GB2343070A - Magnetic amplifier circuit - Google Patents

Abstract

A set-reset magnetic amplifier controlled rectifier circuit comprises a saturable inductor Ls, a positively biased diode D1 and a negatively biased diode D2. A bi-directional current source Ic is provided in parallel with the saturable inductor Ls which is capable of providing a set and reset bias to the core of the saturable inductor Ls. This allows full utilisation of the inductor core, while still allowing low control currents because highly permeable square loop material can be used for the inductor core. The diode D1 may be a Schottky diode. A more detailed circuit is described (Fig 8) having a PWM circuit (10), a magnetic amplifier (20), an output inhibit circuit (40), an output error amplifier (50), a control error amplifier (60), an output voltage error amplifier (70) and a voltage reference circuit (80).

Description

Magnetic Amplifier Circuit This invention relates to a magnetic amplifier circuit.

Magnetic amplifiers or'magamps'are well known devices which use one or more saturable inductors, either alone or in combination with other circuit elements, in order to achieve power gain. The saturable inductors are effectively used as switching elements. One known application of a magnetic amplifier is in a buck-derived rectifier circuit.

A Buck derived circuit is one which is derived from the basic step-down switching power supply topology known as a Buck Converter. The Buck Converter is typically used to generate a regulated dc output voltage which is lower in amplitude than the input voltage. The output voltage is then used to power some form of load. The basic circuit of the Buck Converter is shown in Figure 4 of the drawings. Waveforms indicating the ideal current in switch S I and in freewheel diode D 1 are also shown.

Referring to Figure 4, the switch S1, which is typically a power MOSFET, is periodically switched on. The frequency at which the switch S1 is switched on is not necessarily fixed, but may be variable according to design requirements. When the switch S 1 is switched on, the filter inductor LI is connected to the input voltage Vin and current flows from the input through S 1 to the filter, formed of LI and Cl, and also to the load. When S 1 is opened, the current in LI continues to flow through diode D 1.

In the ideal case, and assuming that the current in LI is continuous, then the voltage at the cathode of D1 takes one of two values, i. e. 1 diode drop below 0V when switch Sl is open (which is approximately 0V), switch on-state volt-drop below Vin when S I is closed (which is approximately Vin). In the steady-state, the average voltage across LI is zero. Therefore, the average voltage at the cathode of D l and at the load is Vin x D, where D is the duty cycle of the switch SI, and represents the average time for which it is closed.

One specific type of Buck-derived circuit is the"Isolated Output"Buck-derived circuit, a circuit diagram of which is shown in Figure 5 of the drawings. Referring to Figure 5, the Isolated Output Buck-derived circuit includes an isolating transformer Tl, but operates in essentially the same manner as the circuit shown in Figure 4. In this case, when the switch S l is closed, the voltage applied to the cathode of D2 is approximately Vin x N2/N1, where NI is the number of turns on the primary winding of the transformer T1 and N2 is the number of turns on the secondary winding of the transformer Tl. When Sl is open, the voltage at the cathode of D2 is 1 diode drop below 0V (i. e. approximately 0V).

Thus, the average voltage at the cathode of D2 and also at the load is V x D, where V=Vin x N2/N1 and D is the duty cycle of the switch Sl.

A magamp-controlled rectifier circuit is another type of Buck-derived circuit, and a known Magamp-controlled rectifier circuit is shown in Figure 1 of the drawings. The rectifier circuit of Figure 1 is essentially the same as the circuit shown in Figure 5, but it includes a second switch in the form of a saturable inductor Ls on the secondary side of the transformer.

Referring to Figure 1, a conventional rectifier circuit comprises a transformer, of which only the secondary winding Tla is shown. The circuit further comprises a positively biased diode Dl and a negatively biased diode D2. A saturable inductor Ls is provided in series with the positively biased diode Dl. the voltage from the rectifier circuit is applied to a filter circuit formed by inductor L1 and capacitor Cl.

The object of the circuit shown in Figure 1 is to allow the average output voltage of the circuit to be controlled. This is achieved by blocking the output of the secondary transformer winding Tla, which is typically a quasi-square waveform, for some of its period using the saturable inductor, so that the average voltage applied to the filter circuit LI, Cl is reduced, and the average output voltage of the overall circuit is reduced. Furthermore, by controlling the amount of time for which Ls blocks the output of T I a in any one period, the average output voltage of the overall circuit can be controlled between approximately zero and a value close to that which it would exist if Ls was a short circuit.

