I've put some notes on the fundamentals of electromagnets here. If you don't recall how electromagnets work, what they look like, or what their magnetic fields look like, this page should be helpful (otherwise skip it).
A solenoid magnet is a cylindrically coiled electromagnet (DC, for our purposes). When current is flowing through the coil, the resulting magnetic field inside the coil is in the axial direction. There is a drawing of this on the electromagnet basics page. We typically use the term solenoid to refer to coiled magnets that are long in the axial direction (length >> width), so that the fringe fields at their ends are small compared to the long axial field inside the coil. Shorter coiled windings in the context of the accelerator are usually called lenses.
A strong (i.e. high current) solenoid magnet will provide a net focusing effect (in a sense) to charged particles passing through it whose velocity has a transverse component relative to the central axis. Depending on the polarity of the magnet and the charge of the particle, it will either exert a "focusing" force, which will cause the particle to revolve helically around the axis, rather than diverging; or a "defocusing" force, causing the particle to spiral out away from the axis. It's not entirely simple to visualize this effect in your head (to me, anyway), but remember that the magnetic field of the solenoid is in the axial direction, and the magnetic force on a charged particle is given by the Lorentz force law F = qv x B, where the direction of the force can be found using the right-hand rule (fingers curl from v to B, thumb shows direction of force). So we know right away that a solenoid won't exert any force on a particle moving in the axial direction, regardless of its offset from the axis; it only talks to particles whose velocity is in a different direction. Because of this, solenoid focusing is considered "weak focusing", because its force is not in the radial direction (it can never reduce a particle's transverse offset with respect to the axis). Consequently, solenoids are used in the accelerator at places where the targeted particles are lower in energy, and for capture purposes rather than actual beam focusing.
Note that since this magnetic force causes particles to spiral about the central axis, a solenoid introduces coupling between the x- and y-planes. Steering in areas with solenoids is therefore very different than steering in other places, because of the inability to steer independently in x and y.
One solenoid that doesn't get mentioned here is the one that may come to mind first: the extremely large BaBar solenoid that surrounds the detector in IR2. The reason it isn't listed here is because it is essentially unrelated to the beam transport lines. Its purpose is detector-related rather than optical. The solenoid field causes particles produced in e+/e- collisions to follow a curved trajectory, and the radius of curvature of a given particle's trajectory can be used to determine that particle's charge-to-mass ratio. Since the BaBar solenoid isn't used for beam transport purposes, any optical effects it has on the beams are residual and unwanted. The x/y coupling effect introduced by the solenoid complicates steering at the IP.
Solenoid windings can be found in various locations around the Low Energy Ring in PEP-II. These appear to be large-diameter insulated wires that are wound around the outside of the entire beam pipe, for several yards at a time, so they are easy to spot. These solenoids were put in to reduce the "electron cloud" effect in the LER: since electrons and positrons attract, it is thought that stray electrons were being drawn from the walls of the beam pipe by the positrons passing through. By virtue of their opposite charges, a solenoid polarized to have a "focusing" effect on positrons will have a "defocusing" effect on electrons. Consequently, positrons passing through a solenoid near its central axis will see nothing more than perhaps some beam rotation, but stray electrons near the edges of the beam pipe will be drawn back toward the walls.
In the injector system, electrons are created by boiling electrons off of a cathode and focusing them into a beam. The mechanisms are different, but this basic principle is the same for both the P-Gun and the T-Gun. Electons boil off the cathode in a sort of directionless shower, and as they do so, they repel one another by virtue of the space-charge effect. Consequently, before the particles can be focused by quadrupole magnets, they must first be "captured" and lumped into some sort of coherent group. This initial electron capture is done by solenoid magnet "lenses" at the source. The solenoid counteracts the space-charge effect by exerting a force on particles whose velocities are in a divergent direction, causing them to spiral about the central axis instead of wandering away.
For both the T-Gun and the P-Gun, electrons are produced in a pulsed way (the P-Gun does this with laser pulsing; the T-Gun does it by pulsing the bias grid), but as far as I know, the solenoid magnets here are continuous DC rather than pulsed. Another type of solenoid magnet in the injector system is the Hemholtz coil, used for P-Gun running only. Additional larger solenoids are also used for further beam focusing in CID. Quadrupole "strong" focusing doesn't come into the picture until further down the line, in Sector 0.
