GB2278746A - A power controller for motors - Google Patents

Abstract

A device to control the speed of induction or shaded-pole fractional horsepower electric motors, when driven by alternating current taken from a mains supply, by connecting the motor to the supply at each zero-crossing by a switch device 9 and breaking the connection part of the way through each half-cycle. To avoid abrupt voltage changes that would otherwise give rise to vibration and Radio Frequency Interference, a capacitor 11, wired in parallel with the motor, discharges its stored energy into the motor after the supply connection is broken. An example of a complete power controller incorporating this principle is shown, together with control circuit 63 - 65 and zero crossing detector 41, 50, 52 to control the duration of the conduction period and hence the power delivered to the motor. Also shown is the circuitry 14, 73 - 77 required to protect the switch device 9 under fault conditions. <IMAGE>

Description

A POWER CONTROLLER FOR FHP MOTORS This invention relates to a method of controlling the speed of an AC electric motor (for example induction or shaded-pole type) by varying the amount of power that it can draw from the public electricity mains supply, without introducing unacceptable degrees of electrical or mechanical noise and at moderate cost.

The need to be able to control the amount of power that a load consumes from the public mains electricity supply has long been recognised and many methods have been used; a well-known example is the domestic lighting dimmer which uses a device known as a TRIAC to reduce the proportion of the full mains voltage cycle that is fed to the lamp. A similar technique is commonly employed to control the power feed to small electric motors of the shaded-pole or induction variety to give a variable speed characteristic. It is a feature of this technique, known as phase-angle control, that large and abrupt changes in voltage and current take place which, experience shows, give rise to Radio Frequency Interference and mechanical vibration, unless expensive and elaborate filtering is used.Other methods of controlling power include the use of resistive droppers, variable transformers and inverters, each of which has its merits and drawbacks.

The method used in this invention is to invert the traditional thinking by using the early part of the voltage cycle, rather than the later, as the conduction period in addition to which, a capacitor is used as an energy store.

The following description considers first the various features that comprise this invention and then they are brought together into a tried and tested practical example.

The description is to be read in conjunction with the accompanying drawings.

Fig.l shows the voltage waveform as it would be seen across the load; Fig.2 shows a simplified method of achieving that effect; Fig.3 shows an elaboration of the principle; Fig.4 shows the beginnings of a practical method; Fig.5 shows an isolating zero-crossing detector; Fig.6 shows another isolated zero-crossing detector having the capability of transmitting useful power; Fig.7 is a complete circuit diagram of a practical implementation of the principle.

Fig.l illustrates the principle in which one complete cycle of the mains voltage is indicated by line 1 which has the form of a sine curve, crossing the zero voltage axis at points 2, 4 and 5. It can be seen that part of the sine curve is shown as a solid line and part is dotted to indicate that the voltage across the load departs from the ideal sinewave in this area. By convention, one complete cycle of the supply voltage is equivalent to an angular change of 360 degrees and this is shown on Fig.l as the distance between points 2 and 5. Thus the half cycle of 180 degrees between 2 and 4 can be divided into two angular portions; the first, between 2 and 3, is called the conduction angle and the remainder the relaxation angle.

During the conduction angle, the voltage across the load follows the same rate of change, in terms of volts per second, as the sinusoidal mains voltage, while during the relaxation angle, the rate of change of load voltage follows the familiar exponential fall characteristic as the energy stored in a capacitor is discharged through the load. There are thus no sharp discontinuities in the voltage supplied to the load and so the angular velocity of the current waveform is sensibly constant. The shaded areas 6 and 7, between the voltage curve and the zero voltage axis, represent the power delivered to the load and it can be seen that the area covered by the shading will vary as the conduction angle is varied.

Fig.2 shows a highly simplified version of the way this is achieved. A control system 10, which is synchronised to the mains voltage cycle, closes a switch 9 at zero-crossing point 2, thus connecting both the load 12 (shown by way of example as an electric motor) and the capacitor 11 to the mains supply represented by an AC generator 8. As the mains voltage follows its positive-going excursion, power is taken from the generator 8 to drive the load 12 and to charge the capacitor 11; this process is indicated by the shaded area 6 which is completed at point 3 as the switch 9 is opened by the control system 10. Between points 3 and 5, the energy stored in the capacitor 11, represented by shaded area 7, is used to continue to drive the load 12 until it is exhausted. During the negative half-cycle, a similar sequence takes place.

Fig.3 shows a more practical arrangement of the principle by allowing unidirectional current flow through the switch 9 while maintaining alternating flow through the load 12 and the capacitor 11; the bridge rectifier 13 makes this possible. This change is necessary to enable a practical semiconductor device to be selected to function as the switch 9 and possibilities include Bipolar Transistors, Metal Oxide Semiconductor Field-Effect Transistors (hereafter known as MOSFETs) and a combination of the two called Insulated Gate Bipolar Transistors (IGBTs). Although in the present implementation, power MOSFETs are preferred for reasons of economy and ease of use, the invention does not depend on the employment of any particular type of switch.

