(1) Field of the Invention
This invention relates generally to loudspeaker systems and relates more specifically to methods and apparatus for computing a metric value for loudspeaker protection.
(2) Description of the Prior Art
It is well known that loudspeakers can be damaged due to overheating when driven at high output levels, especially for audio applications that contain high frequency components. Current practice is to prevent loudspeaker damage by simply hard or soft limiting the maximum power output to the loudspeaker.
As a result the maximum sound pressure level (SPL) that can be produced is severely limited even if the actual temperature of the loudspeaker is below the damage threshold. More advanced systems incorporate a simple thermal model of the loudspeaker which may not be accurate, or a temperature sensor to activate the limiter, which introduces additional complexity and cost.
Therefore it would be desirable to achieve a method for obtaining a threshold value for loudspeaker protection, based on instantaneous voltage and current measurements.
It is also challenge for engineers to achieve loudspeaker protection at reasonable cost without requiring additional devices to the loudspeaker.
A principal object of the present invention is to prevent over-heating of a loudspeaker.
A further object of the present invention is to obtain a threshold value that is representative of an actual maximum allowable temperature of a voice coil of a loudspeaker.
A further object of the present invention is to obtain the threshold value by using auto-correlation and cross-correlation of band-pass filtered current and voltage measurements.
A further object of the invention is to activate a protection circuit if the threshold representative of the maximum temperature of the voice coil is exceeded.
In accordance with the objects of this invention, a method for computing metrics for loudspeaker protection has been achieved. The method invented comprises, firstly, the following steps: (1) providing a loudspeaker, a voltage and a current meter, a speaker protection monitor comprising computational and filtering means, and a speaker protection device, (2) measuring speaker input current and voltage, and (3) filtering by band-pass filters speaker input voltage and current values measured. Furthermore the method invented comprises the steps of: (4) using cross-correlation and autocorrelations equations based on band-passed measurement results to compute loudspeaker protection metrics including a real part of impedance of a voice coil of the loudspeaker, (5) activating speaker protection device if the real part of the loudspeaker impedance exceeds a threshold, and (6) going back to step (2) if input signals of speaker are active.
In accordance with the objects of this invention a system for preventing any damage of a loudspeaker because of over-heating of a voice coil has been achieved, The system invented comprises a speaker protection monitor configured to using cross-correlation and autocorrelations equations based on band-passed measurement results of loudspeaker voltage and loudspeaker current and to compute loudspeaker protection metrics and a speaker protection device lowering, if activated by the speaker protection monitor, a temperature of the voice coil.
In accordance with the objects of this invention a system for preventing any damage of a loudspeaker because of over-heating of a voice coil has been achieved, The system invented comprises, firstly, a loudspeaker comprising at least one voice coil and two signal input ports, a voltage meter measuring a signal input voltage of the loudspeaker and transmitting voltage measurement results to a speaker protection monitor, and a current meter measuring a signal input current of the loudspeaker and transmitting current measurement results to a speaker protection monitor. Furthermore the system comprises said speaker protection monitor comprising: a band-pass filter, an averaging filter, an output port to transmit a signal to a speaker protection device in order to activate the speaker protection device if required, and means to compute cross correlation and autocorrelation calculations comprising computing the loudspeaker threshold. Moreover the system comprises said speaker protection device lowering, if activated, a temperature of the voice coil.
In the accompanying drawings forming a material part of this description, there is shown:
a shows simulation results of the impedance of loudspeaker as function of frequency.
b shows simulation results of the square root (<vv>/<ii>) and of <iv>/<ii> both as function of frequency.
A method and system for obtaining a threshold value for loudspeaker protection, based on instantaneous voltage and current measurements, wherein autocorrelation and cross-correlation of band-pass filtered current and voltage measurements are used to obtain a real part of the complex impedance of the voice coil, are disclosed. An important point of the invention is that it does not require a DC or pilot signal to be added to the audio signal in order to perform the measurement, and that it is not disturbed by capacitive or inductive components in the complex impedance of the loudspeaker.
