Photonic Frontiers: Laser Acceleration: Short pulses speed particles

Intense femtosecond pulses can accelerate electrons, protons, and ions to high energies over very short distances, and could lead to a new class of compact, high-current accelerators.

Jeff Hecht, contributing editor

The concept of laser acceleration of charged particles dates back to 1979 when Toshi Tajima and John M. Dawson predicted that intense laser pulses could create a wake of plasma oscillations that could accelerate electrons to high energy.1 Their computer simulation of the effect attracted considerable interest because it held out the prospect of useful acceleration over much shorter distances than conventional particle accelerators. However, the short-pulse lasers of the day could not generate the peak powers needed for useful laser acceleration.

That changed with the development of chirped pulse amplification, which can generate extremely high peak powers in ultrashort pulses. The past several years have seen remarkable experimental progress, first with electrons and more recently with protons and heavier ions. Early experiments produced particles over a wide energy range, but recent results have narrowed the range of energies, a crucially important feature for applications that require precise control over particle energies.

Particle acceleration basics

Traditionally, charged particles have been accelerated by passing them through long metal tubes in which alternating electromagnetic fields were applied to a series of segments. The applied fields reverse as the particles pass through the segments, so the fields always accelerate the particles. The longer the tube, the more acceleration is applied to the particles, and the higher their energy. Laboratory-scale accelerators have meter-long tubes, but the accelerators used in cutting-edge high-energy physics can be several kilometers long. The maximum acceleration possible depends on the accelerator structure and the power of the alternating field. The upper limit is acceleration that increases energy by tens of megaelectronvolts per meter of tube length, so very high energies require huge accelerators.

Laser acceleration can generate much higher fields, so acceleration distance can be much shorter-typically a millimeter to drive electrons to 100 MeV. Firing powerful ultra-short pulses into a plasma or solid generates extremely intense electric fields, which can reach teravolts per meter at the instant of peak intensity. These fields overwhelm the electric attraction between the positive nucleus and the negative electrons, freeing both electrons and positive ions. The process also generates intense fields that accelerate the particles to high energies over short distances.

“Electron acceleration and proton acceleration are fundamentally different,” says Thomas Katsouleas of the University of Southern California (Los Angeles, CA), so different approaches have been developed for the two. The original laser-wake-field approach proposed by Tajima and Watson deposits energy in a plasma and works best for electrons. Protons and ions are much heavier and better accelerated by firing laser pulses that explode thin-film targets, freeing bursts of charged particles.

Laser-wake-field acceleration

Wake-field acceleration is often compared with surfing. When an intense laser pulse hits a plasma, it creates a density wave of free electrons. The electrons, in turn pull positive ions-protons or heavier nuclei-along with them, creating a density wave as they pass through the plasma (see Fig. 1). The density wave carries the free electrons with it, and the electrons can reach 100-MeV energies within a millimeter, roughly 1/5000th the distance needed in a conventional accelerator.

Like catching an ocean wave to surf, coupling a laser pulse into a plasma is tricky, and the process took time to perfect. The first experiments showed very fast acceleration, but it took time to increase the number of accelerated electrons and focus them in a narrow beam. In 2002, researchers accelerated a burst of 100 million electrons that spread within an angle of only three degrees, but their energy diverged widely.2 That was a concern. “For a lot of applications it’s the ‘holy grail’ to get a monoenergetic beam,” Katsouleas says.

A major advance came in 2004, when three groups reported much narrower ranges of electron energy in papers that appeared in the same issue of Nature.3, 4, 5 The key to their success was finding ways to inject a clump of electrons into a small part of the plasma so the electrons can be accelerated collectively to nearly the same energy. The researchers fired pulses with peak powers of 10 to 30 TW and lengths of 30 to 55 fs into gas jets 2 mm long. By creating plasma channels or adjusting the laser beam to guide the density waves through the jets, they managed to constrain energy spread to no more than 24% for up to a few billion electrons (see Fig. 2).

Laser ion acceleration

Protons and positive ions are too heavy for the wake-field approach to accelerate them effectively. Instead, researchers blow them away by hitting a thin, dense foil with a pulse reaching higher than 1018 W/cm2. The electric field is so much stronger than the nuclear attraction so it blows electrons out the back of the exploding foil at relativistic speeds. The electric charge of the accelerating electrons pulls protons or heavier ions along behind, accelerating the positive ions over micrometer-scale distances. “This is a pure one-stage process,” says Juan Fernandez of the Los Alamos National Laboratory (Los Alamos, NM).

Exploding foils were long known from inertial-confinement fusion experiments. However, they were not seriously considered for laser acceleration until the Lawrence Livermore National Laboratory (Livermore, CA) unexpectedly generated a well-controlled proton beam by firing its Petawatt laser at gold foils.6 Analysis showed that the protons came from impurities in the foil. The beam was intense, but ion energy was distributed over a wide range, and many applications require mono?energetic beams. Medical therapy, for instance, requires uniform energy to ensure the particles all penetrate the same depth.

A pair of experiments reported in January in Nature took a big step toward that goal, using different approaches to generate beams of protons and carbon ions with limited ranges of energy.

A team at Friedrich Schiller University (Jena, Germany) produced the proton beam by focusing 10-TW, 80-fs pulses from a Ti:sapphire laser to an intensity of 3 × 1019 W/cm2 on a 5-?m titanium foil. On the other side of the foil was an array of polymer dots 0.5 ?m thick and 20 ?m across. When a laser pulse hit the metal side of the foil behind a polymer dot, it blew off a cloud of hot electrons on the side of the dot, which in turn pulled protons from the polymer behind them (see Fig. 3). A plot of energy of the ?roughly 100 million laser-accelerated protons showed a narrow peak at 1.2 MeV, which was matched in simulations. The group calculates it could accelerate all 800 million protons in the polymer dot to an energy peak of 173 MeV-suitable for treating deep-seated tumors-if it had a laser that could deliver peak intensity of 10^21 W/cm^2.

Fernandez’s group at Los Alamos concentrated on the more difficult problem of accelerating heavier ions. It focused 30 TW, 600-fs pulses onto 10-um spots on a 20-um palladium foil with a thin graphite layer on the back. The intensity of 10^19 W/cm^2 blew relativistic ?electrons off the surface of the rear of the foil, which in turn accelerated highly ionized carbon atoms. The group found a 17% spread in the mean energy of about 36 MeV for the most abundant ions, C+5.

