The Manchester (e,2e) Experimental Hardware
page prepared by Andrew Murray
Page Updated on 15th March, 2011.
The Vacuum System.
The (e,2e) spectrometer vacuum chamber (figure 1) is manufactured from 304 grade stainless steel, the construction being carried out by Vacuum Generators using a design of Jones, Read and Cvejanovic, full details of which may be found in T. Jones' Ph.D. thesis (University of Manchester (1984)).

Figure 1. The (e,2e) Vacuum System (Simple view). The chamber is constructed of 304 grade non-magnetic stainless steel. The spectrometer hangs from the top flange, as detailed below. The chamber is pumped by a Balzers 500l/s Turbo-molecular pump, which is backed by a 16l/s Edwards low acoustic noise roughing pump. The Chamber is lined internally with 3mm thick mu-metal, and a mu-metal cylinder encloses the chamber externally as shown. This reduces the magnetic field at the interaction region to less than 5mG, an acceptable level for the electron energies used in this experiment.
The experiment which includes
the electron gun,
the electron energy analysers &
the target source
is mounted from the top flange of the vacuum system via four support struts as shown in figure 2.
3mm thick mu-metal magnetic shielding is placed both internally and externally to decrease extraneous external magnetic fields to less than 5mG at the interaction region.
The system is pumped by a Balzers 500l/s turbo-molecular pump backed by an Edwards 16l/s roughing pump, and achieves a background working pressure better than 2.0E-8 torr after 1 week of pumping while baking the system to a temperature of 90°C.
All inputs and outputs to the spectrometer pass through the vacuum chamber via feedthroughs on the top flange.
The Top Flange.
The top flange of the vacuum system is a 25mm stainless steel flange with 11 conflat CF70 flanges arranged around a central 200mm flange designed for locating a hydrogen source.
The CF70 flanges are used for input of electrical signals, the gas target beam and vacuum monitoring systems:
Two flanges are used for the high voltage channeltron and photomultiplier tube inputs and pulse outputs.
A 52 way electrical feedthrough is used for the electron gun voltages
Two CF70 flanges accommodate 19 way electrical feedthroughs for the two analyser voltages.
Three rotary motion feedthroughs allow the analysers and electron gun angles to be changed externally via stepper motors for both analysers and for the electron gun.
A single flange is accessed to rough out the vacuum system prior to engaging the high vacuum pump, whereas
Another CF70 flange accommodates an ionisation gauge.
The ionisation gauge is placed at right angles to the main flange to prevent light from the thoriated iridium filament entering the vacuum system, which would increase noise on the photomultiplier tube.
The ionisation gauge feedthrough is sooted, and a loose cover is placed over the entrance in the chamber flange to prevent stray light from the ion gauge entering the vacuum system.
The target gas enters the system via a CF70 flange which has two 6mm stainless steel tubes welded into the flange. One of these is internally connected to a gas jet which can be set at 45° to the incident electron beam and is used in the perpendicular plane experiments.
The other gas feed is connected to a hypodermic needle set at 90° to the incident beam for measurements in other geometries. These stainless steel feeder tubes are connected to the nozzle using grounded shielded PTFE hose.
The gas jet which is not currently used in the interaction region can be used to non-locally fill the chamber with gas for background counting measurements.
The final CF70 flange is used when the photomultiplier tube is cooled, usually during baking of the system following exposure to atmosphere. This is necessary since the bi-alkali photocathode of the photomultiplier tube degrades rapidly at temperatures in excess of 55°C, and baking of the vacuum system is most effective for temperatures closer to 90°C (this is the maximum that can be obtained with the present heating tapes).
The tube is cooled by passing vapour from a liquid nitrogen source through expansion coils wrapped around the photomultiplier tube housing.

Figure 2. The Spectrometer mounted from the Top Flange. The spectrometer is shown configured in the perpendicular plane, the electron gun being placed in the vertical position with respect to the analysers which span the horizontal detection plane. The electron gun is a two stage non-energy selected gun that produces an incident electron beam current of up to 4 microamps at an incident energy from around 30eV to 300eV. The analysers are identical, and consist of a 3-element electrostatic lens focussing the interaction region onto the entrance of a hemispherical electrostatic energy analyser. The electrons selected in angle and energy are amplified using Mullard 919BL channel electron multipliers. The interaction region defined by the interaction of the gas target beam and the electron beam is focussed onto the photocathode of an EMI 9789QB photomultiplier tube via a lens, 450nm±50nm optical filter and defining aperture. The target gas effuses from a 0.6mm ID hypodermic needle constructed from platinum iridium alloy. The electrons passing through the interaction region which are not scattered are collected in a Faraday cup opposite the electron gun. The gun, photomultiplier tube, gas hypodermic and Faraday cup all rotate on a common yoke with an axis through the centre of the horizontal detection plane, allowing all geometries from the perpendicular plane to the coplanar geometry to be accessed.
The Electron Gun.
The electron gun is a two stage non-selected gun designed by Woolf (Ph.D thesis, University of Manchester) (see figure 3).
The gun consists of a filament, grid and anode source which is focussed using a triple aperture lens through two 1mm defining apertures in the centre of the gun.
The apertures define the electron beam pencil and beam angles, and are focussed onto the interaction region using a second triple aperture lens located between the apertures and interaction region.
There are three sets of deflectors located in the gun. One follows the anode, another corrects the beam direction between the apertures and a third steers the beam onto the interaction region.
11 voltages therefore control the electron beam from the gun.
The electron gun is shielded using advance (constantin) sheet which is grounded to the body of the gun mounting. This shield is coated with aerodag colloidal graphite to reduce patch fields on the surface of the metal, and is made as small as possible to allow the analysers to approach the gun as close as is feasible when non-perpendicular plane geometries are selected (see figure 2).
The gun is mounted onto an x-y translation table to allow placement of the electron beam as accurately as possible to the centre of rotation of the analysers and electron gun rotation axis. The XY translation table is mounted from the rotating yoke, to which is attached the photomultiplier tube.
The gas jet is mounted from the electron gun using a narrow support of 310 stainless steel, and the Faraday cup can be attached to this support.
The original unselected electron gun built by Woolf was completely rewired and a custom built PTFE 25 way inline wire-wrap plug and socket was placed between the 52 way vacuum electrical feedthrough and the electron gun. This allows the gun to be dismounted from the system without the need for rewiring of the feedthrough when the need arises (filament changes, cleaning etc).
The wiring from the 52 way feedthrough to the inline plug is shielded colour coded 30 awg PTFE coated advance wire, and the inline plug and socket connectors are non-magnetic gold plated wire wrap pins. The wiring between the gun and inline socket is also shielded colour coded PTFE coated advance wire facilitating spot welding to the electrostatic lens elements of the gun.

Figure 3. The unselected electron gun. The electron source is a heated tungsten hairpin filament heated by passing around 2A of DC current through the filament. The Grid/Anode extraction region roughly collimates electrons from the filament, this beam being focussed using lens 1 onto the apertures in the field free region. The Anode deflector pair corrects for any misalignment of the filament emission point with respect to the axis of the electron gun. The defining apertures limit the pencil and beam angles of the electron beam following focussing by lens 2 onto the interaction region. Placement of the apertures ensures that the beam angle is close to zero degrees at the interaction region over a wide range of energies, whereas the aperture opening reduces the pencil angle to around ±2 degrees over this energy range. The deflectors between the apertures correct for space and surface charge beam variations inside the field free region, whereas the final deflectors allow the beam to be steered onto the interaction region defined by focussing a 1mm aperture in front of the photomultiplier tube onto the gas target beam.