There are two basic configurations for Ls, namely the reset controlled magamp and the set controlled magamp, which will now be explained in more detail.

Because the coils of a saturable inductor carry an alternating current, the core material of the inductor goes through many magnetic cycles per second. It is therefore well known to plot how the flux density B in the core material varies with the magnetising field H in a magnetic cycle. Initially, when a positive magnetising field is applied, the core material is magnetised until, at saturation, the relative permeability drops to a value close to 1. When the magnetising field is removed, the core material remains magnetised at the remnant flux density. When the magnetising field is reversed, the residual magnetism is opposed, such that each increase in magnetising field causes a decrease in flux density, until the flux density is once again reduced to zero at a value of H known as the coercive force of the core material. If this cycle is plotted as B against H, it forms a closed loop known as a hysteresis or BH loop.

In a reset-controlled magamp, and with reference to Figure 1, at the beginning of the positive half-cycle of the output waveform from the transformer winding Tla, the saturable inductor Ls is in its blocking state. During the positive half-cycle, current builds up in Ls and Dl, until Ls becomes saturated and switches to its conducting state, when its impedance falls to a low level. During the negative half-cycle, current flows in D2, and the flux density begins to drop until the inductor Ls reduces to its remnant flux density and then increases in the opposite direction by the action of a control circuit (not shown). During the next positive half-cycle, the flux density must increase from its negative value (set by the control circuit) up to the saturation flux density before the inductor Ls once again switches into its conducting state. The control circuit can adjust the flux swing through which the core must change before saturation takes place and hence the"the blocking time"of the saturable inductor.

A reset controlled magamp typically uses a saturable inductor having a core material formed of a"square-loop"material, for example, an amorphous metal. The material used in such a saturable inductor is known as"square-loop"material because its BH loop is nearly rectangular in shape. A typical representation of a BH loop for a"square loop"material is shown in Figure 2 of the drawings.

For the avoidance of doubt, if the core material of a saturable inductor is a"square loop"material, it has a very high permeability in its unsaturated state, which gives the inductor a very high impedance in the blocking state, and a low permeability and impedance in the saturated conducting state. The"square loop"also indicates that the remnant flux density Br of the material is very close in value to the saturation flux density Bs. The difference Brs between remnant and saturated flux density indicates the minimum flux excursion through which the material must go when switching from the blocking state to the conducting state. In other words, the value of Brs is directly proportional to the time for which the saturable inductor Ls of the circuit shown in Figure 1 will block the output voltage of the transformer winding T1 a.

The flux excursion Brs required to switch from the blocking state to the conducting state can be increased by applying a negative magnetic field to the core of Ls, so that the flux excursion of Ls can range from a minimum value Brs to a maximum value 2Bs.

There are a number of advantages associated with the use of square loop materials. For example, typical square loop materials have a relative permeability which is substantially greater than 10,000 and a saturation flux density greater than 500mT. This enables operation of Ls with few turns, and hence a low saturated inductance.

Furthermore, many square loop materials have very low core losses, thereby allowing Ls to be operated with high flux swings, and hence a low number of turns. Only negative feedback (to provide the additional negative magnetic field) is required, and high permeability square-loop cores require very low levels of current to achieve biasing of the core, thereby achieving low losses and high efficiency. Finally, reset controlled magamps allow full use of the core material of Ls by allowing operation between the two saturated states.

However, there are also a number of disadvantages associated with reset controlled magamps. For example, square loop amorphous metal cores are significantly more expensive than non-square-loop ferrite cores. Furthermore, the high permeability and low coercive force of the core means that they can be significantly reset by diode leakage currents. In particular, diode Dl in Figure 1 must be a low leakage type with defined leakage current characteristics. Diode leakage reduces the output voltage capability of a reset controlled magamp rectifier. In other words, a high leakage current in D1 will provide a negative bias to the core during the negative half-cycle, so that the time for which the inductor Ls blocks the output voltage from the transformer winding Tla during the positive half-cycle may be substantially increased.

In a set-controlled magamp, and with reference to Figure 1, the core of Ls is operated between Br and Bs when no control signal is applied, and biased towards saturation Bs when reduced blocking is required.

The set-controlled magamp typically uses a low cost, low-loss"power ferrite"core material which has been optimised for power circuit applications. A typical BH loop for such a material is shown in Figure 3 of the drawings.

