High speed electromagnet control circuit

Abstract

A solid state electrical switching circuit of high efficiency is employed to increase the speed with which an electromagnet operated device, such as a relay or a hydraulic valve, can be actuated without increasing the power required to maintain the device in the actuated state. The circuit can also be made to provide the momentary increased electrical current required to obtain a given mechanical force when the magnetic gap between the solenoid core and the movable pole piece is open. Both actuation and de-actuation speed of the electromagnet are increased and rapid deactuation is achieved either by a high reverse voltage applied to the solenoid and the return of its energy to the power source or by a diode and capacitor network which transfers the magnetic energy to a second solenoid which thereupon becomes energized.

Parent Case Text

CROSS-REFERENCE TO RELATED APPLICATION

This is a continuation-in-part of Application Ser. No. 308,268 filed Nov. 21, 1972, now abandoned.

Description

BACKGROUND OF THE INVENTION

1. Field of the Invention

The present invention relates to the art of increasing the speed, force and efficiency of devices which employ an electromagnet to produce a mechanical force or motion in response to electrical signals.

2. Description of Prior Art

An electromagnet type actuator normally employs an electrical winding or solenoid of copper wire over a core of ferrormagnetic material. The core consists of two parts, one over which the solenoid is wound, and the other movable, completing the magnetic path when actuated. The mechanical force between the fixed and movable cores is equal to ##EQU1## WHERE B is the magnetic induction in gauss and A is the cross-sectional area of the gap or pole face in square centimeters. For the ideal case, ignoring fringing magnetic fields, the magnetic induction B is equal to ##EQU2## WHERE N is the number of turns on the solenoid, I is the current in amperes, `w` is the width of the air gap in centimeters, 1.sub.i is the length of the magnetic path in the core (both parts), A is the cross-sectional area of the core (assumed constant for this explanation), in square centimeters, and .mu. is the magnetic permeability of the core. The above expression for the magnetic induction B demonstrates that the permeability .mu. should be as large as possible if large values of B are desired. In a typical electromagnet actuator the value of `w` or width of the air gap is relatively large compared to the term 1.sub.i /.mu., prior to actuation, and for this case ##EQU3## Thus, for a given current I in the winding, the magnetic induction B is inversely proportional to the gap width `w`, and therefore the mechanical force produced itself varies inversely as the square of the gap width. It is therefore very difficult to obtain a constant mechanical force over any appreciable distance of travel of the pole piece and the force is the smallest when the gap is fully open, contrary to what is usually desired.

A further difficulty associated with electromagnet actuators is that a substantial amount of magnetic energy must be delivered to the open gap before the desired force is obtained, this predelivered energy being equal to

F .times. (w) .times. 10.sup.-.sup.7 joules,

where F is the desired force in dynes. The instantaneous power required to achieve this force in a desired time of T seconds is ##EQU4## For example, to provide 15 pounds of force in a gap 0.1 inches wide in one millesecond requires about 175 watts of power. Yet, once the gap is closed (after actuation) the electrical power required to maintain a constant force of 15 pounds depends only on ohmic losses and may be a small fraction of a watt.

Many ingenious methods are available to ameliorate the above difficulties. The magnetic circuit may be provided with shaped pole faces to make the mechanical force more uniform, mechanical linkages may be employed to increase the force with the gap open, or the voltage applied to the electromagnet winding may be made high initially and reduced after actuation. A thermally variable resistance such as an incandescent lamp may be connected in series with the winding to produce a high inrush of current when the lamp is cold and its filament resistance is low and to limit the current when the lamp comes up to temperature and its resistance is high. If the electromagnet actuated device is a relay, a pair of relay contacts may be used to reduce the solenoid current after actuation.