The Electron Energy Analysers.

Figure 4. A photograph of the electron energy and angle selecting analysers. For details, see the text.
The electron analysers are constructed from molybdenum and are mounted from two rotating gear plates controlled externally via the rotation feedthroughs on the top flange (see figure 2). The analysers are mounted onto x-z translation tables that allow them to view the interaction region accurately over the full range of angles accessible.
The input electrostatic lenses for each analyser are three element cylindrical lenses as shown in figure 4. These lenses have a capability for energy zooming of around 10 : 1, and have a maximum entrance acceptance angle of ±3° as defined by the entrance apertures.
The interaction region is focussed onto the entrance aperture of the hemispherical energy analysers with a beam angle of around zero degrees, the image of the analyser entrance aperture being larger than the size of the interaction region. This ensures that the overlap volume accepted by each analyser encompasses the interaction volume for all detection angles as the analysers move in the detection plane.
The energy analyser consists of two molybdenum hemispheres with a mean radius of 25.4mm.
Field gradients around the input and exit apertures are corrected using standard Jost correctors.
The energy selected electrons passing through the exit apertures are finally collected by Mullard X919BL channel electron multipliers which amplify the single electron to produce a current pulse at their output. These devices use an activated lead glass internal surface which has a high coefficient of secondary electron emission when a high voltage (around 3000V) is applied across the multiplier. The saturated gain of these devices yields around 1E8 electrons in the output current pulse for a single electron which enters the input cone.
The Channel Electron Multiplier Pulse Electronics
The output current pulse from the multiplier of approximately 1E8 electrons in a time of around 10ns is sent along doubly shielded Belden RG188A/U silver plated copper coaxial cable before being dropped across a 50W load external to the vacuum system, thereby creating a voltage pulse at the input of a preamplifier.
A principle source of noise on the signal pulses arises from the turbo-molecular pump motor. This noise may be picked up by the channel electron multiplier and the high voltage feed to the multiplier, and is generated internally in the vacuum system.
Adequate precautions must therefore be taken to reduce this pickup. This is accomplished by carefully shielding the analysers in a grounded advance housing, and by using doubly shielded PTFE coated RG188A/U cable from the high voltage electrical feedthrough to the multiplier. Shields are also placed over the feedthrough inputs to help reduce pickup.
The noise from the turbo pump is reduced to ~ 5mV using these techniques, and as the average pulse height from the saturated multiplier is around 20-40 mV, this allows adequate discrimination following amplification.
The pre-amplifiers located external to the vacuum system are commercial 100X preamplifiers with 400pS rise-times, manufactured by Phillips Scientific.
The input and output impedance of these amplifiers is 50W, matching the RG188A/U internal feeder cable and the doubly shielded RG58A/U coaxial cable connecting the preamplifiers to ORTEC 473A constant fraction discriminators located in separate NIM bin crates.
Separate NIM crates are used to increase isolation and reduce common mode pickup. The CFD's are operated in NaI mode as this has a 2ms dead time following the initial pulse, thus reducing the possibility of false triggering due to any mismatching impedance reflections along the delay cable.
The NIM signals produced by the discriminators drive two ORTEC 437 Time to Amplitude Converter's via appropriate delay cables and an ORTEC pulse delay unit.
The main TAC output feeds an ORTEC Multichannel Analyser card located in the PC controlling the experiment, whereas the second monitor TAC drives an external MCA, allowing convenient monitoring of the coincidence signal as the experiment progresses.
The high voltage supplies to the electron multipliers are two Brandenberg 5kV supplies controlled by the main PC via serial DAC's. These EHT supplies are connected via VERO 200 high voltage couplings to high voltage cable connecting to the pickoff circuitry located on the high voltage feedthroughs.
A 12dB/octave passive RC Butterworth low pass filter is placed between the high voltage supply and the input HT feedthrough. This serves three purposes.
Principally the filter is used to reduce pickup noise from entering via the HT cable onto the multiplier output.
Secondly, the filter has a slow response time for high voltage spikes, and reduces the probability of damage to the pulse electronics should the EHT supply suddenly turn on or off.
Finally, the 1nF 6kV capacitor in the filter prior to the 1kW load resistor acts as a short circuit to ground for the CEM pulse current.
A schematic of this filter is given in figure 5.

Figure 5. The channel electron multiplier pickoff and filter circuitry. See text for details
The EHT line is decoupled from the pulse electronics using two 1nF 6kV capacitors.
Capacitor C2 decouples the 1kW load resistor to earth for high frequency pulses from the CEM's, and the voltage appearing across this load is connected via the capacitor C3 to the input of the 6954 preamplifier.
Since the input impedance of the preamplifier is 50W and the CEM is effectively an ideal constant current supply, most of the current passes through to the preamplifier. The effective load presented to the multiplier is therefore ~ 50W, matching the impedance of the internal RG188A/U feed and thereby reducing mismatching reflection problems.
Finally, the 10M resistor prevents the output to the preamp from rising above ground when the preamp is disconnected from the circuit and is therefore a safety device for the preamplifier when initially connected.
The Target Beam Source
Two configurations are used for target source:
The original design placed the atomic beam nozzle and Faraday cup on the same support structure fixed to the body of the electron gun, with the atomic beam direction being in the detection plane when the perpendicular plane geometry is chosen (see figure 2).
A second configuration is adopted for exclusively perpendicular plane coincidence experiments, the atomic beam nozzle being oriented at an angle of 45° to the detection plane so that the gas effusing from the nozzle does not fill the analysers when they oppose the nozzle (see figure 6).
In this configuration the support strut for the atomic beam nozzle is terminated lower than the detection plane, as is the nozzle itself. This allowed the analysers to roam the detection plane over a complete 360° angular range.
For non-perpendicular plane measurements the Faraday cup is once mounted from the support strut and the gas effuses perpendicular to the electron beam direction. The angular movement of the analysers is therefore restricted to the limits ±60° when the gun angle is between 90° (perpendicular plane) and 65.6°, and -60° to +35° when the gun angle is set smaller than 65.6°.

Figure 6. The configuration of the target source hypodermic when perpendicular plane measurements are conducted. This allows the analyser to roam over a complete 360 degrees of the detection plane when in this geometry (apart from the position where the analysers would clash wth each other).
The Photomultiplier Tube.
One difficulty when aligning an (e,2e) coincidence experiment lies in the focussing the incident electron beam and analysers onto the interaction region. The common procedure is to initially maximise the current from the gun onto the Faraday cup, then onto analyser counts as the electron beam is moved through the gas jet. The analysers are then adjusted for maximum signal count and the procedure is iterated until a maximum is found.
There are several problems associated with this technique, and one major disadvantage :
The throughput of electron current as detected by the Faraday cup is insensitive to both position and focusing of the electron beam, since the cup opening is usually made quite large (in this case 6 - 10mm) and the bias voltage applied to the cup to prevent electrons returning to the interaction region allows all electron trajectories within this opening to be detected. Maximising the Faraday cup current is therefore only a crude preliminary alignment of the incident electron gun beam.
