EP1304908A1 - Error correction of reference values in electronic circuits - Google Patents

Abstract

A process to correct signal parameter errors in an operating circuit for gas discharge and halogen lamps or LEDs derives two base values from an error-containing reference value, one of which is compared with a nominal value to find the error. The other is then altered so that the combination of the two is compensated. <??>One base value is derived from reference value without reference to its error. direction and size of base value error is determined by comparison with nominal value. The other base value derived from the reference value, is altered referring to the error of the first base value, so that the influence of the reference value error is compensated for by creation of combined value. Combined value is the product of both base values or proportional to product of them.

Description

The present invention relates to a method for correcting the error of a two basic values for an electronic circuit, the two underlyings from the same one, with an error Reference value can be derived. The electronic circuit is for example, an electronic ballast for operating at least one Gas discharge lamp or fluorescent lamp or around an operating device for LEDs or Halogen lamps.

Electronic circuits often have several separate control loops on, with the help of certain operating parameters of the circuit to a desired Value to be regulated. For electronic ballasts for fluorescent lamps for example, with the help of such control loops DC link voltage and the one flowing over the half bridge of the inverter Current can be set to the power with which the lamp will ultimately operate will keep constant or at a certain value. To regulate the Various operating parameters, a setpoint is generated and with the corresponding current actual value of the circuit compared, the activation of the Switching controller for the DC link voltage and the inverter depending of the deviations of the actual values from the target values.

The setpoint of a control loop is usually derived from a reference value, at a certain point in the electronic circuit - for example with the help a Zener diode - is generated. The reference value is generated with a predetermined accuracy, the actually generated reference value compared to the ideal reference value can deviate upwards or downwards. This unavoidable errors in the reference value also affect e.g. about one Setpoint derived from the voltage divider, the value set using the control loop Operating parameters and ultimately also the lamp power. Will the Setpoints for the control loops from different or independent of each other generated reference values, it can be assumed that for the lamp power the statistically distributed errors of the reference values to one compensate for a certain degree, which is the case, for example, if a Reference value is too high and another is too low for it.

Difficulties arise, however, when the target values of the different Control loops are all derived from a common reference value because the Setpoints then all have the error of the reference value in the same direction are. For example, if the reference value is a little too high, this applies in the above mentioned example of an electronic ballast for the setpoints of DC link voltage and half-bridge current, which - because the lamp power in one certain approximation is proportional to these two operating parameters has that both errors multiply and the lamp power is relatively strong deviates from the desired performance. This problem of error accumulation could can be avoided by having each setpoint from its own reference value is derived, since then - as explained above - the deviations at least partially compensate again, however, this procedure involves an increased technical effort, since several reference values are independent of each other must be generated.

The problem described above applies generally to electronic circuits when two controlled parameters are combined with each other, the setpoints for the relevant control loops can be derived from the same reference.

The present invention is therefore based on the object in an electronic Circuit for operating lamps the increasing error of a Avoid reference deviation or at least significantly reduce it.

This object is achieved by a method which has the features of claim 1 has, solved. According to the invention, one of the two basic values is initially selected derived from the reference value regardless of its error. Then be Direction and magnitude of the error of this underlying by comparison with a Nominal value determined and the other base value derived from the reference value below Taking into account the error of the first-mentioned base value changed in such a way that the Influence of the reference value error in the formation of the combination value is compensated becomes.

According to the invention, when deriving the two base values from the Reference value whose error is taken into account, so that the error of the end result, ie of the combination value is reduced. For example, the combination value the product of the two underlyings, the change of the or one of the Basic values in the inverse direction to the deviation of the reference value. The whole Circuit can be kept very simple with this method, since only one only reference value must be generated. The change in the second underlying can, for example, during a comparison before commissioning the electronic Circuit can be performed.

The method according to the invention can be used, in particular, with an electronic one Ballast for operating at least one gas discharge lamp - in particular a fluorescent lamp - in which the two basic values are used a setpoint for a control circuit for regulating the intermediate circuit voltage and the Form half-bridge current. The combination value corresponds to the lamp power here, essentially through the product of half-bridge current and lamp voltage is formed. The second base value can be modified, for example, with the help a multiple voltage divider, the base value depending on the size and Direction of the previously determined error of the first base value from a specific one Connection point of the multiple voltage divider is derived. Also at electronic circuits for operating other lamps, e.g. LEDs or The inventive method can be used for halogen lamps.

