GB2366005A

GB2366005A - Apparatus and method for controlling a magnet

Info

Publication number: GB2366005A
Application number: GB0005731A
Authority: GB
Inventors: Vincent William Lynch; Vadim Volkovoy; Christopher Neil Fr Waterhouse
Original assignee: Thermo Onix Ltd; Onix Process Analysis Ltd
Current assignee: Thermo Onix Ltd
Priority date: 2000-03-10
Filing date: 2000-03-10
Publication date: 2002-02-27
Anticipated expiration: 2020-03-10
Also published as: GB0005731D0; GB2366005B

Abstract

Apparatus for controlling a magnet comprising a magnet coil L2 includes a search coil L1 which is disposed in a region where it is desired to regulate the magnetic field produced by the magnet. The apparatus includes an amplifier A2 for driving a current in the magnet coil L2 and means A1, R1, C1 to control the amplifier A2 in dependence upon the magnetic field detected by the search coil L1. The apparatus may further include a control means operative to automatically calibrate the apparatus. The magnet may be in a mass spectrometer.

Description

<Desc/Clms Page number 1> APPARATUS AND METHOD FOR CONTROLLING A MAGNE The present invention relates to an apparatus and method for controlling a magnet and particularly, but not exclusively, for controlling the magnet of a mass spectrometer.

Mass spectrometers employ an electromagnet to produce a magnetic field to deflect a beam of charged particles to be analysed. During use, the strength of the magnetic field is adjusted to deflect a species of particle of interest comprised in a beam of particles onto a detector. In order for accurate and meaningful results to be obtained the strength of the magnetic field must be accurately controlled and known, typically to within 0.0001 T.

Conventionally, this requirement has been addressed by providing a Hall probe to measure the strength of the magnetic field. However, there are a number of problems associated with the use of Hall probes. Sufficiently precise and linearised Hall probes are costly. The Hall probe needs to be positioned in the unknown region of the magnetic field which is close to the charged particle beam. This is achieved by positioning the Hall probe in the vacuum chamber of the mass spectrometer to measure the field. This is inconvenient and can lead to contamination of or leakage in the vacuum chamber. The alternative, positioning the Hall probe outside the vacuum chamber, leads to severe mechanical constraints.

It is an object of the present invention to overcome the need for the

X_ use of a Hall probe, and thus the associated disadvantages.

According to an aspect of the present invention there is provided apparatus for controlling a magnet having a magnet coil comprising means for driving a current in the magnet coil, a search coil for detecting a magnetic field produced by the magnet and means for controlling the means for driving a current in dependence upon the magnetic field detected by the coil.

According to another aspect of the present invention there is provided a method for controlling a magnet having a magnet coil comprising the steps of causing a current to flow in the magnet coil, detecting the magnetic field produced by the magnet using a search coil and modifying the current flowing in the magnet coil depending upon the magnetic field detected.

Provision of a search coil overcomes the need for a Hall probe. The search coil is preferably located at or near a region where it is desired to control the field produced by the magnet. Where the magnet comprises a core around which the magnet coil is wound the search coil may advantageously also be wound around the core.

The apparatus is preferably arranged to control a magnet to produce a constant magnetic field. That is, the current flowing in the magnet coil is preferably controlled so as to prevent any change in the field produced by the magnet. The means for driving may comprise an amplifier and the means for controlling the means for driving may comprise an integrator.

To enable the magnetic field to be set to a desired level the apparatus is preferably arranged to allow the field to be increased and decreased. In order that the amount of increase or decrease may be accurately known it is preferred that the apparatus is arranged to allow the field to be increased or decreased linearly with time. In one embodiment the means for controlling the means for driving comprises an integrator comprising an amplifier and provision is made to connect a constant potential to the amplifier thereby to cause the magnetic field to change linearly. Provision may further be made to cause the field to increase or decrease. Such means may comprise a differential amplifier, which may also be the means for driving current in the magnet coil.

The apparatus preferably also comprises a control means which may be a microprocessor. It is preferably arranged to store a value for the magnetic field produced by the magnet and to increase and decrease this value in a corresponding manner to increases or decreases in the magnetic field. The stored value may be indicated on a display.

