EP0833051B1 - Method and device for actioning the interruption of a vehicle engine starter motor - Google Patents

Description

The present invention relates to a method and a device for the automatic shutdown of a starter of motor vehicle.

Usually cutting off the drive from engine by starter is controlled by the user of the vehicle who releases the ignition key when the engine makes a characteristic noise.

However, the tendency to make engines more and more quiet thermal makes it becomes difficult for the user to detect the end of the start-up. This results in more severe solicitation of the starter.

We already know many devices for cut a motor vehicle starter when the engine has started and is sufficiently autonomous to reach its idle speed by itself.

The most efficient of these devices put generally implemented an analysis of the ripples of the starter supply voltage, ripples which are due to variations in current absorbed by the starter when the engine is compressed before the start of it. Such a device is known from FR-A-2393165.

We know that at the end of the training period, when of the first explosions, the engine quickly rises in diet. This is illustrated in the graph in Figure 1, on which we have brought the evolution of the supply voltage of the starter as a function of time during a phase of start-up.

The duration of the undulations - and in particular of the decompression phases - then decreases rapidly. However, the systems proposed to date have a polling window (D ₁ to D ₄ in FIG. 1) corresponding at least to the complete period of the previous ripple and do not take into account the reduced period of the decompression in progress . Thus, it is only at the end of the last time delay (D ₄ in FIG. 1), the duration of which corresponds at least to the duration of the preceding cycle, that one looks at whether a new ripple has occurred or not in this time window and that the decision is made to cut the starter.

It can be understood from this FIG. 1 that this time delay D ₄ is too long; the decision to cut off the starter power supply is much too late compared to the actual starting of the vehicle's thermal engine, indicated by the arrow D _R in FIG. 1.

An object of the invention is therefore to propose a new command which cuts the power to the starter faster.

Furthermore, the voltage signals are generally very strongly affected by parasites from starter switching. That's why we use usually to get filtering particularly effective, an active filter which has the disadvantage to be expensive and to slow down the system response time.

Another object of the invention is to propose a solution that solves this problem.

Thus, the invention provides a method for the vehicle starter cut-off control automobile according to which the ripples of the supply voltage of this starter and we cut the starter when these ripples disappear, characterized in that for each new ripple, generates a polling period of such duration that the said period is intended to end substantially after the appearance of an extremum of said undulation and in that the starter is cut off when no extremum is detected in the last scan period.

So when the engine starts, we cuts the starter power immediately after time when an extremum of tension was expected.

In an advantageous embodiment, the signal on which a voltage extremum is detected starter supply is a clipped signal and opens the polling period when this signal begins to take a non-zero value.

We thus have over the entire period in which this signal is zero of an unpoised signal, which eliminates risk of errors. The start of the scrutiny period is precisely determined.

The invention also provides a device allowing the implementation of this process.

• la figure 1, déjà analysée, est un graphe sur lequel on a porté l'évolution de la tension du démarreur en fonction du temps lors d'une phase de démarrage ;
• la figure 2 est un schéma d'un dispositif de commande de coupure d'un démarreur conforme à un mode de réalisation possible pour l'invention ;
• la figure 3 est un graphe sur lequel on a porté l'évolution en fonction du temps de la tension U analysée par l'unité de gestion du démarreur et qui illustre une variante possible de mise en oeuvre de l'invention ;
• la figure 4 est un logigramme illustrant une mise en oeuvre possible pour le procédé conforme à l'invention ;
• la figure 5 est une représentation semblable à celle de la figure 3 illustrant une autre variante possible pour l'invention.

Other characteristics and advantages of the invention will emerge from the description which follows. This description is purely illustrative and not limiting. It must be read in conjunction with the appended drawings in which:

• FIG. 1, already analyzed, is a graph on which the evolution of the starter voltage has been plotted as a function of time during a starting phase;
• Figure 2 is a diagram of a cut-off control device of a starter according to a possible embodiment for the invention;
• FIG. 3 is a graph on which the evolution as a function of time of the voltage U analyzed by the starter management unit has been plotted and which illustrates a possible variant of implementation of the invention;
• FIG. 4 is a flow diagram illustrating a possible implementation for the method according to the invention;
• Figure 5 is a representation similar to that of Figure 3 illustrating another possible variant for the invention.