Fig. 1

das Schaltbild eines elektronischen Vorschaltgeräts, bei dem das erfindungsgemäße Verfahren zum Einsatz kommt;

Fig. 2a-c

schematische Darstellungen zur Bildung des Kombinationswertes aus zwei Basiswerten;

Fig. 3

eine Schaltung, mit deren Hilfe ein Basiswert in Abhängigkeit von dem Referenzwert-Fehler beeinflußt werden kann;

Fig. 4a

ein Histogramm der Fehlerverteilung für die bei einem elektronischen Vorschaltgerät ohne Fehlerkorrektur eingestellte Lampenleistung; und

Fig. 4b

ein Histogramm der Fehlerverteilung für die eingestellte Lampenleistung bei zuvor durchgeführter Fehlerkorrektur.

The invention will be explained in more detail below with reference to the accompanying drawings. Show it:

Fig. 1

the circuit diagram of an electronic ballast in which the inventive method is used;

2a-c

schematic representations for forming the combination value from two base values;

Fig. 3

a circuit by means of which a base value can be influenced as a function of the reference value error;

Fig. 4a

a histogram of the error distribution for the lamp power set in an electronic ballast without error correction; and

Fig. 4b

a histogram of the error distribution for the lamp power set when the error has been corrected beforehand.

Before the method according to the invention is explained in detail, the task will first be explained using an example in the form of an electronic ballast for operating a fluorescent lamp. FIG. 1 shows the circuit diagram of an electronic ballast which is connected on the input side to the mains supply voltage U ₀ via a high-frequency filter 1. At the output of the high-frequency filter 1 there is a rectifier circuit 2 in the form of a full-bridge rectifier, which converts the mains supply voltage U ₀ into a rectified input voltage for a smoothing circuit 3. This smoothing circuit 3 is used for harmonic filtering and smoothing the rectified circuit 2 rectified mains supply voltage U ₀ and for this purpose comprises a smoothing capacitor C1 and a step-up converter having an inductor L1, a controllable switch in the form of a MOS field-effect transistor S1 and a diode D1.

A corresponding switching of the MOS field-effect transistor S1 generates an intermediate circuit voltage U _z which is present across the storage capacitor C2 connected to the smoothing circuit 3 and which is fed to an inverter 4. This inverter 4 is formed by two further MOS field effect transistors S2 and S3 arranged in a half-bridge arrangement. A high-frequency activation of the two field effect transistors S2, S3 generates a high-frequency AC voltage at their center tap, which is fed to the load circuit 5 with the gas discharge lamp LA connected to it. The mode of operation of such a ballast is already well known and will therefore not be explained further below.

The control of the three MOS field-effect transistors S1-S3 of the smoothing circuit 3 and of the inverter 4 is carried out by a control circuit 6, which generates corresponding switching information and transmits it to a driver circuit 7 connected to the control circuit 6. The driver circuit 7 converts the switching information into corresponding control signals and controls the gates of the three MOS field-effect transistors S1-S3 via the line 12-14. This takes into account the actual value of the intermediate circuit voltage U _z and the actual value of the current flowing across the half bridge. For this purpose, the intermediate circuit voltage U _z is tapped at the intermediate circuit capacitor C2 and fed to the control circuit 6 via an input line 15. The current half-bridge current is measured with the aid of the voltage dropping via a shunt resistor R, which is located at the base of the half-bridge inverter 4, which voltage is likewise fed to the control circuit 6 via a further input line 16. As a result, two control loops are formed, which will be explained in more detail below.

The first control circuit is used to regulate the intermediate circuit voltage U _z , which is supplied to the control circuit 6 via the input line 15. Within the control circuit 6, the intermediate circuit voltage U _z is as an actual value to a first input value for a comparator 8, the second input value is generated by an internal control block 10 and forms the setpoint value for the intermediate circuit voltage U _z. The comparison result between the target value and the actual value for the intermediate circuit voltage U _Z is fed to a control circuit 11, which uses this comparison result to calculate control signals and transmits them to the driver circuit 7.

