WO2009098609A1

WO2009098609A1 - Load detector for a switching device

Info

Publication number: WO2009098609A1
Application number: PCT/IB2009/050338
Authority: WO
Inventors: Marco Berkhout; Martojan Koerts
Original assignee: Nxp B.V.
Priority date: 2008-02-04
Filing date: 2009-01-28
Publication date: 2009-08-13

Abstract

A load detector for detecting the presence of a load (R) at the output of a switching -device ( 600) is provided. The switching device includes a comparator which generates a pulse signal (Vp) based on a information signal (Vin) and which generates an output signal (II) based on the pulse signal (Vp). The load detector analyzes whether the output signal (II) fulfills a predetermined requirement at predetermined moments. Suitably, this predetermined requirement may be whether an average value (il) of the output signal (IL) not exceeds a predetermined limit value (Him). The invention enables to detect the presence of a load (R) at the output of the switching device (600) without knowing further details about the construction of the switching device (600).

Description

LOAD DETECTOR FOR A SWITCHING DEVICE

FIELD OF THE INVENTION

The invention relates to a switching device including a transformer for transforming an information signal into a pulse signal and method for detecting a load at the output of a switching device.

BACKGROUND OF THE INVENTION

In the fields of automotive engineering, car applications need to be ultra-light and highly efficient. Especially audio applications cannot come up to these requirements since their electronic components not only produce large losses but they are also very heavy. One of the most lossy components in car applications is the audio output stage.

An audio output stage utilizes an audio amplifier to amplify electronic audio signals. Such audio amplifiers are divided into a plurality of classes depending on the working principle. A very significant difference in the working principle is, whether the information signal is amplified by an analog design or a switching design.

Analog amplifiers merely amplify electronic audio signals with electronic valve elements like transistors or vacuum tubes. Analog amplifiers are extremely lossy and very heavy if a large amplification and high output power is required.

The counterparts to analog amplifiers are switching amplifiers. These amplifiers transform the information signal into a pulse signal based on a reference signal having a ripple shape. This pulse signal is then amplified. A demodulation filter finally reconstructs the amplified information signal. Since switching amplifiers only amplify the pulse signal, they reach a higher efficiency by a smaller size and a lighter weight in comparison to analog amplifiers. A must have requirement for amplifiers in car applications is the ability to detect whether a load is connected to the output of the audio output stage. In analog amplifiers, the presence of a load is merely detected by applying a test signal and measuring whether the output current is larger than a predetermined limit current. However, in switching amplifiers, the presence of a demodulation filter complicates matters. A conventional method for load detection for switching transistors is proposed in WO 2003/098804. Therein, a switching amplifier comprises a load detector generating an over-current voltage indicating whether the output current of the switching amplifier exceeds a predetermined limit current in case that the above mentioned pulse signal is low. In other words, this over-current voltage is a pulse signal containing all information necessary to determine if a load is connected or not. However, such a load detector requires non-trivial signal processing. Furthermore, knowledge about the above-mentioned ripple current, the switching frequency and the operational voltage need to be known for interpreting the over- current voltage. This makes the conventional load detector in switching amplifiers complicated and inflexible.

OBJECT AND SUMMARY OF THE INVENTION

It is therefore an object of the present invention to provide a switching device including a transformer for transforming an information signal into a pulse signal having a more simple load detector and a method for detecting a load at the output of a switching device in a more simple way.

The object is solved by the features of the independent claims. Further embodiments are subject of the dependent claims. The inventor recognized, that the output signal generated in the demodulation filter of a switching device comprises a portion arising from the information signal and a portion arising from the reference signal. Further, in contrary to the portion arising from the information signal, the portion arising from the reference signal has zero average value. Finally, the portion arising from the information signal can only occur at the output of the switching device, if a load is connected thereto. Therefore, the inventor proposes to analyze the average value of the output signal for determining whether a load is connected to the switching device.