A key advantage of the laser approach is the ability to generate much higher ion currents than can conventional accelerators. Mutual repulsion of ions limits current in a conventional accelerator, but the laser accelerator can produce multiple kiloampere pulses because it produces a neutral beam containing electrons as well as ions, says Fernandez. “You’re shooting a plasmoid” that doesn’t want to fall apart.

Outlook

Laser accelerators are not going to replace the gigantic particle accelerators used in ?particle physics research. In principle, wake-field accelerators might be extended to produce high-energy electrons-but not protons. “You’re never going to make very high-energy proton beams this way,” says Fernandez.

But laser acceleration has two ?other big strengths. Because they use very intense fields to accelerate particles over short distances, they can be made small enough to fit in a laboratory, especially as the size of high-power, short-pulse laser comes down. Laser accelerators can also produce higher-?power beams. This combination makes them attractive for a wide range of applications. One is treating tumors with heavy ions, which deposit little energy until their velocity slows, zapping cancer cells deep inside the body without killing tissue along its path. Laser-accelerated beams might be used for fast ignition in inertial-confinement fusion. And once laser accelerators become available, more applications seem sure to appear.

REFERENCES

1. T. Tajima and J. M. Dawson, Phys. Rev. Lett. 43, 267 (July 23, 1979).
2. V. Malka et al., Science 298, 1996 (2002).
3. S.P.D. Mangles et al., Nature 431, 535 (Sept. 30, 2004).
4. C.G.R. Geddes et al., Nature 431, 538 (Sept. 30, 2004).
5. J. Faure et al., Nature 431, 541 (Sept. 30, 2004).
6. R. A. Snaveley et al., Phys. Rev. Lett. 85, 2945 (2000).
7. H. Schwoerer et al., Nature 439, 455 (Jan. 26, 2006).
8. B.M. Hegelich et al, Nature 439, 441 (Jan. 26, 2006).

Laser Focus World April, 2006

Mr. Beam Version 3.4.2-016

Bernd sent us the new version of MrBeam, it supports 3 different kinds of CCD camera.

a. LU 160 (wavefront sensor)

b. LU080 (beam stabilization)

c. LU160 (FROG).

Boyer, Colgate awarded the 2006 Los Alamos Medal

Laboratory technical staff members Keith Boyer and Senior Laboratory Fellow Stirling Colgate, are recipients of the 2006 Los Alamos Medal. The Los Alamos Medal is the highest honor and most prestigious award the Laboratory can bestow upon an individual or small group. Director Bob Kuckuck will present the medals during a formal award ceremony and reception at 4 p.m. May 23, in the J. Robert Oppenheimer Study Center.

Recipients of the Los Alamos Medal are evaluated based on their exceptionally distinguished achievements that have impacted the success of the Laboratory, either through influencing mission accomplishments or enhancing distinction, making a contribution that changed the course of science and establishing a major direction for Los Alamos and/or the nation.

"Keith Boyer's and Stirling Colgate's distinguished careers at Los Alamos span more than four decades. Their contributions to Los Alamos and the nation have been immense and further underscore the vital importance this Laboratory has played in the past and the excellent science we continue to perform in support of the nation. I am honored and humbled to be able to present the 2006 Los Alamos Medal and I want to say to them, 'thank you for your contributions,' " said Director Bob Kuckuck.

Boyer has served the Laboratory for 55 years and is recognized as being the intellectual force behind Los Alamos’s entry into magnetic fusion, nuclear rocketry, laser isotope separation and inertial fusion. Boyer, who received his doctorate in nuclear physics from MIT, is credited with introducing and leading dramatic advances in science and engineering and has been involved in producing the first neutrons from a thermal plasma, co-inventing the electron beam carbon-dioxide laser and advancing x-ray lasers for high resolution microscopy. After retirement, he continued research and is responsible for a breakthrough in the development of an x-ray laser camera that promises to offer a nanoscale measurement technique comparable in importance to the development of the optical microscope. He is also recognized for his foresight in recruiting and nurturing excellent research people within the Laboratory.

“I was surprised and delighted to hear about my being awarded the Los Alamos Medal,” said Boyer. “I am particularly appreciative of the very generous support from my peers. I again feel that I am a member of Los Alamos Laboratory and have begun to think of new developments for the Laboratory.”

Colgate’s association with Los Alamos began when he was a student at the Los Alamos Ranch School until it was closed by the government in the early 1940s. He obtained his doctorate in physics from Cornell University and worked at Lawrence Livermore and New Mexico Institute of Mining and Technology. He joined the Laboratory in 1976 where he worked in the Theoretical (T) Division. He is recognized for leading the nuclear diagnostics of the nation’s largest weapons test conducted by Los Alamos, and for negotiating the cessation of high-altitude and outer space nuclear tests. Colgate also has inspired the inertial fusion and astrophysics programs at Los Alamos and Lawrence Livermore and contributed basic science to fusion ignition and burn, plasma confinement and shock wave physics. His other accomplishments include seminal work in supernovae and gamma ray bursts. Colgate is a recognized leader in recruiting leading weapons physicists and demonstrates by example that basic and applied science must be partners.

“So many people contribute to one’s career and this is especially true of a scientist,” said Colgate. “In serving two national labs and a state institution of learning I feel all those many people who have contributed to and encouraged me so much should feel associated with this honor.”

Colgate and Boyer join past Los Alamos Medal winners, Conrad Longmire, Nobel Laureate Hans Bethe, former Laboratory Director Harold Agnew, and Laboratory Fellows Nerses “Krik” Krikorian, George Cowan, Francis Harlow and Louis Rosen in this distinction.

(By Sallie Boorman)

Measuring the output energy

The timing delays of DG 535:
A=T+1.44us
B=A-1.28us
C=B+30ns
D=A-1.39us

The final output energy was measured about 350 mJ with the seed pulse energy 25 mJ.

X-ray Anode time delay adjustment

Randy suggested firing the X-ray Cathode and Anode slightly earlier. The previous delay time was set D=B-1.31us, in this condition the Rail Gap Anode firing was missed sometimes. When the firing time was moved forward, the Rail Gap firing became stable.