Further, there is preferably provided a means for calibrating the apparatus so that the stored field value corresponds to the actual field produced, so that the latter can be accurately known. This may be achieved by providing means for detecting whether the potential across the magnet coil is positive or negative, which may be a voltage comparator and by arranging for the control means to perform the following sequence of

operation: increasing the field produced by the magnet so that the field value stored by the control means reaches a value B at which potential across the magnetic coil is positive; linearly decreasing the magnetic field produced by the magnet until the value of the field stored by the control means reaches -13 (equal in magnitude but opposite in direction to B), at which the potential across the magnet coil is negative; whilst the magnetic field is changing from B to -B measuring the time from when the field starts to be reduced until the potential across the magnet coil changes from positive to negative; linearly increasing the magnetic field at least until the potential across the magnet coil changes from negative to positive; whilst the magnetic field is increased measuring the time from when the field starts to increase until the potential across the magnet coil changes from negative to positive; and using the two time periods measured to find the difference between the stored and actual value of magnet field.

This difference can then be used to calibrate the apparatus so that the stored field value corresponds to the actual field value. The magnet is preferably substantially saturated with field when the field is increased to B. The time values are preferably stored and the calculation preferably

performed by the control means.

The above sequence preferably includes an initial step of changing the field from a starting value to a value where the stored field value is -B. This can reduce the influence of hysteresis history of the magnet. The field is preferably maintained at -B for a predetermined period of time, then, when the field is increased to B, it is also preferably maintained at this level for a predetermined period of time. Holding the field at a given value for a period of time reduces the influence of eddy currents.

The control means may provide for a start up calibration sequence during which the above sequence is repeated with increasing values of B until a value is reached at which the magnet is substantially saturated.

Once the apparatus is calibrated, the control means may allow any desired field to be selected. Starting from a known field value the control means may be arranged to calculate the required time for which the field needs to be increased or decreased linearly to reach a desired field, and then adjust the field accordingly.

The invention allows a complex and relatively expensive Hall probe to be replaced with a considerably more economic coil which, by using the above described methods, enables the field to be accurately determined and adjusted.

In order that the invention may be more clearly understood embodiments thereof will now be described by way of example with

reference to the accompanying drawings in which: Figure 1 is a schematic representation of a circuit of apparatus according to the invention; Figure 2 shows the circuit of figure 1 modified to include means for increasing and decreasing magnetic field; Figure 3 is a graph showing how the field may be increased or decreased using the circuit shown in figure 2; Figure 4 shows an alternative improved implementation of the circuit shown in figure 2; Figure 5 shows how a voltage comparator may be added to any of the circuits shown in figures 1, 2 and 4; Figure 6 shows comparative graphs of desired magnetic field and potential across magnet coil during referencing of the magnet; and Figures 7 show various modifications and improvements to the circuits shown in to 10 figures 1, 2 and 4.

Referring to Figure 1 there is shown a schematic circuit diagram of apparatus for controlling the field produced by a magnet, and a magnet. The magnet comprises a magnet coil L2 and a core (not shown) having a gap in which it is desired to control the magnetic field. The apparatus for controlling the magnet comprises a search coil Ll wound around the core of the electromagnet in the vicinity of the gap. Changes in the magnetic field in the gap cause an electrical potential to be induced across the search

coil Ll. The potential is proportional to dF/dt, the derivative with the respect to time of the flux F through the search coil Ll which is related to the magnetic field strength in the magnet gap.

Search coil Ll is connected to an integrator comprising an amplifier Al, resistor R1 and capacitor Cl. The output of amplifier Al is proportional to the magnetic flux F through the search coil Ll, plus some unknown constant. The output of Al is connected to a further amplifier A2 which is used to drive the coil L2 of the electromagnet.

The circuit is arranged so that current flowing in the magnet coil L2 is controlled so that the output of Al is zero and thus that the magnetic flux through search coil L2 remains constant and therefore that the magnetic field in the magnet gap also remains constant.

To enable the magnetic field in the magnet gap to be changed the circuit of Figure 1 may be supplemented with means for influencing the input of amplifier Al and therefore the output of amplifier A2. Such a circuit is shown in Fig.2.

