AU2001263703B2 - Method for negative feedback controlling electrical power and negative feedback controlled power supply - Google Patents

Description

0 Method for negative feedback controlling electrical power and negative feedback controlled power supply

;Z

Cc The present disclosure is directed to a method for negative feedback controlling electrical power delivered to an electrical load, which method comprises generically, and as known to the skilled artisan, monitoring the electrical power delivered to the load, thereby Sgenerating a monitoring signal, forming in dependency of the monitoring signal and of a Cc rated value signal a control deviation signal and adjusting via a controller as e.g. a ,i proportional or proportional-integral controller the electrical power delivered and 1o monitored in function of the control deviation signal.

The present disclosure thereby departs especially from such method and power supply for delivering high power of at least 100 VA to a load.

A need exists to provide such a negative feedback control method and, accordingly, such a negative feedback controlled power supply for superior accuracy of controlled power delivered to the load, with respect to a rated power value to be delivered.

Under a first aspect, the present disclosure departs from a method for negative feedback controlling electrical power delivered to an electrical load as mentioned above, whereat monitoring the electrical power delivered to the load results first in an analog monitoring signal, which is then analog to digital converted so as to result in a digital monitoring signal.

Under a second aspect, the present disclosure departs from a method for negative feedback controlling electrical power delivered to an electrical load as mentioned above, whereat adjusting the electrical power delivered to the load and monitored as performed as a function of the control deviation signal again via a respective controller by means of Pulse-Width Modulation (PWM).

In a method for negative feedback controlling electrical power, whereat the controlled value, namely the power delivered to the load, is digitalized as under the first aspect of the present invention, the overall accuracy of the negative feedback control significantly depends on the accuracy of the analog to digital conversion of the monitored signal. As is well known in the art of analog to digital conversion, noise of the analog input signal [R:\LIBOO]658044.doc:TCW Sleads to the fact that the digital output signal jitters by at least one least significant bit N, (LSB). This problem is customarily resolved by oversampling, i.e. by establishing a ;sampling rate which is considerably higher than necessitated by the spectrum of the Cc analog signal to be converted. In fact, by oversampling a multitude of digital samples are generated in a predetermined time frame, and the respective digitalized output value is formed by averaging the digital samples. Nevertheless, oversampling is always limited Sby the layout of a specifically considered converter not suited to handle sampling rates Cc above a specific maximum rate. For rising the sampling rate over such limit makes it necessary to apply a differently designed A/D-converter which is often much more expensive and more critical to operate.

Therefore, a further need exists to provide for an increased conversion accuracy of A/D-conversion of the monitoring signal, thereby avoiding the above mentioned drawbacks.

According to one aspect there is disclosed a method for negative feedback controlling electrical power delivered to an electrical load comprising: S monitoring electrical power delivered to said load, thereby generating an analogue monitoring signal; S analog to digital converting said analog monitoring signal to generate a digital monitoring signal; 0 forming in dependency of said digital monitoring signal and of a preselected rated value signal a control deviation signal; S adjusting said electrical power delivered and monitored in function of said control deviation signal, characterized by forming said analog to digital converting at least twice in parallel.

Thereby, at least two analog to digital converters, each construed for a predetermined sampling rate, are operated in parallel. This leads to an overall conversion with at least [R:\LIBOO]658044.doc:TCW -3double that sampling rate, wherefore each of the converters is construed for, if both converters as preferred are operated at equal sampling rates.

Under the second aspect of the present invention the following has to be considered: Whenever a method for negative feedback controlling electrical power delivered to an electrical load comprises adjusting the electrical power delivered to the load and monitored by means of PWM, accuracy of adjustment is significantly limited by the predetermined minimum pulse-width adjustment increment by which pulse length may be at all adjusted.

Io Yet another need exists to, at a method as mentioned above, whereat, within the negative feedback loop, adjusting of the electrical power delivered to the load and monitored in function of the control deviation signal is performed by pulse-width modulation with a predetermined pulse repetition period, to improve adjustment accuracy beyond the limit is set by the predetermined minimum pulse-width adjustment increment.