Conventionally, following maximisation of the Faraday cup current, the beam is steered and focussed onto the gas jet as detected by counts in the analysers. This therefore assumes that the analysers are initially focused onto the interaction region, and a slow iterative technique must be adopted where the analysers and electron gun are tuned so as to maximise the signal from both analysers.
This has the disadvantage that the overlap volume is maximised for both atomic and electron beam density, which tends to place the interaction volume as close as possible to the output of the atomic beam nozzle, which is almost certainly not in the centre of the detection plane. In this case a true differential cross section is not obtained, since the scattered and ejected electron trajectories lie in a different plane than defined in the initial experimental alignment.
To eliminate the need for time consuming iterative techniques and the associated errors due to detection geometry, a separate method for tuning the electron gun is adopted that does not depend upon the analysers. An EMI 9789QB reduced photocathode photomultiplier tube is installed in the vacuum system together with optical focussing and defining elements (figure 2).
The interaction between the target and the electron beam not only produces ionisation, but the valence states of the target also are excited with various probabilities depending upon the incident energy of the electron beam. A group of these excited states decay back to lower valence states with the emission of visible photons at wavelengths around 450nm for both helium and argon targets.
These photons are detected by the photomultiplier tube, by placing a 450nm ± 50nm optical filter in front of the photocathode.
The interaction region is therefore defined where an aperture in front of the photocathode is focused by the lens.
This is established by placing a light source (i.e. a torch bulb) in place of the photocathode behind the defining aperture and adjusting the lens position and focus so that the image of the defining aperture is positioned exactly where the analysers, electron gun and gas jet overlap (figure 6). The size of the defining aperture is chosen to yield an interaction volume of ~1cubic mm.
The plano-convex focussing glass lens is covered by a grounded fine tungsten mesh to eliminate surface charging. Tuning of the electron gun to PMT counts therefore guarantees that the electron gun is steered and focussed to the point at the centre of the detection plane independent of the analysers or gas beam density.
The photomultiplier tube may be cooled by circulating nitrogen gas boiled off LN2 through copper coils encircling the photomultiplier tube housing. This is mainly used when the vacuum system is baked at temperatures above 55°C, since the photocathode is rapidly damaged by temperatures exceeding this. The dark noise count is also reduced when cooled, however since the reduced cathode PMT only produces dark counts of less than 20Hz at room temperature, this is not necessary during routine operations.
The Photomultiplier Tube Electronic Circuitry.
The photomultiplier tube used in the experiment is a 9789QB EMI photomultiplier tube with a bi-alkali photocathode peaking in the blue region of the optical spectrum. The tube is a venetian blind dynode construction, and has a reduced photocathode to decrease the noise on the output of the tube to less than 20Hz.
As with the electron multiplier circuitry, it is important to optimise the load of the photomultiplier tube so that reflections along the feeder line to the counting electronics are minimised.
Figure 7 shows the biasing arrangement adopted for the dynode array which maximises saturation of the output current from the tube into the 50W load resistor used to match the RG188A/U line.
The 1nF 6kV capacitors across the dynodes close to the output stage are used as charge storage pumps that quickly refurbish the surface charge on the dynodes, allowing increased counting rates.
The dynode resistors are 1% tolerance metal film resistors as these are more stable with temperature.

Figure 7. The photomultiplier tube bias network. The 1.82M resistor was chosen to optimise the single photon counting efficiency of the tube. The 470k resistors bias the dynodes to achieve saturation of the current pulse at the output of the tube, hence biasing the tube for single photon counting applications. The HT supply is through RG188A/U PTFE coated silver plated 50W coaxial cable, as is the feed to the discriminator external to the vacuum system. The load is set by the 50W resistor across the anode to the HT feed, the pulse being decoupled to ground via the 1nF capacitor. This load ensures that any pulse reflections along the RG188A/U transmission feeder are properly damped, thereby ensuring as fast a current pulse as possible (typically 5ns wide). The 10MW safety resistor prevents the output charging up to a high voltage if the feed is disconnected.
The EHT supply to the photomultiplier tube is a Brandenberg 2kV supply operating at 1500V for this tube.
The negative going output pulses from the tube measure typically 200-500mV in height, and therefore are able to directly drive an ORTEC 473A constant fraction discriminator located in one of the NIM crates.
The HT voltage enters the system via RG-59 cable through a high voltage feedthrough. The internal feed to the PM tube and the internal feed from the pulse circuitry are doubly shielded PTFE coated RG188A/U coaxial cables, terminated in standard aluminium BNC connectors on the photomultiplier tube housing.
All resistors and capacitors for the dynode chain are soldered to the photomultiplier tube base internal to the vacuum system, using standard 5-core solder. The photomultiplier tube base was thoroughly cleaned in an ultrasound following location of these components, to displace any flux residue from the solder joints. It is not found that these solder connections significantly affect the ultimate vacuum pressure in the system.
The External NIM Pulse Electronics.
Typical pulse heights from the electron multipliers in saturation mode are 25-40mV, whereas the typical height of the photomultiplier tube pulses is of the order of 250mV. It is therefore necessary to amplify the channeltron pulses prior to transmission to the discriminators, whereas the photomultiplier tube pulses are sufficiently large to not require further amplification.
The electron multiplier pulses are obtained from the multiplier supply circuitry as explained above. These pulses are amplified by Phillips Scientific 6954 100X amplifiers located external to the EHT feedthroughs connected by 10cm lengths of RG-58 cable. Isolated DC supplies to these amplifiers is obtained from a standard +15V supply, the circuit diagram being given in figure 8.

Figure 8. The 6954 Phillips Scientific preamplifier isolated DC supplies (2 off)
The amplified pulses from the 6954 amplifiers pass along doubly shielded RG-58 cable to the input of the ORTEC 473A discriminators. Similarly the PMT pulses pass to a third ORTEC 473A discriminator.
The two analyser discriminators are located in separate NIM crates to increase isolation and reduce common mode pickup. These crates are further decoupled from the AC mains supply using filter circuitry on each supply line.
The slow rise-time positive outputs from the 473A discriminators pass to three ORTEC 441 rate-meters which allow counting and monitoring of the pulses. Internal to these rate-meters buffer 7400 TTL chips have been added to send to 32 bit counters located in the main PC via RG-58 cable. The power supply to these TTL buffers is derived from the rate-meter supplies via 7805 5V regulators located internally.
The fast NIM pulses from the analyser 473A CFD's pass through appropriate delay lines to two 437 TAC's.
The output of one of these TAC's s sent to an ORTEC multichannel analyser located on the main PC bus, whereas
the second TAC output is sent to an external Tracor Northern multichannel analyser.
This second multichannel analyser allows the operator to monitor the total accumulated coincidence signal during operation.
Figure 9 shows a schematic of the external NIM pulse circuitry used in the experiment.

Figure 9 The External NIM Pulse Circuitry. See text for details.
The Computer Controlled Electrostatic Lens Supplies.
The lens supplies that control the voltages impressed onto the lens elements and deflectors in the experiment were developed over a number of years, and therefore take different forms. Only a brief description of these supplies is given here.
The fully computer controlled deflector supplies are housed in a separate 19" crate.
Three of these supplies feed the deflectors in the electron gun, the other two supply the input deflectors in the analysers. These supplies are fully floating, their mid-point voltage being equal to the lens element voltage in which the deflectors are housed.
Typically these supplies can vary between ±15V with respect to their midpoint voltage.