SUMMARY OF THE INVENTION

The present invention employs switching transistors and diodes to apply a relatively high voltage to the winding of an electromagnet to increase the speed with which the current in the winding will rise to a desired value. Once the desired value of current has bee attained a further increase is prevented by negative current feedback, the current being maintained at near the desired value at very high efficiency by intermittent application of the same high voltage to the winding. To de-energize the electromagnet, the same high voltage is applied to the winding in the reverse direction to produce a cutoff speed equal to the turn-on speed, recovering, in the process, the energy stored in the electromagnet. The addition of a single capacitor to the circuit allows the initial current to rise momentarily to a much higher value than the desired steady-state holding current, this to provide a high mechanical force with the gap of the electromagnet open. When a pair of electromagnets are to alternately energized, energy is transferred from one to the other by a diode and capacitor network at arbitrarily high speed. Except for switching losses, which are low, and ohmic losses in the circuits, a theoretical efficiency of 100% is attainable, the power consumption being only that required for mechanical actuation.

BRIEF DESCRIPTION OF THE DRAWINGS

FIG. 1 is a complete schematic diagram of the high speed electromagnet control circuit.

FIG. 2 is a graph comparing the currents produced in the winding of an electromagnet for both high and low values of applied voltage with the current produced by the circuit of FIG. 1.

FIG. 3 is a graph of the surge current produceable by the circuit of FIG. 1.

FIG. 4 is a simplified equivalent circuit for the circuit of FIG. 1.

FIG. 5 is a schematic diagram employing two circuits such as shown in FIG. 1 to alternately energize two electromagnets.

FIG. 6 is a simplified equivalent circuit for the circuit of FIG. 5 for the interval of energy transfer.

DESCRIPTION OF THE PREFERRED EMBODIMENT

Referring to FIG. 1, the reference numeral 1 indicates an electromagnet having a moveable pole piece 3 which is used to actuate a mechanical device such as a relay or hydraulic valve. The voltage of the DC power source 6 is much higher than that required to provide the current necessary for the electromagnet when this current is limited only by the winding resistance. Transistor 10 acts as a switch responsive to an input control voltage to energize and de-energize the electromagnet 1. When transistor 10 is in the non-conducting state, the electromagnet 1 and the battery are connected together in series with diodes 2 and 4, both of which act to prevent the flow of current to the winding of electromagnet 1. When transistor 10 is caused to conduct by application of a control voltage to the junction of resistors 7 and 8, the electromagnet becomes conntected to the supply 6 in series with transistor 5. If transistor 5 is also caused to conduct, the full battery voltage, less the collector-emitter voltages of transistors 5 and 10, is impressed across the electromagnet winding. The current in this winding rises initially at the rate of E/L amperes per second, where E is the supply voltage of voltage source 6, and L is the inductance of the electromagnet. The winding current would eventually attain the value E/R amperes, where is the winding resistance, a value which, if allowed to continue, would possibly cause overheating of the winding or, if the magnetic circuit is closed by movement of pole-piece 3, would be much higher than that required to provide the desired mechanical force. The relationship between the desired holding current and the excessive current produced by the high voltage energizing of the winding of electromagnet 1 is shown as curve 21 in FIG. 2. Curve 20 of FIG. 2 shows the current waveform produced when the winding is energized from a source of low voltage. It is seen that the low voltage, though it eventually produces the desired winding current, requires a much longer time to do so, whereas the high voltage, though it causes the desired current to be obtained more quickly, would, if continually applied, result in an excessive winding current. The circuit of FIG. 1 acts to provide the rapid current rise of curve 21 to the desired value and to prevent the current from exceeding this desired value thereafter, producing a current waveform such as curve 22 of FIG. 2.

The control of the electromagnet `holding` current to the desired value is achieved by feedback control employing a resistor 11 of relatively low resistance to produce a feedback voltage proportional to the electromagnet winding current. This feedback voltage is compared with a reference voltage appearing across resistor 14 which represents the desired value of steady state current in the winding of electromagnet 1. Resistor 14 is part of the voltage divider network comprised of resistors 12 and 14 providing a desired fraction of the voltage appearing across Zener diode 9 which is energized through resistor 8 by the same input control voltage that is used to bias transistor 10 into conduction. The current feedback voltage from resistor 11 is applied directly to the inverting input of the integrated circuit operational amplifier 17 while the voltage across resistor 14, representing the desired winding current, is applied to the non-inverting input of the same amplifier 17. Resistor 16 from the output of amplifier 17 to the non-inverting input, provides positive feedback of a controlled amount to cause amplifier 17 to operate as a Schmidt trigger. Typically, the feedback is adjusted so that the holding current through the winding of electromagnet 1 is maintained within about 10% of the desired value as shown in curve 22 of FIG. 2. Amplifier 17 drives transistor 24 through its base resitor 18. When transistor 24 is caused to conduct, the switching power transistor 5 is biased into conduction by current flow through its base resistor 19.