The metric is related to temperature. The real part of the impedance increases with temperature, causing a decrease in power.
Furthermore the speaker protection monitor 7 performs calculations to identify situations requiring actions to protect the loudspeaker 9. In order to monitor such situations speaker protection metric signals 8 are provided to activate the speaker protection device 1. These calculations will be outlined in detail below. The speaker protection monitor 7 is a processor.
The impedance measurement will be inaccurate when the input power is below a certain level. Hence the metric is never positive when the power is below Pmin. The Rmin is the minimum real impedance for the metric to become positive, indicating the speaker protection should be activated. For the metric to become positive, R>Rmin and P>Pmin.
The method invented assumes that instantaneous measurements of the current im (t) through and voltage vm (t) over the loudspeaker terminals are available. In general the measured signals are generally not equal to the true current and voltage but will have an instantaneous error signals ei(t) and ev(t) The error signals are caused by noise in the current and voltage sensing devices.
im(t)=i(t)+ei(t) (1)
vm(t)=v(t)+ev(t) (2)
These signals are each passed through a band-pass filter, leading to:
im,bpf(t)=ibpf(t)+ei,bpf(t) (3)
vm,bpf(t)=vbpf(t)+ev,bpf(t) (4)
The band-pass filter can be a combination of a high-pass and low-pass filter, and can be implemented as an analog or digital circuit, for example as an IIR or FIR filter.
From these values the instantaneous products are obtained:
ii(t)=im,bpf(t)·im,bpf(t) (5)
iv(t)=im,bpf(t)·vm,bpf(t) (6)
vv(t)=vm,bpf(t)·vm,bpf(t) (7)
These values are passed through an averaging filter, denoted as < >. This may be low-pass filter, such as leaky integrator filter, or digital filter such as a FIR, IIR or sliding average filter. The result after averaging can be expressed as:
<ii>=<ibpf2>+2<ibpf·ei,bpf>+<ei,bpf2> (8)
<iv>=<ibpfvbpf>+<ibpf·ev,bpf>+<ei,bpf·vbpf>+<ei,bpf·ev,bpf> (9)
<vv>=<vbpf2>+2<vbpf·ev,bpf>+<ev,bpf2> (10)
It should be noted that the current noise <ei,bpf2> should be calibrated in advance. It is a single value, which is equal to the <ibpf2> with an input value of zero.
If there is no correlation between the actual current and voltage and the measurement noise, the cross terms with the noise, shown in the correspondent equation above, can be neglected. However, the autocorrelations of the noise terms will not vanish:
<ii>˜<ibpf2>+<ei,bpf2> (11)
<iv>˜<ibpfvbpf> (12)
If we assume the noise characteristics of the voltage and current measurements are known, then an accurate estimate of the true current and voltage correlations can be obtained as:
<ibpf2>est=<im,bpf2>−<ei,bpf2> (13)
<ibpfvbpf>est=<im,bpf·vm,bpf> (14)
An estimate for the real part of the complex impedance can be computed as:
Rest=ibpfvbpf>est/<ibpf2>est (15)
Due to the power dissipated in the voice coil, its temperature will increase. This will result in a higher resistance. With the proposed method, a threshold value can be set for the maximum resistance (and hence the maximum temperature) that is allowed:
Threshold1(t)=<ibbfvbpf>est−Rmin·<ibpf2>est (16)
Threshold1 defined here is the output of the block SUB1 209 shown in
Threshold2 is the output of block SUB2 210 and is defined:
Threshold2=<ibpfvbpf>est−Pmin
Rmin is a user configurable value. It is related to the maximum allowed temperature, because R(T)=R0+alpha*R0*(T−T0), wherein alpha is the resistivity coefficient of the voice coil at reference temperature T0, wherein the voice coil has a resistance R0.
If threshold(t) is positive, the maximum resistance has been exceeded and the protection circuit 1 should be activated.
