Field of the invention

The present invention relates to a switch controller for the control of a thermally activated superconducting switch of a superconducting magnet. The invention also relates to a power supply in which such a switch controller is integrated, apparatus comprising a switch controller and a method of operating a switch for a superconducting magnet.

Background to the Invention

Superconducting magnets are now used in many different applications, including fundamental research (such as NMR experiments) and medical imaging (particularly MRI systems). Such magnets are designed to provide a very high magnetic field strength, often with high homogeneity, within a working region. Superconducting electromagnets are favoured over resistive electromagnets due to the much higher currents which may be carried per unit cross-sectional area of conductor in superconducting wire. Nevertheless there exist major challenges in operating superconducting magnets since at present the superconductors require a cryogenic environment in which to operate. Thus advanced cryostats and accompanying cooling systems have been developed, these more recently moving away from "wet" systems in which the magnet is immersed in a bath of cryogen, to the more contemporary conductively cooled systems in which a refrigerator (such as a cryocooler) cools the magnet directly through a high thermal conductivity path, the aim being to reduce or eliminate the use of cryogen.

A great advantage of such superconductors is their ability to operate in a "persistent" mode whereby when a supercurrent is caused to flow within the magnet coil (or coils) the provision of a further superconducting connection "across" the coils provides a closed superconducting loop within which, theoretically at least, a supercurrent will continue to flow persistently without the need for external power. In order to initially provide a supercurrent to the magnet coils an external power supply is used which produces a high current as a direct current. During such direct current supply it is not possible for the closed circuit persistent mode to be used and for this reason superconducting switches have been developed which are connected across the magnet coil(s) and provide, in a first mode, a resistive connection across the coils and, in a second mode, a superconductive connection. The change between the modes of the switch is effected by the control of the switch temperature and this is advantageous since there are no moving parts required. The temperature of the switch is typically controlled using a heater which is in thermal communication with the switch, whereby the switch is also cooled by the same cooling system used to cool the superconducting magnet. By controlling the balance between the heat input of the heater and the heat sink of the cooling system it is possible to change the temperature of the switch so as to cause it to switch between a resistive and superconducting (persistent) mode.

Increasing the magnetic field in a superconducting, high field magnet (for example a solenoid), from a zero starting current up to an operational current is a time consuming process known as "sweeping" or "ramping". The rise in field strength is achieved by gradually increasing the current that flows through the magnet from the power supply, an operation which may take a number of hours. During this process it is desired to ensure that the switch is in its resistive mode of operation. This is achieved by the use of the heater. However, it will be understood that superconducting magnet systems are very sensitive to heat inputs and the deliberate introduction of heat to a switch which is necessarily in electrical (and therefore thermal) conductive communication with the magnet carries with it the risk of a "magnet quench" in which the superconducting magnet itself becomes resistive and dissipates its energy as heat. Furthermore, the provision of greater heating power than is necessary adds significantly to the thermal load. There is therefore an ongoing need to improve the control of the temperature of switches in such superconducting magnet systems.

Summary of the Invention

In accordance with a first aspect of the present invention we provide a switch controller for controlling the switch of a superconducting magnet, said switch being adapted in use to provide a superconducting current path across the superconducting magnet when said switch is below its superconducting critical temperature, and to provide a resistive current path across the superconducting magnet when said switch is at or above its superconducting critical temperature; said controller being adapted in use to modulate the power supplied to a heater for controlling the temperature of the switch, in accordance with the heat generated in the switch by a change in the magnetic field of the superconducting magnet, the modulation of the power being so as to maintain the temperature of the switch within a predetermined temperature range above the superconducting critical temperature when providing a resistive current path. We have therefore realised that a significant contribution in terms of power dissipation as heat is provided by the self-induced potential difference across the superconducting magnet coil which is dissipated through the superconducting wire of the switch when it is in its resistive mode (above the superconducting transition temperature). We note here that throughout this description there is a discussion of a superconducting magnet "coil" and it will be understood that the invention is in no way intended to be limited to a single instance of a coil, the term "coil" being used for convenience of illustration and rather being intended to include singular and multiple instances of coils. As will be understood, in general the switch provides a path "across" the magnet in either a superconductive or resistive sense in terms of providing an electrical connection between the respective ends of the magnet coil. Thus, the switch is arranged electrically in parallel with the magnet.