Transistor 5 will be alternately conducting and non-conducting as the output of the Schmidt trigger alternates polarity. While transistor 5 is in the non-conducting state, transistor 10 meanwhile held in conduction continually by application of a control voltage to the junction of resistors 7 and 8, the current through the winding of the electromagnet 1, since it cannot stop abruptly, is switched to diode 4. In this passive state, the only voltage acting to reduce the current in the winding is that due to the winding I .times. R voltage drop, the voltage across resistor 11 and the voltages across transistor 10 and diode 4 which are relatively low. Typically the voltage acting to decrease the current in the winding is a small fraction of that applied to the winding to produce the current so that the current in the winding of electromagnet 1 will decay at a much lower rate than it recovers, exhibiting a waveform such as curve 20 of FIG. 2. Thus the voltage supply 6 is applied to the solenoid only intermittently to maintain the desired value of holding current, typically only 5% of the time. In this manner relatively little power is required of voltage source 6 to maintain the desired current in the winding; only that required to supply the power losses in transistors 5 and 10, diode 4, resistor 11, and the ohmic losses in the winding of the electromagnet 1.

For the electromagnet to remain energized in the manner described, the control voltage appearing between ground and the junction of resistors 7 and 8 must be maintained at a value sufficient both to bias transistor 10 into conduction and to operate Zener diode 9 at its breakdown voltage. When this control voltage is removed, transistor 10 will cease to conduct and the voltage across the Zener diode 9 will fall to zero. Either one of these two conditions, a zero reference voltage to amplifier 17, or transistor 10 rendering non-conducting, would suffice to reduce the current in the winding of the electromagnet 1 to zero. When both conditions are present, as is the case when the control voltage is removed, both transistors 5 and 10 are rendered non-conducting, the former by virtue of a zero reference voltage supplied to amplifier 17, the latter as described above. The current in the winding of electromagnet 1, which cannot cease suddenly, is provided a path now through diodes 2 and 4 in series with the supply 6 which in this instance appears across the winding of the electromagnet with a polarity reversed from that by which the winding was energized. This relatively high voltage, appearing across the winding of the electromagnet in a direction to reverse the current through it causes the current to decay with the same high speed with which the winding was initially energized, but the current is prevented from reversing by both diodes 2 and 4. Thus the current in the winding is reduced to zero with great speed and the energy corresponding to this current is returned to the voltage source. Curve 23 in FIG. 2 showns the current waveform at cutoff.

The optional addition of capacitor 13 across resistor 12, shown by the dotted lines in FIG. 1, will cause the initial surge current in the winding of electromagnet 1 to be much larger than the desired or permissible steady state value. This large initial current is advantageous in providing a high mechanical force in the electromagnet when a maximum gap separates the pole pieces 3 from the core. Momentarily, the full voltage of zener diode 9 becomes the reference voltage corresponding to a higher initial reference current which will decay exponentially to the steady state reference current as shown in curve 26 of FIG. 3. The magnitude of this higher initial reference current and the associated time constant can be adjusted to desired values by the proper choice of capacitance value and suitable modifications of the voltage divider network comprised of resistors 12 and 14 as is well known to those skilled in the art. Once the surge current reaches the reference value it would tend to continue except for a relatively slow exponential decay as shown in curve 25 of FIG. 3, were it not for the fact that at some value of current the pole piece 3 of electromagnet 1 would move to close the gap. The motion of pole-piece 3 under mechanical force of the magnetic field withdraws energy from the field and results in a corresponding reduction in the current in the winding of electromagnet 1, as shown in curve 27 of FIG. 3. After the the current in the winding is reduced to the desired steady state value by whatever means, the voltage across resistor 14 remains constant and the current waveform is as shown by curve 22 of FIG. 2.

