CA1080827A - Method of steering a spacecraft and of regulating its onboard supply voltage - Google Patents

Abstract

ABSTRACT OF THE DISCLOSURE
The invention relates generally to a method of regulating the onboard voltage of a spacecraft and of steering the same by at least two contrarotating motor-generator dynamo-electrical machines. According to the method, there are provided two phases:
- the energy storage mode, during which each machine receives electrical energy from solar generators through a velocity controll-ing generator commanded from a synthesis of the data relating to position and rotor velocity; and - the energy restitution mode, during which each generator supplies energy to a load consisting of the onboard equipment of the satellite. Output monitoring systems effect a synthesis of the data relating to position rotor velocity and additionally control the power switches.
According to the same method, there is also provided on board of said spacecraft, electronic equipment which is divided into two main sections:
- electronic control equipment; and - electronic processing equipment, itself divided into two parts:
control means positioning system and a computing system.

Description

1 The stora~e of energy on a spacecraf-t is rendered necessary by the fact that the solar cells which normally supply the onboard electric current are periodically inopera-tive when an artificial earth sa-tellite i5 in a shadow zone.
It is also known that a satellite can be steered by gas jets or by control:Ling the kinetic moment resulting from the rotation of momentum wheels, or by joint use of the two systems.
When it is required both to store energy in order to be able to restore it in electrical form during periods when the solar cells are inoperative, and to steer a spacecraft by controlling the kinetic moment, there can be provided at least one set of two contrarotating motor-generators, as described in French Patent No. 2,347,716 granted to the applicant on November 3, 1978.

In this manner, the solar cells speed the motors up to their maximum rota-tion speed while the cells are illuminated, whereas the motors operate as generators during the energy restitution phases, their speed being continuously controlled to enable the satellite to be steered.

. . ~
It is not proposed hereinafter to describe the principle of - 20 energy storage by momentum wheel but to describe the theory of operation of dynamo-electric machines capable of performing their functions as motor-generators in the required space enviornment.
During periods of illumination the energy is stored by the momentum wheels of the machines, which regulate it and at the same ~ ;
time absorb the surplus power from the solar generators. These same momentum wheels can be used for steering the satellite by generating moments which can be as small as desired.
During period of occulting, the energy is given up by these same machines, which regulate the voltage delivered. They can continue to be used to steer the satellite by f~rnishing the required kinetic moments, and provide in all cases the necessary gyroscope rigidity.
' '' ,~

3~

Such a pat~erll of operatioll, however9 poses a number o~ sl~ving problems which can be summari7.ed as follows: When the machines are func tioning as motors during the phase when maximum rotation speed is sought, maintaining a constant kinetic moment requires that the rotation speed differential between the machines be slaved.
` According to one method, the speed of the pilot motor increases according to a time-dependent law so determined that the slaved motor fol-low the speed of the pilot motor with a speed differential which is modu-lated by the steering function, whereby it is unnecessary to slave the pilot motor to the theoretical law.
In an alternative method, the two motors follow the same speed incrementing law, and their speed differential provides the kinetic moment.
In this case each control law is modulated by the steering function.
Both these methods, however, require measuring the actual speeds of the two machines and reacting accordingly, which, owing to their high ~; inertia imposed by the energy to be recovered, implies making extremely accurate measurements if it is desired to be able to apply weak perturb-ing moments to the satellite.
These methods therefore have the disadvantage of requiring exten-. .
sive processing prior to actually controlling the machines, which in turn .
calls for highly comprlex electronics unless an onboard computer is avail-able.
.,~ .
` As will be explained in greater detail hereinafter, the present in-vention provides a solution well-suited to the problem involved.
When the machines are operating in the generator mode, the only ... .
parameter available for controlling their speeds is the current they de-~`~ liver.
. .i ~
It is therefore necessary to so adjust the outputs from the two machines that the currents delivered determine their respective speeds, the voltage across the load tarminals being required to remain constant.

.. ' ' :

As will be dPscribed :in greater detail hereinbelow, this invention also provides for a generator meeting the necessary requirements.
The solution proposed by this invention consists in providing two contrarotating dynamo-electric machines forming a momentum wheel, each machine having a permanent-magnet type rotor forming tha field magnet and a stator carrying the armature windings. The stator can be of the type with or without iron, since the control law obtained is the same in both cases.
The heavy parts of the machines can thus be disposed on the rotor, their control being effected by a special method of switching the wind-ings that is one of the essential teachings of this invention.
Knowledge of the physical laws governing the behaviour of the mach-ines makes it possible to establish control laws for producing perturbing moments capable of being very weak.
Generally speaking, the different phases of operation of the two machines are such that, in the motor mode and with no steering correction, a clock controls the incrementing of their respective speeds in acc~rdance with the p~e~established theoretical law.
The commands applied to each machine take account of the speed dif-. . . .
20 ferential imposed by the last steering correction. i Such operation is made possible ~y memoriæation of the com~andedspeed differential.
Memorization of the co~manded speed differential is accomplished by integrating the steering error signal.
With steering correction, control of the machines takes account - both of the pre-established theoretical law and of the instantaneous speed ; differential commanded by the steering function.
A steering moment is applied by accelerating one or the other of the motors, depending on the sense in which the momen~ is to be applied.
In the geneFaeor mode and with no steering correction, a clock con-~ 4 ~ ':

, .`.~ . .

trols decreases in ~he speeds of the ~enerators, in accordance with the pre-es~ablished theoretical laws.
The clock frequency is slaved by the output voltage in such man-ner as to maintain the latter constant irrespective of load variations.
The control inputs to the two machines take account of the speed differential commanded by the last steering correction.
This operation is made possible by memorizing the commanded speed differential, and the memorizing operation is accomplished in the same way as for operation in the motor mode.
With steering correction, control inputs to the machines allow both for the pre-established theoretical law and for the instantaneous speed differential commanded by the steering function.
A steering moment is applied by reducing the current delivered by one machine and increasing that delivered by the other machine, this be-ing done by dividing the commanded speed differential corrections equal-ly between the two machines.
The sense of the moment applied is determined by which of the machines is chosen for the current increase and decrease operat;ons re-spectively.
With such an operating mode, a system devised thus offers a number of advantages:
- - since the perturbing moments are very weak9 the steering function makes only infrequent corrections;
- since the system is autonomous, a temporary failure of the steering - function will not compromise the mission, and this is true both of the energy accumulation and restitution functions and of the moments applied to the satellite;
- similarly, because the speed differential is memori~ed in the con- ~-trol electronics, the steering function does not have to make continuous corrections;

~ 5 -~ lastly, the desi~n o~ the control electron;cs will no~ necessitate revising the electronics should ~he amount of energy to be stored be dif-ferent, for in that case only the informatiorl stored in the read-only memories and the size of the power switches would need to be revised.
The description which follows will enlarge in greater detall on the general concept of the invention and will give; by way of possible forms of embodiment, the means for implementing a system of two contra-rotating wheels having the ollowing leading particulars:
- Energy stored per unit installation weight........................ 35 Wh/kg - Number of charging and discharging cycles.................. Theoretically unlimited - Discharge efficiency for an onboard voltage of 50V................... 8~%
- Power required for electronics....................................... 23W
- Losses in a machine operating in motor mode.......................... 20W
- Losses in machine operating in generator mode........................ 30W
- Modulation depth..................................................... 76%
- Nominal voltage of machine................... ~.... ,.................. 47V
- Nominal speed of machine.......................... ......... ...... 14,400 r~m - Maximum operating speed........................... ......... ...... 42~600 rpm - Minimum operating speed.............................. ~....... 18,000 rpm - Inertia of a machine rotor................................... ...Ø4 kg m2 - Differential kinetic moment created by the machines.......... ....100 Nms - Variation in differential kinetic moment......... ~.................. +25%
- Maximum applicable torque ~........................ .......... 0.036 ~ Nm c 0.18 ~;~
- Speed correction setting......................... 10 bits ~ 1 sign bit - Protection against overloads........................... ............Provided - Perturbing moment in motor mode at nominal load........ ............2 x 10-~ Nm - Perturbing moment in generator mode at nominal discharge... 3 x 10 4 Nm ,~:
- Perturbing moment in generator mode with 25% load ~ariation.............................................. ............3 ~ 10 3 Nm - - Energy recoverable with a system of two contrarotating machines.... -.................. ~................ ............4.3 MJ

:-.:' .
:,. .