The switch controller may be provided as a separate entity for connection to a superconducting magnet system. However, typically the switch controller is integral with a heater power source which is used for powering the switch heater. In addition, or alternatively, the switch controller may be integral with a direct current power supply for providing the superconducting current through the superconducting magnet. Advantageously, such power sources may be provided commercially for retro-fitting to existing magnet systems. Thus, for example, the power to the switch heater may be controlled by either controlling the power from the integrated switch heater supply in known power supplies; or as an "add-on" to control the heater supply output from older generation power supplies; or as a stand-alone controlled supply for systems where no switch heater output exists in the main power supply. As will be understood, typically a heater current of the order of milliamperes is used to heat the switch whereas the current used in the superconducting magnet coil may be between a few amperes and many hundreds of amperes depending upon the application.

The switch controller is preferably arranged to receive as an input some operational data relating to the system from which the heat input dissipated in the switch may be deduced. This may be in the form of temperature data received from a thermal sensor in thermal communication with the switch (and preferably integral therewith). Another source of operational data may be data relating to the current flowing within the magnet coils, particularly the change in such a current as a function of time. This may be provided by the measurement of various electrical properties or simply from data derived from the output of the power supply for the magnet. A knowledge of the physical parameters which describe the switch in particular (by measurement or calculation) may then be used to quantify the heat dissipation in the switch. A combination of information may be used to control the switch heater accordingly. With a sufficient data sampling rate (of the order of a few seconds or fractions of a second) the heater may be controlled with such precision so as to effect real-time control. The switch controller is typically adapted to receive "live" data from the monitoring apparatus such as a resistance thermometer and is therefore responsive to one or more data inputs representing measured physical behaviour of the system whilst the system is in operation and the heater control is being effected. It is also contemplated however that data describing such behaviour may be derived prior to the operation of the system and therefore that the switch controller operates the heater in response to predetermined behavioural data rather than live (contemporaneous) data. Such predetermined behavioural data could represent modelled behaviour with or without being based upon any form of previous experimental or calibration measurements relating to the system.

Furthermore, with particular advantage to a retro-fit power supply, the controller may be provided such that its control mode is switchable between a number of alternative control methods. This may be effected exclusively in software if desired. The controller may also use a PID type controller in accordance with one or more data inputs so as to control the output power of the heater. The controller may be entirely computer-implemented and controlled by software algorithms. It is inferred from the above that the additional source of heating which is taken into account by the controller arises solely from a potential difference generated by the magnet across the switch wire when in resistive mode. It is appreciated that in practice a further source of heat dissipation may be provided in the case that the switch wire is not fully provided in a non-inductive arrangement and therefore that there may be a heating contribution from induced currents in the switch wire itself.

In accordance with a second aspect of the present invention we provide an apparatus for ramping a superconducting magnet comprising:

a power supply for providing a direct current to a superconducting magnet;

a switch having a superconducting critical temperature above or at which it is resistive, and below said superconducting critical temperature it is superconducting and arranged so as to cause a persistent current to flow within the superconducting magnet when in steady state use, whereby said switch is in electrical communication with the power supply and the superconducting magnet;

a heater in thermal communication with the switch, and;

a switch controller in accordance with the first aspect of the invention, so as to control the heater such that the temperature of the switch is maintained within the predetermined range above said superconducting critical temperature during the ramping of the superconducting magnet.