FIG. 2 illustrates a device for controlling the supply of a starter motor D, the electric motor M of which is mounted between a supply terminal B ⁺ at the voltage of the vehicle battery and ground.

This device comprises a contactor 1 mounted between the terminal B ⁺ and the motor M.

This contactor 1 is a relay actuated by a winding 2. One end of this winding 2 is connected to the common terminal of motor M and contactor 1. Its other end is connected on the one hand to the source of a MOSFET transistor 3 and secondly to a coil 5 connected to ground by its other end.

Of course, transistor 3 can be replaced by any other type of switch.

The drain of transistor 3 is connected to the supply terminal B ⁺ by means of an ignition key switch 6.

Its grid is connected to the output of a module command 9 which is itself activated by a control unit management 4, which is for example a microprocessor.

The power supply terminal B ⁺ is connected to a first input of the unit 4 - via the switch 6.

It is also connected - via the 6-to switch a second input from unit 4, via a Zener diode 7 which passes from said input to switch 6. A resistor 8 is also mounted between this second input (connected to the anode of the diode Zéner) and the mass.

The voltage measured at this second input is the voltage U across this resistor 8.

When unit 4 detects the closing of switch 6 on the voltage at its first input, it controls the closing of contactor 1 by module 9.

It then analyzes the voltage U to determine the instant when the actual starting of the heat engine intervenes and immediately order the shutdown of motor power supply M.

Voltage U has been shown in solid lines in figure 3.

It is clipped by the lower values by the Zener diode 7.

The Zener voltage of diode 7 is for this purpose chosen lower than the peak supply voltages of the starter, which correspond to the no-load voltage of the battery decreased by voltage drops in the battery and in cables under the effect of the absorbed intensity by the starter. For heat engines from 1 to 2 liters of displacement and starters with a power of 1 kw, the peak voltage is generally between 10 and 11 volts.

Diode 7 is then for example chosen with a Zener voltage of the order of 9 to 10 Volts.

Thus, the voltage signal U collected at the terminals of resistance 8 is stripped of its component continuous and dips of the alternative component. Of the so throughout the duration of a voltage dip power supply (shown in dotted lines in the figure 3), the signal is no longer parasitized, which eliminates risk of error.

The detection of the actual start time D _{R is} carried out as follows.

When the contactor 1 closes, the unit 4 searches for the appearance of a zero voltage across the resistor 8, then determines the time T ₀ where a positive voltage appears.

The unit 4 then determines the time T ₁ corresponding to the first peak voltage UC ₁ .

This time T ₁ can be determined in several ways, for example by checking the moment when the derivative of the voltage U vanishes - which can be achieved by using an analog differentiator circuit or by calculating the differential of the voltage U. can also more simply compare the maximum value UC _i of a batch of one or more voltage measurements with respect to the maximum value UC _i-1 of a previous batch; when UC _i <UC _i-1 , the peak has just been passed.

Once the time T _{1 has been} determined, the unit 4 waits for the signal to cancel again and then becomes positive again in a time T ₂ .

It then determines whether a new voltage peak UC ₃ occurs between this time T ₂ and a time T _x defined by T ₂ + T _c , with T _c = M. (T ₁ -T ₀ ), where M is a parameter beforehand set with a value between 1 and 2.

If no peak is detected, it means that the engine has started. Unit 4 controls the starter supply cut by the module command 9.

On the other hand, if a voltage peak does occur during this time interval, this means that the engine is not yet started. Unit 4 stores the time T ₃ corresponding to this peak UC ₃ , then resumes the process by replacing T ₀ with T ₂ and T ₁ with T ₃ and T _c with M (T ₃ -T ₂ ).

It thus determines for the following cycle whether a new peak occurs between the time T ₄ + M (T ₃ -T ₂ ).

The process is thus continued until the end of the first observation window where we do not detect new voltage peak.

It will be noted that with such an implementation, the cut-off order occurs after the last dip, as soon as one is beyond the instant T _x before which the new voltage peak was expected.

This gives a response time particularly fast and in particular lower than response time obtained by a solution in which the cut-off order occurs after the last trough at the end of a time window with a duration greater than the duration of the previous compression cycle.

The flow diagram in Figure 4 gives an example of possible implementation for the cutting strategy which has just been described.

Measurements U _n are carried out periodically on the voltage U at different instants T _n successive, the time interval between T _n and T _{n + 1} corresponding to the sampling period of the microprocessor.