The Prometheus has been passivated for about 3 hours with the hight voltage of 10kV and DG 535 time delays of B=A-1.90us, D=A-1.45us. The red light was not changed anymore.

The Prometheus Passivation

Because the output energy of the Prometheus was too low, I decided to passivate it. The chamber was vacuumed this morning, then Helium gas was filled into the chamber about 1 bar. A 16kV H.V. pulse was applied to the rail gap. After filling several mbar fluorine gas, the emission light was changed from white to red. In order to make the railgap firing at the peak time of transformer secondary, I changed the SRS DG-535 channel B timing, from B=A-1.58us to B=A-2.20us. This status has been kept for one hour, then I found a spark on the top 10k Ohm charging resistor. I had to stop passivation, and found the resistor broken. What caused this happened?

Later I replaced the broken resistor, and tried to passivate the Prometheus again. I also measure the Thyratron currents at different timing delays(158us, 190us, 220us).

Cooling the Prometheus gases, ASE too low

The Prometheus ASE was very weak, so I decided to clean the gases by cooling recycle. After cooling, the ASE became bigger, however the ASE energy dropped fast after running for several minutes. The final amplified energy of the system is only about 100 mJ. I want to passivate the Prometheus next Monday to improve the ASE output.

Cleaning the windows

I removed the windows of the long tube, and found some small black grains on the inside surface. I carefully cleaned the window and reinstalled them immediately. After that I passivted the both tube for about one and half hours, the long and short tubes emitted red light. Finally I measured energy of the TWIN excimer with 45 mJ, and the beam pattern looked better than before.

TWIN dirty windows

I tried to measure the laser energy, however I found the output energy of LLG-TWIN was very low. This seed beam could not be amplified up to ~500 mJ after passing the Prometheus. The pre-amplified beam patten looked very bad, I checked the windows and found they are very dirty. So I want to clean them tomorrow.

Testing the Eurothrem 7100A Thyristor Controller

Introduction:

The Eurothrem 7100A Thyristor controller is intended to replace the Eratron SCR controller in the Prometheus laser.The Eurotherm units are speced at 100A 230/240V service with a 5 V analog control input signal.

Breadboard:

The following circuit was breadboarded to test the Eurotherms. A Stanford Research Instrument Delay Generator is use for timing the firing of the Eurothrem and the discharging of the holding capacitor.
Rep Rate 0.2Hz
Delays
A = T+0
B = T+50ms
C = T+1.2s
D = T+1.3s
AB triggers the 900ms flip/flop gating the analog signal to the Eutotherm. CD gates the discharge transistor. The control level is adjustable form 0 to 5V. The full wave bridge output voltage is ~ 160VDC.

Analysis:

The oscilloscope traces shows a built in time delay of ~ 200ms between the rising edge of the control signal input and the firing of the thyristors. If the Eurothrem units are used the 200ms delay will have to figured into the system timing.
Yellow = AB Trigger monitored at F/F input
Blue = Analog Control Signal monitored at Eurotherm input
Violet = E Charge Curve monitored on the holding capacitor
Green = Discharge Signal monitored at the Discharge input
(Written by Michael F. Klawitter)

Trigger channels changed

I tried to trigger channel B (M02) of DG-535 firstly using the SDG II TRIG OUT signal and set up the other channels as below:
A=B+1.57 us
B=T+30 ns
C=B+30 ns
D=B+1.3 us
Under this setup, the laser pulse output is same as the last Friday.

The system timing adjustment

Usually we sent the SYNC OUT signal from SDG II to trigger the Prometheus, it would induce about several hundreds nanoseconds delay time. Today I changed this signal from SYNC OUT to TRIG OUT. Finally I could match the seed beam with the Prometheus ASE.

Timing setup:

SDGII: OUT1=188ns, OUT2=368ns, SYNC OUT=204ns

DG535: A=T+1.60us (Rail Gap), B=A-1.54us (M02), C=B+30ns (M01), D=A-1.31us (X-ray Anode).

Jitter of laser pulse

I set the Thyratron heater and reservoir voltages back to 10.06V and 8.17V. In order to avoid the acring, I inserted a plastic plate between the thyratron and ground plate.

In the afternoon, I tried to send the TWIN laser pulse to the Prometheus. Firstly I wanted to use the method of the first pulse triggered and the second pulse amplified. However, it's very difficult to match these two pulses, I thought there was a time jitter between them. The 4 channels of DG-353 were set as: A=T+250ms 212us 236ns (Rail Gap), B=A-1.66us (M02), C=B+30ns (M02), D=A-1.42us (X-ray Anode). I used a fast photo diode to measure the laser pulse and compared it with the Prometheus trigger signals. From the measurement, I found the laser pulse jitter was more than 2.0us, which might cause the failure of two pulses matching.

Then I used only one pulse to trigger the Promethus and to be amplified. The channel A of DG-535 was changed to A=T+1.71us, and the rest channels were set as above. Please notice the time division, the above is 500ns/div, this one is 100ns/div, the tracing time is same. The jitter time is less than several nanoseconds, that's enough for matching the Prometheus pulse.

However, the time delay of 1.71us was still big, the Prometheus should be triggered early a little bit (about 0.25us). Because the Thyratron (M01&M02) would be triggered early 1.66us than the Rail Gap, it means the Rail Gap delay time must be set more than 1.66us. If we want the pulses matching, the A channel should be set A=T+1.46us. I need time to consider the whole system timing.

The TWIN system running well

I removed the bad connector of the heater, and reconnected them. The TWIN system ran pretty good. However I found there were the arcing around the Thyratron and the ground plate, this is a small problem, I will put a plastic plate between them or tighten them.

Wire Broken

I open the TWIN chamber this afternoon, and found the wire connected to reservoir voltage was broken. I think it might cause the TWIN not running. I will repair the wire tomorrow. The filter connector was not well when we repaired the TWIN on Oct. 20, 2005.

Changing Heater and Reservoir Voltages for TWIN Thyratron

I adjusted the heater voltage and reservoir voltage. These voltages were set at 10.06 VAC(Heater) and 8.04 VAC (Reservoir) respectively. In the beginning, I turn on the system after 10 or 15 minutes warming up, the TWIN ran at 5 Hz for 1 second, then the repetition rate changed to 1 Hz, after 2 or 3 shoots, there was no operation anymore. I increased the V_H to 10.71 VAC, V_R to 8.35 VAC the TWIN did not run. When I increased the V_R up to 8.77 VAC, the TWIN was interlocked as soon as I turn on.