The switch controller may therefore, in accordance with the second aspect, be provided as part of a more general apparatus, including for example the switch which may be thermally connected to the cooled magnet or cold plate reservoir via a weak thermal link. In principle any type of suitable switch may be used, although typically a switch with a non-inductively arranged superconductive wire, co-wound with a resistive heater wire is preferred. It is further preferred if a resistance thermometer (or other temperature sensor) is integrated within the switch so as to provide accurate data regarding the temperature of the wire.

In general current is provided to the magnet from an external power supply. Since the switch is typically connected in parallel with the magnet, the majority of the current provided by the power supply is directed through the magnet instead of the switch for as long as the switch is resistive (i.e. above "Tcs" as referred to later). When the magnetic field strength generated by the magnet has reached a desired level, the heater may be turned off and the temperature of the switch allowed to drop as a result of thermal exchange with the cooled magnet or cold plate reservoir via the weak thermal link such that the switch cools below Tcs and becomes superconducting. At this point, the power supply is switched off such that a persistent current may flow through the superconducting circuit comprising the magnet and the switch. The switch may now be described as "closed" and the desired steady state operational use of the magnetic field may begin. In practice, after the switch heater is turned off the switch then cools. The power supply for the magnet is ramped down allowing the superconducting switch to gradually take the magnet current. The power supply may only be switched off after the ramping down has been completed.

The apparatus may be provided as part of an entire magnet system itself if the magnet and associated cooling system are also provided. Such a cooling system may include a cryostat and a cooler as a cryocooler (mechanical refrigerator) for providing a cryogenic environment for the magnet. The cryocooler may be used to recondense the cryogen in a "wet" system or may be used to directly cool the magnet using thermal conduction.

In accordance with a third aspect of the present invention we provide a method of controlling a switch of a superconducting magnet, said switch being adapted in use to provide a superconducting current path across the superconducting magnet when said switch is below its superconducting critical temperature, and to provide resistive current path across the superconducting magnet when said switch is at or above its superconducting critical temperature; the method comprising modulating the power supplied to a heater for controlling the temperature of the switch, in accordance with the heat generated in the switch by a change in the magnetic field of the superconducting magnet, the modulating of the power being so as to maintain the temperature of the switch within a predetermined temperature range above the superconducting critical temperature when said switch is providing a resistive current path.

As will be appreciated, the method aims to hold the temperature of the switch within a predetermined temperature range. Preferably the temperature control is effected such that the said temperature "range" is substantially a single target temperature. When considering the heat dissipation in the switch the modulating of the power supplied to the heater is preferably arranged so as to minimize the power supplied to the heater whilst keeping the temperature of the switch above the superconducting critical temperature. In practice a temperature as close to the superconducting transition temperature as possible for the switch will provide the optimal reduction in heat dissipation that is desired.

The invention therefore allows the practical provision of a superconducting switch which is in electrical communication with an intelligent, computer controlled power supply and connected in parallel with the superconducting magnet. For example the temperature of the switch may be controlled through the use of a heater which is in thermal communication with the switch, a control system, and a monitoring apparatus embedded within the magnet power supply to monitor the temperature of the switch. This heater is activated for the period of time in which the magnetic field is increased upon ramping up the magnet to a desired operational level. It is known in the art to operate a switch heater at a constant power when in use. An increasing (i.e. changing) magnetic field through the magnet induces a current to flow through the switch. Whilst the switch is resistive, this induced current will further raise the temperature of the switch due to the associated ohmic heating. A rising temperature in the switch could cause sufficient heat to flow from the switch to the magnet such that the temperature of the magnet is increased. If the temperature of the magnet exceeds Tern (see later) for the magnet it could cause parts of the magnet to quench. Quenching results in the apparatus returning to zero field with a corresponding loss of applications data. This invention provides a means of keeping the switch at a stable temperature when in use. The switch heater control system further provides the added benefit of reducing the amount of heater power which is consumed unnecessarily during the ramping process.