First, we check in a test 10 if the value U _n is zero.

If this is not the case, we increment n in a step 11 and we check in a test 12 if the new time T _n is not greater than a time T _x previously chosen, for example equal to 0.3 seconds .

If T _n is greater than this time T _x , this means that U has not taken a zero value in the time interval [0, T _x ] and we consider that there is an incident. Unit 4 cuts the power supply to the starter motor (output 12a).

This incident may for example be due to a contact 1 closed, at a break in the power constituted by windings 2 and 5 or even when the starter is running without load (no meshing of the pinion in the crown, rupture of a mechanical component power, etc.).

If, on the other hand, T _n is actually less than T _x , we resume test 10.

When one of the values U _n of U is zero, it is checked in a loop 13 that U is effectively zero on several values of successive samples, that is to say that the zero value detected for U is not a false detection which would be due to a parasite.

To this end, in a step 14, we replace n by n + 1.

Then in a step 15, it is checked that the new time T _n is still not greater than T _x . Of course, if T _n is greater than T _x , the management unit 4 cuts the supply to the starter (output 15a). This indeed means that the positive ripple did not occur in the time it should have. This incident can have several causes: short-circuit starter, blocked starter, blocked thermal engine, insufficiently charged battery, etc.

If, on the other hand, T _n is less than Tx, in a step 16, J is replaced by J + 1 and U ₁ by the largest value between U ₁ and U _n , J being an increment index whose initial value is 0, U ₁ being a parameter whose initialization value is 0.

In a step 17, it is checked whether the new value of J is or is not equal to JMAX (which is for example chosen equal to 3).

If this is not the case, the process is resumed. from step 14, again replacing n with n + 1.

On the other hand, if J is equal to JMAX, the value of U ₁ is verified in a step 18.

If U ₁ is zero, J is reset to 0 (step 19) and the process is repeated again from step 14, which increments n.

On the other hand, if U ₁ is not zero - which normally means that the voltage U becomes positive again after a plateau where it has taken the value zero - we then replace in a step 20 T ₀ by T _n and J by 0. T ₀ is then the starting point of the positive ripple.

T _{x is} then replaced by T ₀ + T _c , where T _c characterizes the duration of the time window which is opened from T ₀ and in which it is checked whether U takes a peak value or not. For example, T _{c is} initially given the value of 0.3 s.

Then, in a step 21, one then replaces n by n + 1, K by K + 1 (K being an increment index whose initial value is 0), as well as U ₂ by the larger of the two values between U ₂ and U _n , where U ₂ is a parameter whose initial value is zero.

In a test 22, we check whether K has reached or not its maximum value KMAX (for example chosen equal to 3).

If this is not the case, step 21 is repeated by incrementing K and n and modifying the value of U ₂ if necessary.

This looping through the counter K makes it possible to ensure that the measurement of U ₂ has not been distorted by a parasite.

When K has reached its value KMAX, it is checked in a test 23, whether or not U ₂ is less than or equal to U _c , where U _c is a voltage value which characterizes the maximum voltage reached by the voltage U during these oscillations and to which we give the initial value of 0.

If U ₂ is greater than U _c - which means that the voltage U increases - we replace in a step 24 U _c by U ₂ , then we start again from step 21 by checking beforehand whether or not T _n is greater at time T _x as defined in step 20 (test 25).

If T _n is actually greater than the time T _x , this means that no peak has been detected in the time interval T _c and the management unit cuts off the starter power supply (output 25a).

When U ₂ becomes less than U _c within this time interval, this means that we have just exceeded a peak. In this case, in a step 26, T ₁ is replaced by T _n and K, U _c and U ₁ are reset.

In a test 27, it is then checked whether U _n is zero.

If this is not the case, we check in a test 28 if T _n is less than T ₁ + T ₂ , where T ₂ is a fixed time parameter, for example equal to 0.3 s, corresponding to a duration maximum during which the voltage U is positive.

In the case where T _n becomes greater than T ₁ + T ₂ , we consider that there is an incident. This can be due to a blockage of the engine on compression or to an insufficiency of the starter torque (mechanical seizure, internal resistance become too great by heating, the empty battery, ...).

If T _n is actually less than T ₁ + T ₂ , we increment n (step 29) and we repeat test 27 until U _n is zero.

When this is the case, in a step 30, T _c is replaced by M (T ₁ -T ₀ ) and T _x by T _n - T ₂ , then the process is resumed from step 14.

