechanical analysis is greater than the allowable power, then the characteristics of a band rejection filter, centered on the frequency Fa and according to a bandwidth, called band mechanical, corresponding to /; are calculated during a step (690) of mechanical parameterization. The filters calculated during the mechanical and thermal parameterization steps (670, 690) are applied to the signal during a filtering step (695).

The above description and the exemplary embodiments show that the invention achieves the desired objectives, in particular, it makes it possible, particularly in the context of a mobile phone application, to protect the loudspeaker of said telephone when it is used in over-amplification, especially in meeting mode. The selective action on the sound spectrum and the real-time control and anticipation of heating the loudspeaker, allow to keep said speaker at an acceptable operating temperature while preserving the sound volume felt. The fully digital signal processing allows great flexibility in the signal processing and in the selection of the parameterization of the filters applied to said signal.

Claims

Method for controlling the temperature of an electrodynamic loudspeaker (100) comprising a diaphragm (140) excited by an electric motor powered by an excitation signal (221), characterized in that it comprises the steps of :

at. obtaining (610) the spectral distribution of the electrical impedance of the motor at time tj;

b. identifying (620) an impedance peak in a predefined frequency range of the spectral distribution obtained in step a);

vs. determining (630) the frequency difference between the center frequency, Fa, of the impedance peak identified in step b) and the center frequency, Fc ₀ , of an impedance peak, identified in the same range of predefined frequency in a spectral distribution of electrical impedance of the engine obtained at a time t ₀ prior to tj;

e. determining (640, 670) the temperature of the speaker according to the result of step c) and the correlation of step d).

Method according to claim 1, characterized in that steps a) to e) are repeated for each time interval t = (ti-t ₀ ) when the loudspeaker is in operation.

Method according to claim 2, characterized in that step e) comprises the steps of:

ei. determining (640) an instantaneous temperature, T, at time t 1 by the frequency difference between the impedance peaks; eii. determining (650) the power of the excitation signal and determining the energy portion of this signal converted to heat; EIII. determining (660) as a function of the result of step eii), the temperature of the loudspeaker at a time (h + z) such that ε> δί.

Process according to Claim 2, characterized in that it comprises, after step e), a step consisting of:

f. reducing (670, 695) the power of the excitation signal in a frequency band, called the thermal band, if the temperature determined in step e) exceeds a critical threshold.

Method according to claim 2, characterized in that it comprises the steps of:

boy Wut. obtaining (650) the spectral distribution of the power of the excitation signal;

h. from the distribution obtained in step g) determining (680) the power of the signal in a frequency band, called the mechanical band, centered on the center frequency, Fa, of the impedance peak obtained in step b) ;

i. obtain a permissible excitation power for holding the membrane of the loudspeaker;

j. if (685) the power obtained in step h) is greater than or equal to the allowable power obtained in step i), reducing (690, 695) the power of the excitation signal in the mechanical band.

Method according to claim 4, characterized in that it comprises after step c) and before step d) the steps of:

k. obtain a frequency value, F _stab , called stability criterion;

I. comparing (635) the frequency difference obtained in step c) with the stability criterion;

m. perform steps d), e) and f) if the condition (| | Fc ₀ -

\ <F _stab ) is true and return to step a) if this condition is not satisfied.

Method according to claim 5, characterized in that the power reductions of the excitation signal of steps f) and j) are performed by filtering said signal in selective frequency bands and that said method comprises the steps of:

o. determining (670) before step f) the characteristics of a filter capable of selectively reducing the spectral power of the signal during said step f);

p. determining (690) before step j) the characteristics of a filter capable of selectively reducing the spectral power of the excitation signal during said step j).

A method according to claim 7, characterized in that the excitation signal (221) is a digital signal (212) converted to an analog signal to excite the loudspeaker motor and the filters of steps o) and p) are digital filters applied to the excitation signal (21 1) before its conversion into an analog signal.

Method according to Claim 1, characterized in that the spectral distribution of the impedance is obtained in step a) by measuring a transfer function between a signal (340) proportional to the current of the excitation signal (221). ) of the loudspeaker and the delayed digital excitation signal (212) (412) of said loudspeaker.

Method according to claim 1, characterized in that the predefined frequency range of step c) is between 300 Hz and 1000 Hz.

Method according to claim 2, characterized in that ôt is equal to 5.10 ^"3 seconds.

Process according to claim 4, characterized in that the range of Framing frequencies of thermal band is centered on a frequency of 3.10 ³ Hz.

Device (200) adapted to implement the method according to claim 1, on a loudspeaker (100) whose motor is powered by an excitation signal (221), said device comprising:

i. signal processing means (250);

ii. a calculator (240) adapted to modify the characteristics of the processing performed by the signal processing means (250);

iii. means (230) for measuring the spectral distribution of the impedance of the loudspeaker, characterized in that said means (230) comprises a resistor (331) for measuring, having an impedance of less than 1/50 ^th of the impedance of the loudspeaker, mounted in series on the power supply of the loudspeaker motor and traversed by the excitation signal (221), and means (332, 333) capable of delivering a signal proportional to the voltage at the terminals of said resistance.

Sound diffusion apparatus, in particular a mobile telephone, comprising a loudspeaker, characterized in that it comprises a device according to claim 13.

Apparatus according to claim 14, characterized in that it comprises means capable of implementing a mode of amplification of the sound diffusion characterized in that it comprises means able to implement a method according to claim 4, when said boost mode is selected.