According to another aspect there is disclosed a method for negative feedback controlling electrical power delivered to an electrical load comprising: monitoring electrical power delivered to said load, thereby generating a monitoring signal; S forming in dependency of said monitoring signal and of a rated value signal a control deviation signal; S adjusting said electrical power delivered and monitored in function of said control deviation signal by means of pulse-width modulation with a predetermined pulse repetition period; 0 said pulse-width modulation being adjustable with a predetermined minimum pulse-width adjustment increment; characterized by calculating from said control deviation signal a desired pulse-width adjustment increment and applying an integer multiple of said predetermined minimum pulse-width adjustment increments to said pulse-width modulation and/or a pulse of [R:\LIBOO]658044.doc:TCW predetermined length so often in time, that adjustment by said integer multiple and/or said pulse accords, averaged over time, with a pulse-width modulation adjustment with said desired pulse-width adjustment increment.

Thereby, one can exactly deal with any desired adjustment according to any fractions of the predetermined minimum pulse-width adjustment increment.

If e.g. the minimum predetermined pulse-width adjustment increment is of 30 nsec. and an actual control deviation would necessitate a pulse-width adjustment by an increment io only of 10 nsec., then the calculation will reveal in preferred mode that the pulse of predetermined length as in a most preferred embodiment a pulse which is the predetermined minimum pulse-width adjustment increment of 30 nsec. is to be applied in every third pulse repetition period. This leads, time-averaged as by filtering, to the same result as applying a 10 nsec. pulse in every period of PWM.

If e.g. the control deviation signal reveals a desired pulse-width adjustment increment of nsec., then there will be applied in each pulse repetition of the PWM one predetermined minimum pulse-width adjustment increment and, additionally, in every third period an additional predetermined minimum pulse-width adjustment increment, if, as in the most preferred mode, the pulse is selected to be just this increment.

In other words, this inventive technique combines customary pulse-width adjustment technique of PWM with a pulse frequency modulation (PFM) technique, at which pulses of a predetermined fixed length, namely as preferred according to the predetermined minimum pulse-width, are applied at a varying frequency.

In a preferred embodiment of the inventive method and power supply both aspects of the invention as outlined above, namely of inventive A/D-conversion and of inventive adjustment by PWM and superimposed PFM, are combined.

Although the present invention under all its aspects may be used or realized for lower power supply, in a far preferred mode of operation, it is realised for controlling the electrical power of at least 100 VA delivered to a load. Furthermore, monitoring the electrical power delivered to the load may be performed by monitoring the current or voltage delivered to the load.

[R:\LIBOO]658044.doc:TCW O In spite of the fact that the above mentioned A/D-conversion at least twice in parallel, N might be applied in cases too, where both conversions are performed at minimum ;required sampling rates. Additionally, to the inventively performed parallel conversion, Cc each of the A/D-conversions is performed with oversampling.

In spite of the fact that A/D-conversion in parallel could be performed in some cases at Smutually different sampling rates, e.g. at sampling rates of integer ratio, in a preferred Cc embodiment, A/D-conversions are performed in parallel at respective equal sampling rate.

IND

Such an equal sampling rate may vary in time, in some applications, e.g. where the (Ni bandwidth development of an analog signal to be monitored is known in advance.

Nevertheless, in a preferred embodiment, parallel A/D-conversions are performed at respective equal and constant sampling rates. In a further preferred mode the converters are operated synchronously.

In a further preferred embodiment, each of the at least two parallel A/D-conversions is performed with a sampling rate of at least 100 kHz, thereby resulting in an overall sampling rate of two parallel conversions of 200 kHIz.

Turning back to the present invention according to a second aspect or under preferred combination of both of its aspects, in a most preferred embodiment the pulse of predetermined length applied by PFM is selected to be the incremental pulse of the predetermined minimum pulse-width adjustment increment. Although this pulse needs not necessarily be of constant length, this is clearly the preferred mode.

Either once again the instantaneous control deviation may be corrected by an adjustment of the PWM pulse-width by an integer multiple of predetermined minimum pulse-width adjustment increments, or the instantaneous control deviation may only be accurately corrected by adjusting pulse-width modulation as was just explained and by additionally providing for applying pulses, preferably of the predetermined minimum pulse-width adjustment increments, by pulse frequency modulation, i.e. with a variable repetition rate, namely so often in time as necessary to deal with a control deviation which necessitates applying a PWM-adjustment by a fraction of the predetermined minimum pulse-width adjustment increment. Nevertheless, the case may also occur, where no pulse-width modulation adjustment at all is necessary, in these cases, namely [R:\LIBOO]658044.doc:TCW for down to zero adjustments only, the single pulses, preferably of an extent according to the predetermined minimum pulse-width adjustment increment, are applied with the (modulatable) repetition rate set as necessary.