It will be noted that due to the clipping of the Zener diode, the value of the times corresponding to the beginnings of positive undulations (T ₀ , T ₂ , T ₄ in FIG. 3) is particularly precise and not very sensitive to noise.

Other variant embodiments of the invention are of course conceivable.

In particular, the duration of the time window in which a peak must intervene can be determined other than from the duration of the last decompression (which corresponds to the rising phase of positive undulations, that is to say to T ₁ -T ₀ , T ₃ -T ₂ , etc. in Figure 3).

For example, as illustrated in FIG. 5, the time T _c can be determined from the total duration of the previous positive ripple, by applying to this duration a coefficient M of between 0.5 and 1.

This makes it possible to avoid the uncertainties on the determination of the time T ₁ of appearance of the peak, which is relatively imprecise due to the more or less flattened shape of the ripple.

Thus, in FIG. 5, the new time T ₁ determined is that corresponding to the reappearance of the zero voltage. The duration of the ripple being T ₁ -T ₀ , the vertex of the next ripple is approximately 0.5 times T ₁ -T ₀ .

Also again, in all of the above we have describes the preferred case where a detected extremum is a peak of the supply voltage. This extremum could variant be a minimum of said voltage the signal of voltage being for example clipped by values higher.

Voltage peak detection power supply is preferred, the voltage being more regular in the vicinity of ridges only near the minima.

Claims (10)

1. A method of controlling the stopping of a motor vehicle starter (D), wherein the fluctuations in the supply voltage (B⁺) of the said starter (D) are detected and the starter (D) is stopped when these fluctuations disappear, characterised in that, for each new fluctuation, a watching period (T_c) is generated, the duration of which is such that the said period is arranged to terminate substantially after a peak (UC₁, UC₃) of the said fluctuation has appeared, and in that the starter (D) is stopped when no peak (UC₁, UC₃) has been detected in the last watching period (T_c).
2. A method according to Claim 1, characterised in that the signal (U) of which a peak (UC₁, UC₃) in the supply voltage of the starter (D) is detected is a truncated signal, and in that the watching period (T_c) is opened when this signal begins to assume a non-zero value.
3. A method according to Claim 2, characterised in that the duration (T_c) of a watching period corresponds to the duration of time determined on the preceding fluctuation between the instant (T₀, T₂, etc.) at which the said signal begins to assume a non-zero value and the instant (T₁, T₃, etc.) that corresponds to its peak, this duration being multiplied by a coefficient (M) in the range between 1 and 2.
4. A method according to Claim 1 or Claim 2, characterised in that the duration (T_c) of a watching period is determined from the duration, determined on the preceding fluctuation, during which the signal assumes non-zero values, this duration being multiplied by a coefficient (M) in the range between 0.5 and 1.
5. A method according to one of the preceding Claims, characterised in that the peak (UC₁, UC₃) which is detected corresponds to a peak of the supply voltage.
6. A method according to Claims 2 and 5 taken in combination, characterised in that the signal (U) on which a peak (UC₁, UC₃) is detected corresponds to the supply voltage (B⁺) truncated by removal of the lower values.
7. Apparatus for controlling the stopping of a motor vehicle starter, comprising means (4, 7, 8) for detecting fluctuations in the supply voltage of the said starter, together with means (4, 9) for the stopping of the starter when these fluctuations disappear, characterised in that it includes means (4) for generating, on each new fluctuation, a watching period (T_c), the duration of which is such that the said period is adapted to terminate substantially after a peak (UC₁, UC₃) is detected in the said fluctuation, and means (4, 9) for stopping the starter when no peak has been detected in the last watching period (TC).
8. Apparatus according to Claim 7, characterised in that it includes means (7) for truncating the signal on which a peak of the supply voltage of the starter is detected, together with means (4) for opening the watching period when this signal begins to assume a non-zero value.
9. Apparatus according to Claim 8, characterised in that the means for truncating the signal comprise a Zener diode (7).
10. Apparatus according to Claim 9, characterised in that the said means consist of a bridge which includes the said Zener diode (7), together with a resistor (8) connected between the anode of the said Zener diode (7) and earth, the supply voltage of the starter (D) being injected on the cathode of the said Zener diode, and the signal on which a peak of the said supply voltage is detected being the voltage across the said resistor (8).

Images