In a second 19" crate is located the gun lens high voltage supplies together with the high voltage supplies for the Faraday cup. All these supplies are fully computer controlled isolated units with low output impedance to reduce noise and eliminate current loops.
In a third 19" crate is located the analyser voltage supplies together with a set of serial loading supplies that control the electron multiplier EHT supplies.
A further supply controls the current boost circuitry located in the filament constant current supply unit (see below).
The Spectrometer Voltage Interface
The outputs of each computer controlled supply that drives an individual lens element or deflector are connected to a 70 relay switching board. This relay switching unit allows either the voltage or the current on each lens element to be measured by a Keithley Digital voltmeter.
The Digital voltmeter is in turn accessed via an IEEE-488 interface which is addressed by an 8086 slave PC which converses with the central PC which controls the experiment.
The outputs from the relay board connect to the spectrometer via shielded multicore cables which take the gun voltages to the 52 way feedthrough and the analyser voltages to the two 19 way feedthroughs located on the vacuum flange.
The gun supply multicore also carries the filament supply current to the 52 way feedthrough.
The Faraday cup current is measured by the Keithley Digital voltmeter when under computer control but this can also be measured by an external pico-ammeter. The bias supply to the Faraday cup is via the 52 way feedthrough on the vacuum flange.
The Stepper Motor Supplies and Interlocks.
In addition to the analyser electrostatic lens and deflector voltage connections, the 19 way CF70 feedthroughs are used to control signals from internal sensing opto-isolators which control the stepper motor interlocks.
The stepper motors that drive the analysers and electron gun around the detection plane are five phase motors with integral 10X reduction gearboxes driving the rotary motion feedthroughs. This drive is further reduced internally by 10X reduction gears, resulting in a reduction ratio of 100:1.
The intelligent stepper motor drivers are addressed via a serial port located on the controlling PC bus.
The stepper motor drives are contained in a separate 19" rack. An internal unregulated power supply in this unit is dedicated to driving the motors. This supply is separate from the regulated supply that drives the controlling logic. This was found to be necessary to eliminate problems associated with current loops from the 20A peak motor current feeding the common point of the logic circuitry.
An additional interference problem was traced to the RS232 serial line from the controlling PC.
The PC bus was found to be poorly earthed due to very little copper being used for the bus earth. This led to the ground line of the RS232 serial line moving around with respect to the 0V stepper motor reference as drive current to the motors returned along the ground line.
This was cured by decoupling the 0V RS232 line from the main 0V line for the stepper motors, as is shown in figure 10.

Figure 10. The Stepper motor Drive Power Supply and decoupling circuitry required to decouple the stepper motor ground lines from the logic circuitry and RS 232 serial interface to the controlling PC.
The analyser stepper motor supplies have a facility for halting the motors using limit switches which are optically coupled to the internal 68000 logic circuits driving the supplies.
These limit switches were exploited to prevent the analysers from
(a) running into each other,
(b) running into the electron gun when it was in the way of the analysers,
(c) running into the Faraday cup when it was in the line of the analysers and
(d) running into the gas jet support when the gun angle was set from 65° to 90°.
These interlocks were controlled by photodiode/phototransistor coupled pairs inside the vacuum system.
The SD5443-3 phototransistor normally illuminated by the TEMT88PD photodiode during operation ensured a fail-safe mechanism should any of these components fail (this automatically prevent the motors from running).
Should the analysers move into a position as defined by the four criteria (a) - (d) above, a shutter located on the analyser turntables moves between the appropriate photodiode and phototransistor, shutting off the light and thereby shutting down the stepper motor current.
Only motion in the opposite direction is allowed by controlling logic, allowing the analyser to move away from the point of imminent collision but no closer.
The six pairs of optocouplers that prevent the analysers moving into the Faraday cup, electron gun and gas jet support are located on the mounting struts supporting the main body of the spectrometer.
Two optocouplers are located on the side of analyser 2, and two shutters are located on the sides of analyser 1 to prevent the analysers colliding with each other.
Finally a ninth optocoupler is located over the electron gun counterweight which monitors when the gun is at an angle less than 65°, where the sweep angle around the detection plane is reduced.
Figure 11 is a schematic of the logic circuit used to determine the optocoupler status.
a
Fig 11. The optocoupler logic circuit circuit that controls the stepper motors and avoids collisions between the analysers, the electron gun, the Faraday cup and the hypodermic support. See text for details.
The Filament Constant Current Supply and Current Boost Circuitry
The filament constant current supply is a standard 5A constant current supply. Added to this filament supply is a computer controlled current boost circuit that can be used to increase the filament current during electron gun tuning. This supply boosts the current without changing the main supply, thus ensuring supply stability before and after the electron gun is tuned.
The current boost circuit operates from a serial driven supply card as detailed previously. Circuit details for this supply are given in figure 12.
The main filament drive is through six power diodes that feed the current from the constant current supply to the filament. These isolate the constant current supply from the boost supply.
The Darlington transistor T1 feeds additional current to the filament via three parallel 10 turn 100W trim resistors through a 1N4007 diode. The additional current then returns to the 12V DC regulated supply via a second 1N4007 diode.
As the constant current supply acts as an infinite resistance source, the boost circuit supplies additional current only when the voltage at the anode of the feed 1N4007 exceeds the output voltage of the constant current supply. This therefore depends upon both the current driving the filament, as well as the resistance of the filament when hot (typically 4W - 6W).
Finally, the filament bias voltage (and hence energy of the electrons) is set to the midpoint of the filament by two 100W balancing resistors located across the terminals.

Figure 12 The Current Boost Circuit used during tuning of the electron gun. This supply feeds current to the filament without changing the characteristics of the main constant current supply used while operating the spectrometer. This supply is operated only when the signal from the photomultiplier tube is very weak (eg when the (e,2e) cross section is large, but the excitation cross section yielding 450nm spontaneous emission radiation is small).
The Controlling Computer and Bus Electronics
The controlling PC contains the main control logic for the experiment. On board is
the ORTEC multichannel analyser card,
a Blue Chip ADC board which interfaces to a 24 channel buffer card,
a timer counter board which counts the analyser and PMT count rates and
custom built parallel interface and serial interface cards which control the power supplies for the experiment.
In addition to the ADC buffer card, three independent isolated regulated supplies drive REF01 voltage reference sources for measurement of the positions of 10 turn potentiometers located on the analyser and electron gun rotary drive shafts. These high stability reference sources ensure that the analyser and electron gun positions are accurately measured with time.
The angular position of the electron gun and analysers is calibrated when the system is opened using visible laser diodes that define these angles to an accuracy of ±0.2°. A datalogger internal to the controlling PC establishes a cubic spline relationship between the measured angles and the corresponding ADC conversion from the potentiometers.
The 8086 Slave Computer and IEEE Interface
The 8086 slave PC addresses the Keithley voltmeter via a standard IEEE interface. This PC also controls the selection of the relay motherboard for measurement of the required lens element either in voltage or current mode when a request is sent from the controlling computer via an RS232 serial line.
The Keithley DVM connects to the relay motherboard in two ways.
Current measurements are sent directly to the Keithley current input, whereas
voltage measurements connected via a 100MW impedance matching circuit to the voltage input of the DVM. This ensures minimum deviation of the supply voltage when measurements are being made.
Figure 13 shows an overall block diagram detailing the computer control and high speed pulse interfacing to the spectrometer.