The solenoid windings we call "solenoids" are long in the axial direction (length >> width). We therefore consider their fields to be essentially uniform, ignoring fringe fields at the magnet's ends, and so they act on particles with divergent velocities (x',y') rather than on particles with a transverse offset from the central axis (x,y). Lenses are much shorter solenoid windings, and so the fringe fields at the magnet's ends are non-negligible. The single-loop and solenoid drawings on the electromagnets page may help you visualize the effect of the fringe fields. Notice that for a particle just outside the solenoid, the magnetic field lines are radial in direction. This field is strongest near the coil and weakest in the center, so particles with a greater transverse offset from the central axis will experience more force in the radial direction. Whether this force is focusing or defocusing can be determined by the right-hand rule.
A bucking coil is a solenoid magnet whose purpose is to cancel out an obtrusive magnetic field caused by something in the beam line that is not intended to produce an optical effect. The two examples of this are at the P-Gun, and at the BaBar detector.
In the case of the BaBar detector, a large part of the detector known as the DIRC is used to detect Cherenkov light. The DIRC hardware involves a large metal shield, which is inside of the gigantic BaBar solenoid. A magnetic field is induced in this volume of metal by the surrounding magnet. A bucking coil solenoid is placed nearby and driven with the appropriate amount of current to cancel the consequent, unintentional optical effect.
The P-gun also has a bucking coil at the edge of its large metal enclosure (the one you have to duck under when you enter CID), for the same general purpose as that of the BaBar bucking coil: it cancels out stray field effects to make sure there is no net solenoidal field on the gun cathode.
The flux concentrator is a small but extremely thick winding of only a few turns, located just after the positron target. Its purpose is to capture the maximum-possible percentage of the positrons that are sprayed off of the target as it is hit by the scavenger beam. In this sense, it is much like the lenses used to capture the electrons that are boiled off of the gun cathode-- but there are large differences between the two. Electrons are created by gun pulses that are an order of magnitude longer than than the beam pulse once it is bunched down, whereas are produced only at the instant the target is struck by a scavenger beam pulse. The resulting positron spray is small in area, as the positrons spray off from only a small spot, but highly divergent. Thus the flux concentrator needs to exert the maximum possible force on the divergent particles for an extremely short period of time (a few picoseconds). For this reason, the flux concentrator is a pulsed magnet, driven by a powerful network of high-voltage capacitor banks that store a large amount of charge, and thyratrons that act as a switch to dump the charge through the flux concentrator at the appropriate moment, as an extremely high current of short duration. Another reason that the flux concentrator must be pulsed is because of heating considerations. If current of that magnitude were to run through it constantly, it would probably be almost impossible to keep the copper from melting. As it is, the magnet is water-cooled. The physical size of the flux concentrator is not much larger than the beampipe, so that its field lines are highly concentrated in the necessary area. Its pulse time can be adjusted for maximum capture.
Zoe has an excellent write-up on the flux concentrator in the Pulsed Magnets section II.G of this magnets document. Look there for diagrams, quantitative details and additional information.
A dipole magnet for our purposes is an electromagnet consisting of two windings that are placed on opposite sides of the beampipe. You can see a drawing of a dipole magnet's configuration and magnetic field lines here. The dipole's magnetic field is transverse with respect to the beam pipe, so depending on the orientation of the magnet, the dipole may be used to bend the beam's path transversely in either the x- or the y-plane. [Or it could kick in both planes, if the magnet were oriented in a skewed way. We do have some bend magnets that do this (PEP extraction, where the extraction line goes upward and to the right of the linac beampipe), but we don't have any skewed correctors.] The bending power of the dipole depends on the magnitude of current in its windings, and its bending direction depends on the direction of current flow; so we use the SCP to control the current delivered to the magnet from its power supply. All dipole magnets bend transversely, but if it's a small magnet used for small bending effects, to steer a beam's orbit through the beampipe, then it's a corrector; if it's a large magnet used for changing a beam's trajectory to travel in a different direction over a larger distance, then we call it a bend.
Trim windings are smaller magnet coils that are used for fine-tuning the current strength of a large electromagnet such as a quad (in which case it's a QTRM) or a bend (then it's a BTRM). Zoe has some good notes on the three different types of trim windings we commonly see in the accelerator in section 1.b of this magnets document. There are a number of places that trim windings are used for fine-tuning and steering in the accelerator. Here are a few examples of prominent places in which we see them:
One important place we find QTRMs is in the RTLs, where fine-tuning of quad strengths is necessary because the optics there are crucial to proper bunching and setting of energy spread. This isn't something that operators would normally ever tweak, though-- it's more of a set-up thing, done by Rick Iverson or similar. We also have QTRMs in the LTRs, but these are less critical, since the job of the RTL is only to transport the beam from sector 1 to the damping ring; thus the fine steering there is only to maximize throughput, and has no residual effect on the beam that comes out of the damping ring and goes down the linac.