Thereby and in a further preferred mode, at the pulse frequency modulation, the accordingly modulated pulse repetition period is modulated by an integer multiple of the pulse repetition period of the pulse-width modulation.

For addressing the above mentioned need according to a first aspect of the present io invention there is further proposed a negative feedback control power supply. Under its second aspect a negative feedback control power supply addresses the above mentioned need, and in a far preferred mode the needs addressed under both of the aspects of the present invention are at least ameliorated by the digital negative feedback control power supply, which combines both aspects of the present invention.

According to another aspect there is disclosed a negative feedback-controlled power supply comprising: a monitoring unit monitoring electric power delivered to a load and having an analog monitoring signal output, an analog to digital converter unit having an analog signal input being operationally connected to said analog monitoring signal output and having a digital monitoring signal output, a controller unit with a digital signal input operationally connected to said digital monitoring signal output and further having a control output, further comprising a difference forming unit, a first input thereof being operationally connected to said digital monitoring signal input, a second input thereof being operationally connected to an output of a rated value signal source and further having a difference signal output operationally connected to said control output, a power adjusting unit with a power input, a power output and a control input, said power output being operationally connected to an input of said monitoring R:\LIBOO0658044.doc :TCW unit and being operationally connectable to said load, the control input being operationally connected to said control output, characterized by the fact that said analog to digital converter unit comprises at least two analog to digital converters, the analog inputs thereof being operationally connected to said analog signal input and having a digital signal output each being operationally connected to said digital monitoring signal output, further comprising a sampling/converting control unit with at least two sampling/converting control outputs, said at least two sampling/converting control outputs being operationally connected respectively to a sample/converting control input of each of said at least two converters, said sampling/converting control unit generating at said at least two sampling/converting control outputs sample and conversion enabling signals.

According to yet another aspect there is disclosed a negative feedback-controlled power supply, comprising: a monitoring unit monitoring electric power delivered to a load and having a monitoring signal output, a controller unit with a monitoring signal input operationally connected to said monitoring signal output and further having a control output, a difference forming unit, a first input thereof being operationally connected to said monitoring signal input, a second input thereof being operationally connected to an output of a rated value signal source, and further having a difference signal output, a pulse-width control unit with a pulse-width control input operationally connected to said difference signal output and having an output operationally connected to said control output and generating at least said control output a pulse-width modulated control signal; a power adjusting unit with a power input, a power output and a control input, said power output being operationally connected to an input of said monitoring unit and being operationally connectable to a load, the control unit being operationally connected to said control output, said power adjusting unit further [R:\LIBOO]658044.doc:TCW comprising a power-pulse-width modulation unit with a pulse-width control unit operationally connected to said control input and with an input operationally connected with said power input and with an output operationally connected to said power output, said power-pulse-width modulation unit transmitting a signal applied to said power input, pulse-width modulated to said power output with a predetermined minimum pulse-width adjustment increment, characterized by the fact that said controller unit further comprises a calculation unit, one calculation input thereof being operationally connected to said difference signal output, the calculation output thereof being operationally connected to said pulse-width control input and calculating in dependency of a signal applied to said calculation input a desired pulse-width adjustment increment at said pulse-width modulation unit, said calculation unit generating at said calculation output a first control signal controlling adjustment of pulse-width of said pulse-width control unit by an integer number of said predetermined minimum pulse-width adjustment increment according to an integer of the ratio of said desired pulse-width adjustment increment and said predetermined minimum pulse-width adjustment increment, and generating at said calculation output a second control signal controlling frequency of a frequency modulated pulse train with a pulse of predetermined pulse-width according to a fraction of the ratio of said desired pulse-width adjustment increment and said predetermined minimum pulse-width adjustment increment.

