FR2661504A1 - METHOD FOR THE MEASUREMENT, CORRECTED OF DEFECTS, OF AN ELECTRICAL SIZE - Google Patents

Abstract

a) Method for the fault-corrected measurement of an electrical quantity, b) Method characterized in that, immediately one after the other, the electrical quantity and then an auxiliary electrical quantity essentially without drift and whose value is not precisely known, are measured with the same device and are converted into a digital value by an analog / digital converter (6) supplied with an electrical reference quantity, in that to eliminate the reference quantity, the ratio digital values is formed and in that to determine the exact value of the auxiliary quantity (Um), the correction value is determined. c) The invention relates to the method for the measurement, corrected for faults, of an electrical quantity.

Description

"Method for measuring, corrected for defects, of an electrical quantity"

  The invention relates to a method for

  measurement, corrected for defects, of an analogue quantity

  than electrical in the form of numerical values, including a voltage, preferably a voltage

  input of a control unit of a combustion engine

  internal combustion, a process in which a basic electrical quantity, the value of which is known, is used and is compared with its measurement value which

  has been determined, to use as the value of

  rectification of the electrical quantity, deviations

raissants.

  In many fields of electro-

  technical, it is necessary to determine the absolute value of an electrical quantity For this purpose, it is necessary to implement calibrated measuring devices Calibration is necessary because, for

  example because of the tolerances of the component parts

  the devices of a series do not have

constants.

  In the electronics of motor vehicles

  there is also the problem of deter-

  exact quantification of an electrical quantity It is in particular necessary to determine without error, electrical quantities coming from detectors, because

  they are used for the control of

  internal combustion engines for the operation of

  pre-definition of operating parameters

  preferably, control units which control the internal combustion engines of the vehicles are used. At the inputs of these control units, inter alia, electrical quantities dependent on the operating state which are to be apprehended without error of measurement for

  further operation Measurement errors can

  among other things, because of the tolerances of the constituent parts, the drift phenomena (in particular as a function of the temperature) as well as because of additional offset magnitudes.

  to avoid measurement errors, it is carried out

  equation, in a known manner, generally an adjustment of the amplification and shift by interventions

  material inside the control unit.

  This adjustment can, for example, be achieved by the establishment of adjustment resistors. Especially in the mass production, this way of proceeding is however painful and costly and entails

returns high.

  DE-OS 3 844 333 discloses a method for correcting tolerances of component parts during the operation of signals. In order to take into account the influence of these tolerances in the case of an integrator, a microcomputer generates a signal that for error-free processing,

  must result in a known integral value (theoretically

  Because of the tolerances mentioned in the

  these constituents, however, is another integral value that is recorded in the integrator The comparison of the effective integral value with the

  theoretical integral value provides a different value

  which is a measure of the error and which is used as a correction value for the integral value As a result of the integrator, an analog / digital converter is connected, so that the integral value is available under the

numerical size form.

  The process according to the invention is characterized

  in that, immediately one after the other, the large

  electrical converter and then an auxiliary electrical variable essentially without drift and whose

  value is not precisely known, are measured

  with the same device and are converted to

  numerical value by an analog / digital converter

  supplied with an electrical reference quantity

  ence, in that to eliminate the magnitude of reference

  this, the ratio of numerical values is formed and in that to determine the exact value of the magnitude

  auxiliary, the correction value is determined.

  The method according to the invention having the characteristics mentioned above has, compared to the state of the art, the advantage that an error-free measurement of the electrical quantity is

  always possible, even when the parameters required

  for the measurement process do not match

  not exactly to a predefined value, but for example

  ple are not exactly known in absolute value.

  Even when, for example, the reference parameters

  are not stable during a period of

  are subject to these previously unquantifiable drift phenomena, the method according to the invention can be used to

  an exact determination of the electrical quantities.

  According to the invention, the electrical quantity, and then an auxiliary electrical quantity, devoid of drift, but the value of which is not exactly known, are measured immediately one after the other with the same device. In this respect, it should be noted that it is simpler, from the point of view of the branching technique, to dispose of

  of an auxiliary quantity devoid of drift, for example

  with a stabilized voltage, but that, however,

  tolerances of the constituent parts, the difference

  ficulty remains to reproduce this

  in its absolute size To generate in series production apparatus, constantly the same value of the auxiliary quantity considered,

  the disadvantageous adjustment methods mentioned before

  However, in the case of the method according to the invention, this adjustment is superfluous because it is sufficient that the auxiliary quantity is available without drift even if its exact value is not known. Since the measurement of the electrical quantity and the auxiliary electrical quantity are carried out immediately one after the other, it can be based on the fact that a reference quantity necessary for the measurement remains constant inside the time space of these

  two measures and therefore does not drift.

