The invention relates to a method for processing data for an electronically commutated motor, and it relates to an electronically commutated motor for carrying out such a method.
An electronically commutated motor usually has an output stage which is controlled by a driver IC (Integrated Circuit) or a computer and must be switched on and off again as exactly as possible by that driver IC or computer so that a constant rotation speed and quiet motor operation are obtained.
This is difficult to achieve in practice, since a computer such as a microprocessor or microcontroller that controls the output stage must also perform other time-critical tasks, e.g. processing a frequency signal or a PWM (Pulse Width Modulation) signal and/or controlling the motor rotation speed. These signals must also be processed very accurately in order for the motor to run quietly.
There are a number of possibilities for this. For example, the output stages can be controlled very accurately using interrupt operations; the sensing of other signals becomes more inaccurate as a result, however, because accurate sensing of other signals is blocked during an interrupt for controlling the output stage. On the other hand, those other signals could be sensed via interrupt, and the output stages could instead be controlled using a method referred to as “polling”. In such a situation, if the program is currently sensing a signal, simultaneous monitoring of the output stages is not possible. The result of this is that the current in the relevant output stage is switched on or off too late, thereby causing the motor to run unevenly.
Both of the aforesaid possible solutions are therefore unsatisfactory.
A more powerful computer, capable of handling multiple time-critical functions via corresponding interrupts, could also be used. A computer of this kind would then, however, need to have a high clock frequency in order to execute the interrupt routines as quickly as possible, since even with this kind of computer these routines cannot be executed in parallel fashion. This approach would moreover be too expensive for most applications.
It is therefore an object of the invention to make available a method for processing data for an electronically commutated motor, and an electronically commutated motor for carrying out such a method.
According to the invention, this object is achieved by monitoring a ratio between a rotation-speed-dependent time period (TPP) and a current pulse duration (Tcurr) and, as a function of that ratio, selecting either a time interval during a current pulse or a time span outside a current pulse as the time to perform certain calculation operations. A better distribution of the available system time is thereby obtained with simple and inexpensive means, so that time-critical functions can be performed without disruption. Since the time-critical rotational position regions of the rotor in which interrupt operations occur, as well as the placement of the energization blocks, are known in advance, with the method according to the present invention other calculation operations can be shifted into those rotor rotation regions in which no other time-critical signals need to be processed, so that even long calculation operations can be performed with no negative influence on how the motor runs. It is thereby possible, using even a simple microcontroller, to operate an electronically commutated motor reliably and to ensure that the motor runs quietly.
Another way of achieving the stated object is to control operations of the motor with a microcontroller whose program performs the steps discussed above. Using a single simple microcontroller, such a motor can implement numerous functions, e.g. calculating a target value from a delivered signal; measuring a true value for rotation speed; controlling rotation speed; generating an alarm signal in the event of extreme rotation speed deviations; and exact commutation, which results in quiet motor operation.
Further details and advantageous embodiments of the invention are evident from the exemplary embodiment described below and depicted in the drawings, which is in no way to be understood as a limitation of the invention.
In the description hereinafter, identical or identically functioning parts or functions are referred to using the same reference characters, and usually described only once, e.g. current pulses 132, 132a, 132b, 132c, and 132d.
With a HALL signal of this kind it is easy, as depicted in
Time TPP is assumed to be 1 ms=0.001 s. Rotor 50 then requires 4×0.001=0.004 second for one complete revolution, and its rotation speed is
1/0.004=250 revolutions per second.
Since there are 60 seconds in a minute, rotor 50 is rotating at a speed of
(1/0.004)×60=15,000 rpm (1)
Since the time for one complete revolution (or indeed for part of a revolution) for an electric motor 49 having a Hall IC 60 can be measured easily and with very good accuracy, it is preferable, especially in the context of rotation speed controllers for electric motors, to work with time TPP or with a multiple N thereof (N=1, 2, 3, . . .), since this variable can be used directly after it is measured and is also required for controlling commutation of the motor. This time therefore represents, in the context of an electric motor, a more convenient indicator of rotation speed than any of the other variables such as rpm or revolutions per second; and if necessary, TPP can easily be converted into rpm by taking the reciprocal of the time T360° mech required for one revolution through 360° mech. and multiplying by 60, thus:
n(rpm)=60/T360° mech (2).
The time T used here must be in seconds.