- Voltage regulation duri~lg recovery phase with :~30% loacl variation... -~1%
There w-ill lastly be described the electronics section required for operation of th~ machines, this electronics section being divided, in ac-cordance with a teaching of the present invention, into electronic control equipment and electronic processing equipment:, which processing equipment in turn comprises a control-means positioning system and a computing aystem.
For greater clarity in the description which follows, the following notation system will be adopted to designate the various parameters in-volved in the subject ~nethod of this invention for steering a spacecraft and regulating its onboard voltage:
~ 2 to designate the respective angular velocities of the two machines;
; - ~TH to designate the mean theoretical velocity of the machines;
~ 2 to designate the respective velocity differentials of the machines relative to the ~TH;
-~Tl, ~T2 to designate the respective theoretical vslocities of the two machines;
~ A~C to designate the velocity differential demanded by the satel-lite in order to generate the moment.
On the basis of this notation it is possible to determine the pro-cedure for carrying the invention Into practice, to wit:
During the motor phase the requirement is for maximum angular velo-city.
a) To apply the control inputs to the motors . Take account both of the theoretical law ~TH of rise in velo-city and of the instantaneous velocity differentials ~ 2 commanded by the steering function by generating:

TH ~ ~1 and ~2 = ~TH ~ ~2 b) Memorize the velocity differentials ~ 2 by integrating the ; 30 error signal from the steering function.

. .
. ~ '' '` , ' ' ~ ' ' ' .

c~ In order to apply to the macllines control inp~lts that take ac-count o the ~c~ generate:
. A moment SH (c].ock~ise) Wheel of machine 1 tl : A~
t2 ~ ''`)1 +
Wheel of machine 2 tl : ~2 t2 ~2 ~2 . Moment SAH (coùnter-clockwi.se) Wheel of machine 1 tl t2 ~heel of machine 2 t~ 2 t2 : ~2 = ~ c : d) In order ~o synchronize the control input with the angular elocity of the machine, use the clock frequency Fr synchronized with the angulsr velocities ~ 2 of the machines. - .
During the generator phase: : ~
' .

. a) In order to apply control inputs ta the generators~
,; ,; .
: . Take account both of the-theoretical law ~TH and of the in~
stantaneous velocity differentials ~ 2 commanded by the steering ; ~ ~
-. 20 function by generating: .~,....................... . ~
TH + a~l and ~2 = ~TH - h~2 b) Memorize the commanded velocity differentials 4~1~ 4~2 by integra~ing the error signal from the steering function.
c) In order to apply to the machines control inputs that take account of A~c, generate:
. . A moment SH (clockwise) . :
Wheel of machine 1 t~

t2 : ~ 1 2 - Wheel o~ machine 2 t~ 2 t2 : Q~2 2 2 : 8 ~
:' '' ' '~ ' ' . ~ . . .
, ' . . ' , !

. A mometlt S~}l ~counter-clockwise) l~eel of machine 1 tl : Q~l t2 A~l ~ Q~l - 2c ~leel of machine 2 tl : A~l t2 : ~2 = a~2 - 2 d) In order to synchronize the control inputs with the angular .~ velocity of the machine, use the clock frequency Fr synchronized with the angular velocities ~ 2 of the machines.
~ e) Continuously slave the clock frequency N controlling the law 10 of decrease in vslocity to the busbar voltage.
The description which follows with reference to the accompanying non-limitative exemplary drawings will give a clear understanding of how :; the invention can be carried into practice.
: In the drawings: -Figure l is a schematic representation of the basic phenomenon ~-. governing the law of controlg in the motor mode, of a machine according.
: to this inventioni . Figure 2 is a block diagram showing the slaved operation of two ~-contrarotating machines during the energy storage phase;
20 . Figure 3 is a schematic illustration in par~ial perspective of the positioning of the armatures and field magnets of a machine according-to ~ the inven~ion, - - Figure 4 is a schematic representation of the distribution of the magnetic induction in the machine of figure 3;
Figure 5 is a fragmental diagrammatic showing of an elemental part of the machine of figure 3;
Figures 5 and 7 are partial schematic illustrations of the distri-:, . . .
bution of the flux of a field magnet according to figure 5, and portray-..
ing the switching process according to this invention; .~ .
Figure 8 is a graph in which the torque is plotted against the posi-.

. ;` .

' , I . . ~ . . . .. . ., .. , . , " . . " , . .... , ." .. , .. , .. . .. . ., , " , .. , . " .. . . ..... . ... ..

tion of the switcll control;
Figllre 9 is a graph in which angular velocity is plotted against switch control position;
Figure 10 is a diagram for determining the actual velocity in the motor mode;
Figure 11 is a graph showing the instantaneous changes in the total counter-electromotive force in the motor mode;
Figure 12 is a graph showing the instantaneous current variations in a motor devoid of inductance, Figure 13 is a graph showing the actual instantaneous variation in current in the motor mode, i.e. one possessing inductance;
Figure 14 is a diagrammatic illustration of the basic process gov~
erning the law of control in the generator mode of a machine according to -the invention;
Figure 15 is a block diagram showing the slaved aperation of two . ~: -contrarotating machines during the energy restitution phase;
Figure 16 is a diagram showing the positions of the four energy tapping gates and their respective durations, with respect to the instan~
taneous electromotive forces presented on the coils; :
- 20 Figure 17 is a diagram representing the maximum travel for posi - tioning the control;
Figure 18 is a diagram showing the output electromotive forces obtainable from the machine without a filtering operation;
Figure 19 is a diagram showing how the point of operation in the generator mode shifts as the contro~ position is changed;
Figures 20 and 21 are diagrams representing the signals received by the pick-up;
Figure 22 is a diagram showing the manner in which the control gates evolve as the pick up signals vary;
Figure 23 is a diagram showing the disposition of the pick-ups;

b 1 0 ~f~

Figure 24 is a graph depicting thç~ form of the electromotivc force in a single coil;
Figure 25 is a graph portraying the shape of the switching signals;
Figure 26 is a graph portraying the shape of the switching signals subsequent to decoding;
Figure 27 is a diagram showing a possible operating c~cle for the machines;
Figure 28 is an overall block diagram showing the arrangement of the components required for operation,of.the machines;
Figure 29 is a ci.rcuit diagram showing the arrangement of the com-ponents forming the control electronics;
Figure 30 is a circuit diagram of the power selectorsj Figures 31 and 32 are functional block diagrams in the motor and - generator modes respectively, Figures 33 and 34 are diagrams showing the gate leading-edge de-tection means;
Figures 35 and 36 are diagrams illustrating the counter circuits;
Pigures 37 and 38 are circuit diagrams of the delay control means;
Figures 39 and 40 are diagrams of the circuits for converting de-tection into control;
: Figure 41 is the sequencer circuit di.agram; . .
.Figure 42 is the sequencer switching circuit diagram;
Figure 43 is the circuit diagram for the arithmetic unit;
Figure 44 is the diagram for the ~TH computing circuit; :.
.~~ Figure 45 is ~the diagram for the Fo computing circuit;
~ Figure 46 is the diagram for the a~c computing circuit;
., Figure 47 is the circuit diagram for computing changeovers in ~l operating mode;
:: Figure 48 is the diagram for the ~1 computing circuit;
Figure 49 is the diagram for the ~2 computing circuit;