  The drift phenomena mentioned initially

  most often occur indoors

  more important time spaces They are for example

  ple caused by temperature changes The

  mentioned reference parameter (reference quantity

  this) is, in this case, associated with the analog / digital converter which, as usual, operates ratiometrically According to the invention, it is

  provided that this reference quantity is eliminated

  born by forming the ratio of the numerical values of the electrical quantity to be determined and the auxiliary electrical quantity This is possible because, within the short measurement period, it is possible to rely on the constancy of the reference variable, so that its value is eliminated during the formation of the ratio As, however, it must be taken into account that the auxiliary electrical quantity is not the same value for each series production apparatus, but that, as already mentioned , its value is not exactly known, different values may occur due to the amplification dispersions and

  dispersions of copies, as well as the fact of

  the exact value of the

  This is determined by applying a correction value. The determination of this correction value is known per se (DE-OS 38 44 333).

  according to the principle that an electrical quantity

  basis of known value is used and compared with its measured value that has been determined The deviations that occur are determined and processed to obtain a correction value that is applied to the value to be corrected (in this case, the measurement value

auxiliary magnitude).

  In total, the method according to the invention thus allows the exact determination of an electrical quantity in the form of a numerical value without the need to implement an expensive hardware calibration process If in this process , a necessary reference quantity undergoes a drift of long duration, this does not result in measurement errors For this purpose, an additional quantity (auxiliary value) is certainly necessary, but its value does not need to be known exactly It must be free of drift, which may however

  be realized from the point of view of

  with a good approximation by simple means The calibration software process described here

  can be repeated at desired intervals (particularly

automatically).

  According to a complement of the invention, it is provided that the exact numerical value N of the electrical quantity is determined according to the relationship Ni = v i Nm / NH 2 n v i being an exact (ideal) proportionality factor,

  Nm being the numerical value of the electrical quantity

  that, NH the numerical value of the auxiliary electrical quantity, and

  2 N the resolution of the analog / digital converter

than.

  The numerical value N-i can only take

  dre values 0, 1, 2 ,, 2 n-1 The number of

  ranges of possible values, that is, the resolu-

  the analog / digital converter depends on the number of binary digits. For example,

  eight-digit binary system, we have 256 possi-

  For the different numerical values, there is no

  no intermediate value The above developments

  teeth are also valid for all formulas

  following with a similar structure.

  Preferably, the numerical value of the electrical quantity is defined by Nm = vi Um / Urefr 2 n and the numerical value of the auxiliary quantity by NH = vi UH-i / Uref-r 2 n U Hi being the exact value of the auxiliary size

  Uref_r the actual reference quantity.

  In the foregoing, ideal conditions have first been assumed, with regard to the amplification factor, the offset, as well as the auxiliary voltage, only the reference variable is

  supposedly tainted with errors During this request,

  of, we mean by an exact value (or ideal) a value free from error and by a real value,

  a value possibly tainted by an error.

  For the determination of the exact value of the auxiliary quantity, one starts from the relation: NH_r = (vr UH + U of offset r) / Urefr 2 n and one applies the correction factors kl and k 2 of the correction value according to the relationship: N Hr korr = N Hi = kl (N Hr k 2) where v_r is a real proportionality factor, UH the auxiliary quantity (auxiliary voltage) and U_shifts a real offset quantity. The ideal offset amount has the value "zero" By the proportionality factor and the offset amount, the actual values can be corrected both multiplicatively and equally

additively.

  In particular, it is expected that the value of

  rection is determined by measuring two voltages

  Uml and Um 2, the numerical values of these

  Known: Nrl = (vr Uml + Ude shift_r) / Uref_r 2 n N_r 2 = (v_r Um 2 + u_de shift_r) / Uref-r 2 n being measured and processed in a complementary way with the known exact values: N il = vi Uml / Uref_i 2 n Ni 2 = vi Um 2 / uref_i 2 n according to the relation: kl = N_il / (N_rl N_r 2) k 2 = (N il Nr 2 Ni 2 N_rl) / (Nil N_i 2)

  Urefi being the reference quantity.