As
Output stages 37, 39 are controlled by a computer 43, usually a microcontroller (μC), to which HALL signals from Hall IC 60 are conveyed. μC 43 contains, in the form of program modules that are indicated only schematically, a commutation control system 47 “COMM,” a rotation speed controller 48 “n_CTL,” a calculation member 51 “SW_CALCII” for calculating a rotation speed target value TSoll for controller 48, an alarm control system 54 for generating an ALARM signal for situations in which the rotation speed of motor 49 becomes too high or too low, a ROM 55 for storing a program, and an alarm delay counter 56 “AVZ” that coacts with alarm control system 54 which has an output 57 for the ALARM signal. The effect of AVZ 56 is that an alarm is triggered not directly, but only after an alarm condition has continuously existed, for example, for one minute.
Module 51 for target value calculation has conveyed to it from outside, e.g. from an external generator or sensor 58, a corresponding signal that is converted in SW_CALC 51 into a rotation speed target value nsoll or TSoll. This is done preferably by means of a table that can be stored in ROM 55.
This calculation of a target value requires many calculation steps and consequently a great deal of time, and is therefore preferably divided into several shorter parts. What is important is that these calculations must not interfere with the commutation of motor 49, so that it runs quietly. Even the shorter parts of the target value calculation, however, can last so long that they impair exact commutation of motor 49. The same applies to the calculation routines of rotation speed controller 48 and alarm module 54.
Motor 49 that is depicted is, of course, only one very simple example of an arbitrary electronically commutated motor; it serves merely to facilitate understanding of the invention, and in no way limits it.
Flag_FctsEnable
and
Flag_Do_Fcts.
The overall program Main PRG of
The next step S90 contains a routine CALC_Within, which is depicted in FIG. 8 and makes certain settings after the current in one of phases 33, 35 has been switched on.
The program then goes to S92, where it determines the value of flags Flag_FctsEnable and Flag_Do_Fcts. If that value is “1,” the program goes to S94, where these two flags are set to “0” so that at the next pass in step S92, the response is “0” and the program enters a short loop S93, which checks in recurrent steps, e.g. every 100 μs, whether one of output stages 37, 39 needs to be switched on or off.
S94 is followed by a step S98 in which the counter status of a Hall counter Hall_CNT is checked. If that status is even, the program goes into a left branch S99; if it is odd, it goes into a right branch S126.
In left branch S99 the program goes to S100, in which the target value determination SW_CALC is performed.
If the response in S98 is NO, the program goes via right branch S126 to S108 Do_Actual_Speed where the actual value determination is performed, i.e. a value characterizing the instantaneous rotation speed of rotor 50 is measured or calculated. Following S108 in S116 is a controller, e.g. rotation speed controller n_CTL depicted at 48 in
As rotor 50 rotates through 360° mech., the program thus runs through step S98 four times, Hall_CNT successively assuming e.g. the values 1, 2, 3, 4, as depicted in
As
To ensure that the electronic system of motor 49 always “knows” the rotational position of rotor 50, edges 142 of signal HALL must be sensed very accurately, i.e. by way of interrupt operations that are labeled “a” in FIG. 3C. This is the purpose of the Hall interrupt routines of
Another critical aspect in
Time spans b and c should therefore, to the greatest extent possible, be kept unencumbered by other calculation operations, in order to allow clean and exact commutation so that motor 49 runs quietly.
The invention therefore proceeds from the concept of performing necessary calculation procedures within the energization blocks when the blocks are long, and before (or after) the beginning of the energization blocks when the blocks are short, in order to improve the smoothness of motor 49.
This means that the situation
TCurr=TPP/3 (3)
is the point at which the calculation of certain operations should be relocated from one rotor rotation region to another rotor rotation region. This relocation can be accomplished, if applicable, using a switching. hysteresis, and is described in detail below with reference to flow charts.
In
Controller routine S116 in
S150 of
If, on the other hand, the situation as shown in
The value of Flag_Fct_within thus defines where and when certain calculation operations are performed.
Once this matter has been clarified, it is necessary to watch for the arrival of the moment at which those calculation operations can begin at the point defined in FIG. 6. The following conditions are used for this purpose:
If the calculation is to be accomplished outside an energization block 132, 134, etc., it can be started directly after execution of the Hall interrupt. These are points 133, 133′, 133″, 133′″ in FIG. 5.
If the calculation is to be accomplished within an energization block 132, 134, etc., it cannot begin until
a) the Hall interrupt (routine “all” in
AND
b) the energization start operation (routine “b”, in FIG. 3), are complete. These are points 131, 131′, 131″, 131′″ in FIG. 4.
These two conditions are defined by the flags
Flag_FctsEnable
and
Flag_Do_Fcts.