The objective of the protection circuit 1 is to lower the temperature of the voice coil, by either reducing the average power that is dissipated as heat, or by improving the cooling of the voice coil. Examples of such protection circuits are:
Alternatively an additional voltage signal may be added to the loudspeaker terminals that induces large displacements of the voice coil. The large displacement will results in elevated air intake and hence enhanced cooling. If the frequency of the additional voltage signal is very low it will be practically inaudible.
Blocks 200 and 205 are acting as band-pass filters. Blocks 204 and 206 are multiplication blocks. Block 201 is enabled to square. Blocks 202 and 207 are acting as low-pass filters. Blocks 203 and 208 are adders. Blocks 209 and 210 are enabled to subtract, and block 211 is a comparing block determining, which result of blocks 209 and 210 is smaller.
In block 200 and 205 the steps according equations (3) and correspondently equation (4) are computed. Block 201 calculates equation (5) and block 206 calculates equations (6) and (7). Block 202 executes equation (8) and block 207 executes equations (9) and (10). Block 203 calculates equation (11) and block 208 calculates equation (12). Equation (13) is calculated by blocks 204 and 209.
The result of equation (14) is the output of LPF3 shown in
It should be noted that the preferred embodiment of the disclosure is hard-wired, but alternatively an implementation as a program executed by a processor is also possible.
a shows simulation results of the impedance of loudspeaker as function of frequency. Curve 30 shows the total impedance and curve 31 illustrates the real part of the impedance. The peak of 1 kHz is caused by a resonance in the loudspeaker.
The resonance peak will distort the speaker protection method. For this reason the band pass filters 200 and 201 must be chosen such that the resonance peak is not included in the pass band of the filters.
b shows simulation results of the square root (<vv>/<ii>) 32 and of <iv>/<ii> 33 both as function of frequency. Both curves 32 and 33 demonstrate show the very close correlation between the factors sqrt (<vv>/<ii>) 32 and <iv>/<ii> 33 to the impedance curves shown in
While the invention has been particularly shown and described with reference to the preferred embodiments thereof, it will be understood by those skilled in the art that various changes in form and details may be made without departing from the spirit and scope of the invention.
Number | Date | Country | Kind |
---|---|---|---|
12368025 | Sep 2012 | EP | regional |
Number | Name | Date | Kind |
---|---|---|---|
4301330 | Trump | Nov 1981 | A |
4583245 | Gelow et al. | Apr 1986 | A |
4843562 | Kenyon et al. | Jun 1989 | A |
4980920 | Noro et al. | Dec 1990 | A |
5343404 | Girgis | Aug 1994 | A |
5481617 | Bjerre | Jan 1996 | A |
5577126 | Klippel | Nov 1996 | A |
6865274 | Aarts et al. | Mar 2005 | B1 |
7436967 | Neunaber | Oct 2008 | B2 |
20040086140 | Fedigan et al. | May 2004 | A1 |
20040178852 | Neunaber | Sep 2004 | A1 |
20090257599 | Sand Jensen et al. | Oct 2009 | A1 |
20090304127 | Massam et al. | Dec 2009 | A1 |
20110090604 | Butler | Apr 2011 | A1 |
20110182435 | Gautama | Jul 2011 | A1 |
20110228945 | Mihelich et al. | Sep 2011 | A1 |
20140146971 | Marguery et al. | May 2014 | A1 |
Number | Date | Country |
---|---|---|
2 187 607 | Sep 1987 | GB |
2187607 | Sep 1987 | GB |
05264491 | Oct 1993 | JP |
Entry |
---|
Klipel, Nonlinear Modeling of the Heat Transfer in Loudspeakers, Journal of the Audio Engineering Society, vol. 52, pp. 3-25, Feb. 2004. |
European Search Report 12368025.8-1910 Mail date—Feb. 13, 2013, Dialog Semiconductor GmbH. |
“Complete Thermal Protection of an Active Loudspeaker,” by Peter Chapman, Presented at the 108th Convention Feb. 19-22, 2000, Paris, France, AES, An Audio Engineering Society Preprint, 5112 (G-7), pp. 1-21. |
Number | Date | Country | |
---|---|---|---|
20140086418 A1 | Mar 2014 | US |