, : , :~ :

Figu~es 50 and 51 are graphs portraying operation in the motormode; and Figures 52 through 55 are graphs depic~ing operation in the gene-rator mode.
Starting with general considerations, the theory of operation of the machines must first take account of a certain number of parameters.
Thus, in the event of application of the system for storing energy on a satellite of the geostationary type, it should be noted that the eclipse time varies in the course of a year.
Further, the rotation speeds of the machines at the end of the charging period must remain within certain iimits since their maximum speed is limited by their mechanical characteristics and the minimum speed by the energy recovered. But at the same time the steering function in- ;~ -troduces corrections into the system which could cause the rotation speeds of the machines to drift towards limit values.
The different possible charging and discharging cycles of such a system are thus capable of varying constantly, and the slaving function ~ ~ ;
must permit of generating the data concerning theoretical velocity, posi-. tion of control means and limitation against~overcurrents in the energy -;
storage and energy restitution cycies.
Referrine now to figures 2 and 15, which are block diagrams of - the energy storage and restitution system with steering capability accord-. ing to this invention, during the energy storage mode and during the energy restitution mode, respectively, it will be seen that the-machines are iden-,~1 - . , tical and that they function as motors MI,~M2 in the former mode or as generators Gl, G2 in the latter case, and that they rotate in opposite di-~i~ rections about a common axis A.

For greater clarity~ there is represented on figure 2 in solid ~ lines those components which are used in the energy storage function and `~ 30 in broken lines those components which are not used during that phase, .`~ : .

, :

.

::
'.~.'. ~
, and on ~igure 15 in solid lines those components wllich are used in the energy resti~ution ~ullction and in bro~en lines those components which are not used during that phase.
In the cner~y sLora~e mode according to figure 2, each machine ~1 (or M2) receives electrical energy from solar generators 1 through power switches 2 and 3 which, at determinate moments in time, switch the source to the stator windings of the machines, Velocity monitoring systems 4 and 5 control the power switches, and such control is based on a synthesis of the data relating to posi-tion, rotor velocity, and commands from a velocity controlling generator6.
Generator 6 issues commands which impose functional angular velo-cities on the motors, which commands are obtained after processing of the data relating to the misalignment signal Sdp, the rotor velocity and the law of rise in velocity. In the energy restitution mode according to figure 15, each generator Gl (or G2) supplies energy to a load 7 consist-ing of the onboard equipment of the satellite. The power switches 2 and 3 are the same as those used in the energy storage function and are effec-tive in switching the energy source 1 to the windings at any desired mo-~` 20 ment in time.
, .
Output monitoring systems 8 and 9 effect a synthesis of the data i:
relating to position, rotor velocity, and commands from the output control generator~ and they additionally control the power switches.
The output control generator 10 delivers the commands that determine the output from each generator. Such commands are obtained subsequent to processing of the data concerning the misalignment signal Sdp, the voltage across the load terminals, the rotor velocity, and the law for energy re-stitution at constant torque. It should be noted that the voltage across the load terminals is maintained constant regardless of the variations in output imposed by the load.

' ` ' .
, .

~: . .. .
. . . ~

2~

The energy storage mod~ requires th~t the form~lla giving the motor control law be demonstrated.
Referring to figure 1, it can be demonstrated on the basis of the fundamen~al laws of electrical engineerlng, viz.
electromaglletic force F = B x I x Q

. ~ .
newton tesla amp. m and electromotive force E' = B x ~ x v a U X R x I
volts tesla m mlsec V ~ amp.
that, in the motor mode, the engendered velocity v of motion of a conduc-tor through the magnetic field, or velocity v' of motion of the field magnet furnishing that magnetic field, which is equal and directly opposed ~ -to it, is higher when the induction B becomes smaller, by taking ~' to be ;~ - equal to U = Const. by reason of the fact that the product R ~ I can be neglected in this case.
When the instant at which the circuit is opened for a duration T, is advanced before passage of the conductor beneath the field magnet, the .: , flux intercepted by the conductor is thus weaker than that intercepted directly beneath the field magnet, by reason of the spacing of the field magnets and the reversed polarities of the magnetic fields.
Therefore the mean velocity v' of the field~magnet increases if the switching is advanced, while the circuit breaking time T remains un-changed. ' - By suitably adjusting ~he instant of switching of the conductors experiencing passage of the field magnets, it is thus possible to increase and to monitor the speed of such passage provided that the opposing torque is very low, which is in fact the situation in question, as will be ex- -plained here~nafter (presence of magnetic bearings).
~ .` , .
- Referring now to figure 3, wXich schematically illustrates the construc~ion of a motor-generator according to the invention, having si~

- 30 per~anent-magnet type rotating field magnets and four armature-forming : ~
~ - 14 -.

. ;. . , ~ .. . .

fixcd turns of magneti~ rit:ch P, i.t will be seen that if a c~lrrent I is caused to Llow througll ~ile turn al, starting in the position shown in the figure, the l.a~lac:e forces l)roduced in ~he strands will cause field magnets bl and cl to revolve in the direction F provided that they are capable oE revolving together about the axis.
If the distribution of the induction is considered to be trape-zoidal, as shown in figure 4, a n d the current is substantially con-stant, and if the start of yassage of the current is advanced as shown in detail in figures 5, fi and 7, then the force may be expressed as fol-lows in the augmenting zone:f = (~2~o / I)t ~ fO

f = f (1 _ n) ~ n~f ~ 2f ( n) -~ f mean o N2 2N~ o o N o~
= f (1 - ~?
where T is the nominal time of passage of.the currènt, N the number of possible positions during the time T, and n the number of positions cho-sen for the control means~ the lead taken over the nominal moment in time of opening of the switches being consecutively N T.
- It is then possible to demonstrate the torque from the expression~
~ = FR = BIQNbR - K I
, o o where N is the number of conductors and R the radius of the rotor.

However, in the case under consideration, = FR = BIQN [1 ~ (N) R} = KI
where K Ko(l (N) ) (1) ~ The expression for the angular velocity can still be established for, if RI is small, it can be considered that:
. , .
~ ~ UI

KI~ = UI .
U

K
but ~O K
o ~ :
30 w~e~ce ~ = on 2 (2) . 1 ~ (N) ~:
, , . .

., , .... , . ., .. , .. , ...... , . .~ .. , . ... . . , . ~ ... . ... .. .
, , . , , .:
. .

It should l~c noted that ~ and ~ v~ry as ~ function ~ IN~ as shown by the c~1rves in figures 8 alld 9.
It will bc noted a]so on these t~o figures that, when nN tends to-wards unity, the torque tends towards zero, whereas the velocity tends towards infinity.
This in turn leads Lo making certain observations concerning the' limitations of the previously expounded laws. For these laws were estab-lished on the assumption that the current rcmained constant during a switching 'window', which is not entirely true. Further~ the current in-creases with ~, and consequently the voltage drop RI is no longer negli-gible compared to U when ~ becomes very large. Therefore the variation in current absorbed as a function of c~lgular velocity can no longer be neglected.
When the. velocity is stabilized, the-driving torque is equal to the opposing torque.