The invention allows control of the temperature of the switch so as to prevent it from rising beyond a predetermined level, such that the amount of heat flowing into the magnet from the switch is maintained at a constant value that is preferably minimised for the particular magnet. When in use, the temperature of the switch may be controlled by means of a computer control system, whereby said control system adjusts the power of the heater thermally coupled to the switch. There are two illustrated ways to intelligently control the temperature of the switch. The first is to use PID control, based upon a monitored temperature, which is embedded into the power supply to adjust the temperature of the switch. There are a number of ways of monitoring the temperature of the switch: a thermometer, a potentiometer or an ammeter are examples of "real-time" methods.

Alternatively, in a second method, the temperature behaviour of the switch over time could potentially be inferred by means of heat load calculations embedded within the software within a computer controlled power supply. This technique inputs the voltage across the normal state, superconducting components of the switch due to the inductance of the magnet and, a further calculation based upon the normal state resistance of the switch enables control of the switch heater power accordingly with no temperature sensor information being required. For each magnet configuration, this may be corroborated for a particular magnet by means of measuring the temperature of the switch in a previously performed experiment. The control system further comprises means to infer the temperature of the switch and to either increase or decrease the power supplied to the heater, as required, to compensate in order to keep the switch within an optimum temperature range. Brief Description of the Drawings

An example of a switch controller and method according to the present invention are now described with reference to the accompanying drawings, in which:

Figure 1 is a schematic circuit diagram showing the electrical arrangement of the example;

Figure 2 shows the mounting of the switch thermally with respect to the magnet; Figure 3 is a graph showing the operation of the example system;

Figure 4 is a flow diagram illustrating the method; and,

Figure 5 is a block diagram of the controller setting out further details.

Description of Preferred Example Figure 1 is a schematic illustration of the example. A direct current power supply 1 is provided for supplying a direct current to a superconducting magnet coil 2 in the form of a solenoid. The power supply 1 is capable of providing a direct current having a magnitude of hundreds of amperes. The magnet coil 2 is illustrated as a simple single coil in Figure 1 since the winding of superconducting magnet coils is a well understood technique. It will be understood that the present invention is suitable for use in principle with any type of superconducting magnet coil which is "ramped" during use to and/or from a steady state current level. The invention may be used with low temperature superconducting magnets and those using high temperature superconductors. It will be further understood that in a practical application of the invention, the superconducting magnet coil 2 is located within a cryostat or equivalent apparatus which provides a stable low temperature environment so as to enable the superconducting coils to remain in use below their superconducting transition temperature. Typically such cryostats may immerse the coils within a bath of liquid cryogen (typically helium) provided by a so-called "wet" system, although more recent designs utilise conductive cooling in which either no cryogen or only a small flow path of cryogen is present. The present invention is suitable for all such different cooling techniques since the aim is to prevent unwanted heat loads and such heat loads are a problem in all such systems. In the case of a bath of cryogen and unwanted heat load causes additional boil-off of gas (thereby requiring more cooling power and, in some cases, potentially causing the loss of valuable cryogen from the system). In the case of mechanically cooled systems the cooling power available in the system is relatively small (particularly at low temperatures such as 4 kelvin) and therefore and unwanted heat is difficult to handle and requires a low input power of the mechanical refrigerator to dissipate.

Returning now to Figure 1 , the superconducting magnet coil 2 is shown as being connected "across" the power supply 1 , such that respective power leads 3 are provided to each of the ends of the coil. A switch 4 is also shown in Figure 1. This is likewise connected across the power supply 1 via the two leads 3 and is connected in parallel with the superconducting magnet coil 2. The switch 4 is formed from a length of superconductor in the form of switch wire 5 which is wound in a non-inductive path and connects the two "sides" of the power supply. The superconductor switch wire 5 is co-wound with an ohmically resistive wire forming a heater 6. The co-winding provides the heater in a non-inductive configuration and ensures a strong thermal link between the heater 6 and superconductor wire 5. A resistance thermometer 7 may also be integrated into the switch 4 such that the temperature of the switch wire 5 may be monitored.