Although forming a virtually ideal switch for the purpose described in this invention, MOSFETs require careful protection from the potential causes of electrical failure which threaten them, principal causes being over-voltage between source and drain, excessive current flow in the source-drain channel and too much power dissipated in the device. Careful selection of a suitable device is of course necessary but thought has been given to situations which occur in practice which would destroy the most conservatively rated component. It is possible that the capacitor 11 could become disconnected or short-circuited during use; the former situation could lead to over-voltage as the suddenly collapsing current at point 3 causes a back-EMF voltage spike to be generated and the latter would cause excessive current flow.It is also possible that the motor 12 could become disconnected while the capacitor 11 remains in place; this would cause excessive current flow at, or just after, point 2 as the entire stored charge on 11 is quickly dissipated in switch 9.

Fig.4 introduces the use of a MOSFET as switch 9 and outlines the means used to protect it. Fig.4a shows the circuit of Fig.3 adapted to use the MOSFET and Fig.4b shows the relative signal timings. The control system 10 produces a signal 15 which causes the MOSFET 9 to go into its conduction mode at point 2; this is represented by a square voltage waveform 15 on Fig.4b and it can be seen that the signal returns to zero at point 3, thus defining the conduction angle. Signal 15 is known as the gate enhancement voltage of the MOSFET. The current through the MOSFET is detected by inserting a resistor 14 in its source lead and feeding a signal 16 back to the control system 10 which, if it reaches a critical level, will limit the current flow so as to avoid the MOSFET being taken outside its Safe Operating Area. This feedback signal is interpreted in two distinctly different ways by the control system depending on which part of the conduction angle is being performed.

In the present invention, for a fixed period of time 18 after point 2, the system operates as a current limiter and for the rest, 19, of the conduction angle, it functions as an overload trip. The reason for this is that if the overload trip were in operation throughout the whole conduction angle, it is likely that current spikes, that are possible at zero-crossing as a matter of course, would cause a nuisance by stopping the motor unnecessarily. They could be dealt with by inserting a high-powered resistor in series with the load but this would waste power and generate heat. The method used here is to convert the overload detector into a current limiter for a short fixed period immediately following zero-crossing, the time being chosen so that the safe-operating-area of the selected MOSFET is not exceeded.In this way expected current spikes are safely dealt with while fault currents will operate the overload detector if they occur later in the cycle.

When the overload trip is activated, it immediately removes power from the control system 10 which collapses signal 15 to zero voltage in less than 5 microseconds. In one version of this scheme, a manual RESET operation is required to restore power; in another, power is gradually restored automatically after a short delay and the system shuts down again if the fault condition persists, an indicator being provided to warn the user.

Zener diode 17 is chosen to clamp the voltage across the MOSFET 9 to less than its rated maximum and, at the same time, to cause the control system 10 to shut down completely as in the case of the over-current condition.

An essential feature of this invention is the ability to detect the point of zero-crossing accurately and two methods have been devised to do this. In each case, complete isolation from mains voltages is provided and in the second case, useful amounts of power are transmitted across the isolation interface enabling, in many cases, a bulky low voltage power transformer to be dispensed with.

It is a feature of both methods that failure to generate the required synchronising pulses will cause a safe system shut-down.

Fig.5 shows the elements of the simpler method. A full-wave bridge rectifier 31 receives a pair of symmetrical inputs from the generator 8 via identical resistors 30 and delivers unidirectional half sinusoid pulses across the potential divider formed by resistors 32 and 33, the same pulses being used to charge capacitor 36 via diode 34. At any time other than zero-crossing, transistor 35 will be kept saturated by current supplied via resistor 33 and this will hold its collector near to its emitter voltage thus preventing transistor 38 from drawing any current and allowing capacitor 36 to charge.

When the zero-crossing point is reached, the base current supply to transistor 35 is briefly removed and transistor 38 is able to discharge capacitor 36 into the diode section of opto-coupler 39, component values being chosen so that the discharge is complete before transistor 35 is saturated at the start of the next half-cycle. The output transistor of the opto-coupler 39 acts as a fast switch which closes exactly at the zero-crossing and may be used to synchronise a timing voltage ramp, for example, as indicated by the resistor and capacitor network.

The more elaborate arrangement shown in Fig.6 behaves similarly to the above in that a bridge rectifier 41 is used to produce unidirectional half-sinusoids across the potential divider formed by resistors 42 and 43 but in this case, they are at full mains voltage. Here again, at any other time than zero-crossing, transistor 45 is held saturated preventing the thyristor 48 from firing. This leaves capacitor 49 free to charge via diode 44, protection resistor 47 and diode 51, the extent of charging depending on the component values. At the instant of zero-crossing, transistor 45 loses its base current temporarily and its collector voltage rises to fire the thyristor. This very rapidly dumps the energy stored in the capacitor 49 into the input winding of the transformer 50 and transformer action induces a pulse of current in the secondary winding.

This can be used to charge capacitor 53 via diode 52 and sufficient energy can be transferred to power isolated low voltage electronic circuitry. The pulse 56 can be used to synchronise the timing circuits.

Fig.7 is the circuit diagram of a complete power controller incorporating the principles described above plus some additional features. The circuit is divided into two sections separated by an isolation barrier 60. The section to the left of 60 is isolated from mains voltages and so may be safely connected to any other similarly isolated apparatus such as a microcomputer.