~r = ~m = KI
Given a constant opposing torque, then KI must be equal to Ko Io.
; But hence I = - I
~0 This simplified calculation is intended to show that the solut;on provided by the aforementioned expression (2) can indeed be considered only in cases where the opposing torque is very low~ which is in fact the ' situation encountered with the present inventionj which utilizes magnetic - bearings of the kind described in patent application No. 74 00l90 filed by the Applicant on 3 January 1974.
With regard to t'he instantaneous counter-electromotive force, this ~ A!
is known to evolve in the same way as the instantaneous force. Now, the electromotive force induccd in a conductor is given by the formula:
e = B~v where v ~ R~

~ .. ., .. , ., .. , ,, .. ~ ", . , . , ~
~ . . . . .
. . .. .

~s shoull ir. fig~llc; 11~ 12 and 13, t:he total countcr-electromot;.ve fOl'Ce .fl of tlle motor, sllown in solid li.nes ia~ ~i.g~lre 11 (tl~e coullter-elec-tromotive force f2 in ~ coi.l being sllown i.n broken lines thereon), conse~
qucntly varies greatly tllrougl~ time when the control is greatly advanced.
The current, which is given by the formul~
i e - U
-. would conseque.ntly assume very large val.ues when the counter-e].ectromotive force is negative, nL~kin~ it impossible for the motor to function satls-factorily.
In order to overcome this drawback, use is made of the self-induc-tion of each w.inding and an external inductance coil-is provided to smooth the current and thereby obtai.n a low ripple ratio.
Under these conditions, the counter-electrol~otive force experienced by the generator feeding the motor is for all practica]. purposes a DC vol-tage as in the case of a conventional motor.
Determining this inductor by using the first--overtone method en--~ ables the maximum ripple ratio of the counter-electromotive force to be obtained.
After expanded calculations (not perormed here), the following ex-pression is arrived at:

~ o IMl ~3) where [cos 27r (1 ~ N~ 2~ ~ sin N )2 K = BQN R
.- o and ~ = number of switchings per revolution.

Operation of the motor can be viewed in succession through the medium of a lattice oE curves stemming from the relation: :

KI = K~-RE J
j ~ Ku K l~
:, ~ R -R ~ .
according to whi.ch the torque is given as a functLon of angular velocity, tak.ing K as a parameter.

~ ' .
, . .

As shown in figur~ 10, relation (2) above gives the maximllm velo-city atta;nable by t~le Dlo~or, this velocity being obtained when the oppos-ing torque ~r is zcro.
The true angular velocity of the motor is defined by the point of intersection of the curve of o~posing torque ~ith the straight line given by relation (4) above.
~ e curves in figure 10 are thus in fact borne out by the previously made demonstration. ,, - Response of the momentum wheel to a motor drive input can readily 10 be ;nterpreted from the lattice of curves in figure 10. , For if ~ = ~2 - ~1~ and if ~1 remains unchangad, then by using - relation (4) above one obtains Kl IT/r - ~r r After expansion one obtains the following final relation:
a~ = r l ~ K ,~ ,~
`' The velocity ~l will tend towards ~2 along the arrowed path in figure lO. ' The energy restitution phase similarly requires that the formula for the law of control of the generator be demo,nstrated.
By referring to the diagram in figure 14j~it can be shown from the basic laws of electrical engineering, vi~

~lectromotive force Q Q B Q v volt tesla m m/sec that in the generator function the electromotive force supplied to the load R is proportional to the length of the conductor intercepting the ~, mQgnetic induction flux B and to the velocity v with which the conduc- ;
~ tor moves through the magnetic field or to the equal and opposite velo-,' city v~ with which moves the field magnet furnishing that magnetic field.
-The law governing such intercepted~ flux likewise teaches that the value of the electromotive force induced during the time T is given by '~
E = -t - 18 ~
. . .

where ~ re~res~nts the Llu~ ~hit:~l iS swept or i~ltcrcep~ed by the COII-ductor for a time t.
It is consequently possiblc to modula~e ~he electromotive force, and hence tlle intensity, by adjusting the duratiorl of opening of the switches.
` Such method offers the dual advantages of improving the efficl-ency of ~he installation by disyensing with a DC/DC converter and of - simplifying the electronic circuitry by using the same power switches as those used for the motor function.
The instantaneous electromotive force produced is of a trape-zoidal pattern by reason of the alternating distribution of the field - magnets, which are in fact the same as those used in the motor function, as shown in figures 3 to 7. However, its maximum value may greatly ex-ceed the power supply voltage to an extent of as much as E max :, 11 ' :
; such increase in the electromoti~e force resulting from the big inc~ease in angular velocity achieved during operation in the mot~r mode~ whereby it may prove imposslble to suddenly switch a high voltage to a load re-~
quiring a lower voltage, without engendering significant losses.
Here again the inductance of the generator and an external inductor will play a regulating role.
It should be noted that the electrical si~e of such coil is virtu- ;
aliy the same as that required for operation in the motor mode, as emerges from calculations which it is not proposed to demonstrate herein.
Further, in order not to cause an overvoltage and losses upon open-ing of the switch, the same is latched once the current has cancelled it-self naturally.
Figure 16 shows the arrangement for four energy tapping gates Pl, P2, P3 and P4 and their respective maximum durations D , with respect to ~ 3~ instantaneous electromotive forces el, e2, e3~and e~ present on the machine.~ ..
'- 19 -.
,~ , :: ~
., ,~ " . .............. . ..... ~....... . ..... ............ .. .. . . .

coils. ~t is to be noted a]~ that the ~witcll is latched when the electro motive fox-ce is ~ero, ~hereby ensuring that the current in the driven coil is zero at that moment.
The overlapping of the energy tapping gat~s, corr~sponding to the hatched portion of figure 16 is not a disadvantage since the switches never open fully and since they close before the end of the gate, the effect o which is to prevent simultaneou9 control inputs to the switches.
Figures 17 and 18 show the maximum travel Cm for positioning of the control means, and the electromotive force obtained before filtering.
Figure 17 provides an understanding of the mamler of calculating the mean electromotive force (E ) which, taking a first case in which Y mean N ~n E ~ 12 N2 E~ = E ~ ~ n~ d case wherein n ~ N/2, is given by Emean = E N ~2 N) When velocity is maximum, i.e. at the start of operation in the generator mode, the control input to be applied always gives n ~ 0, mak~ ~ ;
ing it possible to operate in the area specified in figure 17.
Since the electromotive force is given by the relation E = E -o whereEo - the no-load electromotive force of the generator rotating at an angu-lar velocity ~O
nominal angular velocity o the generator = sctual angular velocity of the generator, it is therefore possible to determine the two fundamental laws for opera-~ .
tion in the generator mode:
::~
~ n ~ N/2~ Emeall = Eo ~ ~ N~ (6) ; n ~ N/2, Emean ~ Eo Oo N (2 N~

Knowing that K - ~Q it is then possible to express the generator~s 30 coeffi.cient K from relations (6) and (7):