A switching device according to the invention includes two transformers. The first transformer generates a pulse signal based on an information signal. The pulse signal is then re-transformed into an output signal by the second transformer. A load detector connected to the second transformer determines whether the average value of the output signal fulfils at least one predetermined requirement at at least one predetermined moment. The determination of an average value can be performed without any further knowledge about other construction features of the switching device. The at least one predetermined requirement can be used to e.g. define the sensitivity of the load detector. Thus, since no further knowledge about the switching device is necessary to detect the presence of a load, the load detector can be constructed simple and independent from the construction of the switching device. This enables to construct one kind of load detector for a plurality of different kind of switching devices. A further advantage is that no complicated signal processing is necessary to interpret the output signal for detecting the presence of a load. Thus, the load detector of the present invention could be integrated together with the switching device in a very easy way.

Suitably, the switching device is a switching amplifier and includes an amplifier connected between first and second transformer. Therewith, the pulse signal is amplified prior passing second transformer. Switching amplifiers require a connected load during its operation, such that the invention is very suitable for a usage in this technical field. Load detection is used during assembly to verify if all speakers are properly connected and also serves as a diagnostic tool during trouble shooting, e.g. no audio in one channel. Therefore, the present invention is very suitable for an application in switching amplifiers. A suitable embodiment for the predetermined requirement is whether the average value of the output signal not raises a limit value. The definition of a limit value allows to easily define the sensitivity of the load detector. If the limit value is further adjustable, the sensitivity of the load detector can be adapted to the application field of the switching device.

Two non-limiting embodiments are proposed to determine the average value. On the one hand, instantaneous values halfway between two extremes of the output signal include all information about the average value of the output signal. Therefore, these instantaneous values of the output signal may be preferably verified with a limit value. If the output signal exceeds the limit value at the moments halfway between two consecutive extremes, the presence of a load is unambiguously detected. On the other hand, the zero crossings of the output signal also include all information about the average value. Therefore, the distance between two consecutive zero crossings of the output signal may be preferably detected. If the detected distance is higher or smaller than a predetermined distance, the presence of a load is unambiguously detected. Preferably, it may be detected, whether the consecutive distances of the zero crossings keep constant. Even at least three detection cycles are necessary to define the average value, such detection does not need further information to define a limit value for the predetermined distance between the zero crossings.

A third embodiment proposes to derive a verification signal from the information signal prior the first transformer and to verify the output signal with this verification signal. As already mentioned, the output signal includes a portion corresponding to the information signal. Further, this portion needs to be detected in the output signal, to determine the presence of a load. Therefore, the input signal is a suitable basis as a reference for a verification signal to detect the information signal portion in the output signal. Preferably, the verification signal is derived from an integrated sum of the pulse signal and the information signal. This guarantees that the zero crossings of this signal are almost exactly halfway between two edges at the reference signal. Therewith, there is no need to make further calculations for detecting an average signal in the output signal.

In a method for detecting a load signal at the output of a switching device including a transformer for transforming an information signal into a pulse signal, the pulse signal is firstly filtered by at least one inductivity and then analyzed, whether the current through the inductivity fulfils at least one predetermined requirement at a predetermined moment. This method has the same advantage as the above-mentioned switching apparatus in its most simple embodiment.

All above described features for the switching device with the respective advantages may also be part of the preceding method.

BRIEF DESCRIPTION OF THE DRAWINGS

The invention will be described in greater detail hereinafter, by way of non- limiting examples, with reference to the embodiments shown in the drawings.

Fig. 1 is a device according to a first and second embodiment of the invention;

Fig. 2 is a diagram showing the working principle of the comparator shown in fig. 1;

Fig. 3 is a diagram showing the working principle of the low-pass-filter shown in fig. 1;

Fig. 4 is a diagram showing the working principle of a load detector;

Fig. 5 is a diagram showing the working principle of a first and second embodiment of a load detector according to the present invention;

Fig. 6 is a device according to a third embodiment of the invention;

Figs. 7a, b are diagrams showing the transformation of the information signal in fig. 5;

Fig. 8 is a diagram showing the working principle of third load detector according to the present invention;

Fig. 9 is a diagram showing a state diagram of the third load detector; and

Fig. 10 is a diagram showing the output signal of the third load detector; DESCRIPTION OF EMBODIMENTS