Figure 13. The Hardware Interface to the (e,2e) spectrometer, showing the logic and voltage supply connections to the central controlling PC.
The Manchester (e,2e) Experimental Hardware page prepared by Andrew Murray Page Updated on 15th March, 2011. The Vacuum System. The (e,2e) spectrometer vacuum chamber (figure 1) is manufactured from 304 grade stainless steel, the construction being carried out by Vacuum Generators using a design of Jones, Read and Cvejanovic, full details of which may be found in T. Jones' Ph.D. thesis (University of Manchester (1984)). •Figure 1. The (e,2e) Vacuum System (Simple view). The chamber is constructed of 304 grade non-magnetic stainless steel. The spectrometer hangs from the top flange, as detailed below. The chamber is pumped by a Balzers 500l/s Turbo-molecular pump, which is backed by a 16l/s Edwards low acoustic noise roughing pump. The Chamber is lined internally with 3mm thick mu-metal, and a mu-metal cylinder encloses the chamber externally as shown. This reduces the magnetic field at the interaction region to less than 5mG, an acceptable level for the electron energies used in this experiment. The experiment which includes •the electron gun, •the electron energy analysers & •the target source is mounted from the top flange of the vacuum system via four support struts as shown in figure 2. 3mm thick mu-metal magnetic shielding is placed both internally and externally to decrease extraneous external magnetic fields to less than 5mG at the interaction region. The system is pumped by a Balzers 500l/s turbo-molecular pump backed by an Edwards 16l/s roughing pump, and achieves a background working pressure better than 2.0E-8 torr after 1 week of pumping while baking the system to a temperature of 90°C. All inputs and outputs to the spectrometer pass through the vacuum chamber via feedthroughs on the top flange. The Top Flange. The top flange of the vacuum system is a 25mm stainless steel flange with 11 conflat CF70 flanges arranged around a central 200mm flange designed for locating a hydrogen source. The CF70 flanges are used for input of electrical signals, the gas target beam and vacuum monitoring systems: •Two flanges are used for the high voltage channeltron and photomultiplier tube inputs and pulse outputs. •A 52 way electrical feedthrough is used for the electron gun voltages •Two CF70 flanges accommodate 19 way electrical feedthroughs for the two analyser voltages. •Three rotary motion feedthroughs allow the analysers and electron gun angles to be changed externally via stepper motors for both analysers and for the electron gun. •A single flange is accessed to rough out the vacuum system prior to engaging the high vacuum pump, whereas •Another CF70 flange accommodates an ionisation gauge. The ionisation gauge is placed at right angles to the main flange to prevent light from the thoriated iridium filament entering the vacuum system, which would increase noise on the photomultiplier tube. The ionisation gauge feedthrough is sooted, and a loose cover is placed over the entrance in the chamber flange to prevent stray light from the ion gauge entering the vacuum system. The target gas enters the system via a CF70 flange which has two 6mm stainless steel tubes welded into the flange. One of these is internally connected to a gas jet which can be set at 45° to the incident electron beam and is used in the perpendicular plane experiments. The other gas feed is connected to a hypodermic needle set at 90° to the incident beam for measurements in other geometries. These stainless steel feeder tubes are connected to the nozzle using grounded shielded PTFE hose. The gas jet which is not currently used in the interaction region can be used to non-locally fill the chamber with gas for background counting measurements. The final CF70 flange is used when the photomultiplier tube is cooled, usually during baking of the system following exposure to atmosphere. This is necessary since the bi-alkali photocathode of the photomultiplier tube degrades rapidly at temperatures in excess of 55°C, and baking of the vacuum system is most effective for temperatures closer to 90°C (this is the maximum that can be obtained with the present heating tapes). The tube is cooled by passing vapour from a liquid nitrogen source through expansion coils wrapped around the photomultiplier tube housing. •Figure 2. The Spectrometer mounted from the Top Flange. The spectrometer is shown configured in the perpendicular plane, the electron gun being placed in the vertical position with respect to the analysers which span the horizontal detection plane. The electron gun is a two stage non-energy selected gun that produces an incident electron beam current of up to 4 microamps at an incident energy from around 30eV to 300eV. The analysers are identical, and consist of a 3-element electrostatic lens focussing the interaction region onto the entrance of a hemispherical electrostatic energy analyser. The electrons selected in angle and energy are amplified using Mullard 919BL channel electron multipliers. The interaction region defined by the interaction of the gas target beam and the electron beam is focussed onto the photocathode of an EMI 9789QB photomultiplier tube via a lens, 450nm±50nm optical filter and defining aperture. The target gas effuses from a 0.6mm ID hypodermic needle constructed from platinum iridium alloy. The electrons passing through the interaction region which are not scattered are collected in a Faraday cup opposite the electron gun. The gun, photomultiplier tube, gas hypodermic and Faraday cup all rotate on a common yoke with an axis through the centre of the horizontal detection plane, allowing all geometries from the perpendicular plane to the coplanar geometry to be accessed. The Electron Gun. The electron gun is a two stage non-selected gun designed by Woolf (Ph.D thesis, University of Manchester) (see figure 3). •The gun consists of a filament, grid and anode source which is focussed using a triple aperture lens through two 1mm defining apertures in the centre of the gun. •The apertures define the electron beam pencil and beam angles, and are focussed onto the interaction region using a second triple aperture lens located between the apertures and interaction region. •There are three sets of deflectors located in the gun. One follows the anode, another corrects the beam direction between the apertures and a third steers the beam onto the interaction region. •11 voltages therefore control the electron beam from the gun. The electron gun is shielded using advance (constantin) sheet which is grounded to the body of the gun mounting. This shield is coated with aerodag colloidal graphite to reduce patch fields on the surface of the metal, and is made as small as possible to allow the analysers to approach the gun as close as is feasible when non-perpendicular plane geometries are selected (see figure 2). The gun is mounted onto an x-y translation table to allow placement of the electron beam as accurately as possible to the centre of rotation of the analysers and electron gun rotation axis. The XY translation table is mounted from the rotating yoke, to which is attached the photomultiplier tube. The gas jet is mounted from the electron gun using a narrow support of 310 stainless steel, and the Faraday cup can be attached to this support. The original unselected electron gun built by Woolf was completely rewired and a custom built PTFE 25 way inline wire-wrap plug and socket was placed between the 52 way vacuum electrical feedthrough and the electron gun. This allows the gun to be dismounted from the system without the need for rewiring of the feedthrough when the need arises (filament changes, cleaning etc). The wiring from the 52 way feedthrough to the inline plug is shielded colour coded 30 awg PTFE coated advance wire, and the inline plug and socket connectors are non-magnetic gold plated wire wrap pins. The wiring between the gun and inline socket is also shielded colour coded PTFE coated advance wire facilitating spot welding to the electrostatic lens elements of the gun. •Figure 3. The unselected electron gun. The electron source is a heated tungsten hairpin filament heated by passing around 2A of DC current through the filament. The Grid/Anode extraction region roughly collimates electrons from the filament, this beam being focussed using lens 1 onto the apertures in the field free region. The Anode deflector pair corrects for any misalignment of the filament emission point with respect to the axis of the electron gun. The defining apertures limit the pencil and beam angles of the electron beam following focussing by lens 2 onto the interaction region. Placement of the apertures ensures that the beam angle is close to zero degrees at the interaction region over a wide range of energies, whereas the aperture opening reduces the pencil angle to around ±2 degrees over this energy range. The deflectors between the apertures correct for space and surface charge beam variations inside the field free region, whereas the final deflectors allow