Power Supply Polarity Reversal for e e American Linear Collider Physics Group Workshop University of

1
Power Supply Polarity Reversal for e- e-
American Linear Collider Physics Group Workshop
University of VictoriaJuly 28-31, 2004

• R.S. Larsen, P. Bellomo C. M. Spencer
• Stanford Linear Accelerator Center

2
Goals

• 1. Estimate the incremental cost of implementing
  reversible power supplies throughout the positron
  Injection System to enable e-e- operation.
• 2. Make a similar estimate for the Beam Delivery
  area.
• Use available cost estimating data from 1999 and
  2003 to determine the numbers types of power
  supplies involved.
• Since lattices are still being refined quantities
  will be approximate.

3
Background1

• Prior talks at e- e- LC workshops at UC Santa
  Cruz considered options for enabling positrons to
  be alternated with polarized electrons in an e
  injection system.
• Erickson (Ref. 1, 1997) reviewed potential issues
  in the areas of Injection, Main LINAC and Beam
  Delivery
• Injection requires reversal of all magnets
• Main LINAC quads work for either e- or e with
  VH beam profiles interchanged, with a p phase
  shift at injection
• Beam Delivery needs reversal of all magnets
• Conclusions Adding e-e- feature to existing SLC
  rejected due to time cost of manual polarity
  reversal in transport line and modifying
  interaction region. Concluded that e- e- in TeV
  machine must be planned designed up front.

4
Background 2

• Late 1999, permanent magnets were being
  considered for many parts of NLC Injection as
  well as Main LINAC, to save power and/or costs.
• Larsen (Ref. 2, 1999) considered three concepts
  for Injection System
• Polarity Reversal Model
• Direction Reversal Model
• Independent System Model
• 1999 Conclusion Because permanent magnets made
  polarity reversal very difficult, and the
  independent system would be exceedingly costly,
  the Direction Reversal Model was deemed most
  practical.
• Actual costs not estimated Beam Delivery not
  addressed.

5
Direction Reversal Model 1999
6
Direction Reversal Features

• New polarized e- source
• Accelerate e- in pre-linac to 2 GeV
• Build new bypass for e target section, 2 GeV
  linac, pre-damping ring, DR injection line
• Build new injection, extraction lines for e- in
  opposite rotation in DR
• Build new launch and extract lines for reverse
  direction in turnaround
• Add matching components to launch into main linac
• Main advantage No power supply reversal so
  compatible with e permanent or electromagnets
• Main disadvantage New earthworks and transport
  lines, injection and extraction lines cost
  significant but not estimated formally suggested
  further study.

7
Independent Systems Model - 1999
8
Independent Systems Model

• Complete duplicate injection system for polarized
  e-
• e to/from e- operation quickly reconfigurable
  from control room
• Both systems could be brought up, switched
  rapidly
• Would require only modest additional space if
  shared tunneling planned up front
• Operationally ideal solution but not seriously
  considered due to high cost
• 90M for magnets and power supplies alone

9
Background 3

• In late 2002, permanent magnet RD put on hold,
  NLC Baseline Model returned to electromagnets for
  all applications
• Without PMs in picture, Polarity Reversal Model
  was reconsidered.
• Larsen (Ref. 3, 2003 ) Workshop paper suggests
  considering H-Bridge bipolar switching supplies
  throughout e injector complex.
• No detailed technical or cost analysis discussed.

10
Polarity Reversal Model- 2003
11
In the Cold Light of Dawn

• Bipolar H- Bridge supplies not a good general
  solution
• Require large AC line transformer to isolate HF
  switching transformer (switcher normally runs
  directly off the line)
• Adds considerable bulk, cost to every unit
• Control through zero not necessary
• Space requirements increase by 2X
• Solid State Reversing Switches can do the job
  simpler, in less space and cheaper.

12
Technical Cost Considerations

• Main additions to system are polarized e- source,
  transport injection lines, and reversing
  switches.
• New IGBT components make reversing much simpler,
  faster than mechanical switches or manual
  cable-swap.
• IGBT switches can be made very compact and
  low-cost.
• Speed of switching of the order of minutes or
  longer is adequate high-speed not required since
  interleaving of beams not a goal.
• Power supply requirements detailed in 1999,
  updated 2003, are used for determining number of
  switches needed. These still need further
  updating but are fine for calculating approximate
  incremental cost.