Fig. 1 is a device 100 according to a first and second embodiment of the invention. The device 100 includes a first transformer 110 and a second transformer 120 connected to the first transformer 110. Without limiting the invention, the first transformer 110 may be a comparator 110, since comparators are the most easiest way to derive pulse signals. The second transformer 120 may merely include a low-pass filter to retransform a pulse signal into an information signal. In the most simple form, such a low- pass filter function may be realized by e.g. a single inductor L. However, in the present embodiment, a LC-filter 120 should be applied, since LC-filters are suitable for lossless audio applications to maintain a high efficiency of the device 100. However, the invention should not be reduced to low-pass filters using such passive electronic elements. As well known for the skilled person, low-pass filters could also be realized by e.g. active electronic components like operational amplifiers. The comparator 110 receives an information voltage V_1n and a reference voltage V_ref. Both signals are compared to each other. As an output signal, the comparator 110 generates a pulse voltage V_p having a positive amplitude based on whether the information voltage V_1n is larger than the reference voltage V_ref and a negative amplitude in the other case. Next, the LC-filter 120 receives the pulse voltage V_p from the comparator 110 and reconstructs the information voltage V_1n as an output voltage V_out based on the pulse voltage V_p.

In the following, the working principle of the comparator 110 and the LC-filter 120 should be explained in more detail.

Fig. 2 is a diagram showing the working principle of the comparator 110. For a more practical explanation, the information voltage V_1n should be regarded as having a sinus shape. The reference voltage V_ref should be a ripple signal with a triangular shape. The comparator 110 will compare the amplitude of the information voltage V_1n and the reference voltage V_ref. If the information voltage V_1n exceeds the reference voltage V_ref, the pulse voltage V_p will be positive. Otherwise, the pulse voltage V_p will be negative.

The pulse voltage V_p has positive and negative pulses T_n, T_p. A neighboring negative pulse T_n and positive pulse T_p represents one cycle. The ratio between the length of the positive and negative pulses T_n, T_p within one cycle length is named duty cycle. The cycle duration of the pulse voltage V_p is the same as the cycle duration of the reference voltage Vref. The duty cycle of the pulse voltage V_p varies. The length of each pulse T_p, T_n depends on the level of the information voltage V_1n. The higher the level of the information voltage V_1n, the longer is negative pulse T_n and the shorter is the positive pulse T_p. Thus, the pulse voltage V_p includes information about the shape of the information voltage V_1n. Such a procedure is known as pulse width modulation (PWM).

Fig. 3 is a diagram showing the working principle of the low-pass filter 120. Exemplary, the low-pass filter comprises an inductor L and a capacitor C connected in series. The pulse voltage V_p is supplied to the inductor L. The output voltage V_out is gripped from the connection between the inductor L and the capacitor C. The other side of the capacitor is connected to ground. Both, the inductor L and the capacitor C should be regarded as ideal electronic elements.

Depending on the positive or negative pulse of the pulse voltage V_p, each pulse T_p, T_n generates a constantly raising or constantly falling inductor current I_L through the inductor L. In other words, the inductor current I_L has a triangular ripple shape with rising and falling slopes. Since the length of each pulse T_p, T_n depends on the shape of the information voltage V_1n, the rising and falling slopes of the inductor current I_L are not equally long. Therewith, the average value of the inductor current I_L is not zero. The average value follows the shape of the information voltage V_1n. It could also be said that the inductor current I_L is a superposition of a current having a shape similar to the reference voltage V_ref and a current having a shape similar to the information voltage V_1n.

The capacitor C blocks signals with low frequencies and lets pass signals with high frequencies. In other words, the high frequency component of the inductor current I_L is grounded, whereas the low frequency component is blocked by the capacitor C. Therefore, the capacitor C grounds the ripple component of the inductor current I_L and generates an output voltage V_out depending on the average value of the inductor current I_L and therewith on the information voltage V_1n. This output voltage V_out can be gripped from the port between the inductor L and the capacitor C.