Every time an edge 142 of the HALL signal occurs—which is also referred to as a “Hall change” because the Hall signal then changes either from 0 to 1 or from 1 to 0—this causes a Hall interrupt S160 that is depicted in FIG. 7.
In S162 a variety of steps are performed, e.g. steps necessary for commutation; once they are complete,
Flag—FctsEnable=1
is set in S164 because Condition 1 (as explained above) has been met.
If the calculations can now be started, Flag_Fct_Within has a value 0 (cf. S154 in FIG. 6), and the response in S166 is therefore “0” and Flag_Do_Fcts is set in S168 to “1.” The routine then goes to S170 Return. The calculation operations can thus begin at points 133, 133′, etc. of FIG. 5.
Both flags are thus set, and in the main program (
If, however, Flag_Fct_Within has a value of “1” in S166 of
Flag—Do—Fcts=0
is set in S172; i.e. the response in S92 of
Step S178 inquires whether the current in the relevant phase is presently switched on. If NO, the routine goes directly to S180 Return, and monitoring to determine whether the current should be switched on is continued.
If the response in S178 is YES, S182 then checks whether both flags
Flag_FctsEnable (S164 in
Flag_Fct_Within (S152 in
If NO, the program goes to S180 Return. If YES, it goes to S184, where
Flag—Do Fcts=1
is set, i.e. both conditions are now met in S92, and the calculation steps that are to be performed at that time in
In this case, therefore, the calculations below S92 (
Flag—FctsEnable=1
AND
Flag—Do—Fcts=1
are met. In
In
Assuming that controller n_CTL defines a control output TCurr of 1.35 ms (since little energy is required here), there remains before each current pulse 132c, 134c a period of approximately 4 ms in which calculations can be performed, for example the calculations in S100 of
Assuming a control output TCurr (from controller n_CTL) of 3 ms=3000 μs, what remains available for calculation operations is, for example, 2900 μs. The calculation operations in S100 of
The operations in the lower part of
A preferred type of commutation by means of polling is described in detail in DE 200 22 114.0 U1=PCT/EP01/15184 =WO 02-054567-A2 published 11 Jul. 2002=U.S. Ser. No. 10/433,139 filed May 29, 2003, which is therefore incorporated by reference in order to avoid excessive length. Commutation can be accomplished in a variety of ways known to those skilled in the art, commutation in accordance with DE 200 22 114.0 U1 and U.S. Ser. No. 10/433,139 being preferred.
Many variants and modifications are of course possible in the context of the present invention. A number of possibilities for further embodiments and refinements of the inventive concept can result from consideration of additional variables, for example the nature, duration, and priority of the calculations that need to be performed at a particular moment.
Number | Date | Country | Kind |
---|---|---|---|
101 61 688 | Dec 2001 | DE | national |
This application is a section 371 of PCT/EP02/13772, filed 5 Dec. 2002 and published 26 Jun. 2003 as WO 03-052920-A1, claiming priority from German application DE 101 61 688.0, filed 15 Dec. 2001. This application incorporates by reference commonly assigned U.S. Ser. No. 10/433,139, BERROTH et al., filed 29 May 2003 as the U.S. phase of PCT/EP01/15184.
Filing Document | Filing Date | Country | Kind | 371c Date |
---|---|---|---|---|
PCT/EP02/13772 | 12/5/2002 | WO | 00 | 8/25/2003 |
Publishing Document | Publishing Date | Country | Kind |
---|---|---|---|
WO03/05292 | 6/26/2003 | WO | A |
Number | Name | Date | Kind |
---|---|---|---|
4901366 | Rottger | Feb 1990 | A |
5552685 | Young et al. | Sep 1996 | A |
5590235 | Rappenecker et al. | Dec 1996 | A |
5675231 | Becerra et al. | Oct 1997 | A |
5712550 | Boll et al. | Jan 1998 | A |
5847523 | Rappenecker et al. | Dec 1998 | A |
6078152 | Dieterle et al. | Jun 2000 | A |
6167338 | De Wille et al. | Dec 2000 | A |
6249096 | Shin | Jun 2001 | B1 |
6307338 | Kuner et al. | Oct 2001 | B1 |
Number | Date | Country |
---|---|---|
2346118 | Apr 2002 | CA |
198 45 626 | Apr 2000 | DE |
100 24 636 | Nov 2000 | DE |
0 986 855 | Mar 2000 | EP |
WO 97-25768 | Jul 1997 | WO |
Number | Date | Country | |
---|---|---|---|
20040160204 A1 | Aug 2004 | US |