In the linac bunch compressor chicane in sector 10, BTRM magnets are used for fine steering through the chicane. Here they basically serve the purpose of correctors, making finer corrections to the beam's orbit through the beampipe, while their associated large bend magnets make the necessary directional changes to its path.
Another place trim windings can be very useful is in the A-line, where the majority of steering is usually done with only four correctors (two in x, two in y) whose power supplies are old school. The electronics inside these power supplies have a reversing network that takes a fair amount of current to activate, so for currents near zero (i.e. asking the corrector to make any small correction), the power supply hits a "dead band" in which small changes to the BDES value do not necessarily have the desired effect on the BACT. Luckily these correctors all have associated trim windings that can be used to correct for the dead-band error.
Correctors are smaller dipole magnets used to change the beam's orbit through the beampipe by giving the beam a small transverse kick as it passes through the magnet. All steering is done with correctors. The corrector introduces a small angle change in the beam's trajectory, rather than a transverse offset. To steer with a corrector, then, we look at the beam's position on a BPM downstream, with enough distance between the magnet and the BPM that the corrector can act as a decent "lever arm". Manual steering consists of changing a corrector's current directly, either with a diagnostic knob or by changing its BDES and trimming. Steering packages are also used to calculate corrector values necessary to produce a certain orbit through an area and implement it. Other correctors are controlled by orbit feedbacks, whose job is to change corrector strengths as needed to maintain a given orbit through an area.
The typical linac sector has 16 total correctors: eight in the x-plane (XCORs) and eight in the y-plane (YCORs). The pattern of magnets in the linac is very regular for all areas in which the beam does nothing but go straight ahead. There are eight quadrupole magnets in the typical sector, numbered 201, 301, …, 901. Each of these "standard" QUADs is located on a drift section of beampipe, at the place where the disk-loaded waveguide fed by one klystron terminates, and another begins shortly thereafter. I'll bring this up again later in the QUAD section, but here it is relevant only to say that each of these QUADs is directly followed by one XCOR and one YCOR, whose respective unit numbers are n02 and n03 respectively.
Elsewhere the patterns of correctors are less easily defined. For instance, in bendy places such as the PEP rings, there are a great many correctors. But they always serve the same purpose (steering!).
Bends are large dipole magnets used to change the direction of the beam's travel. Bend magnets are therefore found wherever the beam pipe does anything other than travel in a straight line. Some bends are pulsed and some have steady-state currents, depending on whether or not that bend is used to make different beam pulses go different places. Pulsed bend magnets are called kickers, and many of these are discussed individually in the pulsed magnets section below.
The required strength of a bend magnet is determined by the bending radius of the beampipe and the energy of the beam passing through it. Since the bending radius of the beampipe is a fixed variable (it's already been built, and it doesn't move), the strength of a bend magnet sets the energy of the line, in the sense that for a given bend magnet strength, only particles of a specified energy range will be able to pass through it. Particles with too little energy will be over-bent by the magnet and will crash into the wall at the inner [w.r.t. the curve] edge of the beampipe, and particles with too much energy will be under-bent, crashing into the outer edge. The strength of a bend magnet's field is therefore a very crucial measure.
The easiest way to understand why lower-energy relativistic particles get bent more than higher-energy ones in a bend magnet, it is useful first to consider the non-relativistic case, in which a higher-energy particle is one whose velocity is greater. The radius of curvature of a particle in a magnetic field B is given by r = mv/qB. Thus the radius of curvature of a non-relativistic particle's trajectory is directly dependent on its velocity; a particle with velocity four times greater will travel in a circle four times larger. In other words, the higher-energy particle gets bent less by a given B-field. The case for relativistic particles is analogous. The velocities of two highly-relativistic particles are, for all practical puproses, exactly the same-- but higher-energy particles have greater momentum, which translates to higher magnetic rigidity. So like their non-relativistic counterparts, higher-energy relativistic particles are bent less by a given B-field. The accelerator physics textbooks show this in equation form.
Another important point about bend magnets is that they create dispersion. More on this later (if I ever really get it).
A wiggler magnet is essentially a string of small dipole magnets placed one after the other along a short span of beampipe, with alternating polarities. A beam passing through the wiggler will therefore follow a zig-zag path. In doing so, the beam emits synchrotron light, and thus the overall energy of the beam is reduced by some amount. The wiggler also changes the emittance of the beam, in the plane of wiggling, which I'm sure has been shown with some complicated physics by someone somewhere.