I have to shut down the TWIN system. Several minutes later, I turn on the system and warmed up, set the V_H 11.38 VAC, V_R 8.40 VAC, the repetition rate 5Hz. This time the TWIN ran normally, then I increased the repetition rate to 6Hz, everything was ok! Then turn off the system, set the repetition rate up to 8 Hz, the TWIN was not running, after a few seconds, it ran at low rate then up to 8 Hz. After the system ran for a couple minutes, I turn it off again and increased the repetition rate up to 10Hz, this time the TWIN ran normally after 1 minute. However, when I decreased the rate down to 5 Hz, the TWIN was dead again.

I have to shut down the TWIN again and left it not run for a couple of minutes. Then I set up the V_H 11.35VAC, V_R 8.35VAC, the TWIN has been run well for about 30 minutes. Then the TWIN excimer became unstable, there were one miss fire after 10 shoots (@repetition rate 10Hz). I switched off the TWIN and waited for several minutes, then turn it on, but just run 1 minutes, the TWIN became unstable either at 5Hz or 10Hz.

I found the TWIN excimer was sensitive to the reservoir voltages, when V_R under 8.35 VAC, the TWIN could be run; when the reservoir value up to 8.35 VAC, the power supply would be interlocked. I don't know why.

The TWIN excimer laser not working

Finally we repaired the Prometheus, I tried to run the TWIN for synchronizing. However, I found the TWIN excimer did not run anymore. I changed the trigger mode to internal, there was not any output yet. I think the trigger system does not work.

Burks Oil Pump Overheating

I found the red light was on when I turn on the oil pump switch. The power light was on when I pressed the reset button on the relay, however the motor did not run any more. The overload light was on after several seconds. After trying several times, I found the smoke from the motor, I think it's burnt:(.

BURKS OIL PUMP OVERHAUL
TROUBLESHOOT

• Overheating motor may have burnt out winding. 3 phase amps check will show unequal values on 3 lines.
• Bad noises are probably motor bearing. Replace them immediately, or motor will be ruined.
• Oil pooling in the recess within the motor adaptor (item 5) indicates a worn shaft (ok if a new seal leaks for a very short time).
• Special tools and parts are in the closet near rm 2159.
• Standards tools: 3/8” ratchet with 9/16” + 7/16”sockets, and 12” extension. One large flat - blade screwdriver. 3/16” hex key.

GUTS OUT

1. Electrical: Shut off motor switch. Label Switch, “LEAVE OFF”.
2. Disconnect wiring.
3. Drain: Close all oil manidold valves.
4. Uncork 1/2 “Poly Flow plug at top of oil manifold, for air vent.
5. Tap oil into clean beaker under pump drain valve.
6. Transfer oil (If clean) to drip reservoir. Repeat until drained.
7. Replace Poly Flow plug.
8. Push “Pump Prop” disk (in special tools box) under pump housing.
9. Unbolt pump body from motor / adaptor (bell housing): (4 bolts, item 6). Don’t loosen bolts that join motor to adaptor housing. Pull motor and adaptor housing away. Some oil will gush.
   

SHAFT SEAL OUT
Note: No left-hand threads are used on this entire pump.

1. Located rubber doughnut “Suction sleeve” (Item 12) and remove it.
2. Remove “Diffuser” (cone ? shaped, item 9); just 2 small screws.
3. Remove O ring (Lab unit) or paper gasket (basement unit). If paper gasket is used, meticulously clean the sealing surfaces.

Impeller Removal:

1. Clamp special big pinch clamp around bronze impeller, with the stud sticking through a housing hole.
2. Unscrew impeller clamp bolt and brass beveled washer.Note: If screw is socket head, it has Loctite on threads. I tapped motor shaft with wrong (Course) thread, so it needs extra grip.
3. Reapply Loctite when reassembling this motor.
4. Pop snap cap off motor rear.
5. With large screwdriver, remove impeller by “unscrewing” motor shaft.
6. Catch impeller. Note flat brass washer behind impeller. Leave special pinch clamp on impeller, for reassembly.

Adaptor Housing Removal:

1. Remove the 4 housing bolts, and pull off housing. You are fighting the drag of the rotary seal part that grips the shaft. If it’s really stuck, use special “motor bell puller”, with the CPI nut to protect the shaft threads.
2. Press stationary seal out of adaptor housing.Note rubber “slinger” washer on shaft. It needs replacing if it does not grip the shaft.
3. Begin reassembly unless motor bearing need replacement.
   

MOTOR BEARING OUT
Motor Dismantle:

1. Pull off slinger.
2. Mark motor shell and ends for alignment, if not already.
3. Unscrew 4 long screws from rear of motor.
4. Pull front end off, with armature.
5. Pull bearing: Note: It’s not necessary to remove front bearing (long shaft end), and protect shaft end from damage.
   
6. Use shop puller to pull bearing off. Note: There may be a loose piece, a motor bearing clamp, laying loose behind bearing. Don’t forget to reinstall it.
7. In like manner, remove rear bearing. Note: If damage to rear bearing is no concern, you may omit adaptor.
   

New motors don’t have the required internal threads in the motor shaft end, necessary to keep the impeller from spinning off. With the rear bearing remove, you can now chuck the armature shaft in the lathe, with the long end sticking out, and drill and tap for 1/4” -28 (fine) X 1/2" deep.

NEW BEARING IN

Note: Details of pressing techniques are not discussed here. One caution: Work with shaft only. Do not support or press armature body or fan.

1. Clean long shaft of any rubber seal residue.Press bearing onto armature shaft short end, all the way to shoulder.
2. Slip front bearing clamp (a piece of the motor), if used, onto long shaft end. Using “motor bearing installer”, press bearing onto armature shaft end, but stop when bearings measure an outside separation distance of 8.190”.
3. Reassemble Motor as you took it apart. If long screws are stubborn, remove screws, sight through holes and give ends corrective rotational taps.
   

NEW SEAL IN

Note: It is reasonable to assume that a seal that has little wear and soft rubber parts should work fine if carefully dismantled, cleaned, silicone ? greased, and reassembled.