Fig. 4 is a diagram showing the working principle of a load detector. As shown above, the pulse voltage V_p and the inductor current I_L depend on the shape of the information voltage V_1n. However, while the pulse voltage V_p merely depends on the presence of the information voltage V_1n, the inductor current further depends on the presence of a load R connected to the capacitor C in parallel. In detail, if e.g. a constant positive information voltage V_1n is supplied to the device 100, the comparator 110 generates a pulse voltage Vp having a duty cycle lower than 50%. Due to the further presence of the capacitor C, this would generate a positively shifted inductor current I_L with a triangular shape at the inductor L (ideal inductance). However, the direct portion of the inductor current I_L can not discharge over the capacitor C. Therefore, the presence of a load R can be detected by supplying a DC- voltage as information voltage V_1n and measuring whether the inductor current I_L has a DC-portion.

The alternating portion of the inductor current I_L is shifted 90° forward in respect to the reference voltage V_ref. Therefore, the following condition can be formulated for the inductivity current I_L without a DC-portion. If the reference voltage V_ref rises, the inductor current I_L must be negative, and vice versa. Conventional load detectors use this condition and generate an over-current voltage oc if the inductor current I_L is negative and falling. To further consider tolerances, the over-current voltage oc is only generated, when the inductivity current I_L further falls below a predetermined negative limit value I_lim. If no load R is present (namely, if the inductor current I_L does not have a direct portion), the duty cycle of the over-current signal oc must not exceed a predetermined value.

However, the determination of this predetermined value for the duty cycle of the over-current signal oc requires some none-trivial signal processing. Further, knowledge about the frequency of the reference voltage V_ref and the inductance of the inductor L is necessary. In case of further amplifying the pulse voltage V_p, also the supply voltage for the amplifier must be considered.

The present invention therefore proposes a method for detecting the presence of a load R without determining a predetermined value for the duty cycle of the over current signal oc.

Fig. 5 is a diagram showing the working principle of a first and second embodiment of a load detector according to the present invention. The diagram of fig. 5 is an elongated section of the diagram shown in fig. 4. According to fig 5, the present invention proposes to determine the average value I_L of the inductor current I_L. AS already outlined, the inductor current I_L is a superposition of a current having a shape similar to the reference voltage V_ref and a current having a shape similar to the information voltage V_1n. However, since the reference voltage Vref is a signal having zero average value, the average value I_L of the inductor current I_L must be similar to the information voltage V_1n. In a first embodiment, the present invention proposes to determine the average value I_L half-ways between two consecutive edges of the inductor current I_L. Since the periodic component of the inductor current I_L has a triangular shape, the periodic component of the inductor current I_L is zero half- ways between two consecutive edges. Therefore, only the average value I_L of the inductor current I_L become effective at this moments in time.

In a second embodiment, the present invention proposes to determine the zero crossings of the inductor current. Since the zero crossing of an average-free triangular shaped current must have a constant distances to each other, the occurrence of varying distances between the zero crossings of the inductor current I_L unambiguously indicates the presence of a further load current and therewith the presence of a load R.

It should be outlined, that a plurality of further methods for determining an average value in a signal are known to a skilled person in the art. Examples are the determination by integrating the signal or by analyzing the spectrum of a signal. Therefore, the above examples should not be used in any limiting purpose.

Fig. 6 is a device 600 according to a third embodiment of the invention. The device comprises three voltage-to-current transformers 610, 630, 640, two integrators 620, 660, a voltage controlled current source 650, a comparator 670, an amplifier 680 and a low-pass filter 690. A first voltage to current converter 610 transforms an information voltage V_1n into an information current I_1n. The information current I_1n is superposed with a pulse current I_p resulting from an amplified pulse voltage V_p to be described later. A first integrator 620 integrates the sum of the information current I_1n and the pulse current I_p and makes sure that first comparator voltage Vi steadily changes. The first comparator voltage Vi is fed into the comparator 670 and further transformed into a first current Ii , which is added to a reference current I_ref outputted by the voltage controlled current source 650. The resulting sum is then integrated by a second integrator 660 making sure that a second comparator voltage V₂ is shifted based on the information voltage V_1n. The second comparator voltage V₂ is also fed into the comparator 670. Based on the levels of the first and second comparator voltage Vi, V₂, the comparator 670 outputs a pulse voltage V_p, which is then amplified by an amplifier 680. In the present embodiment, a digital amplifier is shown. However, it should not be used in a limiting scope. As already mentioned, the amplified pulse voltage Vp is then fed to the input of the device 600 and into the low pass filter 690 working according to the same principle as already discussed in fig. 3. At the output of the low pass filter 690, the amplified information voltage V_1n can be gripped as output voltage V_out. To more fully understand the third embodiment, the principle of the two integrators 620, 660 should be explained in more detail.