- 20 ~

~7 ;~ n _ N/2, K - Ko ~ + N (8) n _ N/2, K Ko N 2 ~ N (9) In order to attenua~e the current fluct:uations resul~ing from the non-DC electromotive force, it is necessary as already stated to connect in series an inductor of virtually the same electrical siæe as the one used for the motor func~ion.
For operation in the generator mode, it must be recalled that the electronics must perform a dual role consisting, firstly, of maintaining the output voltage constant irrespective of changes in angular velocity and, secondly, o~ obtaining identical torques produced by the two machines.
To this end, variation of the position of the machine control means is effected at regular time intervals, the value of each interval being:
~T =

where T = total discharge time ' ":
M = number of control steps.
The time intervals being regular, calculations-(not made here) in~
dicate that control-must be modified for constant energy differences.
The rela~ionship between current and velocity must take account of two imperatives, to;w;t, maintaining the output voltage constant irre-spective of velocity modifications and obtaining identical torques on thetwo momentum wheels.
These requirements then impose two basic relations~
il + i2 ~ I ~10) Kl il = K2 i2 ' (11) where il, 1~ are the currents delivered by the two momentum wheels, I the current drawn by the load Kl, K2 are coefficients characteristic of the momentum wheels and given by relations (8) and t9) above.
Expanded calculations (not given here3 prov;de a simplified pre sentation of the torque quantity given by the useful torque hyperbola Hcu . .

which is indica~cd on figul-e 19 and determined from ~e ~elalion ~ ~
- ~ , where P is the powe~ clelivcred and ~1 and ~2 the resl~ective veloci~ies of tllc wo m~chines.
Figure 19 also shows llow the torque applied to the momentum wheel evolves when a fresh valuc of K i9 applied, the demonstration oE wh;ch is similar to ~hat made for the motor, and it is to be noted that modi-fying the con~rol corresponds to a change of strai~lt line and that dur-ing the energy recovery cycle the device will sweep the set oP straight lines Kl, K2. . . O
The generator which delivers the signals for producing the switch-ing windows is an essential element of the channel, for it ls on the basis c~f the leading edge of these signals that positioning of the control in-~ put-means is established.
- This generator could be formed, for instance, by an oscillator having its inductor positioned near the rotor. Three metal pieces 13, 149 15 fast with the rotor, pass through the air gap of ~wo rings 11, 12 form-ing the inductor and modify their quality coefficient, as shown in figure 23.
The voltage produced by this oscillator, which oscillator is con-tained i-n each output control unit 8 and 9 and in each velocity control unit 4 and 5, is filtered and then applied to an error detector which converts the signal into a logic command, as shown in the cliagrams in `~ figures 20 and 21.
It should be noted that, because of the control means positioning accuracy required, the maximum permissible error in the triggering thresh-old must be less than 20 mV for machines rotating at about 700 rps.
However, the result must be weighted by the fact that the switch-ing windows Fl, F2, F3 and F4 are generated on the basis of two signals Sl and S2 outphased by one window (represented by one-~welfth of a revolu-tion.

~ 22 -, .

.

:, .

Conseq~lently, as showll i.n figure 22, any variation, in one sense, in the detection insl:~n~s is compensated for by an opposite variation in the last two windows, which is tantamount ~o saying that these differ-ences musttheoretically cancel out in one re.volution of the rotor, and : that in reali.ty the accuracy required in the triggering thresholds could be less stringe~lt.
Reference to figures 23 through 26 and to figure 3 shows the angu-lar disposition of the rings and the manner in which the switching sig-nals are processed.
In the case of a machine with six magnets, the metal parts 13, 14, 15 passing through the air gap in rings 11 and 12 are arranged to form three 60-degree sectors, the said rings being spaced 30 apart.
Referring to figure 24, it will be seen that under such condition.s :
the electromotive force induced in a coil during a complete revolution of ~
the rotor will be of trape~oidal shape, in accordance with the induction ~ -laws expounded precedin~ly.
At the same time, the rings 11 and 12 have generated signals of the shape shown in figure 25. When these signals are decoded they assume ~
- the shape shown in figure 26 and are then usable directly for controlling ~ :
the switching commands applied to each coil.
B fore proceeding any further with the description, it is proposed.
to briefly recall the basic features of the subject method of this lnven-tion for steering a spacecraft and regulating its onboard voltage by means of LWo motor/generator dynamo-electrical machines contrarotating about : ;
the same axis. These basic features are as follows:

. - In the energy storage function, each machine, which then functions .~ as a motor, receives electrical energy from solar generators through the - medium of power switches which, at specific instants in time, switch the -~ power source onto the stator windings of the machines.

` 30 Further, velocity monitoring systems have as their function to COII-''` :

~ ' :
.. : ... . ..... .. .... . ... ... . ... ..

trol the powe~ switches~ such control being effected Ervm synthesizcd data relat;ng to position, ro~or angular vclocity and commands fram a velocity con~rol input generator.
- In the energy recovery function, eacll machine, which then f~mc-tion.s as a generator, supplies energy to a load formed by the satelllte~s onboard equipment, the power switches being the same as those used in the energy storage f~mction and performing the function of switching the ener-~ gy source onto the windings at each required nwment in time.
; Further, output monitoring systems are effective in synthesizing data relating to position, rotor angular velocity and con~ands from the output control generator, and furthermore control the power switches.
It is now proposed to describe the electronics required for opera-tion of said machines, which e].ectronics is separated, in accordance with a teaching of this invention, into control electronics and processing electronics, itself comprising a control means positioning unit and a com--;~ puting unit.
In designating the various parameters used in the subject regulat~
. ing method of this invention, reference will be made more particularly, -~ in what follows, to the notation,system referred to in the preamble.
T h e. graph ABCDEF in figure 27, in which angular velocity ~ is represented along the y-axis and time t along the x-axis, schematically portrays the different possible forms of charging during motor phases AB, ~ `
CD, EF, and of discharging during generator phases BC, DE, feasible on - the basis of given starting conditions such as velocities ~MIN and ~MAX
or charging times tl, t2, t3.
This highlights the need to generate the theoretical veloci.ty datum ~ ~T~I--which varies between a minimum velocity ~MIN limited by the recovered .~1 energy, and a maximum velocity ~AX l;.mited by the mechanical character-; istics of the machlnes--and the fact that it is additionally necessary to take account of the steering unction which may introduce corrections ~ .

.' -~ -' . .. ~ . . - . . ., ........ ~ . ~ ... . ...... . . .. . . ........ . . . . ... . . . . .

into t~ sys~enl that could cause ~hc maclli.ne vel.oci.L;.e3 to drif~ towards these li.mit values.
Re~crence is nex~ Imd to fi.gure 28 for a block diagram of ~he over-all arrangemcnt of ~le vario~ls components for pcrfor~ling ~hc functions needed to carl~y the invention into prac~ice.
In accordance with th:is invention, the overall installation is shown in thi.s block diagram as being divided into ~wo main parts des:igna-ted by reference numerals 100 for the control e].ectronics and 200 for the pr-ocessing electronics, respecti.vely, the latter part being subdivided into a machine momentum-wheel positioning unit 201 and a computi.ng unit 230.
More specifically (see figures 28 and 29), the control electronics portion 100 basically includes power switches generally designated 2 and