1. Reinstall slinger washer, nominal 1/8” from motor face.
2. Use steel wools and acetone to clean the seal recess in the adaptor housing.
3. Press the stationary part of the seal into the adaptor housing, rubber lit first, using “stationary seal installer”, with the cardboard disk as a protector for the ceramic face. Even, straight, firm hand pressure is enough.
4. Dab some silicone oil on the ceramic seal face.
5. Grease motor shaft.
6. Bolt adaptor housing onto motor.
7. With the steady pressure, push spring ? loaded part of seal onto shaft, graphite first; don’t turn. Push all the way, only from the spring. As per disassembly, reinstall flat brass on shaft, then Impeller (snugly), special brass beveled washer, and shaft bolt (Loctite, only if ? - 20 threads).
8. Remove impeller clamp tool.
9. Replace motor rear snap cap.
10. Reinstall diffuser.
11. Grease and install new paper gasket or # 2- 263 O ring, as req.
12. Reinstall suction sleeve on end of diffuser.
    

GUTS IN

1. Bolt motor adaptor to pump housing, keeping flanges evenly gapped.
2. Retrieve 7/16” pump prop.
3. Reconnect wiring.
4. Set valves for full oil flow.
5. Waite 10 minutes for oil to flood the pump.
6. Remove “LEAVE OFF” label.
7. Turn on pump.
   

Redrawing the system timing schematics

I combined the front end laser and the Prometheus to run. I drew the timing schematics to understand the system control easily. Then I tried to synchronize the seed laser pulse and the Prometheus. I filled new gases into the Prometheus and ran it at 20kV. This time the Prometheus was trigger by the seed laser pulse. When I adjusted the time delay between the Railgap and the X-ray anode around 1.93 microseconds, I measured the maximum UV light. However, I could not synchronize the seed beam with the Prometheus very well. Another problem is the oil recycling system, the oil pump could notbe run, the light show it's overheat.

Eurotherm Firing Angle Control for Prometheus Operation

1. Introduction

Recently UIC purchased a Eurotherm 7100A SCR Controller to replace the no longer supported Eratron FPE 202-D30 SCR Controller. This Technical Note will discuss ways in which to utilize not only the Euroterm Controller but also any generic SCR controller for the charging of the main Thyratron Capacitor Banks. One goal of changing to a commercial controller is to eliminate much of the troublesome Firing Angle and PC Board Control cards that were utilized for the Eratron unit. The Eurotherm unit as ordered is a basic unit without “alarms” or “soft start” features that are useful in driving inductive loads like the HV Transformer. More advanced units incorporate a “safety ramp” that involves progressively increasing the Thyristor firing angle in order to apply the voltage (and current) to the load smoothly and thus reduce the start-up current of loads which either have a low resistance when cold or are inductive. For inductive loads, the “safety ramp” prepares the initial magnetization of the transformer to avoid saturating transformers on power up that can lead to large in rush currents to the primary. This note will show via simulations that it is relatively straightforward to control the present Eurotherm unit and incorporate “soft start” features.

2. Circuit Model Description
This circuit is essentially the same as that shown in the note “Firing Angle Control of SCRs during Charging”, 26 March 2006. The main addition to that circuit is the generation of the SCR control voltages (E1 and E2) by the Eurotherm unit, which is driven by a 0 to 5 Volt DC control signal. V1 is the 208 VAC rms line voltage and the “ideal” HV transformer has been modeled as having a 32kVAC rms output voltage (turns ratio of 154) with a secondary resistance, R5, of 2K ohms. R1 is the 0.1-ohm peak inrush current limiting resistor to protect the SCRs. The diodes are made up of three SCH20000 diodes in series and labeled as D1 thru D4. R3 is the series-charging resistor of 100K made up of ten; 10K 225-watt wire wound power resistors. The North and South capacitor banks are labeled as C1 and C2 and interconnected with a 3K resistor. R2 is the 70-ohm resistor for purposes of monitoring the pulsed charging current (Imon). The actual Pulse Modulator (PM) HV Schematic of Prometheus (see D-2151) shows a Transzorb 5KP6.0A that clips the Imon signal at a voltage ranging from 6.67 to 7.37 volts. This is done in order to accommodate the Pulse Integrator of the PM HV Control Board (see C-2154). This is not necessary for use of the Eurotherm unit and no feedback current via Imon is needed.

3. Results of Simulations

In the event of the SCRs becoming shorted or commanded to full on, the resistor charging string is the only limit to the charging current pulses and the time that it takes for the system to reach a full voltage of 32kVAC rms x 1.414 = 45 kVDC. The purpose of the PM Voltage Control Timing Circuit (see C-2155) is to limit the charging time available with the SCRs commanded to full on such that the Thyratron Bank Voltage as indicated by Emon never exceeds 30kV although the Anode rating of the CX1622 Thyratrons is 35kVDC. This time should include the “soft start” feature of about 8 full SCR firing cycles or 0.133 seconds. The results of a simulation showing the SCR Firing Angle (set at 180 deg), the Imon current pulses, the Emon charging voltage, and the first 0.2 seconds of the Transformer Primary Voltage, gives a maximum desired charge time of 0.820 seconds. The Top trace (Red) shows the Firing Angle ramping up to 180 degrees in 0.133 seconds with the corresponding Transformer Primary Voltage being given a “soft start” over 8 full cycles as shown in the Bottom trace (Black). The Imon current pulses are shown in the 2nd trace (Magenta) and charging resistor limited to 420ma with the Emon voltage of the 3rd trace (Blue) indicated at the complete charge of 30kV in 0.82 seconds.