Referring to Figure 2 a resistor R2 is connected to the junction of Rl, Cl and Al and to switches Bl and B2. Closing either BI or 62 connects R2 to either a positive or negative potential, u or -u respectively.

Closing switch Bl connects R2 to the positive potential u, causing a current to flow in resistor R2, and affecting the output voltage from Al. The tendency of the remainder of the circuit is to compensate for this, to maintain a

zero output potential from amplifier Al. This results in the output of amplifier A2 c,'--.3nging to cause a change in the magnetic flux produced by the magnet so that a constant negative potential U(L1) is produced across the search coil L1 . To produce a constant potential across the search coil Ll the magnetic flux through the coil must change linearly with time. The magnitude of the potential will depend upon the rate of change of the magnetic flux through it. This depends on the potential U applied to the resistor R2, Rl/R2 ratio and number of turns N in the search coil L1, i.e. (dF/dt) = (Rl/R2)* u/N. If the time for which Bl is closed is T, then the overall change in the magnetic flux equals (Rl/R2)* u*T/N.

On closing either switch Bl or B2 the circuit will produce a linear change in the magnetic field produced by the magnet, irrespective of the required change in potential across coil L2 to achieve this.

Closing switch B2 will cause the magnetic flux to change in the opposite direction.

Figure 3 shows graphs comparing the potential applied to resistor R2, the resulting change in the magnetic field B and the potential this induces across the search coil L1.

For reasons which will become apparent below it is desirable that the values of u and -u, the potentials used to increment and decrement the magnetic field, are well known so that the rate of change of magnetic field is well known. To obtain this one source of potential, instead of two, can

be used, along with an analogue switch SW1 and a differential amplifier A3 shown in the modified circuit of Figure 4. With the analogue switch SW1 in its first position closing switch 131 will increment the magnetic field and in its second position, will decrement the magnetic field.

It will be appreciated that the circuits shown in Figures 2 and 4 can only change the magnetic field produced by the magnet with respect to some unknown field. In order to calibrate the magnet it is necessary to determine the magnitude of that constant field which typically comprises three components: 1) current in the magnet coil (dB,); 2) residual magnetisation of the magnet core (dBr,,); and 3) external magnetic field.

In practice the magnitude of external magnetic fields is so small as to be negligible, so this component can be ignored. The constant field can therefore be represented as dB=dB,+dBm. dB can be determined using a referencing or calibration procedure involving varying the magnetic field between positive and negative values and monitoring the current in the magnet coil L2.

The current in the magnet coil L2 can be measured by measuring the voltage across the coil. To do this a voltage comparator COM1, shown in Figure 5, is connected to the magnet coil L2. The voltage U across the magnet coil L2 equals Ua + Ur, where Ua is the active

and Ur is the reactive component of U. If current in the magnet coil L2 equals I then Ua = R*I and Ur= L*(dl/dt) = (dF/dt)*NL2,where R is the coil resistance, L the coil inductance, F the magnetic flux in the coil and NL2 the number of turns in the coil.

The magnet control circuit further comprises a programmable digital microprocessor (not shown). This receives the output from the voltage comparator operates the circuit to cause the magnetic field to be incremented and decremented and also generates and stores a 'desired' value for the magnetic field which it increments and decrements simultaneously with the actual magnetic field. This can be achieved because the rate of linear increase and decrease of the magnetic field is accurately known. The stored value may be displayed on a display. The processor is also arranged to allow a desired field to be selected.

dB is therefore the difference between the stored and actual field strength. Once dB is determined the apparatus can be calibrated by subtracting dB from the stored field strength so that it corresponds to the actual field strength.

Referring to Figure 6 the basic procedure for obtaining dB is as follows: the field is increased to a level at which the stored field is B, at which the potential across the magnet coil L2 is positive, and the core of the magnet is substantially saturated with field;

the field is then linearly decreased by the circuit until the stored value is -B (equal in magnitude, but opposite in direction to B) at which the potential across the magnet coil L2 is negative; and whilst the magnetic field is changing from B to -B the processor measures the time T, from when the field starts to be reduced until the potential u across the magnet coil L2 changes from positive to negative, as indicated by the change in output of the voltage comparator COM1.

