Method and circuit system for calibrating voltage and temperature deviations of the effective current of hydraulic valves in a pwm drive

Abstract

The present invention relates to a method for reducing deviations between the effective current and the measured current in a pulse-width-modulated current control, in particular for electronic brake control units of motor vehicles, wherein the measured current is determined at a certain predefined time within an actuation period and a compensation occurs by means of temperature-responsive and/or supply-voltage-responsive compensation variables which are added to the measured current such that a corrected nominal current is available for current control. The invention also relates to a circuit arrangement for actuating several inductive loads and comprises a circuit for PWM control of the load current. The method of the invention is implemented as a program in a microcomputer or microcomputer system that is electrically connected to the PWM circuit.

Description

TECHNICAL FIELD

The present invention relates to a method for reducing deviations between the effective current and the measured current in a pulse-width-modulated current control, in particular for electronic brake control units of motor vehicles, and a circuit arrangement for driving several inductive loads comprising a circuit for the PWM control of the load current.

BACKGROUND OF THE INVENTION

It is known that significant differences between the regulated nominal current and the effective current occur in the coil in a valve actuation by means of a pulse-width-modulated current (PWM current control), at least when the ratio between PWM frequency and the time constant of the coil is unfavorable. It is further known in the art that there are dependencies on external parameters such as supply voltage and temperature.

For example, the (maximum possible) current I_100%=V_REFx/(R_L+R_DSon-LS).  (1) flows through a permanently activated inductive load (e.g. valve coil).

This current consequently depends on

• • the voltage at the top side of the valve, and, thus, indirectly, on the battery voltage available in the motor vehicle at terminal KL30B,
  • on the coil resistor R_L and (to a lower degree) on the on-resistor R_DSon-LS of the semiconductor element(s) used to actuate the load(s). Both resistors are highly temperature-responsive: variations of approximately 0.4% per 1° C. for the load resistor (this is e.g. the temperature coefficient for copper, real coils have a somewhat lower dependency) und 0.5% per 1° C. for R_DSon-LS (e.g. Power-MOSFETs, provided on one chip) are typical values.

BRIEF SUMMARY OF THE INVENTION

An object of the invention is to disclose a method and a circuit arrangement for driving loads, reducing deviations from the nominal current and the effective current that flows in the load.

This object is achieved according to the described method and the described circuit arrangement.

A compensation variable according to the invention implies a compensating current ΔI, which can adopt both positive and negative numerical values.

In the method of the invention for reducing deviations between the effective current I_RMS and the measured current I_meas in a pulse-width-modulated current control, it is preferred to determine the measured current I_meas in the middle of the switching time t_on during an actuation period t_PWM.

In a preferred manner, the supply voltage dependency is compensated by the extraction of a valve-related table from defined discrete reference points, wherein the discrete reference points are especially favorably formed of pairs of values produced from the nominal current I_nominal and the supply voltage V_KL30B. Further, it is preferred that values lying between the discrete reference points are determined by interpolation.

Favorably, the valve-related table is stored in a data memory that is preferably a non-volatile data memory in which data is preserved even after the ignition's switch-off.

It is preferred that the compensation variables are determined separately for each load, in particular for each valve coil, or stored in a table.

BRIEF DESCRIPTION OF THE DRAWINGS

FIG. 1 exhibits the difference between the current measurement in PWM actuation and the average current and the effective current.

FIG. 2 shows the difference between the measured current and the effective current for a typical hydraulic valve.

FIG. 3 shows the difference between the measured current and the effective current for a typical hydraulic valve relative to the difference that prevails with a supply voltage of 12 volt and at a temperature of 25° C.

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

The current variation at a valve coil is plotted as a function of time t in FIG. 1. With a current control by means of PWM actuation, an average current of I_AVG=DC*I_100%=DC*V_REFx/(R_L+R_DSon-LS)  (2) develops, with DC indicating the pulse-duty factor (Duty Cycle) of the PWM actuation. The mode of operation of a PWM control that can be implemented according to the invention has been disclosed in international patent application PC/EP 0 115 040. Strictly speaking, equality applies only with an actuation by means of straight line 1 or with ideal e-functions.