If the SCRs or the Controller fails such that the SCRs stay on, then the only protection to stopping the charging cycle of the capacitor banks is to command open the AC Contactors that drive the input to the HV transformer. This means that if the Emon voltage continues to increase after 0.82 seconds, the contactors must open before 1.25 seconds, which is the time that the system will reach 35kVDC on the Anodes of the Thyratrons. Some sort of comparator circuitry for Vref and Emon would appear to be needed to actuate the opening of these HV transformer contactors. Although a version of this circuitry exists on schematic C-2154, there are simpler ways to implement this fault protection. One way that is already being used is that the Thyratron banks are being fired just after the time when the charge is complete.
Although it might seem complex to generate a ramp for the firing angle control, a simple pulsed RC circuit (time constant about 0.2 seconds) driven by the present Output Transistor of the PM Voltage Control Timing Circuit (see C-2155) could be utilized. The results are shown on the bottom four plots on the next page and can be compared to those at the top of the page for the linear ramp of the “soft start”. Note that any additional complexity in ramping the top of the firing angle voltage is not necessary since ALL SCRs and Diodes are being operated safely within their limits.
(Written by Randy)

Clipped Imon Signals

Randy sent an email talking about the I_mon Clipped signals.
=======================

I looked more carefully at the Schematic D-2151 and just noticed that the Transzorb is actually a 5KP6.0A which has a breakdown voltage between 6.67 and 7.37. The scope traces that you sent show the Imon peaks clipping at about 6.7 volts and so this all makes sense now. This explains to me why the current peaks have always looked near constant in amplitude and yet the Emon voltage charges exponentially. The "true" Imon pulses are probably decaying exponentially in time as I have calculated in previous models. This also means that the circuitry of schematic C-2154 is not driven by what the Imon current is really doing; the Imon pulses are acting more like an "enabling" gate. This circuit is not as smart with Imon feedback as I had thought but nevertheless I still need to understand what it does in more detail.

In summary, I am looking at this circuit (C-2154) more carefully to understand what it really does. I will also model and make some comments on what the Eurotherm Controller should do for a DC command input and how the system should charge versus various but constant phase angles (this is very simple to implement). In my mind, the ideal Firing Angle input should have a "soft start" ramp for about 8 AC cycles followed by an "average" Phase Angle that has an upward ramp to somewhat increase the area of the charging pulses and therefore enable the Thyratron Capacitor Bank to charge more near linear versus exponential. Although this process can be more efficient and quicker in time, the actual charging of the banks is of little consequence as long as no SCR or Diode ratings are exceeded. The ratings are safely limited by the 100k resistor charging string other than for the initial turn on that is governed by the magnetizing current of the transformer core.

Randy

Computer control SDG II

The SDG II can be controlled with a standard RS-232 serial port. I made a standard 9-pin D-sub male/female extension cable for hookup. Only three pins are used for serial communications:

Pin	Function
2	SDG II trnsmit data, computer receive data
3	SDG receive data, computer transmit data
5	Gound

Through the serial communication, I can control the Hurricane laser repetition rate acuurately. By sending command "set:rate 0127", the rate was fixed at 507Hz/127=3.99Hz.

Synchronizing the seed pulse

Previously we used the pulse generator to trigger the Prometheus, actually the trigger pulse should be generated from the front end excimer laser system. So I tried to send the laser pulse to trigger the Prometheus today. I set the laser repetition rate at around 4 Hz, then it was divided by 10 times, this 0.4 Hz pulse was the main pulse to trigger the Prometheus. I used a photodiode to measure the front end laser pulse. There was a jitter about ~4 microseconds between the Thyratron trigger and the laser pulse.

Timing for whole system

It's very import to synchronize the seed laser pulse with the Prometheus. In our old system, a 10 ms delay time has been added before the seed pulse was sent to Promtheus, which made it very easy to trigger the gas nozzle when the laser beam reached. However, in the new system it's only about 2 microseconds was given. We had to trigger the nozzle using the former laser pulse, which made the nozzle time delay more than 2.5 seconds when the system were run at 0.4Hz. I changed the system timing shown as the right diagram. The Prometheus will be fired when the next seed laser pulse comes, which will give the later equipment enough time(~250ms @ 0.4Hz) to be adjusted. I think this is better than 2.5 seconds time delay.

Moving the time delay forward

I measured the voltage and current of H.V. transformer when I changed the time delay. Here is the scope traces when the delay changed from 1280ms to 1580 ms @ 10 kV. Please note the current wave form @ delay less than 1480 ms, there was a current later than the charging current.

There were no miss fire when I set the delay time less than 1480 ms when I increased the high voltage up to 20 kV. After I fix the Prometheus timing, I will try to send the seed beam and adjust the time to synchronize them.

Calibrating the charging voltage

We calibated the charging voltage at increments of 10kV, 15kV, 20kV and 25kV using a Fluke high-voltage probe connected to the Thyratron capacitor bank and read with a Fluke digital meter. The results of these measurements indicated that the system charged to about 19kV when the control panle meter actually readout a voltage 0f 24.5kV. This error is merely caused by an error in the high voltage divider at the output of the transformer that needs to be fixed. Unfortunately, this also meant that the maximum voltge that the system could be charged to was 19kV.

During these measurement, it was noticed that erratic charging occured at the higher voltages above 15kV and, in fact, the sytem operated correctly only every other pulse. The pulse rate was then slowed down to one pulse every 5 seconds. Although the initial timing adjustment seemed logical and worked up to 15kV, further insight from the above trace indicated that it is desirable that there be a longer recovery period after thyratron firing before commanding the system to charge. The Thyratron should recover very rapidly (less than 1ms)but the induced transient noise in the system may need a much longer recovery period in order for the present firing angle electronics to work properly. This can be seen in the slight negative signal right after Thyratron firing and to last about 0.2 seconds. Rather than increasing the Charging Delay to about 2.5 seconds as had been done, the delay should have been reduced to about 0.8 seconds. This would have resulted in a charging delay of 0.8 seconds, a charge time of 0.9 seconds, and a charge hold time of 0.8 seconds for a total cycle time of 2.5 seconds. As mentioned previously, the system was finally run at 0.4Hz at 15kV for about one hour with no adverse heating of the charging resistors being noticed. Although the evidence presented somewhat confirms that timing has been the main problem during the past 4 months, only operation at higher voltages and longer periods of time can guarantee that the charging resistor problem is “really” fixed. (from Randy's report)

The below is the experimental data.

DIAL	H.V. Probe(kV)	Control Panel Meter (kV)	C-2153 J1 (V)
6.16	11.10	14.0	3.35
7.10	13.00	16.5	3.91
8.72	15.25	19.6	4.6
10.00	19.00	24.5	5.7

Randy put a Rogowski coil to measure the current pass from the Thyratron to ground. When we carefully adjusted the time delay bewteen the rail gap and the Thyratron firing, we certainly found the current closed to 0. This result totally matched Randy's simulation.

Adjusting the time delay 2498 ms

After review of the timing diagram (shown as the left schematics), we decided to adjust the time delay between the firing angle and the thyratron firing.