Preferred embodiments of the inventive negative feedback control power supply will additionally become apparent from the following detailed description. Further, the inventive power supply, especially combining both aspects, is applied to magnet-supply of a synchrotron, thereby resulting in an inventive synchrotron system. In the detailed description, preferred embodiments of the present invention will be described as examples and referring to figures. These figures show: Fig. 1 By means of a simplified functional block signal flow diagram an inventive negative feedback controlled power supply under the first aspect and operating according to the inventive method under its first aspect; [RAL1B001658044.doc:TCW 8a Fig. 2 qualitatively, control input trains applied to a parallel operated analog to digital converter as provided in the embodiment of fig. 1; Fig. 3 in a simplified functional block signal flow representation according to the representation of fig. 1, a further preferred embodiment of the present invention under its second aspect, namely of a further embodiment of an inventive power supply operating according to an inventive method; Fig. 4 in a simplified functional block signal flow diagram, a possible structure of a calculation unit as incorporated in a controller unit as of the embodiment of fig.

3; [R:\LIBOO]658044.doc:TCW WO 01/86368 PCT/CH01/00376 9 Fig. 5 a time diagram of superimposed pulse-width modulation and pulse-frequency modulation as realized by the embodiment according to the figs.

3 and 4; Fig. 6 a preferred form of superimposed pulse-width modulation and pulse frequency modulation as preferably realized in the embodiment according to figs. 3 and 4, and Fig. 7 in a representation according to fig. 4, a today's most preferred embodiment.

In fig. 1 there is shown, by means of a simplified functional block signal-flow diagram a preferred embodiment of the negative feedback controlled power supply according to the present invention and operating according to a preferred variant of the inventive method, both under the first aspect of the present invention.

According to fig. 1 uncontrolled electrical power line Pe is fed to a power input El of a power adjusting unit i.

From a power output A, of unit 1 the adjusted or controlled power line Pa is led to an electrical load 3. A monitoring input E 5 of a monitoring unit 5 is operationally connected to power line Pa, thereby monitoring load current and/or load voltage. Generically the adjusted power on line Pa at the output A, of adjusting unit 1 is monitored by monitoring unit 5 as the measured controlled value Xm of the negative feedback control loop to be further described.

The output signal X and respectively the output A 5 of monitoring unit 5 are operationally connected to a difference-forming unit 7, whereat the signal from A 5 is subtracted from a rated value signal W, which is selected WO 01/86368 PCT/CH01/00376 10 at a rated value generating source 9. In spite of the fact that this rated value source 9 is shown in fig. 1 incorporated within controller unit 7, a rated value signal W may clearly be fed to that control unit 7 from an external signal source.

At the output All of difference forming unit 11, whereat, from the said rated value signal W and from the controlled value signal X a control deviation signal A is formed, which is fed to a controller 13, the output thereof being operationally connected to a control output A 7 of controller unit 7. Output A 7 is operationally connected to control input Ci of adjusting unit 1i.

According to the embodiment of fig. 1 and thereby according to the present invention under its first aspect, the input

E

5 of the monitoring unit 5 with the analog monitoring signal Xm is operationally connected to the analog inputs and E15b of at least two analog to digital converters (ADC) 15a and 15b of monitoring unit 5. The digital signal outputs Ai 5 a and A15b of the ADCs 15a, 15b are operationally connected to superimposing unit 17.

The monitoring unit 5 further comprises a sampling/converting control unit 19 generating at two outputs A19a, A19b respective pulse trains which are fed to sample/conversion control inputs CL5a and C15b at the respective ADCs 15a and As shown in fig. 2 the sampling/converting control signals Scisa and SC1Sb, and especially their edges as of e are preferably in synchronism, so that the two ADCs 15a and which are operated in parallel, perform sampling simultaneously..

WO 01/86368 PCT/CH01/00376 11 Nevertheless, it is also possible to operate the converters staggered in time and/or at time varying sampling rates and/or at different sampling rates in some cases.

Thereby, the sampling rates at the at least two ADCs are most preferably selected to perform oversampling.

At the output of superimposing unit A 17 the conversion results of the at least two ADCs 15a, 15b appear as respective digital words.

The output A 17 and thus the conversion result signals are fed to an averaging unit 21, which performs digital averaging of the digital signals provided from A 17 If sampling and especially oversampling each of the converters 15 provides for n digital values, due to the fact that parallel conversion is performed, 2 n or more generically k n digital values are averaged.

Thus, the overall sampling rate is risen by the number k of parallely processed conversions, without the need of resolving at a single ADC the problems for accordingly rising the oversampling rate.

Again in a simplified functional block signal flow diagram, fig. 3 shows a negative feedback controlled power supply according to the present invention under a second aspect and operating accordingly to the second aspect of the method according to the present invention.