This process is then repeated in the reverse direction. The magnetic field is linearly increased at the same rate it was previously decreased at least until the potential across the magnet coil L2 changes from negative to positive and the time T2measured from when the field starts to increase to the time when the potential across the magnet coil L2 becomes positive.

Because the rate of change of the magnetic field is linear and constant T, effectively represents the value of magnetic field which has to be subtracted from the field, when the stored value is B, for the potential across the magnet coil L2 to become negative. Likewise, T2 represents the value of magnetic field which needs to be added to the field when the stored value is -13 for the potential across the magnet coil L2 to become positive. dB can therefore be calculated by the formula clB = (Tj -T2)* R/2 where R is the rate of change of field.

This method is able to account for both components, clB, and dBm, of dB because, although clB is constant, the magnitude of its components

changes as the magnet is magnetised. As the magnet reaches saturation the magnitudes of dBm and dB, tend to zero and clB respectively. The difference between the theoretical and actual field values can therefore be attributed entirely to a residual current in the magnet coil L2 and thus calculated using the above method.

This method also takes account of hysteresis in the magnet. Because the measurements are made both with a decreasing and an increasing field the effect of hysteresis will be the same on both T, and T2 and when their difference is found will be cancelled out.

The actual algorithms employed by the processor for referencing the field and for setting desired fields are described further below. First, however, a number of further features of the magnet control circuit are described. These are concerned with reducing error in setting the magnetic field.

There are five main reasons for the actual magnetic field value to differ from the stored and hence desired one. First is a drifting of the magnetic field due to offset voltage and offset current of Al and thermo- induced voltages in L1, SW1.1, SW1.2, R1 and wires connecting Ll to the circuit. Second is an error in setting the magnetic field due to instability of the voltage U, time period T, ratio (131/112), and non zero active resistance of L1. Third is an error of the referencing process caused by domain noise during the process of magnetisation and the presence of ambient magnetic

fields. Fourth is a shift in the magnetic field value due to switching of SWl. Fifth is a non-linear temperature dependent relationship between magnetic flux through the coil Ll and magnetic field in the magnet gap.

1. Reducing drift of the magnetic field.

The offset voltage of Al comprises two components, static offset voltage, which is normally temperature dependent, and input voltage noise. The offset current of Al also comprises two components, static offset current and input current noise. Input offset current I produces equivalent input offset voltage U = R1 * 1. Therefore, to reduce influence of the input offset current the value of the resistor Rl should be reduced until the equivalent input offset voltage (Rl * 1) is much smaller than the offset voltage of Al. Static offset voltage of Al can be compensated by using a differential amplifier Al with the non-inverting input connected to a voltage Vref equal to the input offset voltage. See Fig.7. In practice the offset voltages of Al will differ from one amplifier to another so Vref should be adjustable. To simplify the hardware though Vref is set to a non-adjustable value which will exceed any possible offset voltage of Al under any possible working conditions. The excess of Vref over the offset voltage is compensated by closing Bl for short periods of time. The average voltage shift Us then equals Us = U * (t/T), where t is the pulse duration and T is time between pulses. Care must be taken to keep t and T short enough to not cause any significant deviation in the magnetic field.

To reduce drift of the magnetic field caused by thermo-induced voltages in the coil Ll and wires connecting it to the circuit the coil, wires and connectors should be made out of materials with as low a T.E.C. as possible. Also, temperature gradients in the coils, wires, connectors and the circuit board should be kept to a minimum.

2. Reducing error in setting the magnetic field.

The voltage U applied to the button 131 and ratio (131/112) should be as stable as possible. Error in measuring time T of closing B1 should be kept to a minimum. In practice the coil Ll and the switch SW1 always have non- zero active resistance, which should be added to R1 in calculating R1/R2 ratio. Because it is very difficult to control the resistance of L1 and SW1 a buffer amplifier A3 is added to the circuit. See Fig.8.

3. Reducing the error of the referencing process.

In practice reactive component Ur of the voltage U (1-2) across the coil L2 during change of the magnetic field has a significant value of noise voltage caused by uneven magnetisation of the magnet core, i.e. domain noise. This noise voltage produces significant error in measuring times T1 and T2 in the process of referencing. To reduce error a simple RC circuit based on R3 and C2 is used. See Fig.8. Increasing time constant (R3*C2) reduces the influence of domain noise but slows down the process of referencing. Therefore a compromise in choosing values R3 and C2 has to be found.