Figure 1 illustrates the electrical operation of the switch in two different temperature regimes (see the "box" enclosed by dashed lines). As is shown in the upper part of the dashed box (see path 6), when the temperature of the switch is below the superconducting critical temperature of the switch ("Tcs"), the switch wire 5 is in a superconductive ("superconductive regime") and provides a "zero" impedance short circuit across the superconducting magnet coil 2. If the coil 2 is carrying a current when the switch 4 is below Tcs then a closed circuit of superconducting material is provided which causes the current to flow endlessly in the coil 2 without the need to receive external power. Such a mode of operation of the coil 2 is called "persistent" mode and is an effectively steady state mode of operation during which nuclear magnetic experiments and measurements such as imaging or spectroscopy may be performed in the magnetic field generated by the coil 2. Conversely, and as shown in the lower part of the dashed box (path 7), when the superconducting switch wire 5 is above Tcs the switch wire 5 behaves as an ohmic impedance and is in a "resistive regime". The transition between superconducting and resistive behaviour in the wire 5 is effectively at a single temperature such that switching between the two regimes is effectively a binary operation without a notable temperature dependent transition.

Figure 2 is a schematic illustration of the arrangement of the switch 1 with respect to the superconducting magnet coil 2. The heater 6 is shown for clarity purposes as adjacent to the switch 4. The switch 4 is weakly linked thermally to the magnet coil 2 in such a way that a different temperature may be effected in the switch 4 and magnet coil 2 during operation without causing an overload to the cooling system (such as a cryocooler) and without causing sufficient heat to enter the magnet coil 2 so as to cause it to "quench" and therefore become resistive. This is achieved by the provision of a common base 8, formed from a highly thermally conductive material (such as copper or aluminium), to which each of the superconducting magnet coil 2 and switch 4 are thermally linked. A strong thermal link is provided between the coil 2 and the base 8. A weak thermal link 9 (such as a length of copper wire) is positioned between the switch 4 and the base 8. When in use the heater 6 may be operated to elevate the temperature of the switch wire 5 above Tcs and the weak thermal link 9 causes a small heat input into the magnet coil 2 via the base 8 allowing the superconducting magnet coil 2 to be maintained by the cooling system below the superconducting transition temperature Tern for the superconducting wire of the magnet coil 2. When the heater 6 is switched off, the switch 4 is cooled via the base 8 and weak thermal link 9 so as to adopt a similar temperature as the superconducting magnet coil 2. The operation of the heater 6 raises the temperature of the switch wire 5 above Tcs and the deactivation of the switch allows the cooling system to cool the switch wire (via the base 8 and weak thermal link 9) to below Tcs. Thus the "resistive" and "superconductive" modes of operation of the switch may be effected by the combination of the cooling system, the heater 6 and the weak thermal link 9. As will be understood, the actual thermal response of the switch and the difference in temperature between the switch 4 and the magnet coil 2 which can be safely sustained is a function of the particular switch, magnet and cooling system designs.

In practice, since the superconducting transition in the switch wire 5 is virtually a point transition as a function of temperature, ideally only sufficient heat is provided by the heater to hold the temperature of the switch 4, and in particular the switch wire 5, slightly above Tcs when the "resistive" mode of operation of the switch is desired. This is however problematical in practice not least since there are thermal lags in the system and different parts of the switch and the magnet coil may each experience temperatures which in practice deviate from the representative temperatures which may be measured.

We have found that a significant problem in the control of the switch is caused during the "ramping" of the superconducting magnet coil 2. The "ramping" of a superconducting magnet is understood to be the manner in which the current in the superconducting magnet coil 2 is changed from an "off condition, in which no current is flowing in the coil, to an "operational" condition in which a desired predetermined current is caused to flow, as is the case for example when the "persistent" mode is effected. Such ramping may be in the sense of an increase in the current in the superconducting magnet coil 2, or a decrease, indeed in general it may be thought of as a change (in either manner) in the current.