3, interconnection switches 101, 102, 103 and 104, two control voltage generators 105 and 106, and diodes 107, 108, 109, 110, 111, 112, 113 and 114, the electrical connections being made on the one hand to a solar generator 1 and the satellite's power supply 7 and, on the other hand, to the general current distribution busbar 16.
. Further, the stators of machines 120 and 130 are energized as - 20 shown in respect -of their windings Ll, L2,.L3, L4 and L'l, L~2, L'3, L'4, and two inductors L' smooth the current in order to obtain a low ripple ratio, as described in the main patent. :
: Control means positioning system 201 basically includes two sub-systems respectively designated 202 for machine MlGl and 212 for machine M2G2 and common mode-changeover~ switching and sequencing functional units 204, 205 and 206 respectively. Subsystem 202 of machine 120 com-prises functional unlts for delay control 207, for start-of-gate detec-tion 20~ and for converting detection into control commands 209, and a counter 210 and read-only memories 211. Simi.larly subsystem 212 of 30 machine 130 comprises functional units for delay control 217, for start-~ 25 -~. . , ' , ' , ' , , ' . .
~:. , :

of-gate cletection 2l8 and for converting detection into control comman(Ls 21~, as wall as a counter ~20 and read-only ~nemories 221.
ComputiIIg system 230 includes an arithmet;c unit 231 and functional units 232 Eor computi~ T~I, 233 for processing the frequency Fo, 234 for computing ~C,235 for computing ~1, and 236 for computing Q~2.
~ s explained prececlingly, in the matter of de~ecting the machine rotor/stator position, signals Sl-S2 are given by pick-ups in logic form, and these pick-ups are contained in units designated 121 and 131 fast with the machines forming the systems 120 and 130. The signals Sl-S2 are conveyed via the picl~-up channels A and B to delay control means 207, 217 as sho~ in figure 28. The control electronics must in any event include generation means of the analog circuits mainly comprising the power switches and its associated circuitry. As shown in figure 30, the power switches-- `
which are contained in units designated 2 and 3 in figure 29, and themselves respectively designated 2a, 2b, 2c, 2d and 3a, 3b, 3c, 3d--include four cascade-connected high-voltage transistors Tl, T2, T3, T4, and the circuit is devised so that the whole of the current passing through each switch should also pass through each coil.
Diodes Dl, D2 eliminate all return voltage on the switch term nals.
Further diodes D3, D4 connected around the control transistor T5 prevent application of a return voltage to the base-emitter junction of said tran-sistor.
; ~s shown in figure 30, the manner of connection to the power supply bus 16 is such that each switch is controlled off the voltage generators - 105-106 that either generate 60V from a source c for application during l operation in the generator mode, or 5V from a source b for application , , durlng operation in the motor mode, or from a source a for speeding up the machines during the acquisition phase.

Furtherrnore~ .irrespective of whether the machines are in the charg-ing (motor) mode or in the discharging (generator) mode, it is necessary . , .
.

~ ~3 ~

to avoicl ally anom.lly linbJc ~o cause a ~oo high current output through tlle machi~es. A curlen~ linliting CilCUit is accordingly provided that operates on the san-e l~rinciple as the machine, that is, in either a motor or a generator conEiguration.
~ voltage proportional to the current flowing through each machine is accordingly tapped off the terminals of resistors Rl-R2 and applied to a trigger circuit included in each generator 105 and 106. Thus when the current exceeds the maximum value set by a reference voltage, the logic output state of the trigger circuit disables the switches controlling the coils of each machirle, for a precletermined time period.
If the anomaly persists, disabling of the switches wi]l occur cycli-cally.
As shown in figure 29, the machines are connected into a bridge - circuit, enabling the same power switch to be used both in the motor and the generator mode.
The machines are thus so connected eleotrically that switches 101, ., -~ 102, 103, 104 should be conclucting while the machines operate in the motor ~ .
- mode~ and diodes 111, 112, 113, 114 should be conducting while the machines operate in the generator mode.
Further, diodes 107 and lOCJ isolate solar generator 1 during opera-tion in the generator mode, and diodes 108 and 110 isolate one machine .-from the other if either of them experiences a failure.
The processing electronics performs all the operations shown on the block diagrams in figure 31 relating to the motor mode and in figure 32 relating to the generator mode.
The processing electronics is split up into ten elemental parts functionally designated in figure 28 as follows:
Ref. numeral 202- Control means positioning for machlne 120 Ref. numeral 212: Control means positioning for machine 130 Ref. numeral 206: Sequencer ~

. . . : . .
.
. . ; , , 2~

~ef. numeral 23~: ~rithmctic unit Ref. Ilumeral 234: Comrutation of ~t~( Ref. numera] 235: Computation of ~ol Ref. nul~eral 236: Computation o~ ~2 Ref. numeral 204: Change oE operating mode Ref. numeral 232: Computation of ~'l~
Re. numeral 233: Processing oE frequency Fo.
; Each control meall~ positioning system includes:
- for machine 120, a subsystem 202 comprising:
. A delay control unit 207 . A start-of-gate detection unit 208 ; . A detection/control conversion unit 209 . A counter 210 . A read-only m~mory 211 - and for machine 130, a subsystem 212 comprising: ;
. A delay control unit 217 . A start-of-gate detection unit 218 . A detection/control conversion Ullit 219 . A counter 220 . A read-only memory 221 The system is completed by a sequencer switching means 205.
The elements listed hereinbefore perm;t processing by means of predetermined laws based on studies of the dynamic behaviour of the machines. These studies have enabled correspondence tables to be estab-lished between various physical quantities that must be present during the different processing phases~ and these tables are stored in read-only memori.es 211 and 221.
The data to be stored are obtained from calculations performed on a computer.
Such a method does not call for highly sophisticated processing, ~ 28 -.~ , .

:' ' ' ''' :' ' '''''~' ` '' ' '' ''' ' ' ;'' ' making it possible to usc a wir~d Jogic for i.ts im~lelllelltatioll while minimizing the nulllber of c;rcuits used~ Only the vol-ulle of the data to be stored in the read-only memories remains large, with more than twelve K-words of twe]ve bits eacll in this instance.
Regarding tlle perorlDance charactelistics of such a system, it is to be noted that it behaves perfectly in the region oL the initial opera1-ing conditions but that performance deteriorates somewhat as one departs from these nominal conditions.
However, any other convenient self-adaptive type of processing could be resorted to, but would involve extensive calculations and would require to be performed with a microprocessor.
In the energy storage cycle, the 'theore~ical angular velocity ~TH' datum is generated on the basis of a set of counters Cl, C2, C3, as shown in the block diagram in figure 5. The clock f~mction is provided by a multiplier circuit which modulates the frequency Fo as a function of two 'charging velocity' data, one of which is provided by the satellite and the other from read-only memory MM4bM.
This read-only memory and read-only memory MM4aM receive upon their -~
address inputs the start-of-charging velocity stored in a register. A
further function of memory MM4aM is to ini~ialiæe the counters.
As already recalled on figure 27, modulation of the clock frequency and initialiæation of the counters permit of performing charging cycles with totally different initial conditions.
; Thus memory MM3aM delivers the theoretical velocity ~TH datum on the basis of the time variable. `~
With regard to generation of the 'control means position' datum, the i ~elocity correction ~c commanaed by the steering function is memoriæed ,, ;:
by an integrating circuit j and is applied to two summing units ~ which receive the theoretical velocity datum ~TH on their second inputs~.
A multiplier circuit X reduces the information produced by the .

: ; .
,' ' .
' ''.'': '' '~ ' ' . , .~: . : ' SU11Ul-i~lg ~IUi t to a t~n~e-moclulated dat~ml, all(l this i.6 pel'tOrmed 011 tlle basis of a fl-equellcy F{ sy~-chrollizecl with the angular reloci,tics of the mach;,nes.
Presetting of this estimated velocity datum obtained thus on the address inputs of read-only memories ~2aM and ~2bM allows obtaining the required control-means posi.tion as an output. In the event that the cur-rent absorbed by a machine should exceed a preset maY.imum value, protec-tion circuits would react im~lediately.
As shown in the ~lock diagram in figure 32~ in the energy restitu-~0 tion cycle the velocity datum ~TH is likewise generated from a set of counters Cl, C2, C3, but here an additional constraint arises because, in addition to equalization of the torques, it is necessary to match the machine outputs to load variations.
- In this case the clock circuit is a multiplier circuit which modu-.
lates its output frequency as a function of a tdischarging velocity' datum.
This 'discharging velocity' datum is siaved to the output voltage ~' ' and a double-threshold type of comparison is made between the voltage delivered and a reference voltage. The resulting error signsl is sampled, then added to the number 'n' characterizing the pre-correction discharge.
,` 20 Adjusting the sampling frequency thus permits of optimizing the ; slaving system's-response.
~he correction process does not permit of rapidly-compensating for , a change in the position of the control means following a variation in the load. Modificatioll of the control means position is effected with a correction to the contents of the time counter.
The operating principle is similar to that of the clock frequency ;, monitoring function9 except that the comparison thresholds are different.
Since it is-impossible to know exactly the end-of~charging velo-city, it is~necessary to initialize the counters so as not to introduce ~' 30 perturbations in the satellitc at the start of the discharging cycle.