Figs. 7a, b are diagrams showing the generation of the comparator voltages V₁, V₂. Fig. 7a shows the generation of the first and second comparator voltage V₁, V₂ if no information voltage V_1n is supplied to the device 600, and fig. 7b shows the generation of the first and second comparator voltage Vi, V₂ if a positive information voltage V_1n is supplied to the device 600.

Turning first to fig. 7a, if no information signal V_1n is present, the duty cycle of the pulse voltage V_p is 50%. Further, only the pulse voltage V_p is integrated at the first integrator 620. Therefore, the first comparator voltage Vi has a triangular shape wherein the rising and falling slopes are identically steep. For generating the second comparator voltage V₂, the sum of the reference current I_ref and the first current Ii (generated based on the first comparator voltage Vi) is fed into the second comparator 660. The reference current I_ref is a pulse current with a duty cycle of 50%. That is, after integrating the reference current I_ref results into triangular voltage having identically steep rising and falling slopes. Due to the control loop, this voltage is shifted and results into the second comparator voltage V₂ crossing the intersection points of the first control voltage Vi and the pulse voltage Vp. This shift is zero in fig. 7a, since no information voltage V_1n is present.

In fig. 7b, it is assumed, that a positive information signal V_1n is fed into the device 600. In this case, the duty cycle of the pulse voltage V_p is less than 50%. The control loop therefore forces, that the rising slopes of the first comparator voltage Vi are less steep than the falling slopes. Therefore, the intersection points of the pulse voltage V_p and the first comparator voltage Vi move closer to the V=0-axis. However, this causes that the second comparator voltage V₂ will be shifted upwards, since the second comparator voltage V₂ must cross the intersection points of the pulse voltage V_p and the first comparator voltage V₁.

Fig. 8 is a diagram showing the working principle of third embodiment of load detector according to the present invention. The over-current signal oc is generated in the same way as generated in the prior art. However, as already mentioned, the present invention enables an interpretation of the over-current signal oc without having detailed knowledge about the reference current I_ref, the inductance of the inductor L and the amplification in the amplifier 680. This is achieved by verifying whether the average value I_L of the inductivity current I_L exceeds a predetermined limit current. In the present embodiment, it is therefore proposed to verify the inductivity current I_L with a verification signal having definitely zero average value. As already mentioned, such a signal is the first comparator voltage V₁. The first comparator voltage Vi is point symmetric to the intersection points with the V=O axis. Thus, a verification signal sgn may be derived, which indicates whether the first comparator voltage Vi is positive.

In a first step, it should be assumed, that limit value I_lim=0. The inductor current I_L is shifted 90° in time in respect to the falling and rising slopes of the pulse voltage Vp. Thus, if the inductor current I_L does not have an average value I_L, the first comparator voltage Vi must be larger than zero if the inductor current I_L is lower than zero. Thus, in such a case, the over-current signal oc would rise together with the verification signal sgn.

That is, if a load R is connected to the device 600, the over-current signal oc will rise prior or together with the verification signal sgn, since the inductor current I_L has zero average value I_L.

However, to consider tolerances for the average value I_L of the inductor current I_L, the limit value I_lim should be set to a value below zero. Therefore, the presence of a load R will not be indicated until the average value I_L of the inductor current I_L falls below the limit value I_lim.

