Laser Power Supply

Laser Power Supply

Second Semester Report

Spring 2007

Josh Cook

Prepared to fulfill the requirements for

EE 402

Department of Electrical and Computer Engineering

ColoradoStateUniversity

Fort Collins, Colorado80523

ABSTRACT

In an academic environment where we design and build our own lasers to meet the needs of specific applications, it is often impossible to find a high voltage, commercially available, power supply which can produce the exact electrical requirements necessary for its operation. It is then necessary to design and build a specialized power supply to for these unique requirements. This project is to build, test and implement two pulsing units for the 46.9 nm capillary discharge high repetition rate soft x-ray lasers employedhere at CSU, extrapolating from and improving on a previous design.

The bulk of the design in this project is not in the circuit itself but in the implementation of the connections between the components and the support structures within the pulsing unit. The electrical connections with in the unit have been altered from the original unit to be more uniform, smaller, easier to fabricate and easier to connect and disconnect. The support structure within the pulsing unit which consists of half inch Plexiglas panels has been altered to be less complicated, more stable, and to facilitate cooling oil flow to components which may overheat. All of this has been done bearing in mind that all components and connections will be carrying high voltages and currents.

TABLE OF CONTENTS

Title1

Abstract 2

Table of Contents 3

IIntroduction4

IIReview of Previous Work6

IIIPhysical Design

Electrical Connections7
Internal Support Structure8

IVCosts 9

VResults and Future Work10

Acknowledgements 13

LIST OF FIGURES

Figure 1Barrel connector7

Figure 2Lengthwise Dividers8

Figure 38Ω Resistor Supports9

Figure 4Pspice Circuit11

Figure 5Pspice Output11

Figure 6 Unit 1 Output12

Figure 7Unit 2 Output12

LIST OF TABLES

Table 1Cost10

CHAPTER I -Introduction

For this project we will need to take in a DC voltage of approximately 50 kV and provide a pulse of 80-90 kV for a very short time of 1 micro second. We will also need to provide a pre-ionization voltage to the lasers cathode prior to the main voltage pulse. We have available to us a commercial high voltage DC power supply that can produce 50kV, we will use this to supply the pulsing unit and spark gap switches to pulse it. The step input is then sent through an RLC resonance circuit that will provide overshoot which we will use to charge a second capacitor bank, which is external to the pulsing unit, to approximately 80kV. The second capacitor bank will then be switched by a second spark gap switch and discharge directly to the laser.

The unit is housed in a Steal box that has internal dimensions of 20 and 5/8 inch by 15 and 1/8 inch by 7 and ½ inches deep. The box is water tight and has water tight fittings on the back for wire and tubing penetrations. The lid is ½ inch thick aluminum with an o-ring groove and a connection for the oil out. Inside the steal box is a Plexiglas box which is used to electrically isolate the entire circuit from the grounded outer steal box. The only penetrations through the Plexiglas box are in the back for wire and tubing which must go to external equipment, and in the bottom for the grounding bar stands, which are welded to the steal box. Inside the Plexiglas box the circuit is divided into three sections by ½ inch Plexiglas dividers. The left section is for the pre-ionization circuitry, the center holds the 60nF capacitor bank, the bus bars and the spark gap switch, and the right sections will hold the trigger circuitry on the bottom level and the inductor and output resistors on the top level.

The main issues when physically assembling this circuit are controlling the extremely high voltage, and providing cooling to the components that may overheat. Both problems are solved by immersing the circuit in high impedance oil. Air will conduct at about 30kV per inch, so using voltages as high as 90kV can cause problems with arcing. Arcing will not only interfere with the operation of the circuit but may also cause damage to it. By submerging the circuit in transformer oil, which has a dielectric strength of 100 kV/cm, all of the wires and connections will be sufficiently isolated from ground and from lower voltage components such that no arcing will occur. Still it is important to keep components of different voltages a distance of at least 1cm apart to maintain a sufficient margin of error (at least 1 inch from ground was also maintained).

The dielectric oil is also used to cool the circuit by being constantly circulated through the circuit and an oil chiller. The two components which are most in danger of overheating are the 1 MΩ pre-ionization resistor and the two 8 Ω output resistors. In order to achieve this, the oil inlet tubing has been routed within the pulsing unit directly to these components and the Plexiglas dividers designed to guide the flow of the oil to these vital cooling loads. This was achieved simply by not allowing other paths for the oil to flow and opening paths of low resistance for the oil to flow past the vital cooling loads. In the second unit the 1 MΩ resistor came from the template unit and already had an oil tube around it so the oil is directly routed in to the tube.

The support structures and the internal components are not attached to the insulating Plexiglas box or the external steal box except at the grounding points for the grounding bar. This power supply is not meant for use in situations where it may be jarred or moved while in use so this will not put additional strain on components or connections. This is done to make it easier to disassemble for quick access to internal parts in order to facilitate maintenance, repairs and trouble shooting. This being said, the components inside do not move freely because the internal support structure is designed to be a tight fit.

CHAPTER II – REVIEW OF PREVIOUS WORK

Prior to the current two units that are being produced now there have been several units built by different people. The only ones that have been available for reference on specific design improvements during this project the one that I was given as a template and the one built by Scott Heimberg who is the graduate student I am working with on this project. This is due to the fact that they are all currently in use. I was also provided with a binder containing the dimensions and specifications of many of the parts I would need to fabricate, which was made by David Braley.