The scope traces are shown below. The lower trace shows that the charging current pulses are nearly constant in amplitude. This does not lead to constant power charging of the capacitors as evidenced by the exponential rise of the voltage in the middle trace. This exponential rise is further aggravated by the fact that the firing angle of the top trace is decaying in time whereas it should be probably be increasing in time.

Invesigating the main charging system

We put the good Eratron SCR controller connected to H.V. transformer, then we set the charing voltage 10 kV to measure the charing voltage and transformer secondary current. It was shown in the right picture, these wave forms are similar to the scope traces recorded on 9 March, 2006. Note the second step in the firing angle indicating that the charging cycle has been disrupted by the Thyratron firing too soon. It was hypothesized that the extra current drawn before and during the Thyratron firing caused the 10K resistors of the charing string to overheat.

Assuming the repitation rate is 0.4 Hz, the Stanford Box will send a trigger pulse per 2.5 second. According the old timing setup, the firing angle should be sent after 2.1531 seconds. From the left timing diagram, the Thyratron was triggered during the capacitors were charging. The time delay between firign angle and thyratron firing was around 347 ms, which matched the scope traces. The timing must be adjusted to move the charging time backward or forward such that the charging cycle started after or before the thyratron firing.

Removing ERATRON from X-ray Anode control system

By analyzing the x-ray anode trigger circuit, we concluded that we can remove the ERATRON and SCR power block from control unit and drive the system directly with Hiptronics H.V. power supply. As shown in the right picture, we removed the SCR and connected 208 VAC directly with step-down 240/120 autotransformer to drive Hiptronics 30 kV/5 mA power supply. All but the voltage meter were removed from the X-ray anode PCB, the remainder of the board served no function.

In the beginning, we could not drive the Hiptronics to 30 kV, the output was only 15 kV. However, the output could reach 30 kV when we energized it with regular 120 VAC power. So we tried to exchange 208 VAC phases, the Hiptronics ran normally after phase exchanging. I could detect the x-ray flux.

The RC charge time of the Hipotronics power supply is dominated by the 2.5M series charging resistor and the three 0.1uF capacitors in parallel (total value of 0.3uF), which are located in the X-ray Anode oil tank. The calculated time constant is 0.75 seconds. Since the system has a cycle time between pulses of 2.5 seconds, this means that the Hiportronics will charge to about 95% of the set value. It was found that good x-rays were generated for a set voltage of about 25kV (this also corresponded to a meter reading on the main system control panel of about 150kV). A setting of 20kV was too low and no x-rays external to the machine were measurable. (by Randy Carlson)

Randy measured the Primary resistance of H.V. transformer by measuring the current of a DC voltage around the transformer. The value is about 30 miliOhm. Unfortunately we were not successful in measuring the inductance with General Radio Bridge.

Installing the New Grounding System

Randy and John came to help for maintaining the Prometheus. They brought the scanned and clean schematics of the system.

We also changed the old system, this time the ground is connected by the copper sheet, which will provide a good grouding for the Prometheus. Randy and I discussed to try running the X-ray anode without the SCR controller, which would simplify the control system, we will test it tomorrow.

Time of Flight Mass Spectrometer

We already installed TOF. The right picture shows the grids and the shell. The below is the schematic of TOF in our lab. We planned to put another MCP on the right side to measure electrons or ions. The disk shows the connector for TOF.

Ordering EUROTHERM 7100A SCR controller

We found a SCR Power controller similar to our SCR controller. The product is from EUROTHERM.

The Model 7100A is a new range of economic SCR Power Controllers for use with resistive, infrared or inductive loads. This unit features integral heatsinks with analog voltage or current inputs for precise control.

Our ordering code is 7100A/100A/240V/230V/XXXX/MSFU/PA/XXXX/0V5/ENG/NONE

Transformer for X-ray Gun H.V. Power Supply burnt

Mike replaced the IC chips and transistors already, then we send a control signal pulse to measure the SCR output. The waveforms looked well, so I install the SCR controller in the X-ray Anode control box.

We tested the whole system in the afternoon, unfortunately the x-ray gun could not be run correctly. The x-ray anode high voltage was not monitored, while the high voltage connected to the Thyratrons was pretty good. After running a few minutes, I smelled the smoke. Then I found the transform connected between SCR and X-ray anode H.V. power supply was burnt.
===========================
AUTOTRANSFORMER P-8634
Primary ---- Line Cord
230V 50-60Hz

Secondary ---- Receptacle
115V@400VA

STANCOR Chicago, IL

Data acquiring system

We received the Tek 2024 scopes this afternoon, John helped me to install them in the rack. I connected these 2 scopes and TDS 520 to the computer with GPIB cables. The data can be acquired in real time, even it's not as fast as the scope response. I think it should be enough to obtain the data at 0.4 Hz. Actually we don't need to acquire the data for every shoot, we only need some special data for diagnostics. Previously Keith only used two channels scope to track the signals, which could provide enough information. Now we are using 10 channels to measure the signals.

ERATRON FPE 202 D30 Circuit board

Mike already removed all IC chips and put the sockets for pluging easily. He wants to replace all the IC chips from IC1 to IC8 and transistors Q1 to Q3. Until now we received the chips except IC4 and IC5.

No.	Chip/Transistor	Function
IC1	TL084CN	Low Noise JFET Quad OpAmp
IC2	CD4093BE	14-Dip Quad 2In Schmit Trgr
IC3	LM741CN	Operational Amplifier
IC4	MC14027B	Dual JK Flip-Flop
IC5	CD4046BE	CMOS Micropower Phase-Locked Loop
IC6	TC084CN	Low Noise JFET Quad OpAmp
IC7	LM555CN	Timer
IC8	CD4082BE	CMOS Dual 4-input AND gate
Q1	J113	N-Channel Switch
Q2	MPS A13	NPN Darlington Transistor
Q3	MPS A13	NPN Darlington Transistor

Changing the Beam Stabilization Setup

It's difficult to recover the beam stablization setup if we move it, so I changed the setup. Kevin helped to make the plate and the posts. I aligned the beam into the beam expander and make sure the expand beam finally enter the target chamber correctly. This time we only used one mirror to steer the beam to the CCD camera, I will test the imaging system soon.