To reduce the sensitivity of the referencing process to external mains generated magnetic fields the process should be synchronised to the frequency of the mains power supply.

4. Reducing error caused by switching of SWII.

To reduce parasitic change of the magnetic field due to switching this should be done only when the whole circuit is settled and Al output voltage is as close to zero as possible. Also, the input offset voltage of A2 should be as low as possible. Reducing the value of C1 reduces the error caused by switching of SW1 but increases the error caused by non-zero response time of Al during closing and opening of B1. Therefore a compromise has to be found to keep the overall error to a minimum.

5. Reducing error caused by a complex relationship between magnetic flux through the coil Ll and magnetic field in the magnet gap.

The relationship between magnetic flux through the coil Ll and magnetic field in the magnet gap is non-linear and depends on many parameters including form and position of the sensing coil L1, mechanical stress in the magnet core, temperature of the core, and the hysteresis history of the magnet. To reduce the influence of these parameters temperature compensation, special form of the coil L1, and procedure for reducing the influence of hysteresis are used.

To reduce the complexity of the magnet control circuit power supply and interconnecting cables a single polarity power supply is used. To do this

the output stage of the magnet control circuit modified. See Figure 10. Where A2 has differential output and COM1 has differential input. To reduce. the influence of possible problems due to the connections between L2 a(..., A2, resistors R3 and R4 are connected directly to the leads of L2.

The field calibration and setting procedures will now be described in further detail. The following basic calibration or 'referencing' algorithm is used, 'referencing' because the object is to reference the field to zero. The stored field is set from its current value to -B by changing the actual field. This reduces the influence of the hysteresis history of the magnet on the referencing procedure stability by saturating the whole magnet with a negative magnetic field.

The field is left at the -B value for 100ms. This reduces the influence of eddy currents in the magnet construction on the magnet magnetisation.

The apparatus waits for the mains synchronisation signal to go from 0 to 1. This reduces the influence of mains related external magnetic fields on the referencing procedure.

The field is linearly increased until the stored field value reaches + B and maintained at + B for 1 OOms.

The apparatus waits for the mains synchronisation signal to go from 1 to 0. Change of the polarity of the mains synchronisation is necessary to get actual reduction of the mains influence.

The field is linearly decreased until the stored field value reaches -B and the

time from the start of changing the field until the comparator output goes from 1 to 0 is measured.

The field is maintained at -B for 1 OOms.

The apparatus waits for the mains synchronisation signal to go from 0 to 1. 10) The field is linearly increased until the stored field reaches 0 and the time from the start of changing the field until the comparator output goes from 0 to 1 is measured. (in practice, the comparator output will always go from 1 to 0 before the stored filed value exceeds 0, which is a convenient place to stop.) Next, calculations are made to convert the time difference from step six and step nine into a magnetic field value using the formula dB = (T1- T2)*R/2. Where dB is the difference between stored magnetic field and actual magnetic field, T1 and T2 are time intervals from step 6 and 9 respectively, and R is the field change rate. Then, the dB value is used to correct the current stored value.

A special case of the use of the referencing algorithm is when the apparatus is first switched on. The value of the magnetic field is then uncertain and in practice for a mass spectrometer magnet capable of producing a maximum field of about 1T could be any value from -0.05 to +0.05 T. Therefore several basic referencing procedures with different

values for B are used one after another.

First the stored field value is set to zero, its value therefore differs from the actual magnetic field by maximum of +/-0.05T. A first referencing procedure is then carried out using the above algorithm using 0.6T as value B. This ensures that current through the magnet coil will not reach its maximum value. The calculated dB value is then subtracted from the stored field value, thus reducing the field setting error from 0.05T to approximately 0.005T. The procedure is then repeated using 0.8T as value B, again with simple subtraction of the calculated dB value at the end. The field setting error is now reduced to approximately 0.001T. The procedure is then repeated again using the maximum 1.OT as value B. The field setting error is now reduced down to approximately 0.0001 T. The same procedure is then repeated reducing the error to the required 0.00001 T.