For current control, it is necessary to measure the present coil current at a defined time, illustrated by the symbol @ (“at”), e.g. after half the switching time t_on. Consequently, the controller adjusts a measured current I_meas of I_meas=I(@t_on/2)=I_nominal  (3).

The measured current I_meas corresponds to the average current I_AVG only when actuation takes place by way of straight lines. With an actuation with ideal e-functions (corresponds to a coil without an iron core), the current I_meas measured at time t_on/2 is higher that the average current I_AVG. In current control of a valve, however, the effective current I_RMS is of interest that is still somewhat lower than the average current I_AVG. Saturation effects (hysteresis) will additionally be encountered in a valve that can be illustrated in a simplified manner as a coil with an iron core, with the result that non-linearities occur, as becomes apparent from the variation of the current curve 2. From this results a further deviation between the effective current I_RMS and the measured current I_meas. Thus, I_nominal=I_AVG=DC*V_REFx/(R_L+R_DSon-LS)  (4) applies in approximation. This equation is the more precise the higher the PWM frequency is.

FIG. 2 shows the difference between the measured current I_meas and the effective current I_RMS for the electromagnetic valve of an electronic brake control unit, plotted by way of the nominal current I_nominal, for different voltages at KL30B and different coil temperatures. The difference decreases with a rising nominal current I_nominal: This results from the fact that the current controller starts reaching saturation (that means the Duty Cycle amounts to 100% approximately).

A first compensation is still relatively simple in order to eliminate the dependency on the nominal current I_nominal. A current difference to be taken from the diagram is added for a defined nominal current I_nominal. This is successful only for a defined voltage and a defined temperature. Example: nominal value compensation at V_KL30B=12 volt and T=25° C. (curve 3). To reach an effective current I_RMS=1 A, a nominal current I_nominal=1 A+62.5 mA is predetermined.

FIG. 3: To detect voltage and temperature dependencies, it is appropriate to illustrate the deviations of the curves of FIG. 2 from a reference curve (at V_KL30B=12 volt and T=25° C.) (see illustration 3). It can be seen that e.g. at a nominal current I_nominal=1.1 A, there is a maximum voltage dependency of −37.5 mA/+29 mA at a voltage ranging from 9 volt to 16.5 volt approximately over a constant temperature of 25° C. On the other hand, with a temperature variation of roughly −40° C. to roughly 180° C. with respect to a constant voltage of 12 volt, it is possible to read a maximum temperature dependency of +10.5 mA/−25.5 mA with a nominal current I_nominal=1.1 A. These two dependencies do not add simply linearly because in the two corner points {17 volt, −40° C.} and {9 volt, 180° C.} only deviations of +30.5 mA/−49.5 mA are reached for 1.1 A. However, the influence of the voltage is significantly greater than the influence of the temperature.

A (valve-related) table is produced for a compensation of the voltage dependency in FIG. 3. Defined discrete reference points are used for this purpose, which respectively comprise a pair of values {I_nominal, V_KL30B}, and one compensation current ΔI is stored with respect to each pair of values {I_nominal, V_KL30B} for the adaptation of nominal values. Intermediate values are determined by means of interpolation. For example, it is desired to achieve an increase of the nominal current I_nominal of 200 mA to 1000 mA at a voltage of 9 volt at the valve and a temperature of 180° C., corresponding to curve 19 in FIG. 3. To achieve the nominal current I_nominal=1000 mA, a compensating current ΔI=−45 mA is added to the nominal current I_nominal=1000 mA. However, as the valve, due to its time constant, follows the specification of the nominal current with delay, initially only a compensating current ΔI=−10 mA is predetermined for compensation. This corresponds to the current compensation at I_nominal=200 mA. The current variation at the valve is thereby adapted to the variation of the curve 19. Further, the compensating current ΔI is adjusted corresponding to the course of the curve 19 until the nominal current I_nominal=1000 mA is reached.

To compensate variations or abrupt changes in the supply voltage (e.g. at KL30B), an averaging operation by way of the present voltage measurement and previous values is preferred.