Power ferrites have much lower relative permeability than amorphous materials, typically 2,000-3,000. The remnant flux density Br is also much lower than Bs, typically 20-30%. In other words, these materials do not have a"square-loop" characteristic.

Bs is typically 400-500mT at 25 C, dropping to 250-300mT at 100 C, and power ferrite materials require a significantly larger magnetic field strength than that required by an amorphous core to achieve saturation. Typically, the magnetic field strength at saturation may be greater than 100A/m compared to around l OA/m for an amorphous core. In the set controlled magamp circuit, the core of the inductor Ls is designed to give full blocking capability for a flux swing from Br to Bs, i. e. a flux swing of Brs. As stated above, the core is biased towards saturation by a control signal to reduce blocking and switch the inductor to its conducting state.

There are advantages associated with the set-controlled magamp, such as because it requires positive feedback to reduce blocking and achieve an output, the core reset effects caused by diode leakage can be offset. This is especially useful when operating at high frequencies, because Schottky rectifiers can be used. Furthermore, the set controlled magamp is particularly suited for operation at high frequencies since its peak to peak flux swing is relatively low (typically less than 200mT).

However, there are also a number of disadvantages associated with the set-controlled magamp. Firstly, because the allowed peak to peak flux swing is relatively low, more turns are required for a given core size than would be required with the reset, control method utilising a"square loop"core. The semi-saturated and saturated inductance is relatively high, and a lower output current is available. Furthermore, power ferrite cores have a lower permeability and a high magnetic field strength when saturated so that higher control currents are required than those required for the highly permeable "square-loop"cores.

Because the set-control technique requires positive feedback to achieve an output, it is necessary to provide some type of"start-up"circuitry in order to to'start-up'operation of the rectifier circuit and produce an output, for example, an auxiliary voltage could be generated initially. Finally, the set-control technique poorly utilises the core of Ls since it is only operated between Br and Bs.

In the case of the magamp controlled rectifier circuit shown in Figure 1, both the saturable inductor Ls and the switch S 1 (not shown) on the primary side of the transformer must be simultaneously conducting in order to produce an output voltage.

Thus, several outputs may be independently regulated by independently controlling each magamp in a multi-output Buck-derived circuit.

Referring to Figure 6 of the drawings, there is shown a circuit diagram of a dual-output Buck-derived power supply with independently regulated outputs. The output (Output, Output2) of each magamp is independently regulated or controlled from an error signal in order to reduce the effective"on-time"of switch S 1 as seen by the filter circuits. In one arrangement, TL431 shunt regulators are used as the controlling devices. These devices supply base-drive to PNP transistors which provide a reset current to control the magamps independently of one another, thereby controlling their respective isolated output voltages.

It is an object of the present invention to provide a magnetic amplifier circuit which seeks to provide the advantages of both the set-and reset-controlled magamp arrangements described above, but to overcome all of the disadvantages. We have devised an arrangement which achieves the following advantages: 1. Full utilisation of the inductor core with controllable flux swings from zero to 2Bs 2. Low control currents are possible because it is possible to use the highly permeable square loop material for the inductor core 3. Low saturated inductance of the"square loop"material 4. Low loss Schottky diodes may be used, despite high leakage currents.

As described above, in the reset-controlled circuit described above, only reset bias of the core can be achieved, whereas in the set-controlled circuit, a controlled current or voltage source is provided which can only provide a set bias to the core. Thus, whichever type of circuit is chosen, there will be some disadvantages, as described above. In other words, there is always a trade-off between the advantages and disadvantages of each circuit, and the type chosen will depend on the particular application.

In accordance with the present invention, there is provided a switching circuit comprising a saturable inductor and input means for providing energy to said saturable inductor, wherein the circuit includes a bi-directional current source capable of providing a set and a reset bias to the core of said saturable inductor.

The circuit of the present invention is particularly useful in a magnetic amplifier circuit, such as a magamp-controlled rectifier circuit.

It is because the bi-directional current source can provide both set and reset bias to the core of the saturable inductor that the advantages of the both the set and reset techniques described above can be realised.