The magnet is then left at the resulting field for 100 seconds and the last referencing procedure repeated again. The resulting dB value is again simply subtracted from the variable containing current field value. This dB value is used to calculate the time period T, described above in relation to the reduction of drift in the magnetic field. T is calculated according to the formula: T = (Bp/dB) * 100sec, where Bp is a magnetic field change 'per each B1 closing pulse'.

Therefore, at the end of the last basic referencing procedure the actual magnetic field in the magnet gap equals the stored field value within

the desired minimum error and the time period T is set to a value which corresponds to minimum drift of the magnetic field.

To keep error in setting the magnetic field to a desired level the referencing procedure has to be repeated regularly. A maximum time interval Tr between the subsequent procedures can be calculated from: Tr < Be / Re, where Be is the maximum allowed error in setting the magnetic field and Re is the rate of drifting of the magnetic field caused by voltage and current noise.

The actual field values used during referencing will depend on the specific magnet being controlled, the aim is that the magnet is substantially saturated when the maximum value for B is used.

The basic referencing procedure described above is also used as the main referencing procedure, during use of the magnet. The only difference from the start-up procedure is that the value dB, before being used for calculating the new period of compensating pulses, is first multiplied by a coefficient less than 1.0. This multiplication reduces the influence of the error of the referencing procedure on the compensating algorithm but increases the error caused by magnetic field drift. Therefore a compromise has to be found. The actual value of the multiplication coefficient depends on the parameters of the hardware used. For an embodiment where 0.5 is used instead of the formula described above in relation to the start up referencing algorithm the following is used: (Tnew-Told) = (0.5*Bp/dB)*Tm

where Tnew is the new period of the compensating pulses, Told is the old period of compensating pulses Bp is magnetic field change per 'each B1 closing pulse' and Tm is the time passed from the last referencing procedure.

Once the apparatus has been referenced the magnetic field can be set to a desired level. The processor is arranged to allow an operator to input a desired field level, Bnew. The processor then calculates the time period Tnew for which the field will need to be incremented or decremented using the formula Tnew=(Bnew-Bold)/R where Bold is the current field value stored by the processor and R the rate of change of field. Depending on the sign of Tnew switch SW1 is then set to increment or decrement the field and the switch B1 closed for time period Tnew.

The apparatus allows for accurate control of the magnetic field using relatively inexpensive components and, in particular, without the need for a Hall probe.

The above embodiments are described by way of example only, many variations are possible without departing from the invention.

Claims

CLAIMS 1 Apparatus for controlling a magnet having a magnet coil comprising means for driving a current in the magnet coil, a search coil for detecting a magnetic field produced by the magnet and means for controlling the means for driving a current in dependence upon the magnetic field detected by the coil.
2. Apparatus as claimed in claim 1 comprising a magnet to be controlled and wherein the search coil is located at or near a region where it is desired to control the field produced by the magnet.
3. Apparatus as claimed in either claim 1 or 2 comprising a magnet to be controlled comprising a core around which the magnet coil is wound and wherein the search coil is also wound around the core.
4. Apparatus as claimed in any preceding claim arranged to be operable to automatically control a magnet to produce a constant magnetic field.
5. Apparatus as claimed in any preceding claim, wherein the means for driving a current in the magnet coil comprises an amplifier and the means for controlling the means for driving comprises an integrator.
6. Apparatus as claimed in claim 5, wherein the means for driving current in the magnetic coil comprises a differential amplifier.
7. Apparatus as claimed in any preceding claim arranged to be operable to enable the magnetic field produced by a magnet being controlled

<Desc/Clms Page number 22>

to be set to a desired level.
8. Apparatus as claimed in claim 7 arranged to be operable to enable the magnetic field produced by a magnet being controlled to be increased or decreased linearly with time.
9. Apparatus as claimed in claim 8, wherein the means for controlling the means for driving comprises an integrator comprising an amplifier and provision is made to connect a constant potential to the amplifier thereby to cause the magnetic field to change linearly.
10. Apparatus as claimed in any of claims 6 to 9 comprising a control means arranged to store a value for the magnetic field produced by a magnet being controlled and to increase and decrease this value in a corresponding manner to increases or decreases in the magnetic field produced by the magnet.
11. Apparatus as claimed in claim
12 comprising a display operative to indicate the magnetic field value stored by the control means. 12. Apparatus as claimed in either claim 10 or 11 comprising a means for detecting whether the potential across the magnet coil of a magnet being controlled is positive or negative and wherein the control means is arranged to calibrate the apparatus by performing the following sequence of operation: increasing the field produced by the magnet so that the field value stored by the control means reaches a value B, at which potential