Electric power line Pe is fed to the power input E 30 for uncontrolled power of a power adjusting unit 30. The output

A

30 thereof for controlled power is operationally connected WO 01/86368 PCT/CH01/00376 12 to a power monitoring unit 35 and to an electrical load 3.

Power monitoring unit 35 monitors load current and/or load voltage. Monitoring unit 35 needs not be construed as was explained in context with fig. 1. In most generic terms, monitoring unit 35 does not even provide for analog to digital conversion.

The output A 3 5 of the monitoring unit 35 is operationally connected to input E 37 of the controller unit 37. A control output A 37 of unit 37 is operationally connected to control input C 30 of adjusting unit The adjusting unit 30 is construed as a power pulse-width modulation unit. A control signal applied to control input

C

30 controls pulse-width within the pulse repetition period Tp of the pulse-width modulation PWM. Additionally, every pulse applied to C 30 leads to transmission of power from E 30 to A 30 during the time extent of such pulse. Pulse-width modulation of power delivered to load 3 from power adjusting unit 30 may be said to be controlled by a control signal CpwM defining for a controlling pulse-width modulated signal, which is generated within controller unit 37 by a pulse-width control unit 39. In block 39 the control signal CpWM is shown as a pulse-width modulated analog signal, also it will be represented in preferred mode by a digital control signal.

Due to the specific layout of PWM power adjusting unit and/or possibly pulse-width controlling unit 39, pulsewidth Tmod of the pulse-width modulated power signal may only be changed by an integer number m of predetermined minimum pulse-width adjustment increment 5min. This predetermined pulse-width minimum adjustment increment 6 min WO 01/86368 PCT/CH01/00376 13 limits adjusting accuracy of the power output to the electric load 3 and monitored by monitoring unit Still according to the invention under this second aspect, the output signal of monitoring unit 35 is operationally connected via input E 37 of controller unit 37 to a difference forming unit 41, as controlled signal X. At difference forming unit 41 from the controlled value X and a rated value W generated by a rated value source 43 a control deviation signal A is generated. The control deviation A is fed via a controller (not shown here) to a calculating unit 43.

Fig. 4 shows in more details, what operations are principally performed inventively at calculation unit 43.

In function of the actual control deviation signal A calculation unit 43 determines as schematically shown in fig. 4 in block 45 a desired adjustment 6des which would have to be applied to the control input C 30 of power adjusting unit 30 to negative feedback control the power Pa delivered to the electric load 3 exactly to match the desired value W so that A becomes at least approximately zero. Nevertheless, and as was explained, power adjusting unit 30 and/or pulse-width modulation control unit 39 are limited with respect to the minimum predetermined adjustment increment §min with which pulse-width Tmod may be changed. Therefore, the present invention under this aspect considers that although no pulse-width adjustment increments may be applied which are smaller than frequency with which a pulse-width adjustment is applied is an additional parameter which allows controlling the effect on the power output at A 30 WO 01/86368 PCT/CH01/00376 14 As will be explained, the drawback that pulse-width modulation provides for a predetermined minimum pulse-width adjustment increment 6min is bypassed by introducing additionally to pulse-width modulation PWM, pulse frequency modulation PFM. We understand by pulse-frequency modulation PFM applying pulses of predetermined fixed length and varying the repetition rate of these pulses.

Following this principle, the calculation unit determines the repetition rate of pulse frequency modulation with pulses preferably with an extent of 6in so as to adjust the power at A 30 by less than is possible by varying Tmod by a single increment 6 min.

As schematically shown in fig. 4 by block 47, calculating unit 43 determines as a possible form of realization the ratio R of the desired adjustment increment 5min and the predetermined minimal adjustment increment 5 min. Let's assume this ratio results in a number R 2.6. Unless the desired adjustment 5 des accords exactly with an integer multiple of the predetermined minimum pulse-width adjustment increment 6,mn, R may be written as R I 100 F 10-1 as of 2.6 with I 2 100 and F 6 10-1. Thus, the result of ratio forming in unit 47 is split in an integer I and a fraction F. The integer number I controls directly adjustment of pulse-width Tmod of the pulse-width modulation by the according number of increments I 6min in each pulse repetition period This accords to customary pulse-width modulation control.