An embodiment of the present invention will now be described by way of example only and with reference to the accompanying drawings, in which: Figure 1 is a circuit diagram of a conventional magamp-controlled rectifier circuit ; Figure 2 is a schematic representation of the BH loop for a"square loop"core material ; Figure 3 is a schematic representation of the typical BH loop of a power ferrite; Figure 4 is a circuit diagram of a conventional Buck Converter, together with a schematic waveforms representing the ideal current in the switch Sl and the diode Dl Figure 5 is a circuit diagram of a conventional"Isolated Output"Buck-derived circuit, Figure 6 is a circuit diagram of a conventional dual-output Buck-derived power supply with independently regulated outputs; Figure 7 is a schematic diagram of a magamp-controlled rectifier circuit according to an embodiment of the present invention; Figure 8 is a complete control circuit diagram of a magamp-controlled rectifier circuit according to an embodiment of the present invention; and Figure 9 is a circuit diagram of a power supply module for connection with the control circuit of Figure 8.

Referring to Figure 7 of the drawings, a magamp-controlled rectifier circuit according to the present invention comprises a transformer, of which only the secondary winding Tla is shown, a saturable inductor Ls, a positively biased diode D1 and a negatively biased diode D2. The rectifier circuit feeds a filter formed of LI and C1 and also a load (not shown).

A controlled bi-directional current source Ic is provided in parallel with the negatively biased diode D2 and the load, and is connected between the saturable inductor Ls and the positively biased diode D1. The bi-directional current source can provide both set and reset bias to the core of the saturable inductor Ls. As a result, the following advantages are associated with the present invention: The topology of the present invention can operate the core of the saturable inductor Ls with approximately zero flux swing (by holding the core in saturation using Ic) to obtain the highest possible output voltage or with any flux swing to 2 x Bs.

Furthermore, the circuit of the present invention has a smaller minimum flux swing than the reset circuit described above, thereby giving a shorter blocking time than would be achieved using the same saturable inductor and the reset control method. As a result, higher frequency operation can be achieved.

The circuit allows a larger peak-to-peak flux swing than the set-controlled circuit described above, thereby allowing a saturable inductor having fewer turns and therefore a lower saturated inductance to be used. Such a saturable inductor has a lower saturated impedance such that higher power levels can be achieved than would be achievable using the same core and the set-controlled method of biasing. In fact, the circuit of the present invention can use a saturable inductor having a low-loss, high permeability "square-loop"core, and still use a diode having a high leakage current, e. g. a Schottky diode, for diode D1 because the leakage current can be shunted away from Ls by Ic.

However, the circuit also allows"non square-loop"materials to be used for the core of Ls because the minimum flux swing is not constrained by Brs.

Referring to Figure 8 of the drawings, a magamp-controlled rectifier circuit according to one embodiment of the present invention is made up of a number of separate modules, as follows: a pulse-width-modulation (PWM) comparator and synchronisation circuit 10, a magamp current sense amplifier circuit 20, an output over-voltage protection circuit 30, an output inhibit circuit 40, an output circuit error amplifier circuit 50, a control error amplifier circuit 60, an output voltage error amplifier circuit 70, and a voltage reference circuit 80. The configuration of each of these circuits is shown in Figure 8. However, a detailed description of each individual component and its manner of operation will not be given here because this would be clear to a person skilled in the art from Figure 8 and the foregoing description. The circuit shown in Figure 8 may be driven by a power supply module such as that shown in Figure 9 of the drawings.

The bi-directional biasing technique of the present invention can be used in any application where ac signals are controlled and switched with saturable inductors (or reactors).

The present invention has been described above purely by way of example, and modifications can be made within the spirit of the invention. The invention also consists in any individual features described or implicit herein or shown or implicit in the drawings or any combination of any such features or any generalisation of any such features or combination.

Claims (8)

1. CLAIMS: 1. A switching circuit comprising a saturable inductor and input means for providing energy to said saturable inductor, wherein the circuit includes a bi-directional current source capable of providing a set and reset bias to the core of the saturable inductor.
2. 2. A circuit according to claim 1, the circuit being a magnetic amplifier circuit.
3. 3. A circuit according to claim 2, the circuit being a magnetic amplifier controlled rectifier circuit.
4. 4. A circuit according to any preceding claim, wherein the current source is provided in parallel with the saturable inductor.
5. 5. A circuit according to any preceding claim, wherein the core of the saturable inductor is formed of a"square loop"material.
6. 6. A circuit according to claim 5, wherein the"square loop"material is an amorphous metal.
7. 7. A circuit according to any preceding claim, comprising one or more Schottky diodes.
8. 8. A magnetic amplifier circuit substantially as herein described with reference to Figures 7 to 9 of the accompanying drawings.