Repairing the SCR controller

Mike contacted PLC Center requesting the circiut drawing of SCR controller, unfortunately the company did not have any drawing. Like Mike doing, they only can replace the chips one by one to test the logical output. Up to now, Mike already ordered the chips and transistors for the circuit repairing.

beam stabilization system

Regarding the installation of the LLG beam stabilization system at UIC it can be
stated that
1. The optimizations works properly.
2. The closed loop control works in a standard configuration i.e. beam control is achieved with a stationary 500 or 4 Hz pulse train respectively.

However, some problems encountered during the installation schedule have to be solved:
1. The motherboard of the PC delivered by LLG is defect.
The fastest solution, we suggest is that you buy a identical mainboard and replace the old one on your own. Regarding the compensation of your financial effort, we should discuss with Klaus Mann and Uwe Wachsmuth next week.
2. The contrast ratio of the pre-amplified pulses and the frontend ones are approximately one to two orders of magnitude too high. Thus, at an attenuation level sufficient to perform the closed loop control of the frontend beam,the CCD chip gets damaged after illumination with very few (app. 10) amplified pulses.
Actually the chip shows 7 damaged pixel, but this virtually does not hit the performance,as we use only a 1/16 part of the chip, so a non damaged part can be selected.
The best solution we think would be the use of a variable attenuator, based on a mechanical shutter. This shutter should have a contrast ratio of 100, and could easily be triggered by the 4Hz Trigger of the TWIN amplifier without any software support. This could be done in Goettingen, but doing it by yourself in Chicago will be much faster.
3. The closed loop control based on switching the frondend from a 4Hz to a 500Hz mode for a short time interval between two consecutive 4Hz TWIN pulses does not work. Part of this failure seems to be the high overexposure mentioned above. Another reason is probably some internal logical software bug. - However, the mode switching works, and closed loop control has been achieved without amplifier, although the software fails to set the correct values the the pulse timing schedule.

This problem should be solved in Goettingen during the next weeks. An updated version of the program may then be tested by Dr Song, who will have my full support via email or phone.

(Reported by Bernd Schaefer)

Testing computer control system

We connected SDG-II to computer with RS-232 port.

Installing laser beam stabilization equipment

Dr. Bernd Schaefer came to help us install the equipment of laser beam stabilization today. I will assist him. I ran the TWIN excimer amplifier, its output energy reached ~30 mJ. By sending less than 1% transparent beam behind a reflected mirror to the equipment, we could monitor the laser beam profile. The problem is after the same ND filter the transparent of seed beam is too weak compared to the pre-amplfied beam.

The computer was sent from Germany, we could drive it by 115V power supply. I ordered a 110V 500W power supply for computer.

Exchanged SCR Phase Controller

Because the X-ray anode preionizer system used the same high voltage control design as thyratron discharging. We exchanged the SCR phase controller ERATRON FPE 202 D30 for troubleshooting.

This time, I can control the H.V. pulse for thyratrons, but cannot control the x-ray anode H.V. pulse. We confirmed that SCR controller cannot be run well. When the H.V. pulse modulator set scroll was tuned from 3.0 to 6.0, the current waveform at transformer primary was followed to change. The voltage meter showed the H.V. changed from 5 kV to 15 kV.

I also monitored the transformer secondary current I_mon and SCR firing angle. We finally concluded that SCR firing phase controller was not well, Mike helped to replace some problematic chips. ERATRON has been not existed, it's impossible to buy the same controller. The similar controller may be found to replace. A company named PLC Center provide repairing service. We would like to send this block to them for repairing if we cannot do it.

Failure of H.V. pulse control

Like before, we failed to control the high voltage output. When I turned on the H.V., even without the modulating pulse (it means no firing angle signle applied to SCRF-1), we could measure the high current through the transformer primary. So I disconnect the primary of big transformer and connected them to a step down (40:1) small transformer. Then we measured the wave form from the secondary. We always saw the wave form (see the picture) which could not be adjusted.

SCRF-1 pin 12

We received the 451L chip this morning. Then we installed it on circiut board. The right picture shows the firing angle wave forms with/without H.V. pulse modulation. When I tuned the voltage amplitude, the firing angle voltage was followed to change.

Unfortunately, there was not any output yet. I checked the wires and found pin 12 on SCRF-1 was losing. Maybe this caused no high voltage.

Frequency voltage converter AD 451L

Today we combined the C-2154 circuit board and SCR together to check, we found pin 13 at C-2154 was not connected to common point. This caused the output of firing angle floating. After we fixed it, the firing angle pulse was changed. The test result was shown on the right picture, the wave 1 is the pulse without out common, wave 2 is the pulse after fixing.

Then I installed C-2154 and SCR back into the system for testing, unfortunately the high voltage could not be measured yet. The difference is that there was no high voltage output even after running for a couple of minutes. So we wanted to remove the C-2154 card to check again, however, we accidentally took off the card with the power on. We found the power was still on even I turn off the switch. Then we found the frequency-voltage converter AD-451L does not work, it must have been damaged when we removed the circuit board. I already ordered 451L from Newark. We will change the chip and fix the switch next Monday.

Check SCRF-1 and SCR independently

I removed SCR Firing Circuit SCRF-1 (ERATRON FPE 202 D30) and SCR (Powerex CM431290) from M03 board. Mike connected them (circuit shwon as yesterday) to check them independently. The firing angle input was a pulse with amplitude of 0~5 V and duration of 140 ms, pin 7 of SCRF-1 connected to ground. An 1:4 step down transform was connected to SCR out port. When we tuned the firing angle amplitude, we measured the transform secondary wave form. Actually we have not found anything wrong of these measured parts, the output could be adjusted by tuning the firing angle amplitude. We will send pulses generated by C-2154 board to SCRF-1 for checking the output tomorrow.

I also filled the new gases into the LLG-TWIN excimer and ran the front end system.

SCR gate 2 and gate 1

We firstly disconnected the transform primary to measure the SCR gates G1 and G2. The picture of wave forms is shown below. During measurement we found the firing angle was changed from pulse to continued voltage. I thought this might be caused by the feedback.

In order to send a correct control pulse to SCR controller, we sent a pulse generated by the function generator. This time we connected the transform primary, the waveform was shown as left bel. We found only SCR gate 2 pulse was modified, the gate 1 was not changed anymore.