the beam to be steered onto the interaction region defined by focussing a 1mm aperture in front of the photomultiplier tube onto the gas target beam. The Electron Energy Analysers. Figure 4. A photograph of the electron energy and angle selecting analysers. For details, see the text. The electron analysers are constructed from molybdenum and are mounted from two rotating gear plates controlled externally via the rotation feedthroughs on the top flange (see figure 2). The analysers are mounted onto x-z translation tables that allow them to view the interaction region accurately over the full range of angles accessible. The input electrostatic lenses for each analyser are three element cylindrical lenses as shown in figure 4. These lenses have a capability for energy zooming of around 10 : 1, and have a maximum entrance acceptance angle of ±3° as defined by the entrance apertures. The interaction region is focussed onto the entrance aperture of the hemispherical energy analysers with a beam angle of around zero degrees, the image of the analyser entrance aperture being larger than the size of the interaction region. This ensures that the overlap volume accepted by each analyser encompasses the interaction volume for all detection angles as the analysers move in the detection plane. The energy analyser consists of two molybdenum hemispheres with a mean radius of 25.4mm. Field gradients around the input and exit apertures are corrected using standard Jost correctors. The energy selected electrons passing through the exit apertures are finally collected by Mullard X919BL channel electron multipliers which amplify the single electron to produce a current pulse at their output. These devices use an activated lead glass internal surface which has a high coefficient of secondary electron emission when a high voltage (around 3000V) is applied across the multiplier. The saturated gain of these devices yields around 1E8 electrons in the output current pulse for a single electron which enters the input cone. The Channel Electron Multiplier Pulse Electronics The output current pulse from the multiplier of approximately 1E8 electrons in a time of around 10ns is sent along doubly shielded Belden RG188A/U silver plated copper coaxial cable before being dropped across a 50W load external to the vacuum system, thereby creating a voltage pulse at the input of a preamplifier. A principle source of noise on the signal pulses arises from the turbo-molecular pump motor. This noise may be picked up by the channel electron multiplier and the high voltage feed to the multiplier, and is generated internally in the vacuum system. Adequate precautions must therefore be taken to reduce this pickup. This is accomplished by carefully shielding the analysers in a grounded advance housing, and by using doubly shielded PTFE coated RG188A/U cable from the high voltage electrical feedthrough to the multiplier. Shields are also placed over the feedthrough inputs to help reduce pickup. The noise from the turbo pump is reduced to ~ 5mV using these techniques, and as the average pulse height from the saturated multiplier is around 20-40 mV, this allows adequate discrimination following amplification. The pre-amplifiers located external to the vacuum system are commercial 100X preamplifiers with 400pS rise-times, manufactured by Phillips Scientific. The input and output impedance of these amplifiers is 50W, matching the RG188A/U internal feeder cable and the doubly shielded RG58A/U coaxial cable connecting the preamplifiers to ORTEC 473A constant fraction discriminators located in separate NIM bin crates. Separate NIM crates are used to increase isolation and reduce common mode pickup. The CFD's are operated in NaI mode as this has a 2ms dead time following the initial pulse, thus reducing the possibility of false triggering due to any mismatching impedance reflections along the delay cable. The NIM signals produced by the discriminators drive two ORTEC 437 Time to Amplitude Converter's via appropriate delay cables and an ORTEC pulse delay unit. The main TAC output feeds an ORTEC Multichannel Analyser card located in the PC controlling the experiment, whereas the second monitor TAC drives an external MCA, allowing convenient monitoring of the coincidence signal as the experiment progresses. The high voltage supplies to the electron multipliers are two Brandenberg 5kV supplies controlled by the main PC via serial DAC's. These EHT supplies are connected via VERO 200 high voltage couplings to high voltage cable connecting to the pickoff circuitry located on the high voltage feedthroughs. A 12dB/octave passive RC Butterworth low pass filter is placed between the high voltage supply and the input HT feedthrough. This serves three purposes. •Principally the filter is used to reduce pickup noise from entering via the HT cable onto the multiplier output. •Secondly, the filter has a slow response time for high voltage spikes, and reduces the probability of damage to the pulse electronics should the EHT supply suddenly turn on or off. •Finally, the 1nF 6kV capacitor in the filter prior to the 1kW load resistor acts as a short circuit to ground for the CEM pulse current. A schematic of this filter is given in figure 5. Figure 5. The channel electron multiplier pickoff and filter circuitry. See text for details •The EHT line is decoupled from the pulse electronics using two 1nF 6kV capacitors. •Capacitor C2 decouples the 1kW load resistor to earth for high frequency pulses from the CEM's, and the voltage appearing across this load is connected via the capacitor C3 to the input of the 6954 preamplifier. •Since the input impedance of the preamplifier is 50W and the CEM is effectively an ideal constant current supply, most of the current passes through to the preamplifier. The effective load presented to the multiplier is therefore ~ 50W, matching the impedance of the internal RG188A/U feed and thereby reducing mismatching reflection problems. •Finally, the 10M resistor prevents the output to the preamp from rising above ground when the preamp is disconnected from the circuit and is therefore a safety device for the preamplifier when initially connected. The Target Beam Source Two configurations are used for target source: •The original design placed the atomic beam nozzle and Faraday cup on the same support structure fixed to the body of the electron gun, with the atomic beam direction being in the detection plane when the perpendicular plane geometry is chosen (see figure 2). •A second configuration is adopted for exclusively perpendicular plane coincidence experiments, the atomic beam nozzle being oriented at an angle of 45° to the detection plane so that the gas effusing from the nozzle does not fill the analysers when they oppose the nozzle (see figure 6). In this configuration the support strut for the atomic beam nozzle is terminated lower than the detection plane, as is the nozzle itself. This allowed the analysers to roam the detection plane over a complete 360° angular range. For non-perpendicular plane measurements the Faraday cup is once mounted from the support strut and the gas effuses perpendicular to the electron beam direction. The angular movement of the analysers is therefore restricted to the limits ±60° when the gun angle is between 90° (perpendicular plane) and 65.6°, and -60° to +35° when the gun angle is set smaller than 65.6°. •Figure 6. The configuration of the target source hypodermic when perpendicular plane measurements are conducted. This allows the analyser to roam over a complete 360 degrees of the detection plane when in this geometry (apart from the position where the analysers would clash wth each other). The Photomultiplier Tube. One difficulty when aligning an (e,2e) coincidence experiment lies in the focussing the incident electron beam and analysers onto the interaction region. The common procedure is to initially maximise the current from the gun onto the Faraday cup, then onto analyser counts as the electron beam is moved through the gas jet. The analysers are then adjusted for maximum signal count and the procedure is iterated until a maximum is found. There are several problems associated with this technique, and one major disadvantage : •The throughput of electron current as detected by the Faraday cup is insensitive to both position and focusing of the electron beam, since the cup opening is usually made quite large (in this case 6 - 10mm) and the bias voltage applied to the cup to prevent electrons returning to the interaction region allows all electron trajectories within this opening to be detected. Maximising the Faraday cup current is therefore only a crude preliminary alignment of the incident electron gun beam. •Conventionally, following maximisation of the Faraday cup current, the beam is steered and focussed onto the gas jet as detected by counts in the analysers. This therefore assumes that the analysers are initially focused onto the interaction region, and a slow iterative technique must be adopted where the analysers and electron gun are tuned so as to maximise