The fractional number F, in the above example of 0.6, is used to apply additionally to the pulse-width modulation a pulse train, which is pulse frequency modulated PFM and WO 01/86368 PCT/CH01/00376 15 with a preferably fixed pulse, preferably of time extent 6min. Thus, such additional pulse is not anymore applied in succeeding pulse repetition periods Tp, but just as often as necessary to have, averaged over time, the same effect on the power at A 30 as if in every pulse repetition period Tp the pulse length would be changed by F 65n, which is not feasible. If we take Tp as time unit, the time period with which an impulse of extent Smin is to be applied becomes TpFM 1/F, thus for our example Tpm 1,67. Thus, instead of applying a respective adjustment of pulse-width in every time frame according to every pulse repetition period Tp, the repetition period prevailing for applying a single additional pulse of extent 5mi, is modulated to Tp.

In fig. 5 this superposition of pulse-width modulation adjusting and of pulse frequency modulation adjusting is exemplified. In every pulse repetition period T 2 pulse width adjustment is performed with an adjustment according to I 6 min according to the above example by 2 in.

Additionally, there is superimposed the pulse frequency modulated signal consisting of fixed-extent pulses as of 8min applied with a rate according to a repetition period of TPFM Thus, in Fig. 3 the control signal to becomes Cpw CFM.

Although this procedure is absolutely possible, it is not optimum in view of the fact that pulse-width modulation is clocked at a predetermined fixed rate according to Tp and that procedure according to fig. 5 may lead to pulse-width modulated signals overlapping with the frequency-modulated WO 01/86368 PCT/CH01/00376 16 impulse train. This situation would necessitate additional efforts to compensate for such occurrences.

Therefore, in a preferred embodiment the frequencymodulated signals with pulses of the extent 6 min are applied during F N of succeeding pulse repetition periods Tp wherein N stands for a number of succeeding pulserepetition periods of pulse-width modulation. According to the above mentioned example and taking N e.g. to be this means that the single pulse of Smin is applied to 6 pulse repetition periods Tp out of Thereby, in fact frequency modulation is performed in discrete steps of pulse repetition period Tpm which step being Tp. This preferred realization form is shown in fig.

4 by dashed lines.

This latter approach is further exemplified in fig. 6, where according to the above example out of 10 pulse repetition periods Tp of the pulse-width modulation, in 6 an additional pulse of 8min is applied. Averaged over time the output power Pa of power adjusting unit 30 is adjusted accurately according to the desired pulse-width adjustment increment 6des which accords to the instantaneous control deviation A.

In fig. 7 a most preferred form of PMW and PFM realization is shown as incorporated preferably in calculation unit 43.

The instantaneous control deviation A is fed via the controller again to a unit 45, where, according to the instantaneous control deviation A, a desired pulse-width adjustment increment 6des is calculated.

WO 01/86368 PCT/CH01/00376 17 From this desired pulse-width adjustment increment 5 des, as was explained in context with fig. 4, unit 47 determines the ratio R of 6 des and the predetermined minimum pulsewidth adjustment increment 8min- From the resulting ratio R (6des/6min) the integer part I is stored and controls the adjustment of pulse-width of the pulse-width modulation by I 6des. Thereby, the control deviation A drops so that R becomes smaller than unity according to the fraction F of the previously calculated value. The fraction F of the ratio R at the output of unit 47 is integrated or summed over time in a summing unit 48, preferably for each subsequent period Tp. The output of summing unit 48 is added as schematically shown in fig. 7 by adding unit 49 to the integer output of unit 47. Thus, as soon as time integral or the sum of the remaining fraction F of ratio R at the output of unit 48 reaches unity, the value I is incremented by one, and, in a single pulse repetition period Tp of pulse-width modulation, the prevailing pulse is additionally adjusted by the one predetermined minimum pulse-width adjustment increment min- By this adjustment in that period Tp considered, the prevailing value of the control deviation and of the ratio R again drop. By resetting unit 48 as schematically shown at Rs in fig. 7 as soon as its output signal reaches unity, in the subsequent periods Tp pulse-width is kept adjusted by I' As shown in dashed line in fig. 7, instead of checking whether the sum of the subsequent values F of the ratio R reaches a value for incrementing I by unity, it is WO 01/86368 PCT/CH01/00376 18 absolutely possible to integrate or sum, preferably after the controller, the control deviation A.