Should the pole-piece 3 fail to be actuated during the high initial surge current through the winding of the electromagnet 1, the high current in the winding, shown as curve 25 in FIG. 3 will tend to persist but with no further drain upon power supply 6 until the value decays to the reference steady state value. Thus a high mechanical force remains to actuate pole-piece 3 for an appreciable length of time, a desirable circumstance. If the pole-piece fails to move, the energy it would have taken is dissipated as heat in the winding of the electromagnet.

When two solenoids are to be operated alternately, as is often the case, for example to achieve a bidirectional mechanical motion or to control a three-way solenoid valve, the energy of one electromagnet being de-energized can be used to produce the high voltage to energize the second electromagnet at high speed without the aid of a separate high voltage source. To clarify the mechanism by which the energy transfer from one electromagnet to another is made to take place, a simplified equivalent circuit of the circuit of FIG. 1 is shown in FIG. 4, where, for greater ease of comprehension, the switching transistors 5 and 10 of FIG. 1 are represented respectively by on-off switches 5 and 10. When switches 5 and 10 in FIG. 4 are open, as shown, no current flows in the solenoid 1. To energize solenoid 1 continuously, switch 10 is closed continuously and switch 5 is closed intermittently as required to bring and maintain the current in solenoid at a desired value as described previously. When solenoid 1 is to be de-energized, both switches 5 and 10 are opened and the current in the solenoid 1 is then required to flow through diodes 2 and 4, rendering them conducting and thereby applying the full voltage of voltage source 6 across the solenoid 1 in a polarity opposing the flow of current.

In FIG. 5 two circuits, such as shown in FIG. 4, are employed to control two solenoids, circuit 35 controlling solenoid 1 with simplified circuit elements 2, 4, 5 and 10 as described above, and a second similar circuit 36 controlling a second solenoid 31 with simplified circuit elements 32, 34, 33 and 30 corresponding in function respectively to circuit elements 2, 4, 5 and 10 of circuit 35. The two solenoid control circuits 35 and 36 are coupled to a voltage source 6 through diode 28 and are shunted by capacitor 29. Not shown in FIG. 5 are the input signals and control circuitry which control the operation of switches 5, 10, 30 and 33, these being as shown in FIG. 1. Though with the circuit of FIG. 5, solenoids 1 and 31 may be operated independently, it is required for energy transfer from one to the other that they be operated alternately, and in this respect when an "on" signal is applied to circuit 35, an "off" signal is assumed applied simultaneously to circuit 36, a mode of control most easily obtained by driving circuits 35 and 36 with the complementary outputs of a flipflop circuit. In this "either-or" mode of operation, where either one of two electromagnets is to be energized but not both, one of the two circuits 35 or 36 will be in the "on" state, say circuit 35, and in this state switch 10 will be closed, switch 5 will operate intermittently under current feedback control to maintain the current in solenoid 1 at the desired value, and switches 30 and 33 of circuit 36 will remain open, solenoid 31 thereby remaining un-energized. When solenoid 1 is to be de-energized and solenoid 31 simultaneously energized, switches 5 and 10 are opened under the action of the effecting control signal, switch 30 is closed and switch 33 is actuated under current feedback control to bring and maintain the current in solenoid 31 to the desired value.