The base circuit design has not changed since the original unit, but there have been various improvements along the way. For example in the original unit the circuit and Plexiglas box was attached to the lid of the steal box, effectively making it upside-down in the box and very difficult to take apart for repairs. The units now rest inside the outer box such that the circuit can be accessed easily without draining the oil, just by taking off the lid. Also the means by which the high voltage wires that enter and exit the box have been change form an L type connector pipe that came out of the lid to a straight water proof fitting that comes out of the back of the box. This makes it easier to remove any damaged wires and replace them relatively quickly as well as more space efficient in the rack where the unit is usually stored.

CHAPTER III – PHYSICAL DESIGN

PART A – ELECTRICAL CONNECTIONS

In the unit that was being used as a template there are a large number of different types of connections. Some were simple barrel connectors and some were complicated, using stands and brass pieces that were welded together and some had specially fabricated parts. In this version all connections for wires are made using either standard ring connectors or barrel connectors as shown in figure 1. The barrel connectors are small easy to make and provide for a large surface for astrong electrical connection. These barrel connectors are made from ¼ inch brass rod stock and drilled down the center for the wire to be inserted and soldered. A ¼ inch hole is then drilled into the large brass discs and a set screw hole is tapped perpendicular into the side of the hole. A set screw is then used to hold the barrel connector in place. Ring connectors were used in places where barrel connectors would not fit or would be inconvenient. By using only ring and barrel connectors the construction of this unit is significantly less complicated and more easily taken apart for maintenance or repairs. It should be noted that in the second unit some of the specially fabricated connection pieces from the template pulsing unit were reused along with much of the template units components in order to save time and money.
B – INTERNAL SUPPORT STRUCTURE

The internal support structure is used both for mechanical support and to electrically isolate lower voltage components from higher voltage components. There are two large dividers that divide the circuit as describe in the introduction, these dividers are 19 and 5/8 inches by 6 and ½ inches and have sections cut out to allow wires to pass through and to facilitate oil flow. There is a hole in the right divider for an oil inlet line to be run to in order to cool the output resistors and inductor directly.

There are several other smaller pieces in the unit which support individual pieces of equipment. The spark gap has two Plexiglas pieces which are 9 and 1/8 inches long by 6 and ½ inches tall that support it completely via holes in the center which the brass input and output connectors for the spark gap fit into. On both of the outside divisions there is an upper and lower level. On the left side the lower level contains the 1MΩ pre-ionization resistor which is supported with two Plexiglas pieces, one on each end; these pieces also support the platform for the upper level. The piece that supports the 1MΩ resistor near the back of the unit has one 5/8 inch hole in it for an oil line to supply cooling oil directly to the resistor and the one on the front has two large holes (one above the resistor one below) to allow the oil to flow freely. On the right side the trigger circuit is on the bottom and the capacitors for the trigger circuit are suspended by the Plexiglas dividers which supports the top level. The right side top level contains the two 8Ω output resistors which are supported together by two Plexiglas pieces which each contain four additional large holes to allow the cooling oil to flow freely to the oil return line and ensure adequate cooling.

CHAPTER IV – COSTS

Due to the nature of high voltage equipment and the fact that most of the items were bought one at a time rather than in bulk, many of the pieces of equipment that have been purchased for this project are quite expensive. Things like screws, ring connectors and 100kV cable are not included because many of the components were found already in the lab. I was able to save money by using the scrap materials found in the machining room next to room B-310 at the ERC for the raw materials necessary to fabricate many of the components. Also the steal box and the associated fittings were already completed before I started this project, though the bill for having the lid machined and the matching hole pattern put into the box and lid is on the list under CNC Steal Work. I do not expect that there will be many more large purchases as we have all materials necessary to complete the second unit and finish the project next semester. Table 1 is a list of all of the major expenses that have occurred while I was working on the project. If you include the cost of the capacitors, the spark gap, and the other resistors the cost of a single unit is about $5,500. This can very by several hundred dollars depending on whether the resistors are orderedin bulk or not. Ordering in bulk can reduce the cost to around $100.00 apiece, and ordering them one at a time will raise the cost to about $425.00 per resistor.

Item / Bought From / Date / Cost
Tube Fittings / Denver Valve and Fitting / 07/07/06 / $52.00
HV Cable / McMaster-Car / 07/07/06 / $101.00
CNC Steal Work / Acra-Tech / 07/18/06 / $400.00
Reducing Union / Denver Valve and Fitting / 07/21/06 / $56.00
Rubber Tubing / McMaster-Car / 8/14/06 / $36.75
500Ω Oil Impregnated resistors x2 / U.S. Resistor / 08/15/06 / $532.36
Oil System Components / Denver Valve and Fitting / 08/18/06 / $256.20
Male Elbow, Reducing Union / Denver Valve and Fitting / 09/08/06 / $105.30
Ball Valve / Denver Valve and Fitting / 09/22/06 / $160.90
TOTAL / $1700.51

Table 1

CHAPTER V – RESULTS AND FUTURE WORK

Testing was performed on the first unit successfully producing the expected results. Testing did take longer than expected because we did not have the pre-ionization circuit properly grounded causing it to act as a very large capacitance and giving the circuit an extremely over damped response. This had nothing to do with the unit itself, it was a problem with the way we were testing it. The results of testing for the first unit are shown in Figure 6, the expected results from the Pspice model are shown in Figure 5 and the circuit is shown in Figure 4. Testing for the second unit went much smoother and only took about 45 minuets once the equipment was set up. The results for the second unit are shown in Figure 7.

Figure 4

Figure 5

Figure 6

Figure 7

With both units now completed and tested the support structures for the laser itself are currently being constructed and the discharge head being assembled so that the units can be tested, actually firing a laser at full power.

ACKNOWLEDGEMENTS

This project has been funded and supported by Colorado State University Extreme Ultraviolet Engineering Research center and Jorge Rocca. Guidance on the process and techniques used came from Scott Heinbuch as well as Dale Martz and Michael Grisham.