There is a horizontal wiggler magnet on the LER beampipe in region 4. This wiggler was originally put in to increase the LER emittance to match that of the HER (the HER emittance being larger by design). But it turns out that the electron cloud effect blows up the LER beam to a larger emittance than its design, and so the wiggler is no longer used. When it was turned on, the wiggler reduced the energy of the LER beam (and by extension, the center-of-mass energy at the IP) by 5MeV.
There is also a horizontal wiggler on the FFTB beamline. Synchrotron light created by this wiggler can be collected and used for diagnostic purposes, to reveal some properties of the beam that created it. I'm also told it may have some use in beam optics, but don't know the details.
The undulator used by SPPS is another type of wiggler, made of permanent magnets rather than electromagnets. So it can't be turned off, but it can be removed. When in place, its effect on the beam can be altered by physically moving the undulator, or by steering the beam that passes through it.
Pulsed dipole magnets are bend magnets whose current can be turned on or off quickly, controlled by the timing system. Such magnets are more commonly referred to as kickers. A kicker magnet is required anywhere in the beamline that different beam pulses may need to go different places, and the beampipe therefore splits with some sort of Y-junction into two different transport lines that diverge from one another. In some cases, whether or not a particular kicker is required to fire on a given beam pulse depends on the pulse's beamcode in the timing system. For instance, in Sector 19 there is a kicker magnet that fires on beamcode 10 to bend the scavenger beam pulse into the positron extraction line, but does not fire on beamcode 3, allowing the FFTB beam to pass through unaffected. In other cases, a kicker may be required to fire on a single beamcode at some times but not at other times. For instance in the damping rings, a stored beam pulse passes through the inactive extraction kicker many times before the kicker is fired to extract that beam from the ring.
The timing system controls when a given kicker fires. The master pattern generator (MPG) broadcasts a particular modifier to the pulsed delay unit (PDU) CAMAC module associated with that kicker, and the PDU then sends the trigger the kicker's switching mechanism at the appropriate time. Some kickers use thyratrons as a switching mechanism, which act as extremely high-powered and very fast switches. Others have a solid state switching system, which involves a type of transistor known as an IGBT (insulated gate bipolar transistor). High-power solid state switching is newer technology, and presumably is a lot cheaper to use than a huge, high-power thyratron-- but it can't do the job as quickly as a thyratron can or handle as much power, so thyratrons are still required in places that the kicker firing must be extremely fast. Thinking about the requirements of different places that we have kickers, you can make an educated guess as to whether the kicker there requires thyratron switching or not. For example, beam pulses go through the BAS-1 kicker only as often as a beam pulse is created by the gun-- that's 30Hz for current running conditions, so the time between beam pulses at the BAS-1 is around 33ms. On the other hand, the damping ring kickers need to be fast enough to ramp up fully, and then ramp down fully in the space of one damping ring turn, since the kicker has to have no effect on a beam pulse passing through one turn earlier, and then fire on that same beam pulse the next time it comes around. The damping ring has a revolution frequency of 8.5MHz. So time between the beam pulses passing through a damping ring kicker is about 100ns, five orders of magnitude smaller than the time between pulses through the BAS-1 kicker. The damping ring kickers require thyratrons; the BAS-1 kicker doesn't.
In a lot of places, a kicker works in conjunction with a defocusing quadrupole directly following it, which helps exaggerate the kick (that is, defocusing in the plane of kicking). And then in most cases, the kicker is also followed by a septum magnet, which helps to spread the two transport lines apart faster (i.e. in a shorter distance). More details on this in the individual kicker descriptions below, and also in the subsequent septum magnet section.
See Zoe's write-up on the damping ring kickers in the Pulsed Magnets section II.A-F of her magnets document for a comprehensive description of kickers in general and details on the kicker set-ups for the north and south damping rings.
The sector 19 kicker is used to kick the scavenger beam (beamcode 10) out of the linac into the scavenger extraction line, which then transports that beam to the positron target. Although this kicker is not required to fire extremely quickly (since beam passes through sector 19 at a maximum rep rate of 30Hz for current running conditions), it nonetheless uses thyratron switching rather than solid-state. This is an artifact from SLC days, when all beams passed through sector 20, and the pulse needed to be fast enough to not disrupt the travel of the production electrons and positrons. Its power supply is EP01 LGPS 1, also called the SCXL KICKER (scav extraction line kicker). This kicker works in conjunction with the EP01 Lambertson horizontal and vertical septum magnets (HLAM and VLAM) to force the divergence of the extraction line beam from the straight-ahead linac beam.