In the interval of transition, during which solenoid 1 becomes de-energized and solenoid 31 becomes energized, diode 28 prevents the return of energy from solenoid 1 to the voltage source 6, whereupon the current from solenoid 1 is discharged into capacitor 29, raising its voltage to a value which may be as high as desired for a suitably small value of capacitance of capacitor 29, and which voltage is impressed across solenoid 31, likewise energizing it as quickly as desired within the voltage breakdown limits of the circuit elements employed. Since, during the interval of transition, diodes 28, 32 and 34 are non-conducting and may be replaced by open circuits, and correspondingly transistors 5 and 10 may also be replaced by open circuits and the elements represented by numerals 2, 4, 30 and 33 are all conducting and may be represented by short circuits, the mechanism of energy transfer from one solenoid to another is most readily made apparent by recourse to the simplified equivalent circuit of FIG. 6 where e is the instantaneous voltage across the capacitor 29, the latter having a capacity of C farads, L is the inductance in henries of the two solenoids 1 and 31, each assumed, for purposes of illustration to have an inductance of L henries, and the voltage of the voltage supply is assumed to be E volts. The two currents in the solenoids 1 and 31 flow in the direction shown and are denoted respectively by the variables i.sub.1 and i.sub.31, the initial value (at the instant that the process of energy transfer begins) of i.sub.1 being I amperes, the desired value of current in either of the solenoids in the "on" state, and the initial value of i.sub.31 being zero amperes. The dynamic behavior of the simplified circuit of FIG. 6 is easily obtained by solving the relevant differential equations for the initial conditions and circuit element values given, the voltage e and the currents i.sub.1 and i.sub.31, being as follows: ##EQU5## where ##EQU6## is the angular frequency in radians per second and t is the time in seconds. Since it is obvious from the above that i.sub.1 + i.sub.31 = I, the current i.sub.31 in solenoid 31 will reach the value I when the current i.sub.1 in solenoid 1 reaches the value zero. The time T, in seconds, required for this complete transfer of energy from solenoid 1 to solenoid 31 can be obtained by setting i.sub.1 equal to zero in the above equations, resulting in the following expression for the switchover time T: ##EQU7## For an extremely large value of C, or in effect a very large capacitor 29, we may use the approximation that for small angles, the trigonometric tangent of an angle is equal to its value in radians, for which case the switchover time T can be given explicitly as ##EQU8## which, to give an example, would be 4.167 milleseconds for typical values of supply voltage E of 12.0 volts, solenoid inductance of 0.1 henries and solenoid current of 0.5 amperes. This switchover time, using a very large capacitor, is the same switchover time that would be obtained with the circuit of FIG. 1 for one solenoid having a voltage supply 6 of 12.0 volts. For smaller values of capacitance C of capacitor 29, the switchover time T may be made as small as desired, provided safe voltage limits of the circuit elements are not exceeded. Thus, for a desired switchover time of one millesecond, the first expression given above for the switchover time T can be solved for the requisite value of C from the known values of circuit parameters I, E, L and T, which, for the values assumed in the present example, results in a value C for capacitor 29 of 4.0 microfarads.

When a small value of capacitance C is used to obtain a short switchover time T from one solenoid to another, it is important that the peak voltage appearing across capacitor 29 not be excessive. The peak value e.sub.peak of the voltage e may easily shown to be given by the following expression: ##EQU9## For a very large value of C, this voltage will be equal to the voltage E of the voltage supply 6, and for values of E, L and C of the above example, the value of e.sub.peak would be only 57 volts. To produce a switchover time of one millesecond in this example with the circuit of FIG. 1 only, would have required a voltage supply 6 of 50 volts, yet it was produced with the circuit of FIG. 4 with a supply voltage E of only 12.0 volts and with a peak voltage not much in excess of the above mentioned 50.0 volts. Thus the addition of one capacitor 29 and one diode 28 to a pair of basic circuits 35 and 36 provides for energy transfer from one solenoid to another at high speed using only a single low voltage supply. An advantage of the lower supply voltage is the reduced probability of damaging the switching transistors in short circuit conditions and the greater ease with which the switching transistors may be protected; further, since all transistors are employed in a switching mode they dissipate relatively little power and low power transistors can be used to control relatively high power solenoids.

Claims

I claim:

1. In combination with an electromagnetically operated device, a circuit for producing rapid current growth in the coil of said device, and for maintaining a predetermined level of current in said coil, an electrical circuit comprising:

a source of relatively high voltage capable of effecting rapid current growth in said coil;

semiconductor switching means responsive to external control for applying said high voltage source to energize said coil, said semiconductor switching means operating in a switching mode as a substantially non-dissipative element;

first means for sensing the current in said coil to open said semiconductor switching means when a predetermined coil current value is reached, and for closing said switch intermittently thereafter, to maintain said predetermined coil current value;

and second means subject to external control for applying said high voltage source to said coil in opposite polarity, thereby accelerating the rate of current decline in said coil.

2. Apparatus according to claim 1 in which a shunt diode path is provided to permit continued current flow through said coil, said diode being poled so as not to conduct in response to forward energization from said source, but to conduct in response to the reverse coil voltage extant between applications of said source intermittently to said coil by said first means.