It will be recalled that the wire of the switch heater 6 and superconducting switch wire 5 are wound in a non-inductive configuration. When there is a change in the substantial magnetic field generated by the superconducting magnet coil 2, whilst this does not induce a significant current in the heater 6 or switch wire 5, the change in current does, due to self-inductance, generate a significant potential difference across the terminals of the superconducting magnet coil 2 itself. This drives a current through the switch wire 5 since this is connected across the superconducting magnet coil 2 (as well as the power supply 1). When the switch wire 5 is in "resistive" mode above Tcs the current driven through the wire causes localised heating in the switch which is additional to the intentional heating provided by the heater 8. The additional heat input into the switch must ultimately be removed by the cooling system and there is a risk that, if sufficient heat is input into the superconducting magnet coils 2, then they may "quench". A quench is something which should be avoided in so far as is possible since it causes a significant down time in the operation of the system and requires a significant use of energy in returning the system to an operational state. The additional power dissipation in the switch from the coil is dependent upon the rate of change of the current in the coil and therefore this problem is particularly significant at high "ramping" rates.

The graph of Figure 3 illustrates the above-described effect in more detail. In Figure 3 the ordinate is temperature in kelvin (and magnetic field strength in Tesla) and the abscissa is time in hours. Illustrative curves are provided for the switch temperature (Curve A), that is the temperature of the switch wire 5, the strength (Curve B) of the magnetic field generated by the superconducting magnet coil 2, and the temperature (Curve C) of the superconducting magnet coil 2.

As is shown in Figure 3, initially the temperature of the switch 4 is similar to that of the superconducting magnet coil 2, this being denoted T1 in this example which is the operational base temperature of the system.

Referring to the flow diagram of Figure 4, the operation of the switch control begins with the system in this state at step 100. Before the ramping of the magnet can begin the switch 4 is required to be in its resistive mode by attaining a temperature above the superconducting transition temperature Tcs for the switch. Therefore at step 200, a controlled current is driven through the switch heater 6 which causes a rapid rise in temperature of the switch 4. This can be seen in curve A. As the temperature in the switch heater 6 rises in comparison with the base temperature T1 , the heating is counteracted at step 300 by cooling through the weak thermal link 9 until the temperature no longer rises. The power supply 1 is then operated at 400 to begin to drive a direct current of increasing magnitude through the superconducting magnet coil 2. As is illustrated by the curve B, the current in the magnet, in this example at least, is increased generally linearly causing a corresponding increase in the magnetic field strength. The change in magnetic field has an inductive effect upon the superconducting magnetic field coil 2 itself which in turn causes a potential difference to be generated across the ends of the coil 2. Referring back to Figure 1 this potential difference is existent across the (presently resistive) switch wire 5 and therefore a current flows within the wire 5 which would not otherwise flow (since the coil 2 as a superconductor otherwise effectively "shorts" the circuit across the wire 5).

The heat dissipated by the current flowing through the resistive wire 5 can be observed in Figure 3 to produce a climb in temperature. Meanwhile the temperature in the superconducting magnet coils 2 passes through a local maximum as flux pinning effects and the additional heat dissipated in the heater warm the coil 2 via the weak thermal link 9 and base 8. This is shown in Figure 3 as curve C. Such behaviour may be exhibited by a conductively cooled system or a wet system in which T1 is below the boiling point of the coolant cryogen. In the case of a wet system an increase in the heat input to the magnet causes an increase boil off in coolant gas. As illustrated, during the ramping of the magnet, the temperature of the magnet is above the temperature T1. In the case of a conductively cooled magnet this may be particularly problematical since Curve C illustrates an average temperature of the magnet and there may be areas which are local hot-spots which are particularly prone to causing a magnet quench. At the point in time illustrated by X in Figure 3, the control system causes a reduction in the heater 6 current so as to take into account the additional heat load on the system caused by the ramping of the current. The modulating of the heater power may be effected by a number of different control techniques.