<Desc/Clms Page number 23>

across the magnetic coil is positive; linearly decreasing the magnetic field produced by the magnet until the value of the field stored by the control means reaches -B (equal in magnitude but opposite in direction to B), at which the potential across the magnet coil is negative; whilst the magnetic field is changing from B to -
13 measuring the time from when the field starts to be reduced until the potential across the magnet coil changes from positive to negative and storing this time period; linearly increasing the magnetic field at least until the potential across the magnet coil changes from negative to positive; whilst the magnetic field is increased measuring the time from when the field starts to increase until the potential across the magnet coil changes from negative to positive and storing this time period; and computing the difference between the stored and actual value of magnet field from the two time periods stored. 13. Apparatus as claimed in claim 12 wherein the control means is arranged to change the field produced by a magnet being controlled from a starting value to a value where the stored field value is -B prior to performing the sequence of operation outlined in claim 12.
14. Apparatus as claimed in claim 13 arranged to maintain the field produced by a magnet being controlled at a constant level for a

<Desc/Clms Page number 24>

predetermined amount of time when the stored field value reaches -13 and B during the sequence of operation.
15. Apparatus as claimed in any of claims 12 to 14, wherein the control means is arranged to repeat the sequence of operation for calibration with increasing values of B.
16. Apparatus as claimed in any of claims 10 to 15, wherein the control means is operable to enable a desired field to be selected by being arranged to calculate the required time for which the field needs to be increased or decreased linearly to reach a desired field from a starting field, and then to adjust the field accordingly.
17. A method of calibrating apparatus comprising apparatus as claimed any of claims 10 to 16 arranged to control a magnet including a magnet coil comprising the steps of: increasing the field produced by the magnet so that the field value stored by the control means reaches a value B, at which potential across the magnetic coil is positive; linearly decreasing the magnetic field produced by the magnet until the value of the field stored by the control means reaches -13 (equal in magnitude but opposite in direction to B), at which the potential across the magnet coil is negative; whilst the magnetic field is changing from B to -B measuring the time from when the field starts to be reduced until the potential across the

<Desc/Clms Page number 25>

magnet coil changes from positive to negative and storing this time period; linearly increasing the magnetic field at least until the potential across the magnet coil changes from negative to positive; whilst the magnetic field is increased measuring the time from when the field starts to increase until the potential across the magnet coil changes from negative to positive and storing this time period; and computing the difference between the stored and actual value of magnet field using the two time periods stored.
18. A method as claimed in claim 17, wherein the magnet is substantially saturated with field when the field is increased so that the stored field value is B.
19. A method as claimed in either claim 17 or 18 comprising an initial step of changing the field from a starting value to a value where the stored field value is -B.
20. A method as claimed in claim 19, wherein the field is maintained at -B for a predetermined period of time, then, when the field is increased to B, it is maintained at this level for a predetermined period of time.
21. A method for controlling a magnet having a magnet coil comprising the steps of causing a current to flow in the magnet coil, detecting the magnetic field produced by the magnet using a search coil and

<Desc/Clms Page number 26>

modifying the current flowing in the magnet coil depending upon the magnetic field detected.
22. A method as claimed in claim 21 comprising modifying the current flowing in the magnet coil to prevent any change in the field produced by the magnet.
23. A method as claimed in claim 21 comprising the step of modifying the current flowing in the magnet coil to cause the field produced by the magnet to increase or decrease linearly with time.
24. Apparatus for controlling a magnet substantially as herein described with reference to Figure 1 of the accompanying drawings.
25. Apparatus for controlling a magnet substantially as herein described with reference to Figure 1 and any or all of the other Figures of the accompanying drawings.