the signal from both analysers. •This has the disadvantage that the overlap volume is maximised for both atomic and electron beam density, which tends to place the interaction volume as close as possible to the output of the atomic beam nozzle, which is almost certainly not in the centre of the detection plane. In this case a true differential cross section is not obtained, since the scattered and ejected electron trajectories lie in a different plane than defined in the initial experimental alignment. To eliminate the need for time consuming iterative techniques and the associated errors due to detection geometry, a separate method for tuning the electron gun is adopted that does not depend upon the analysers. An EMI 9789QB reduced photocathode photomultiplier tube is installed in the vacuum system together with optical focussing and defining elements (figure 2). The interaction between the target and the electron beam not only produces ionisation, but the valence states of the target also are excited with various probabilities depending upon the incident energy of the electron beam. A group of these excited states decay back to lower valence states with the emission of visible photons at wavelengths around 450nm for both helium and argon targets. These photons are detected by the photomultiplier tube, by placing a 450nm ± 50nm optical filter in front of the photocathode. The interaction region is therefore defined where an aperture in front of the photocathode is focused by the lens. •This is established by placing a light source (i.e. a torch bulb) in place of the photocathode behind the defining aperture and adjusting the lens position and focus so that the image of the defining aperture is positioned exactly where the analysers, electron gun and gas jet overlap (figure 6). The size of the defining aperture is chosen to yield an interaction volume of ~1cubic mm. •The plano-convex focussing glass lens is covered by a grounded fine tungsten mesh to eliminate surface charging. Tuning of the electron gun to PMT counts therefore guarantees that the electron gun is steered and focussed to the point at the centre of the detection plane independent of the analysers or gas beam density. The photomultiplier tube may be cooled by circulating nitrogen gas boiled off LN2 through copper coils encircling the photomultiplier tube housing. This is mainly used when the vacuum system is baked at temperatures above 55°C, since the photocathode is rapidly damaged by temperatures exceeding this. The dark noise count is also reduced when cooled, however since the reduced cathode PMT only produces dark counts of less than 20Hz at room temperature, this is not necessary during routine operations. The Photomultiplier Tube Electronic Circuitry. The photomultiplier tube used in the experiment is a 9789QB EMI photomultiplier tube with a bi-alkali photocathode peaking in the blue region of the optical spectrum. The tube is a venetian blind dynode construction, and has a reduced photocathode to decrease the noise on the output of the tube to less than 20Hz. As with the electron multiplier circuitry, it is important to optimise the load of the photomultiplier tube so that reflections along the feeder line to the counting electronics are minimised. •Figure 7 shows the biasing arrangement adopted for the dynode array which maximises saturation of the output current from the tube into the 50W load resistor used to match the RG188A/U line. •The 1nF 6kV capacitors across the dynodes close to the output stage are used as charge storage pumps that quickly refurbish the surface charge on the dynodes, allowing increased counting rates. •The dynode resistors are 1% tolerance metal film resistors as these are more stable with temperature. • •Figure 7. The photomultiplier tube bias network. The 1.82M resistor was chosen to optimise the single photon counting efficiency of the tube. The 470k resistors bias the dynodes to achieve saturation of the current pulse at the output of the tube, hence biasing the tube for single photon counting applications. The HT supply is through RG188A/U PTFE coated silver plated 50W coaxial cable, as is the feed to the discriminator external to the vacuum system. The load is set by the 50W resistor across the anode to the HT feed, the pulse being decoupled to ground via the 1nF capacitor. This load ensures that any pulse reflections along the RG188A/U transmission feeder are properly damped, thereby ensuring as fast a current pulse as possible (typically 5ns wide). The 10MW safety resistor prevents the output charging up to a high voltage if the feed is disconnected. The EHT supply to the photomultiplier tube is a Brandenberg 2kV supply operating at 1500V for this tube. The negative going output pulses from the tube measure typically 200-500mV in height, and therefore are able to directly drive an ORTEC 473A constant fraction discriminator located in one of the NIM crates. The HT voltage enters the system via RG-59 cable through a high voltage feedthrough. The internal feed to the PM tube and the internal feed from the pulse circuitry are doubly shielded PTFE coated RG188A/U coaxial cables, terminated in standard aluminium BNC connectors on the photomultiplier tube housing. All resistors and capacitors for the dynode chain are soldered to the photomultiplier tube base internal to the vacuum system, using standard 5-core solder. The photomultiplier tube base was thoroughly cleaned in an ultrasound following location of these components, to displace any flux residue from the solder joints. It is not found that these solder connections significantly affect the ultimate vacuum pressure in the system. The External NIM Pulse Electronics. Typical pulse heights from the electron multipliers in saturation mode are 25-40mV, whereas the typical height of the photomultiplier tube pulses is of the order of 250mV. It is therefore necessary to amplify the channeltron pulses prior to transmission to the discriminators, whereas the photomultiplier tube pulses are sufficiently large to not require further amplification. The electron multiplier pulses are obtained from the multiplier supply circuitry as explained above. These pulses are amplified by Phillips Scientific 6954 100X amplifiers located external to the EHT feedthroughs connected by 10cm lengths of RG-58 cable. Isolated DC supplies to these amplifiers is obtained from a standard +15V supply, the circuit diagram being given in figure 8. Figure 8. The 6954 Phillips Scientific preamplifier isolated DC supplies (2 off) The amplified pulses from the 6954 amplifiers pass along doubly shielded RG-58 cable to the input of the ORTEC 473A discriminators. Similarly the PMT pulses pass to a third ORTEC 473A discriminator. The two analyser discriminators are located in separate NIM crates to increase isolation and reduce common mode pickup. These crates are further decoupled from the AC mains supply using filter circuitry on each supply line. The slow rise-time positive outputs from the 473A discriminators pass to three ORTEC 441 rate-meters which allow counting and monitoring of the pulses. Internal to these rate-meters buffer 7400 TTL chips have been added to send to 32 bit counters located in the main PC via RG-58 cable. The power supply to these TTL buffers is derived from the rate-meter supplies via 7805 5V regulators located internally. The fast NIM pulses from the analyser 473A CFD's pass through appropriate delay lines to two 437 TAC's. •The output of one of these TAC's s sent to an ORTEC multichannel analyser located on the main PC bus, whereas •the second TAC output is sent to an external Tracor Northern multichannel analyser. This second multichannel analyser allows the operator to monitor the total accumulated coincidence signal during operation. Figure 9 shows a schematic of the external NIM pulse circuitry used in the experiment. Figure 9 The External NIM Pulse Circuitry. See text for details. The Computer Controlled Electrostatic Lens Supplies. The lens supplies that control the voltages impressed onto the lens elements and deflectors in the experiment were developed over a number of years, and therefore take different forms. Only a brief description of these supplies is given here. The fully computer controlled deflector supplies are housed in a separate 19" crate. Three of these supplies feed the deflectors in the electron gun, the other two supply the input deflectors in the analysers. These supplies are fully floating, their mid-point voltage being equal to the lens element voltage in which the deflectors are housed. Typically these supplies can vary between ±15V with respect to their midpoint voltage. In a second 19" crate is located the gun lens high voltage supplies together with the high voltage supplies for the Faraday cup. All these supplies are fully computer controlled isolated units with low output impedance to reduce noise and eliminate current loops. In a third 19" crate is located the analyser voltage supplies together with a set of serial loading supplies that control the electron multiplier EHT supplies. A further