Howard has a good write-up on the PEP injection scheme here . Zoe has further information near the end of this magnet document. Each PEP ring has two injection kickers, each having its own thyratron, and the timing between these thyratrons can be adjusted by PEP physicists for maximum injection efficiency. The basic scheme of the PEP injection kickers is this: The first kicker is placed on the storage ring beamline prior to the injection point. When fired, this kicker kicks the stored beam upward. This stored beam diverges upward from its normal path. It then passes through a vertically focusing quadrupole, which causes its trajectory to "straighten out", so that it is traveling parallel to the normal stored-beam trajectory but with a vertical offset. It then passes through the lower channel of the injection septum magnet, which should not have any B-field, so that it passes straight through. At the same time, the injected beam comes in at a downward angle toward the stored beam, and enters the upper channel of the septum, which contains a B-field that bends its trajectory to be parallel with that of the stored beam in the lower channel. At the far side of the septum, both beams pass through a vertically focusing quadrupole, which bends both beams downward toward the normal stored-beam trajectory. The resultant conglomerate beam then passes through the second injection kicker, which kicks it back on-axis to complete the job.
Unfortunately, it is impossible for this process of off-axis injection to make both the stored beam and the injected beam end up at exactly the same place-- if you look at Howard's diagram, you can think about why this must be true. To neutralize part of this effect, the ring is designed to have maximum betatron amplitude at this location. We also try to maximize its efficiency by tweaking septum bumps and so on. The reason that we typically see PLIC in the septum area when the injection kickers fire is that the closer the stored beam is to the injected beam once they exit the septum, the better-- so we push the stored beam up to the very top of the septum's lower channel, and it scrapes a little bit. The off-axis injection is also responsible for the transverse ringing-down we can see on BPM buffered data that sees an injected pulse. It takes a lot of turns for the orbit perturbation to completely subside.
The injection kicker thyratrons receive triggers from the MPG only when an injected pulse has been requested by either the BIC or an by injection diaglist. Additional logic insures that the injection kicker will not fire if the SBDL module is telling the single beam dumper to fire (more on this below, in the single beam dumper section). Note that for every beam pulse that comes down one of the PEP injection lines, either the injection kickers are fired (if an injection request has been broadcast), or the single beam dumper kicker fires and sends the incoming beam to the tune-up dump. One or the other must always occur; if neither the injection kickers nor the single beam dumper fired, the incoming beam would enter the storage ring and almost immediately crash.
Beam analyzing station 1, or BAS-1, is a beam diagnostic tool located near the beginning of sector 1 in the linac, just after the sector 1 chicane and just before the first sector 1 klystron (1-2). The end of the sector 1 chicane is where positrons first enter the linac, coming from the positron return line, so BAS-1 is located at a place were both e- and e+ travel through the same beampipe. The beam analyzing station consists of an auxiliary piece of beamline that diverges from the linac, containing a profile monitor followed by a beam dump. When beam is kicked out to this dump by the BAS-1 kicker, operators can use the beam image on the profile monitor to analyze and tweak beam parameters such as injector energy spread. These days, this set-up procedure is generally only done after accelerator down-times, or after major work has been performed in the injector. I'm told that in past times, operators customarily used BAS-1 to set CID master phase. The BAS-1 kicker uses solid-state switching, and can only fire at a rate of up to 10Hz.
The 2-9 dump is a "scrapper" dump near the very end of sector 2 in the linac. We kick the beam out to this dump whenever it is the case that we want the injector to continue to produce a beam pulse of a given beamcode, but we do not wish to accelerate that beam down the linac. Because we may sometimes wish to scrap some beamcodes while continuing to accelerate others, it is necessary to use a pulsed kicker that is triggered independently on different beamcodes. The beam is relatively low energy at this point in the linac, so we can park beams on the 2-9 dump for long periods of time without fear of excessive radiation. The 2-9 kicker (also known as LI02 pulsed BEND 009) is controlled with solid-state switching. Special buttons on the SCP make it easy to deact or react 2-9 kicker triggers for different beamcodes.
You have to be careful not to get the beam steering screwed up at the end of sector 2, because a badly steered beam will receive the wrong kick from the 2-9 kicker, and go crashing into the wall instead of cleanly onto the dump. We don't have a video signal on the 2-9 dump to watch how the beam is hitting it, so during set-up we have to watch linac analog statuses for sector 2 dump temps.