" , .~ , Tllc role o~ rcad-only mc~l~ory ~ AG is therefore ~o ;nitiali~e tlle COUllt~' ers; tllus it receives on i~s acldrec;s input the star~-of-discharging ve]o-city and this datum is stored in a registcr after each comput-ing cycle.
Gencration of the 'control means position' datum is similar to gene-ration of the dat~lm in ~he energy storage cycle. The velocity correction commanded by the steering function is memorized by an integra~ing circuit and the datum is applied half each to the two machines in order to allow for the fact that the total current furnished by the two generators must remain constant.
~ The fact of applying the control input for varying the velocity dif-ferential half each to the two machines causes an increase in the velocity of the one and a decrease in that of the other in order to obtain the re-quired moment.
:
Further, because variations in the currents are of opposite sign ; on the two machines, the power delivered to the load is maintained con-~ stant.
- Limitation against overcurrents is accomplished by the same circuits as in the motor mode, and only the preset maximum current values are dif-ferent.
As will readily be understood from the-block diagrams in figures 31 and 32, the same are absolutely identical both in the energy stbrage and in the energy restitution cycles. This was a design goal so as to -~ ~ minimize the number of circuits required for performing the functions.
Reference to figure 28 shows that each control means positioning functional system, to wit system 202 for machine 1 and system 212 for machine 2, comprises a number of component parts for the processing elec-,, ! ' tronics. ;~
For greater clarity in what follows, reference will be had, in de-scribing each of said component parts, to whichever of figures 33 through 49 portrays its functional diagram.
;~; .
~ ~ 31 -,. .

,~ .
.
.: . .
, . . . .

Furtllc!r, it is to be noted thlL, ilL ordcr to simplify a rca~ing of these diaglams, use is m~de thereon of the fo'llowi.ng'notation or desig~
nating tlle conventional elect:ronic components shown thereon:
C -- Counter A - Adder P = N~ND or NOR gate MM - Read-only'memory I = Changeover switch ~ = Reg:ister OR = OR gate D ~ Div:i.der : B = Trigger Mu - Multiplier = Monostable circuit More specifically:
Detection of the gate leading edge by component 208 (figure 33) for machine 1 and 218 (figure 34) for machine 2, produces a pulse which writes into counters 210 (figure 35) and 220 (figure 36) the delay value to be applied to each control input.
In switching to the zero state, each counter delivers a pulse .~. which controls the delay system included in the delay control components 207 (figure 37) and 217 (fi.gure 38).
The principle whereby the detection is converted into a control input to be applied to each machine will now be described with reference 20 'co the diagrams in figures 39 and 40.
The signals ~l and B' from delay analysers 207 and 217 pass through the circuits of components 209 and 219 and are received at 2a, 2b? 2c, 2d . :
.
' and 3a~ 3b, 3c, 3d Otl the base of transistors TS (see figure 4) controlling high voltage transistors Tl through T4, referred to precedingly, in power switches 2 and 3.
The components can be disabled by a-.circuit for inhlbiting command inputs from component 204..
~ The:read-only memories of elements 211 and 221 are of a type well known per se and are connected between the registers con~ained in the arith~etic unit and counters 210-220.

- 32 ~ :

, .

.'.' ~ .
,: , .. : :

Sequellcer ~0~; and SWitCI-.illg urlit 205 mllS~ perform the followin~
ca]c~lla~iorls in ~espce~: c.L eael~ achine:
- Read ~C in series -- Xnte~rate ~ C and compute ~ o2 - ~ead ~TH in the re~isters of the arithmetic unit - Detect generator/motor mode changeovers oL each machine.
The sequencer proper accordingly includes a counter with a control frequellcy of 750 kHz and read-only memories controlled by that counter and delivering the necessary commands.
Activation o~ the sequencer is controlled by the start-of-gate pulses. ' Aritlm~etic unit 231 (figure 43) basically includes an adder, a trigger, a double multiplexer, two registers s, a register r and a divider for dividing by N in order to provide beat between two consecutiye posi-tions. Such a dividing circuit is rendered necessary by the fact that, ~ for example, out of 4096 possible positions of the çontrol means ~12 bits) ; only 256 (8 bits) are used, whereas the precision required in respect of ; the perturbing moments produced by shifts in the control means requires use of the 4096 positions.
If a beat is set up between two consecutive positions of the 256 usable positions, it is possible to create fic~itious positions.
If the space be~ween two consecutive usable positions is divided into sixteen parts, there will then be 4096 positions obtained, 256 of them real and 3840 fictitious.
For instance~ if the two consecutive positions are 123 and 124 and if position 123 is com~anded every other time and position 124 every other time, chis is equivalent to a position in the middle of the space included between positions 123 and 124.
Register s determines the 2S6 usable positions with an accuracy of twelve bits.

. .
.:-~ rz ~

l~egisl-er r determilles t11e bcat fullctioll, toLal capscity of which is 64 ~iCtitiQU3 pOSitiOllS WiLh a precision o~ 4 bits. A~C is calcula~ed by co1nponent 234 (fi~ure 20).
~ C relresents the difference which tlle satellite requires in the angular velocities of the machines in order lo generate a moment.
~ C will be written into the register in the series mode respons-ively to a commancl rom the sequencer.
At the start of each sequence, A~C is recorded with its small weight first and sign last.
The mamler of calculation of ~l, determined by component 235 and schematically depicted in figure 4~, and of ~2, deterMined by cornponellt 236 and schematically depicted in figure 49, will be better understood by reference to the graphs in figures 50 through 55.
It should be remembered once more that:
- ~TH is the mean theoretical velocity of the two machines - ~l is the difference between ~T~ and ~Tl ~2 is the difference between ~T~ and ~T2 - ~Tl is the theoretical angular velocity of machine l - ~T2 is the theoretical angular velocity of machine 2.
In the motor mode~ a moment can be engendered only by speeding up one of the machines.
If the moment to be engendered is cloclcwise (SH), it will be ob-tained (f:igure 50) by accelerating Rl (machine 1) in such manner that ~l increases and ~2 ~ ~2 If the moment to be engendered is counter-clockwise (SAH), it will be obtained (figure 51) by accelerating R2 (machine 2) in such manner that ~2 increases and ~2 - ~2 - ~C
' , In tile generator mode, ~l~C is applied simultaneo-lsly to th~ two machines buL is halved in cach case.
As shown in figures 52 and 53, if the moment to be engendered i9 clock~ise, it will be ob~ained by increasing the G2 output in such malmer that ~2 decreases and ~2 = ~2 ~ ~2C
and by decreasing the Gl output in such manner that ~1 increases and ~1 = A~ 2 If the mornent to be engendered is counter-clockwise (figures 54 and 55), such monlent will be obtained by increasing the Gl output in sucll man-ner that ~1 decreases and ~;
1 ~ 2-and by decreasing the G2 output ln such manner that ~2 increases and A~2 ~ ~2 2 ~;~
It should be noted that a single graph could be used for figures 52-53 and 54-55 since application is made simultaneously.
~TH is computed by component 232, as diagrammatically depicted in figure 44.
The system consists of a twelve-stage reversible co~mter and four 1024 xil2-bit read only memories.
The frequency Fo is processed by component 233, as sho~m in figure 45.
In the motor mode9 the system is designed for achieving ~TH in a twenty-hour velocity build-up, and Fo is determined at ~ 178 k~z.
H~wever, in the case of certain periods of eclipss, the whole of the ~Th curve is not described. In this case one of the two groups of component 233 (see figure 45) indicates at what velocity the machines must operate, while the other group indicates the ~TH counting rate--'lineari~ed' for ~he available sunlight time.