Part of the second control loop is another one in the control circuit 6 arranged comparator 9, the first input value of which via the shunt resistor R falling and supplied with the help of the input line 16 is voltage Provides information about the current actual value of the half-bridge current. The setpoint for this current is also generated by the control block 10 and second Input signal supplied to the comparator 9, which in turn the comparison result passes the control circuit 11. The control circuit 11 then generates corresponding ones Switching information that is sent to the two MOS field-effect transistors via the driver circuit 7 S2 and S3 are forwarded.

By means of these two control loops, the half-bridge current and the intermediate circuit voltage are separately kept constant or regulated to a certain value in order to keep the power at which the lamp LA is ultimately operated at a predetermined value. An error in the two setpoints for the control loops formed by the control block 10 has a direct effect on the intermediate circuit voltage U _z and the half-bridge current and thus also on the lamp power. If both setpoints are derived from a common reference, there is a risk that the errors in the intermediate circuit voltage U _z and the half-bridge current will multiply, so that the lamp power ultimately set is associated with a very high error. In order to avoid this, the two setpoints are derived from the common reference according to the method according to the invention, which will be explained below with reference to FIGS. 2a-c.

FIG. 2a first shows the ideal case in which the reference value has no errors. In this case, the reference value R which corresponds exactly to its setpoint is first generated within the control block 10. This reference value R can be generated, for example, with the aid of a tens diode or the like. From this reference value R, two basic values a and b are then derived, which form the target values for the two control loops. For this purpose, the reference value R is multiplied by a suitable factor α or β, which is done, for example, using a voltage divider. Accordingly, the following applies to the two basic values: a = α R b = β R

The combination value c is finally formed from the two basic values a and b; In the example of the electronic ballast, the basic values a and b correspond to the actual values for the intermediate circuit voltage U _z or the half-bridge current, the combination value c corresponds to the lamp power. Since the lamp power is essentially proportional to the product of half-bridge current and DC link voltage U _z , the two basic values are thus multiplied to the combination value in the present example and the following applies: c = K · a · b or c = K · α · β · R 2 where K is a proportionality factor corresponding to the circuit.

If the reference value R corresponds exactly to its target value, the lamp power set later corresponds exactly to the desired value c. However, it must be assumed that in reality the reference value R has a certain error ΔR and therefore the following applies to the actually generated reference value R *: R * = R + Δ R

The base values a ₁ and b ₁ derived from this error-prone reference value are then also subject to errors Δa and Δb: a 1 = a + Δ a 1 = α · ( R + Δ R ) b 1 = b + Δb 1 = β · ( R + ΔR )

Since the combination value c ₁ formed from the base values represents the product of the two defective base values a ₁ , b ₁ , this combination value c _{1 also} has an error Δc ₁ , ie c 1 = c + Δ c 1 where: Δ c 1 = α · β · Δ R · (2 R + Δ R ) · K

In the just-explained example in which the combination value substantially representing the product of the two base values, this has the result that the error of the underlying .DELTA.a ₁ and .DELTA.b ₁ increase in the same way so that the combined value actually produced c ₁ a relatively has high deviation Δc ₁ from the desired ideal value c.

In order to avoid the accumulation of the error in the formation of the reference value just described, the derivation of the second base value b ₂ is therefore influenced according to the invention, specifically as a function of the direction and size of the error Δa _{2 of} the first base value. This takes place, for example, in that the reference value R + ΔR is now not multiplied by the factor β as in the ideal case to form the second base value, but instead is instead multiplied by a multiplication factor shifted by a correction value Δβ. So now applies. b 2 = (β - Δβ) · ( R + Δ R )

The other base value a ₂ corresponds to the base value a _{1 of} the example in FIG. 2b. The correction value Δβ is shifted in particular in the inverse direction to the deviation Δa _{2 of} the first basic value a. Accordingly, the derived second base value b ₂ again has a shift .DELTA.b ₂ compared to the ideal value b, but now in the opposite direction to the shift of the reference value or the first base value. The result of this is that both base values a ₂ and b ₂ derived from the incorrect reference value R + ΔR are shifted relative to their ideal or nominal values a and b, the errors, however, being so opposed to one another that the error of the combination value increases a large part or almost completely compensated.

For example, applies to the displacement factor Δβ = 2 · β · Δ a 2 α · R . the error Δc _{2 of} the combination value c ₂ Δ c 2 ≈ α · β · Δ R 2

The combination value formed at the end thus has a significantly lower error, although it was formed from two base values, both of which were derived from the same error-prone reference. If the error Δa _{2 of} the first base value is thus known, then a suitable shift factor for the multiplication factor for deriving the second base value can be selected, with which the error of the combination value is completely or at least substantially reduced.