In the present example a computer controller is integral with the proprietary power supply 1 (which powers each of the magnet coil 2 and heater 6 independently). The controller takes the temperature of the switch wire 5 from a resistance thermometer in the switch as an input and seeks to modulate the power to the heater 6 so as to maintain the switch wire 5 at a substantially constant temperature. The effect of this control can be seen in the dotted temperature illustrated by Y in Figure 3 where Curve A is modified by intelligent switch control and a relatively constant temperature is attained. The beneficial effect of this control may be seen in the dotted line illustrated by Z in Figure 3 where Curve C is modified and throughout the energising of the magnet, the temperature of the magnet coils drops towards the initial base temperature T1 which substantially reduces the risk of a quench occurring. Figure 3 in particular shows the onset of the switch control at point X. In practice this control is put into effect as soon as the heater had caused the temperature to pass Tcs.

As will be appreciated, a responsive system may use temperature sampling (for example a number of times per second) in order to control the power to the heater 6 in real time. Such real time control is advantageous in that different ramping regimes may be used so as to change the rate of increase (or decrease) of the magnetic field in the coil 2 as a function of time. It is noted here that Figure 3 shows a linear increase for clarity purposes, whereas this is not essential. Indeed, as will be understood, the majority of magnets are ramped using multiple ramp rates so as to limit peak temperatures. Furthermore similar control may be used when decreasing the current on "powering down" the system for maintenance. Returning to Figure 4, once the magnet has achieved its operational magnetic field strength, then at step 600 the ramping is ceased and the power to the heater 6 is taken down to zero. The cooling of the switch 4 through the weak thermal link 9 then occurs and the temperature of the switch drops below Tcs. At this point, with the magnet no longer changing in field strength, the switch "closes" and a persistent supercurrent is established in the coil 2 at step 700, after which the power supply 1 is powered down. At step 800, the magnetic field generated by the coil 2 is used for its intended purpose such as for performing gathering NMR or MRI data or many other applications. Figure 5 shows a schematic block diagram of a controller 10 (enclosed in the dashed box) which may be provided as an integral part of a power supply 1 for a superconducting magnet. The controller uses a programmable software algorithm that allows the system to control the power of the switch heater 6 in order to optimize the temperature at which the switch is run. There are two alternatives provided by the controller depending upon the input information used by the system.

A PID control 1 1 is provided in the upper part of Figure 5. This measures and controls the actual temperature of the switch 4 (and in particular the superconductor wire 5 within the switch) using a Proportional band to set the gain of the control, and an Integral band to maintain the thermal momentum of the system, thus protecting it from overreacting to spurious changes. It is also possible to include a Differential band to optimise the control for speed. All three parameters for these behaviours are software programmable. In the case of the PID control, the Set Point determines the temperature at which the switch should operate when in the resistive mode, which should be configured to be slightly above the superconducting threshold Tcs of the switch 4. In the case where there is no temperature sensor in the switch 4, an algorithm selector 13 allows for a method which provides a constant power output of the switch, based on measured values from the heater power output, and the magnet's inductive voltage. In this case, the algorithm, within a power estimator 12, takes as "switch parameters" known values for the switch resistance, and thermal link (seen as the target power output), to calculate the required power to keep the heat switch just above a desired temperature. The measured parameters are used to calculate the actual power being dissipated by the system. Finally, as will be understood, provided the behaviour of the system is known in terms of the effect of the ramp rate, then for an implemented ramp rate (which may be set rather than actively measured) it is possible to modulate the heater accordingly to take into account the additional heating effect produced, if such ramp rate dependent behaviour of the system has previously been modelled or otherwise evaluated.