In a preferred embodiment of the inventive power supply and of the inventive method, both aspects according to fig. 1 and directed to accurate analog to digital conversion of the monitored power and according to the figs. 3 to 6 with respect to inventively adjusting the power fed to the electric load 3 in a negative feedback power control loop are combined to result in digital negative feedback control of the power supplied to a load 3.

By realizing such combined technique with two 16 bit analog to digital converters 15 according to fig. 1, each operated at a sampling rate of 100 kHz and by realizing pulse-width modulation with superimposed pulse frequency modulation as was described especially with the help of fig. 4 and 6 there was achieved an accuracy of resolution of at least ppm and even down to a resolution accuracy of 1 ppm. This for a power supply for at least 100 VA power used for power supplying the magnet arrangement of synchrotron magnets.

Claims (22)

1. A method for negative feedback controlling electrical power delivered to an electrical load comprising: monitoring electrical power delivered to said load, thereby generating an analogue monitoring signal; analog to digital converting said analog monitoring signal to generate a digital monitoring signal; forming in dependency of said digital monitoring signal and of a preselected rated value signal a control deviation signal; adjusting said electrical power delivered and monitored in function of said control deviation signal, characterized by performing said analog to digital converting at least twice in parallel.

2. A method for negative feedback controlling electrical power delivered to an electrical load comprising: monitoring electrical power delivered to said load, thereby generating a monitoring signal; forming in dependency of said monitoring signal and of a rated value signal a control deviation signal; adjusting said electrical power delivered and monitored in function of said control deviation signal by means of pulse-width modulation with a predetermined pulse repetition period; said pulse-width modulation being adjustable with a predetermined minimum.pulse-width adjustment increment; WO 01/86368 PCT/CH01/00376 20 characterized by calculating from said control deviation signal a desired pulse-width adjustment increment and applying an integer multiple of said predetermined minimum pulse-width adjustment increments to said pulse-width modulation and/or a pulse of predetermined length so often in time, that adjustment by said integer multiple and/or said pulse accords, averaged over time, with a pulse-width modulation adjustment with said desired pulse-width adjustment increment.

3. A method according to claim 2, characterized by generating by said monitoring an analog monitoring signal; analog to digital converting said analog monitoring signal to generate a digital monitoring signal; performing said analog to digital converting at least twice in parallel.

4. The method of claim 1 or 3, characterized by performing each of said analog to digital convertings with oversampling.

5. The method of claim 1, 3 or 4, characterized by the fact of performing said analog to digital convertings at a respective equal sampling rate.

6. The method of claim 5, characterized by performing said analog to digital convertings at a constant sampling rate.

7. The method of claim 1, 3 to 6, characterized by performing said analog to digital convertings synchronously. WO 01/86368 PCT/CH01/00376 21

8. The method of one of claims 1, 3 to 7, characterized by performing each of said analog to digital convertings with a sampling rate of at least 100 kHz.

9. The method of claim 2 or 3, characterized by selecting said predetermined length equal to said predetermined minimum pulse-width adjustment increment.

The method of one of the claims 2, 3 to 9, characterized by applying said pulse of predetermined length as a pulse-frequency modulation with only a variable pulse repetition frequency.

11. The method of claim 10, characterized by modulating pulse repetition period of said pulse-frequency modulation by an integer multiple of said pulse repetition period of said pulse-width modulation.

12. The method of one of claims 2, 3 to 10, characterized by adjusting pulse-width of said pulse-width modulation by an integer multiple of said predetermined minimum pulse- width adjustment increment, integrating the resulting control deviation in time or adding the resulting control deviation in subsequent pulse-repetition periods, and applying said single pulse of predetermined length, whenever the result of said integrating or of' said adding accords with a control deviation which necessitates adjusting said pulse-width by a predetermined minimum pulse-width adjustment increment.

13. The method of one of claims 1 to 12, characterized by controlling electrical power of at least 100 VA.

14. The method of one of claims 1 to 13, characterized by monitoring said electrical power by monitoring current or voltage delivered to said load.