Since many of the features shown here relate directly to matters already discussed, they have been allocated the same reference numbers to aid identification; for example, the MOSFET has the number 9 associated with it as it has throughout. To avoid clutter, although the diagram shows all components required to make a practical working version, only those essential to an understanding are numbered.

64 is a comparator whose output terminal is at a level determined by the relative levels of a ramp 63 derived from the zero-crossing detector (see Fig.6) and a control voltage 62 generated by operational amplifier 61 and its associated circuitry. Whenever the + input of 64 is at a lower voltage than the - input, the optocoupler 67 is energised and the conduction angle is performed. The ramp voltage is forced to its lowest level at zero-crossing thus marking the start of the conduction angle, the end of which is the point at which the ramp voltage rises above the control voltage 62. Thus by changing the level of 62, the extent of the conduction angle and hence the power to the load, can be controlled.

An INPUT voltage, 0 to 5 volts in this case, is able to vary voltage 62 between a minimum value set by 66 and a maximum value set by 65.

Powering for the isolated part of the system is provided as described in Fig.6, a zener diode 55 typically stabilising the supply at 12 volts. The MOSFET control and protection circuit, which forms the rest of the arrangement, is powered by a low-current supply provided by a network of components 69 and a reservoir capacitor 70; here again, the voltage level is stabilised by a 12 volt zener diode.

The MOSFET 9 is brought into conduction by a rise in voltage at its gate terminal; the energising of the optocoupler 67 short-circuits the emitter-base junction of transistor 68 and signal 15 rises towards supply level at a rate determined by the values of resistor 71 and capacitor 72. The current limiter is formed by transistors 73 and 74 with their associated components; this acts during the time period 18 (see Fig.4b) by holding down the enhancement voltage to maintain a maximum channel current flow in the MOSFET. During period 19 the current limiter is disabled and if an excessive current causes potential difference across resistor 14 to rise high enough, thyristor 75 will be triggered into conduction. Since this is connected across the power supply, the voltage will collapse to zero very rapidly and remove the enhancement signal 15.

Reservoir capacitor 70 discharges very rapidly through LED indicator 76, causing a bright flash to act as a fault warning. The discharge of capacitor 70 deprives the thyristor 75 of its holding current and it ceases to conduct, allowing the capacitor to recharge once more through the network 69. The charging current also passes through the diode of the optocoupler 77, the effect of which is to pull down control voltage 62 so far that the ramp cannot activate optocoupler 67, this state of affairs lasting until capacitor 70 has charged fully. Control voltage 62 is now free to rise slowly to its normal level but as it does so, the conduction angle increases gradually from zero and if the fault condition still exists, the thyristor 75 is triggered once more to repeat the cycle. In this way, power dissipation in the MOSFET under fault conditions is kept within safe bounds and the system will repeat the fail/reset cycle continuously and harmlessly.

It can be seen that any current flow through zener diode 17 due to excessive voltage across the MOSFET, also passes into the gate terminal of thyristor 75 and the fault cycle is triggered as described above.

Claims (12)

1) A power controller for Alternating Current driven devices wherein early conduction angle phase control is combined with the use of an energy store to eliminate abrupt voltage changes across the load.

2) A power controller as in Claim 1 wherein a capacitor is used as a convenient means of energy storage.

3) A power controller as in Claim 2 wherein energy stored in the storage means from the supply during the conduction angle is used to continue to provide power to the load during the relaxation angle.

4) A power controller as in Claim 3 wherein the chosen load is an electric motor and the capacitance value of the energy store is chosen empirically to give an acceptably low degree of mechanical vibration in the motor.

5) A power controller as in Claim 4 wherein means may be provided to protect the system from stresses brought about by faulty load conditions.

6) A power controller as in Claim 5 wherein the system returns to normal operation automatically upon the removal of the fault condition, the protection means safely and periodically re-testing for the continued presence of the fault condition.

7) A power controller as in Claim 5 wherein a manual reset action is required to restore normal operation upon the removal of the fault condition.

8) A power controller as in any of the preceeding claims wherein a sensibly linear ramp waveform, precisely synchronised with the zero-crossing points of the AC supply, is used, in conjuction with a control voltage, to establish the duration of the conduction angle and hence the relaxation angle.

9) A power controller as in Claim 8 wherein the Alternating Current is provided by the public mains electricity supply (typically, though not exclusively, at 240 volts 50 Hertz) and wherein the low voltage (typically 12 volts) required for the circuits that perform the control of conduction angle may be isolated from mains voltages to facilitate connections to other apparatus.

10) A power controller as in Claim 9 wherein the isolated low-voltage supply is provided by pulses of energy at the zero-crossing points of the AC supply transmitted by a pulse transformer, simultaneously effecting synchonisation and power transfer.

11) A power controller as in any of the previous claims wherein the power delivered to the load may be varied between an upper and lower limit in response to a control input.

12) A power controller as in any of the previous claims wherein the variation in power delivered to the load may be made proportional to the difference between a control input and a set point.