For the compensation of the temperature dependency, the temperature is indirectly detected by way of the Duty Cycle adjusted by current control. From equation (4), R_L+R_DSon-LS=(DC*V_REFx)/I_nominal  (5) follows. This formula implies that for the present Duty Cycle only the coil resistor R_L (and the on-resistor R_DSon-LS) is responsible; the coil temperature appears only indirectly. Therefore, it is initially suitable to convert the data in illustration 3 to a dependency of the coil resistor R_L (and the on-resistor R_DSon-LS) R_L(T)=R_L(@T_reference)*(1+α_coil*(T_present−T_reference)) or R_DSon-LS(T)=R_DSon-LS(@T_reference)*(1+α_Ron*(T_present−T_reference)).  (6)

In equation (6), the temperature-responsive values of the coil resistor R_L(T) and the on-resistor R_DSon-LS(T) are determined in consideration of known resistor values R_L(@T_reference), R_DSon-LS(@T_reference) at a reference temperature T_reference. To this end, the known resistor values R_L(@T_reference), R_DSon-Ls(@T_reference) are multiplied with coefficients of correction. These coefficients of correction are basically composed of temperature coefficients (α_coil, α_Ron) and a temperature difference between the present temperature T_present, which is determined from the present Duty Cycle, and the reference temperature T_reference. In this respect, α_coil describes the temperature dependency of the coil material used and α_Ron describes the temperature dependency of the on-resistor R_DSon-LS. The on-resistor R_DSon-LS represents the parasitic resistance of a switch, with said switch being realized in the shape of a MOSFET-transistor provided on a semiconductor chip. It is also possible to achieve this required switching function in another way, i.e. by means of relay technology, bipolar technology, etc. To relate the dependency to R_L+R_DSon-LS rather than to the temperature offers the additional advantage that different temperatures in the valve and in the semiconductor chip are detected correctly because these different temperatures are implicitly contained in the present Duty Cycle.

Each one ΔI is in turn stored for pairs of values {I_nominal, R_L+R_DSon-LS} in a table for the purpose of adaptation of nominal values. An additional calibration is suitable in this respect because equation (4) applies only in approximation. It is advisable to measure the Duty Cycle with a specific valve and reference values (e.g. I_nominal=1 A, temp.=25° C., V_KL30B=12 volt) and to convert the table by means of a corresponding offset.

Claims

1-12. (canceled)

13. Method for reducing deviations between the effective current and the measured current in a pulse-width-modulated current control, in particular for electronic brake control units of motor vehicles, wherein the measured current is determined at a certain predetermined time during an actuation period and a compensation is executed by way of compensation variables in response to temperature and supply voltage, which are added to the measured current so that a corrected nominal current is available for current control.

14. Method as claimed in claim 13, wherein the supply voltage dependency is compensated.

15. Method as claimed in claim 13, wherein the compensation variables are stored in a table, in particular in a data memory.

16. Method as claimed in claim 13, wherein several loads are driven, and the compensation variables are fixed individually for each load, in particular for each valve coil.

17. Method as claimed in claim 15, wherein an interpolation is carried out for temperatures lying between two table values in order to determine the optimal compensation variable.

18. Method as claimed in claim 15, wherein an interpolation is carried out for supply voltages lying between two table values in order to determine the optimal compensation variable.

19. Method as claimed in claim 13, wherein an averaging operation is executed by way of the present nominal value and previous nominal values to compensate abrupt changes in nominal values.

20. Method as claimed in claim 13, wherein the temperature is determined indirectly by way of the Duty Cycle adjusted by current control.

21. Method as claimed in claim 19, wherein the sum of the coil resistor and the resistor of the connected semiconductor component for driving the load is taken into consideration for the determination of temperature.

22. Method as claimed in claim 19, wherein the Duty Cycles of several PWM periods are averaged for temperature measurement or the determination of the indirect temperature value.

23. Method as claimed in claim 19, wherein the nominal resistance value of the coil is used at the presently measured or estimated temperature of the control unit for the average value of the indirectly determined temperature quantity directly after the switching on of the ignition, in particular after the ignition's re-start.

24. Circuit arrangement for driving several inductive loads comprising a circuit for the PWM control of the load current, wherein the method as claimed in claim 13 is implemented as a program in a microcomputer or microcomputer system which is electrically connected to the PWM circuit.

25. Circuit arrangement for driving several inductive loads comprising a circuit for the PWM control of the load current, in particular according to claim 24, wherein the method as claimed in claim 13 is realized at least in part by digital logic.