  Preferably, the electrical quantity or the auxiliary quantity is applied to the input of the analog / digital converter by means of a switch. The electrical quantity and / or the auxiliary electrical quantity and / or the electrical

  reference are, preferably, electrical

  c. It is clear from all this that thanks to the ancillary size devoid of drift but whose absolute value is not necessarily known

  exactly, an exact measure of the electrical

  that is possible, even when the reference

  this, due to drift phenomena that can not be quantified beforehand, is unstable to eliminate the errors of amplification and carryover constantly intervening in the case of a series production and because of the dispersion of the copies, and for to determine the exact value of the auxiliary quantity which is certainly stable, but the absolute value of which is not exactly known, it is determined by comparative measurements with two known voltages, correction factors of a correction value, thanks to which without process calibration equipment, a

  exact determination of the similar electrical

  in the form of a numerical value, is possible.

  The invention will be explained in more detail

  hereinafter with the aid of the attached figures.

  FIG. 1 is a block diagram,

  FIG. 2 is a block diagram

pleted,

  Figure 3 is a block diagram.

  In electronic control devices

  internal combustion engines, for example to

  In the case of diesel injection systems, it is necessary to measure an electrical quantity 3 at an input 1 of a control device 2, since the absolute value of this quantity is used as a control variable for the determination of the operating parameters. of the internal combustion engine (for example the injection time) delivered by the control device 2 The electrical quantity 3 is in the form of an analog voltage Um It is applied to an adaptation circuit 4 of the apparatus of

  command 2 which also assumes a function of pro-

  The output 5 of the matching circuit 4 is connected to an analog-to-digital converter 6 to this analog / digital converter 6 operating

  ratiometrically, a reference quantity is applied

  This is, in this case, a voltage of

  Uref reference Output 8 of the analog converter

  that / digital 6 is connected to a microcomputer 9

  which is connected with a programmable memory 10.

  The memory 10 is preferably made in the form of

  A data link 11 goes from the microcomputer 9 to a serial interface 11 'to which a calculator is connected.

  12 According to another example of

  alisation, not shown, the function of the calculator

  12 is assumed at the same time by the micro-

  calculator 9, so the control calculator

12 can be deleted.

  The circuit arrangement of FIG. 1 has the following function:

  To be able to understand the analogical tension

  that Um exactly in the form of a numerical value, therefore free of error, the measured value is assigned a correction value to eliminate the errors that occur for example due to dispersions

  between copies during mass production.

  In the ideal case, that is to say in the case where there is no

  no mistake, the voltage Um would be, through the

  of the adaptation circuit 4, applied to the conver-

  analog / digital weaver 6 and converted by

  ci in the numerical value: Ni = vi Um / Urefi 2 n with Ni = 0, 1, 2 ,, 2 n-1

  According to the resolution of the analog converter

  The number of possible numerical values is determined by the number of its binary digits. For example, for an eight-digit binary system, 256 numerical values are possible. In the preceding formula, v_i represents an ideal proportionality factor and Uref_i, the

  ideal absolute value of reference quantity 7.

  Since, however, the aforementioned dispersions between the copies occur and result in measurement errors, it is necessary to start, not from ideal values but from real values.

  most often multiplicative and additive errors

  Therefore, instead of starting from the exact numerical value N i, in practice the actual numerical value which results from N_r = (vr Um + U_of offset r) / Uref_r 2 n Ude of offset_i = O

  Because of the multiplicative and

  However, the accuracy of the measurement is unacceptably altered. Up to that point, according to the state of the art, a calibration of the amplification was performed.

  postponement by material interventions to the

  of the control device 2, for example by setting up calibration resistors This

  material intervention can be avoided thanks to the

  For this purpose, according to what is called a calibration mode, two known voltages il Uml and Um 2 are applied to the input 1 of the control device 2, these voltages resulting in the values following numerals: N_rl = (v_r Uml + U_de_range_r) / Uref_r 2 n N_r 2 = (v_r Um 2 + U_de_set_r) / Uref_r 2 n By the index "r", it is clearly indicated that it is in this case of real values and not

  ideal exact values that carry the index "i".