13
e Injection System Magnets by Type(Quantities
Rev. 071404 by C. M. Spencer)
Legend PPLePre-LINAC, PPDRePre-Damping Ring,
PDReMain Damping Ring, LTRLinac-to-Ring,
XFERTransfer, BCBunch Compressor
14
Typical Power Supply Types
Corrector MCORs Bipolar
20-90 kW EMHPs
2 MW String Supply
2.5-17 kW Intermediates Solid State Inverters
290kW Bulks
15
Typical PS Types 2
Rear Panel Outputs
Intermediate Rack-mounts
200kW Chopper Module
Bulk 200 kW Chopper Racks
16
Reversing Switch Proposal

• Goal Operate e-/e switchover entirely remotely
  from Control Room (no wrenches).
• Design family of switches to match the various
  classes of power supply family
• For smaller rack-mount supplies, package
  air-cooled switches in separate 3U high
  rack-mount chassis located adjacent to supply.
• For larger free-standing supplies, can easily
  mount water-cooled switches inside typical large
  cabinets.

17
Reversing Switch Components

• IGBT Switches are packaged in 1, 2 or 4 units
  depending on power ratings.
• Superior on/off control, low gate drive c.f. SCR
  or GCT (Gate Commutated Thyristor) switches.
• Assembly consists of
• 4 IGBT switches in H-Bridge configuration
• 4 Gate Drivers
• Heatsink (air or water cooled)
• Copper bus and hardware
• Control Interface
• Enclosure and Connectors

18
Cost Estimating Procedure 1

• Power Supply Systems bottom-up cost estimates
  originally done for DoE Design Review in 1999,
  revised by C. M. Spencer in 2003
• Designed estimated cost for every unique magnet
• Assigned power supply from commercial in-house
  look-up table
• Calculated cooling, AC requirements, cable plant,
  IC
• Summed all component costs by major sub-areas
• Gun area, e target, pre-linacs
• Pre-damping ring, transport lines
• Main damping ring, post linac, turn-around
• These data used to obtain numbers and types of
  power supplies
• Numbers then used to calculate incremental
  Reversing Switch costs.

19
Cost Estimating Procedure 2

• Created conceptual design, cost estimate for
  seven (7) Reversing Switch models (P. Bellomo,
  2004).
• Added columns to original 1999 power supply
  spread sheet rollup.
• Estimated Reversing Switch cost increment for
  fully redundant e system
• Power Supply redundancy model taken from 1999
  estimate.
• Estimates made for both Injection and Beam
  Delivery.

20
Reversing Switch Estimate Example
21
Reversing Switch Model Summary
22
Reversing Switch Cost Summary Injection Area
23
Reversing Switch Cost Summary Beam Delivery Area
24
Assumptions, Caveats, Disclaimers

• Costs cover completed Reversing Switch hardware,
  mounted either inside purchased supply or in
  separate rack-mount chassis.
• Only half Pre-Damping Ring included since e- beam
  traverses only half the PPDR on its way to Main
  Damping Ring (JS).
• Not estimated
• Polarized electron source, associated transport,
  controls.
• Spin rotators, polarimeter, phase shifter
• Injection extraction kickers
• EDI, installation, system integration, checkout,
  commissioning, contingency, GA, sales tax,
  foreign exchange, VAT, beer, wine...

25
Discussion Conclusion

• IGBT Reversing Switches are a practical low-cost
  solution to polarity switching for e- e-, costing
  of the order of 5M for additional hardware for
  Injection and Beam Delivery areas.
• For comparison, the total is 5.5 of the 93M
  total for power supplies for the Injection Region
  alone, assuming full redundancy.
• A comprehensive technical study, cost estimate
  and operational model is needed if e- e- is ever
  to become part of the ILC plan.
• An important side-benefit of polarity reversal is
  that e rings could be tested initially with
  electrons (J. Sheppard).

26
References Acknowledgments

• 1. Erickson, R, e- e- Collisions in a TeV
  Collider Built for e e- Operation,
  SLAC-PUB-7768, March 1998.
• 2. Larsen, Raymond S., e- e- Switchover in the
  NLC LINAC, Intl J Modern Phys A Vol.15 No. 16
  (2000) 2477-2483
• 3. Larsen, R. S., Revisiting e- e- Switchover in
  the NLC LINAC, SLAC-PUB-10452, February 2004.
• The original models for Injection alternatives
  were suggested by J. Sheppard, Head of Injection
  Systems for the NLC Department, SLAC.
• C.M. Spencer and P. Bellomo provided the Magnet
  Systems cost data and the Reversing Switch design
  and cost estimates respectively.
• These contributions are gratefully acknowledged.