Fig. 9 is a diagram showing an asynchronous state diagram of the load detector 900. Thereafter, the load detector 900 has four states S0-S3. States So and S3 indicate the absence of a load, whereas states Si and S₂ indicate the presence of a load. Therefore, in states Si and S₂ a signal detect is outputted by the load detector 900. The load detector starts in the state So.

If the verification signal sgn precedes the over-current signal oc the state machine jumps to state S3 and remains there until both, the verification signal sgn and the over-current signal oc will be low again and the load detector 900 will be transferred back into state So. In other words, state S3 represents the case, that the average value I_L of the inductor current I_L fall below the limit value I_lim.

If on the other hand the over-current signal oc precedes the verification signal sgn, the state machine will jump to state Si and the signal detect goes high. Therefore, state Si represents the moment, if the average value I_L of the inductor current I_L exceeds the limit value I_lim. From state S₁, the load detector 900 could only jump to state S₂ if both, the verification signal sgn and the over-current signal oc are low again. Therefore, state S₂ is analogue to state So. However, in state S₂, the signal detect is high. Therefore, analogue to state So, the load detector 900 can jump to state S3 or back to state Si depending on which of the verification signal sgn and the over-current signal oc goes high first.

The load detector 900 is shown as an asynchronous state diagram. Based on the input values oc (derived from the inductor current I_L and the average value I_L) and sgn (derived from the first comparator voltage Vi), a technical realization of such a state diagram as a finite state machine is well known to skilled person. The output signal detect of the load detector 900 is the result and could be used for further purposes like automatically disabling the switching device, if the absence of a load R is detected. Popular examples for finite state machines are electronic circuits realized with flip-flops or virtual state machines realized as programs in a microcontroller. However, these examples should not limit the invention.

Fig. 10 is a diagram showing the output signal detect of the load detector according to the third embodiment. It is shown, that the signal detect is high as long as the average value I_L of the inductor current I_L exceeds the limit value I_lim.

The present invention proposes a load detector for a switching device. In a switching device, a load cannot be detected by merely supplying a test signal to the output, since a switching device steadily outputs a signal. The present invention therefore proposes to detect the presence of a load by measuring the average value of the output signal generated by a low-pass filter included in the second transformer of the switching device. Therewith, it is not necessary to consider other characteristics of the switching device for the load detection.

Claims

CLAIMS:

1. Switching device including a first transformer (110, 670) for transforming an information signal (V_1n) into a pulse signal (V_p), further including: - a second transformer (L) generating an output signal (I_L) based on the pulse signal (V_p); and

- a load detector (900) connected to the second transformer (L) and adapted to determine, whether an average value of the output signal (I_L) fulfils at least one predetermined requirement at a predetermined moment.

2. Switching device of claim 1, wherein the switching device (100, 600) is a switching amplifier and further includes an amplifier (680) connected between the first transformer (110, 670) and the second transformer (L) and adapted to amplify the pulse signal (V_p) prior passing the second transformer (L).

3. Switching device of claim 1 or 2, wherein the predetermined requirement is whether the instantaneous average value (I_L) of the output signal (I_L) not exceeds a predetermined limit value (I_lim).

4. Switching device according to one of the claims 1-3, wherein the predetermined moments are halfway between two extremes of the output signal (I_L).

5. Switching device according to one of the claims 1-3, wherein the predetermined moments are the zero crossings of the output signal (I_L).

6. Switching device of claim 5, wherein the predetermined requirement is a predetermined distance between the zero-crossings of the output signal (I_L).

7. Switching device of claim 6, wherein the predetermined distance is constant.

8. Switching device according to one of the claims 1-3, wherein the output signal (I_L) is verified with a verification signal (sgn) having a predetermined duty cycle.

9. Switching device according to claim 8, wherein the verification signal (sgn) is derived from the integrated sum of the pulse signal (V_p) and the information signal (V_1n).

10. Method for detecting a load signal at the output of a device including a transformer for transforming an information signal into a pulse signal, the method includes the following steps:

- filtering the pulse signal by at least one inductor; and

- detecting whether the current through the inductor fulfils at least one predetermined requirement at a predetermined moment.