~:,........ .

l.n t~Ie geI-ler.ltor mode ttIC sys~erlI is inteIlcIed for a ma~;m~Im utili~a-tion time of 72 mi~ te~i, with a poss;bîlity of delivcring Q tkrc~e times higher power at tin~es, and ]~o is accordi.ngly set at ~ 186 kH~.
III realjty ~o will be 187.5 k~I~, or a quarter of 750 kUz7 the pilot frequency Eor sequencer 206.
The element 204, schelIlatized in figure 47, allows changing Erom the : generator to the motor mode.
For it is not possible to switch from the generator to the motor mode or vice versa witIIout taking precautions, for itl order to generate a :~C moment for e~amp].e~ as shown in the graphs in Eigures 24 and 26, ~ then, as already stated:
A~Il ~ Q~ C in the motor mode Q~l = Q~ 2C in the generator mode The switch is consequently efected in several stages:
- Detecting the switch i.n order to inhibit control inputs - Modifying ~TH
- Modifying the place of the resh control inputs, i.e. two com- -.~ plete sequences - Eliminating inhibition of control inputs.
The foregoing description has revealed.a possible separation oE
: the electronics required for applying the control laws according l:o the - main patent.
Said electronics has accordingly been divided into two main sec-.~ tions:
- the control electroni.cs I - the processing electronics which in turn comprises:
.l . two positioning systems9 one for machine 1 and one for machine 2 .:,j .
. a computing system.
However, an alternative self-adapti.ve type of processing could be substituted for the read-only memories oE the control means positi.oning : -',~': ' ' ' '' - . ': . ' '-systllns and or Lhe conll~uLillg systcm.
On the other hll1~1~ f;UCh adaptcltiOll rC:~Illi.rC9 the USC o l IlliCrO-processor, which is not envi.saged as part o~ ~he p-resent invention.
It goes withouL say:ing t:hat changes aild substiLutions may be ~lade in the embodi~ients hereinbefore descri.bed without departing fro~n the scope o the invention as set forth in the appended c1aims.

` ~ ;.` ':

.:

.' . ' i~ ' ' ~ .
, . .

, .

Claims (7)

The embodiments of the invention in which an exclusive property or privilege is claimed are defined as follows:

1. A method of regulating the onboard voltage of a spacecraft and of steering said spacecraft by an inertial device having at least two contra-rotating dynamo-electric machines operating during periods of illumination as motors powered by solar generators through power switches, and during periods of occultation as generators connected to an onboard electric load through said power switches, said method comprising the steps of:
in the motor mode of the machines, applying switching control signals on the power switches to adjust the switching to take place in advance with respect to the nominal opening time of said switches in order to change the rotational speed of the machines;
in the generator mode of the machines, applying switching control signals on the power switches to adjust the switching duration of said switches to maintain the onboard voltage constant notwithstanding variations in speed;
said control signals applied so as to affect the machines in such manner that, without a steering correction, when the speeds of the machines are equal, said speeds follow a pre-established theoretical law of speed, when the speeds of the machines are different, said speeds vary while maintaining the speed difference imposed by the last steering correction produced by the steering device of the spacecraft, after said speed difference had previously been memorized, upon the appearance of a steering correction order,
1. Claim 1 continued....
   
   said speeds of the machines vary both as a function of the pre-established theoretical law of speed and of the instantaneous speed difference imposed by the steering device;
   and further including obtaining the instantaneous speed difference by inte-grating an error signal provided by the steering device;
   and controlling the pre-established theoretical law of speed by a clock having its frequency slaved to the onboard voltage when the machines operate in the generator mode.
2. 2. An inertial device for regulating the onboard voltage of a spacecraft and for steering said spacecraft comprising at least two contra-rotating dynamo-electric machines with alternating magnetic fields and having motor and generator modes of operation, solar generator means to power said machines, power switching means connected to said machines to connect said solar generator means to said machines during said motor mode of said machines, velocity control means also connected to said machines during said motor mode, said power switching means also connected to said machines to connect an onboard electrical load of the spacecraft to said machines during said generator mode of said machines, and control means for determining the time of opening of said power switching means.
   
   3. The inertial device as claimed in claim 2 wherein each of said dynamo-electric machines includes a rotor supported in magnetic bearings and having at least six rotating field magnets each formed by a permanent magnet with
3. Claim 3 continued...
   said magnets being spaced a magnetic half-pitch apart and said alternating magnetic fields being radial, and a stator having at least four coils in the form of turns set at a magnetic pitch and having at least one strand opposite said permanent magnets, said turns being spaced one-half magnetic pitch apart, with said strands of said turns intercepting said magnetic fields to produce upon switching of said coils, in said motor mode, a torque given by the relation ? = B I ?Nb R [1 - (?)2]
   
   where:
   B is the magnetic induction intercepting said strands, I is the mean current flowing in each of said turns, ? is the active length of a strand, N is the number of said strands, R is the positioning radius of said strands, ? is the ratio of the number of commanded advance positions to the total number of possible positions;
   and in said generator mode, an electromotive force E, [or a current I associated thereto,] determined by the relations:
   E mean = Eo ?o [? + ?] for n ? N/2 E mean = Eo ?o [2 + ?] for n ? N/2 where:
   ? is the ratio of the number of commanded opening positions to the total possible number of positions, .omega.o is the nominal angular velocity of the generator, .omega. is the true angular velocity of the generator, and Eo is the no-load electromotive force of the machine at the angular velocity .omega.o.
4. 4. The inertial device as claimed in claim 2 wherein said control means includes two sensors each comprising an oscillator means whose self-inductance is constituted by a ring fastened with a stator of each of said machines,metallic segments fastened with a rotor of each of said machines passing through an air gap of said rings to modify the quality factor thereof, the voltage produced being filtered and applied to a comparator which provides a logic control signal.
5. 5. The inertial device as claimed in claim 4 wherein each metallic segment has a length of one magnetic pitch, said segments being spaced from one another by one magnetic pitch, and said rings are spaced from each other by one magnetic half-pitch.
6. 6. The inertial device as claimed in claim 2 wherein said power switching means includes power switches each comprising firstly, high voltage transistors cascade-connected to pass the total current passing through each said switch also through each of the coils of said machine it controls, and, secondly, associated diodes which eliminate all reverse voltage at the terminals of said switch, each of said switches being controlled from voltage generators which generate a 60 V supply for application when said machines operate in said generator mode, and a 5 V supply applied during operation of said machines in said motor mode, and a supply of determinate voltage for speeding up said machines during an acquisition phase.
   
   7. The inertial device as claimed in claim 2 wherein interconnection switches, diodes and blocking diodes are connected to said solar generator means and to the onboard electrical load of the spacecraft,
7. Claim 7 continued....
   
   said interconnection switches being conducting during operation of said machines in said generator mode, and said block-ing diodes isolating said solar generator means during operation of said machines in said generator mode and isolating one of said machines from the other in the event of breakdown of one of said machines.