  The known ideal values corresponding to the numerical values N_rl and Nr 2 are as follows: Nil = vi Uml / Urefi 2 n Ni 2 = vi Um 2 / Uref_i 2 n To pass from real values to ideal values, it is necessary that the microcomputer 9 or the

  control computer 12 calculates a value of

  correction which consists of correction factors kl and k 2 This states: kl = Nil / (Nrl Nr 2)

  k 2 = (N_il N_r 2 Ni 2 N_rl) / (Nil N_i 2).

  If these correction factors are

  the calibration mode can be abandoned and normal operation can be approached, in which all the measured values N_r are then corrected according to the following relation: N_r korr = kl (N_r k 2) = Ni

  In this way, we can avoid

  these dispersions mentioned between the copies.

  The reference variable 7 necessary for the analog / digital converter 6 and which is formed from the reference voltage Uref, is most of the time not stable in practice, because the drift phenomena are not quantifiable at However, according to FIG. 2, it is possible to indicate a method with which an analog electrical variable to be measured can be apprehended exactly in the form of a digital value. For this purpose, it is provided in FIG. 2, in addition to FIG. arrangement according to FIG. 1, a switch 13 between the matching circuit 4 and the analog / digital converter 6, so that either the voltage

  to measure Um or an auxiliary electrical quantity

  re 14 can be applied to the analog converter

  that / numeric 6 The auxiliary quantity 14 is preferably

  fence, formed from an auxiliary voltage UH.

  While in the process described above, the reference voltage Uref is reached at its exact value, this is not the case for the voltage

  auxiliary voltage UH This auxiliary voltage UH

  be available without drifting but its absolute value

  read does not need to be necessarily exact

  In practice, such a voltage can be

  made in a simple way with a circuit of hand-

  of a constant voltage, which makes available an absolute value in practice free of drift, but which, however, for example because of the dispersion of the constituent parts in mass production, has

  not constantly the same absolute value.

  For the implementation of the process,

  then in two steps First, as described above, the voltage Um is measured and immediately thereafter, the voltage UH It should be based on the fact that the reference voltage Uref does not change between these immediately following two measurements Assuming ideal conditions for the factor

  amplification and lag, as well as for the

  auxiliary value, we have for the absolute value of

  Re: N = vi Um / U Hi 2 n From the measured values (numerical values): NM = vi Um / Uref r 2 n NH = vi U Hi / Urefr 2 n we have: N = vi Nm / NH 2 not

  Since in the case of

  As already mentioned above, it is possible to start from amplification and carryover errors due to the dispersions of the specimens, etc., and that the auxiliary voltage UH is certainly stable but that its absolute value is possibly not exactly known. the measurement value determined in this way must in practice be further corrected by means of the corresponding correction value as already explained for the measurement value Um. This means that, first of all, in the

  calibration method, the correction factors correspond to

  laying kl and k 2 must be determined, and that

  As has already been stated, a correction must be

  be performed to obtain the ideal values.

  The method is again explained in the functional diagram in FIG. 3 Starting from the starting field 15, it is determined in the field 16 whether the operation must be carried out according to the calibration mode A or according to the normal operation No Si calibration mode A is used, then in step 17 the application of the voltage Uml, the value of which is known, is known. In step 18, the corresponding digital value Nrl is measured. Next, in step 19 the second known voltage Um 2 is applied to the input 1 of the control unit 2 and to the step

  , the corresponding numerical value Nr 2 is measured

  In step 21, the correction factors k 1 and k 2 are determined. These correction factors k 1 and k 2 are deposited in step 22 in the memory 10 of the control device 2. They are available for the normal operation No for which in step 23 the real value Nr is measured and using the correction factors kl and

  k 2, the correction is made in the next step 24.

  If, as in the case of Figure 2, one works with an auxiliary variable 14 (voltage

  UH), it also occurs, according to the

  3, starting from the starting field 25, first of all the voltage Um is applied (step 26) and the corresponding digital value Nm is measured in step 27 As soon as the voltage UH is applied to the analog / digital converter 6 (step 28) and at step 29 the corresponding numerical value NH is measured. The correction previously explained in FIG.

  the auxiliary voltage UH is then cor-

  corresponding to steps 17 to 22 already described.