Another possibility is to also derive the second base value from the reference value R + ΔR without taking ΔR into account, but the second base value obtained in this way, taking into account the error Δa _{2 of} the first base value a ₂ later with a correction factor B _c to be modified so that: b 2 = β · ( R + ΔR ) - B c

Is selected as correction factor B _c : B c = 2β · Δ a 2 · R a + Δ a 2 the same applies to the error Δc _{2 of} the combination value c ₂ Δ c 2 ≈ α · β · ΔR 2

The change in the second base value is preferably carried out as part of a comparison before the ballast is started up. For this purpose, FIG. 3 shows a circuit which can be used to influence the formation of a setpoint for one of the two control loops in the manner just described. This circuit essentially consists of a multiple voltage divider 17, the output signal of which forms the setpoint for the comparator 8 of the control loop. The input signal for the multiple voltage divider 17 forms the reference voltage generated by a DC voltage source 18, which, depending on the state of the various switches S ₁₁ to S _{16, is converted} into a specific input voltage for the comparator.

In the course of the adjustment, for example, the first base value is first derived from the reference value in a normal manner, that is to say changed until it is within a specified tolerance window. This comparison, especially if it is carried out digitally with a limited resolution, never leads to a perfect result, but merely to the fact that the first base value comes within the permissible tolerance window. In a subsequent step, the error or the deviation of the derived first base value from its nominal value is then determined. Because of this deviation and in particular taking into account the sign of this deviation, one of the switches S ₁₁ to S _{16 of} the multiple voltage divider 17 is then set in accordance with a defined assignment.

The success achieved with the method according to the invention can be seen in FIGS. 4a and 4b be removed. Both figures show a histogram of the error distribution for the over the two control loops for the DC link voltage and the Half-bridge current set lamp power P, Fig. 4a for the case without Error correction and Fig. 4b with previously performed error correction. Here results a standard deviation in the conventional method without error correction of about 4.7 W compared to the ideal power of 100 W, while with the The method according to the invention achieves a standard deviation of only 2.4 W. becomes. The accuracy of the regulation of the lamp power can thus by The inventive method can be increased significantly.

The present invention thus offers the possibility of correcting the error caused by the Regulation of the half-bridge current and the DC link voltage To keep lamp power as low as possible, although the necessary for the control Setpoints are derived from a common reference value. The one for this necessary circuitry is extremely low. However, the present invention is not limited to the example of an electronic ballast shown, but can generally be used in electronic circuits for operating Illuminants - e.g. LEDs or halogen lamps - are used where two controlled parameters are combined with each other, the setpoints for the Control loops can be derived from a common reference.

Claims (6)

Method for correcting the error (Δc) of a combination value (c) formed from at least two basic values (a, b) for an electronic circuit for operating gas discharge lamps, LEDs or halogen lamps, in which the values are signal parameters, such as amplitude values, the two Base values (a, b) are derived from the same faulty reference value (R + ΔR),
characterized in that one of the two base values (a) is derived from the reference value (R + ΔR) regardless of its error (ΔR),
that the direction and magnitude of the error (Δa) of this base value (a) are determined by comparison with a nominal value,
and that the other base value (b) derived from the reference value is changed taking into account the error of the first-mentioned base value (a) in such a way that the influence of the reference value error (ΔR) in the formation of the combination value (c) is compensated for. Method according to claim 1,
characterized in that the combination value (c) corresponds to the product of the two basic values (a, b) or is proportional to the product of the two basic values (a, b), the change in the basic value (s) in inverse direction to the deviation of the reference value (ΔR). Method according to claim 1 or 2,
characterized in that the change of the base value or values (a, b) is carried out during a comparison before the electronic circuit is started up. Method according to claim 3,
characterized in that the base value (s) (a, b) are derived from the reference value (R) with the aid of multiple voltage dividers (17). Method according to one of the preceding claims,
characterized in that the electronic circuit is an electronic ballast for operating at least one gas discharge lamp (LA). Method according to claim 5,
characterized in that the two basic values (a, b) each form a setpoint for a control circuit for regulating the intermediate circuit voltage (U _z ) and the half-bridge current, the combination value (c) corresponding to the lamp power.

Images