supply controls the current boost circuitry located in the filament constant current supply unit (see below). The Spectrometer Voltage Interface The outputs of each computer controlled supply that drives an individual lens element or deflector are connected to a 70 relay switching board. This relay switching unit allows either the voltage or the current on each lens element to be measured by a Keithley Digital voltmeter. The Digital voltmeter is in turn accessed via an IEEE-488 interface which is addressed by an 8086 slave PC which converses with the central PC which controls the experiment. The outputs from the relay board connect to the spectrometer via shielded multicore cables which take the gun voltages to the 52 way feedthrough and the analyser voltages to the two 19 way feedthroughs located on the vacuum flange. The gun supply multicore also carries the filament supply current to the 52 way feedthrough. The Faraday cup current is measured by the Keithley Digital voltmeter when under computer control but this can also be measured by an external pico-ammeter. The bias supply to the Faraday cup is via the 52 way feedthrough on the vacuum flange. The Stepper Motor Supplies and Interlocks. In addition to the analyser electrostatic lens and deflector voltage connections, the 19 way CF70 feedthroughs are used to control signals from internal sensing opto-isolators which control the stepper motor interlocks. The stepper motors that drive the analysers and electron gun around the detection plane are five phase motors with integral 10X reduction gearboxes driving the rotary motion feedthroughs. This drive is further reduced internally by 10X reduction gears, resulting in a reduction ratio of 100:1. The intelligent stepper motor drivers are addressed via a serial port located on the controlling PC bus. The stepper motor drives are contained in a separate 19" rack. An internal unregulated power supply in this unit is dedicated to driving the motors. This supply is separate from the regulated supply that drives the controlling logic. This was found to be necessary to eliminate problems associated with current loops from the 20A peak motor current feeding the common point of the logic circuitry. An additional interference problem was traced to the RS232 serial line from the controlling PC. The PC bus was found to be poorly earthed due to very little copper being used for the bus earth. This led to the ground line of the RS232 serial line moving around with respect to the 0V stepper motor reference as drive current to the motors returned along the ground line. This was cured by decoupling the 0V RS232 line from the main 0V line for the stepper motors, as is shown in figure 10. •Figure 10. The Stepper motor Drive Power Supply and decoupling circuitry required to decouple the stepper motor ground lines from the logic circuitry and RS 232 serial interface to the controlling PC. The analyser stepper motor supplies have a facility for halting the motors using limit switches which are optically coupled to the internal 68000 logic circuits driving the supplies. These limit switches were exploited to prevent the analysers from •(a) running into each other, •(b) running into the electron gun when it was in the way of the analysers, •(c) running into the Faraday cup when it was in the line of the analysers and •(d) running into the gas jet support when the gun angle was set from 65° to 90°. These interlocks were controlled by photodiode/phototransistor coupled pairs inside the vacuum system. The SD5443-3 phototransistor normally illuminated by the TEMT88PD photodiode during operation ensured a fail-safe mechanism should any of these components fail (this automatically prevent the motors from running). Should the analysers move into a position as defined by the four criteria (a) - (d) above, a shutter located on the analyser turntables moves between the appropriate photodiode and phototransistor, shutting off the light and thereby shutting down the stepper motor current. Only motion in the opposite direction is allowed by controlling logic, allowing the analyser to move away from the point of imminent collision but no closer. The six pairs of optocouplers that prevent the analysers moving into the Faraday cup, electron gun and gas jet support are located on the mounting struts supporting the main body of the spectrometer. Two optocouplers are located on the side of analyser 2, and two shutters are located on the sides of analyser 1 to prevent the analysers colliding with each other. Finally a ninth optocoupler is located over the electron gun counterweight which monitors when the gun is at an angle less than 65°, where the sweep angle around the detection plane is reduced. Figure 11 is a schematic of the logic circuit used to determine the optocoupler status. a •Fig 11. The optocoupler logic circuit circuit that controls the stepper motors and avoids collisions between the analysers, the electron gun, the Faraday cup and the hypodermic support. See text for details. The Filament Constant Current Supply and Current Boost Circuitry The filament constant current supply is a standard 5A constant current supply. Added to this filament supply is a computer controlled current boost circuit that can be used to increase the filament current during electron gun tuning. This supply boosts the current without changing the main supply, thus ensuring supply stability before and after the electron gun is tuned. The current boost circuit operates from a serial driven supply card as detailed previously. Circuit details for this supply are given in figure 12. •The main filament drive is through six power diodes that feed the current from the constant current supply to the filament. These isolate the constant current supply from the boost supply. •The Darlington transistor T1 feeds additional current to the filament via three parallel 10 turn 100W trim resistors through a 1N4007 diode. The additional current then returns to the 12V DC regulated supply via a second 1N4007 diode. •As the constant current supply acts as an infinite resistance source, the boost circuit supplies additional current only when the voltage at the anode of the feed 1N4007 exceeds the output voltage of the constant current supply. This therefore depends upon both the current driving the filament, as well as the resistance of the filament when hot (typically 4W - 6W). •Finally, the filament bias voltage (and hence energy of the electrons) is set to the midpoint of the filament by two 100W balancing resistors located across the terminals. •Figure 12 The Current Boost Circuit used during tuning of the electron gun. This supply feeds current to the filament without changing the characteristics of the main constant current supply used while operating the spectrometer. This supply is operated only when the signal from the photomultiplier tube is very weak (eg when the (e,2e) cross section is large, but the excitation cross section yielding 450nm spontaneous emission radiation is small). The Controlling Computer and Bus Electronics The controlling PC contains the main control logic for the experiment. On board is •the ORTEC multichannel analyser card, •a Blue Chip ADC board which interfaces to a 24 channel buffer card, •a timer counter board which counts the analyser and PMT count rates and •custom built parallel interface and serial interface cards which control the power supplies for the experiment. In addition to the ADC buffer card, three independent isolated regulated supplies drive REF01 voltage reference sources for measurement of the positions of 10 turn potentiometers located on the analyser and electron gun rotary drive shafts. These high stability reference sources ensure that the analyser and electron gun positions are accurately measured with time. The angular position of the electron gun and analysers is calibrated when the system is opened using visible laser diodes that define these angles to an accuracy of ±0.2°. A datalogger internal to the controlling PC establishes a cubic spline relationship between the measured angles and the corresponding ADC conversion from the potentiometers. The 8086 Slave Computer and IEEE Interface The 8086 slave PC addresses the Keithley voltmeter via a standard IEEE interface. This PC also controls the selection of the relay motherboard for measurement of the required lens element either in voltage or current mode when a request is sent from the controlling computer via an RS232 serial line. The Keithley DVM connects to the relay motherboard in two ways. •Current measurements are sent directly to the Keithley current input, whereas •voltage measurements connected via a 100MW impedance matching circuit to the voltage input of the DVM. This ensures minimum deviation of the supply voltage when measurements are being made. Figure 13 shows an overall block diagram detailing the computer control and high speed pulse interfacing to the spectrometer. •Figure 13. The Hardware Interface to the (e,2e) spectrometer, showing the logic and voltage supply connections to the central controlling PC.