  On the basis of the process just described, a calibration is thus carried out in the form of

  software that does not require any intervention in the in-

  The implementation of this process is not only possible in the finishing of the devices, but also in service, for example for recalibration, or, as far as there is in the field of operation of the device points of

  operation with a measurement voltage known to be exactly

  for example when the measuring voltage

  a path measurement signal for a

  position adjustment, and that operating points

  exactly known can be regulated by

  mechanical stops, this implementation is even possible

  during normal operation.

  On the basis of the invention, it is in

  sequence possible to enter exactly a voltage value to be measured absolutely in a

  control unit with an analogue converter

  6, in which the transfer errors and the amplification errors are eliminated by a calibration implemented in a software way. For this purpose, as already explained, two known voltages Uml and

  Um 2 are applied to the input of the communication device

  2 and from them, measured values are determined. By comparison with the corresponding ideal values, a value can be determined.

  multiplicative and additive correction for trans-

  forming the actual measured values into ideal measurement values The correction factors of the correction value are deposited in a programmable memory 10 of the control device 2 and are then available for the correction of the future measurement values. the reference voltage Uref of the digital analog converter 6

  is drifting, it is used in addition to

  an auxiliary voltage source UH which does not

  although not drifting, but whose absolute value

  This source of voltage is likely to be realized with

  simple means to determine the precise magnitude

  the auxiliary voltage, it is measured, while for the elimination of measurement errors,

  we proceed as previously mentioned.

Claims (7)

  A method for the fault-corrected measurement of an electrical analog quantity in the form of digital values, in particular of a voltage, preferably an input voltage of a control apparatus of an internal combustion engine , a method in which a basic electrical quantity, the value of which is known, is used and compared with   its measured value, which has been determined for use   as the correction value of the electrical quantity   that, appearing differences, process characterized in that, immediately one after the other, the electrical magnitude (3) and then an auxiliary electrical magnitude substantially without drift, and whose   value is not precisely known, are measured   with the same device and are converted into digital values by a powered analog-to-digital converter (6) with an electrical reference variable, in that, to eliminate the reference quantity (Um), the ratio of the digital values is formed and in that for determining the exact value of the auxiliary quantity (UH), the correction value is determined.   2. The process according to claim 1, wherein   in that the exact numerical value (Ni) of   the electrical quantity is determined according to   tion: Ni = vi Nm / NH 2 n where v_i is an exact proportionality factor (ideal) Nm is the numerical value of the electrical quantity NH is the numerical value of the auxiliary electrical variable and   2 N is the resolution of the analog / digital converter merica.   3. Process according to any one of   claims 1 and 2, characterized in that the   numerical value of the electrical quantity is fixed at Nm = vi Um / uref_r 2 n   and the numerical value of the electrical   link: NH = vi U Hi / Urefr 2 n U Hi being the exact value of the auxiliary quantity   Uref_r being the actual reference quantity.   4 Process according to any one of   Claims 1 to 3, characterized in that for   to determine the exact value of the auxiliary quantity, one starts from the relation: NH_r = (vr U Hi + U_de offset_r) / Uref_r 2 n and introduces the correction factors K 1 and k 2 of the correction value, according to the relationship: N Hr korr = NH i = kl (N Hr k 2) where vr is a real proportionality factor, UH is the auxiliary magnitude and U_ is an offset of a magnitude of actual offset.   5. Process according to any one of   Claims 1 to 4, characterized in that the value   correction is determined by the measurement of two known voltages, Uml and Um 2, the numerical values of these known voltages Nrl = (v_r Uml + U_de_set_r) / Uref_r 2 n Nr 2 = (vr Um 2 + Ude offset_r) / Urefr 2 n   being measured and being treated in a complementary way.   with their exact known values Nil = vi Uml / urefi 2 n N i 2 = vi Um 2 / urefi 2 n according to the relation: kl = Nil / (Nrl Nr 2) k 2 = (Nil Nr 2 Ni 2 Nrl) / (Nil Ni 2)   Uref_i being the exact reference quantity.   6. Process according to any one of   Claims 1 to 5, characterized in that the   electric generator or the auxiliary   be applied to the input of the analog converter than / digital by a switch.   7. Process according to any one of   claims 1 to 6, characterized in that the   electric generator and / or the auxiliary electrical   re and / or the reference variable, are electrical voltages.