WO 01/86368 PCT/CH01/00376 22 A negative feedback-controlled power supply comprising: a monitoring unit monitoring electric power delivered to a load and having an analog monitoring signal output, an analog to digital converter unit having an analog signal input being operationally connected to said analog monitoring signal output and having a digital monitoring signal output, a controller unit with a digital signal input operationally connected to said digital monitoring signal output and further having a control output, further comprising a difference forming unit, a first input thereof being operationally connected to said digital monitoring signal input, a second input thereof being operationally connected to an output of a rated value signal source and further having a difference signal output operationally connected to said control output, a power adjusting unit with a power input, a power output and a control input, said power output being operationally connected to an input of said monitoring unit and being operationally connectable to said load, the control input being operationally connected to said control output, characterized by the fact that said analog to digital converter unit comprises at least two analog to digital converters, the analog inputs thereof being operationally connected to said analog signal input and having a digital signal output each being operationally connected to said digital monitoring signal output, further comprising a sampling/converting control unit with at least WO 01/86368 PCT/CH01/00376 23 two sampling/converting control outputs, said at least two sampling/converting control outputs being operationally connected respectively to a sample/converting control input of each of said at least two converters, said sampling/converting control unit generating at said at least two sampling/converting control outputs sample and conversion enabling signals.

16. A negative feedback-controlled power supply, comprising: a monitoring unit monitoring electric power delivered to a load and having a monitoring signal output, a controller unit with a monitoring signal input operationally connected to said monitoring signal output and further having a control output, a difference forming unit, a first input thereof being operationally connected to said monitoring signal input, a second input thereof being operationally connected to an output of a rated value signal source, and further having a difference signal output, a pulse-width control unit with a pulse-width control input operationally connected to said difference signal output and having an output operationally connected to said control output and generating at said control output a pulse-width modulated control signal; a power adjusting unit with a power input, a power output and a control input, said power output being operationally connected to an input of said monitoring unit and being operationally connectable to a load, the control input being operationally connected to said control output, said power adjusting unit further comprising a power-pulse-width modulation unit with a WO 01/86368 PCT/CH01/00376 24 pulse-width control input operationally connected to said control input and with an input operationally connected with said power input and with an output operationally connected to said power output, said power-pulse-width modulation unit transmitting a signal applied to said power input, pulse-width modulated to said power output with a predetermined minimum pulse- width adjustment increment, characterized by the fact that said controller unit further comprises a calculation unit, one calculation input thereof being operationally connected to said difference signal output, the calculation output thereof being operationally connected to said pulse-width control input and calculating in dependency of a signal applied to said calculation input a desired pulse-width adjustment increment at said pulse-width modulation unit, said calculation unit generating at said calculation output a first control signal controlling adjustment of pulse- width of said pulse-width control unit by an integer number of said predetermined minimum pulse-width adjustment increment according to an integer of the ratio of said desired pulse-width adjustment increment and said predetermined minimum pulse-width adjustment increment, and generating at said calculation output a second control signal controlling frequency of a frequency modulated pulse train with a pulse of predetermined pulse-width according to a fraction of the ratio of said desired pulse-width adjustment increment and said predetermined minimum pulse- width adjustment increment.

17. The negative feedback control power supply according to claim 16, characterized by the fact that said monitoring WO 01/86368 PCT/CH01/00376 25 signal output is an analog monitoring signal output and said monitoring signal input is a digital monitoring signal input and further comprising an analog to digital converter unit interconnected between said analog monitoring signal output and said digital monitoring signal input and comprising at least two analog to digital converters, the analog inputs thereof being operationally connected to said analog monitoring signal output and having a digital signal output each being operationally connected to said digital monitoring signal output and further comprising a sampling/conversion control unit with at least two sampling/converting control outputs, said at least two sampling/converting control outputs being operationally connected respectively to a sample/converting control input at each of said at least two converters, said sampling/conversion control unit generating at said at least two sampling/converting control outputs sample and convert enabling signals.

18. The power supply according to one of claims 15 to 17, said power adjusting unit adjusting a power delivered to said load of at least 100 VA.

19. The power supply according to one of claims 15 to 18, wherein a power at said power output is controlled with a resolution of at least 10 ppm, preferably with a resolution of at least 1 ppm.

A synchrotron system with electron beam-controlling magnets, said magnets being electrically supplied by power supplies according to one of the claims 15 to 19. -26-

21. A method substantially as described herein, with reference to any of the accompanying drawings.

22. A negative feedback-controlled power supply substantially as described herein with reference to any of the embodiments, as that embodiment is illustrated in any of the accompanying drawings. Dated 31 May, 2006 Paul Scherrer Institut Patent Attorneys for the Applicant/Nominated Person 1o SPRUSON FERGUSON [R:\LIB00]658044.doc:TCW