Power Supply Upgrade for ATLAS Liquid Argon Front End

Download Report

Transcript Power Supply Upgrade for ATLAS Liquid Argon Front End

Power Supply Upgrade for ATLAS
Liquid Argon Front End Crate
• Motivation for Upgrade
• Potential Companies
• Work to date and future work
Motivation for Upgrade
• Reliability – Power supply made by Modular
Devices Inc has been made more reliable but is
still not at a comfortable point yet.
• The upgrade environment is harsher in terms of
radiation requirements (MDI supply is not
qualified)
• Different voltage/current/distribution
requirements (no information yet)
Environmental Qualifications in Fingers
•
•
•
•
Anticipated Radiation Environment for Upgrade (including safety
factors)
TID 4500 Gray
NIEL 7.7 x 1013 neutrons/cm2 for 1MeV neutrons
SEU 2 x 1013 cm-2 of 20 MeV or greater protons/neutrons
These numbers will change when actual data is available.
Magnetic Field Environment
• 70 Gauss when shielded with top plate. This number is changing
(increasing) at this point but should not pose any problems
Approach
• RELIABILITY
Whatever power supply approach is selected the # 1 priority must be
reliability.
• Other configurations than our present system should be considered.
For instance
1) Centralized or Distributed (i.e. board level) system or perhaps a
combination.
2) AC or DC input. If DC what is the lowest voltage we can supply
from outside?
Constraints. There are some things which can not easily be changed
from the present configuration and need to be incorporated in any
upgrade.
Constraints in Approach
• The only practical way to deliver power from the outside
is in higher voltage/lower current to keep the size and
mass of the cables reasonable
• Very difficult to change existing cooling system
• Existing cabling should be used, specifically the input
cable (280 Volt) leading down to the tile fingers
• Monitoring and control configuration should be
maintained
• Limited volume requires a high power density in the
converters
Three Potential Paths at this point
•
Vicor – commercial products with no radiation
hardening
•
Wiener – produces power supplies for existing
applications at Cern. Units have been tested in
radiation environments but are not specifically
designed for a radiation environment.
•
AEi Systems – engineering firm experienced in dc-dc
converter design and components used in radiation
environments. Also industry leaders in WCCA (Worst
Case Circuit Analysis)
Vicor Advantages and Disadvantages
•
•
•
•
Advantages
Very high power density equal to the best products available,
good noise characteristics
Very flexible architecture which should make it possible to easily design for
custom applications. Parts are readily available for testing. Power density
allows
•
•
Disadvantages
Products not intended for radiation environments requiring parts changes
and multiple iterations of testing of changes. This could be very expensive
Company is only mildly interested in making the product line radiation hard.
Most of the work is on our side. However, we do have a non-disclosure
agreement in place with them and we are getting information
“Factorized” architecture uses a low voltage (48-70 volts) which would
require new cabling or a front stage converting the 280 volt to 48 volt.
•
•
.
PRM (Pre Regulator Module)
+ VTM (Voltage Transformation Module) Combination
Provides Regulation and Isolation/Transformation
(slide courtesy of Vicor)
PRM
VTM
Regulation
Wide range
DC bus
e.g. 36–75 Vdc
100% effective duty
cycle at any Vout
Isolation & Transformation
NP
L
NS
K = NS / N P
Factorized bus
Regulated load
voltage
Vicor module configuration as might be used in our
present requirements
• 7 voltages with
redundancy using 24
pairs of PRM + VTM
combination.
+70 V
INPUT
• 70 volt input
• Would require
replacing the existing
cable from USA 15
with a larger diameter
cable for current
requirements or using
a front stage
converting 280 volts to
70 volts.
VICOR -4
PRM + VT M
VICOR +6
PRM + VT M
VICOR +4
PRM + VT M
VICOR +6
PRM + VT M
VICOR -4
PRM + VT M
VICOR +6
PRM + VT M
VICOR +4
PRM + VT M
VICOR +6
PRM + VT Me
VICOR -4
PRM + VT M
VICOR +6
PRM + VT Me
VICOR +4
PRM + VT M
VICOR +6
PRM + VT Me
VICOR -4
PRM + VT M
VICOR +6
PRM + VT M
VICOR +4
PRM + VT M
VICOR +6
PRM + VT M
VICOR -4
PRM + VT M
VICOR +6
PRM + VT M
VICOR -7
PRM + VT M
VICOR -7
PRM + VT M
VICOR +11
PRM + VT M
VICOR +11
PRM + VT M
VICOR +7
PRM + VT M
VICOR +7
PRM + VT M
+6(100AMP)
+4 (130AMP)
-4(180AMP)
+6(150AMP)
-7(16AMP)
+11(20AMP)
+7(16AMP)
Vicor VTM module
(figure courtesy of Vicor)
• Voltage divider (or
current multiplier)
with isolation but no
regulation
• VTM at 50 – 60 krad
due to failure of the
control chip (PWM)
failure mode was
determined by Vicor
• VTM MOSFETs were
tested separately and
failed at 3.5-4.5 Mrad
(good news)
Other Vicor modules tested
• PRM (non-isolating buck converter) was
not tested independently but along with
VTM to 50 – 60 krad. No failures observed
• BCM (Bus Converter Module) – works as
a functional combination of PRM and
VTM. Voltage input shorts after 35 - 40
krad. Reason is not known.
Vicor ( future work)
• The MOSFET in the VTM module will be tested for
Single Event and displacement damage effects.
• The control chip (PWM) in the VTM fails at a dose of 50 60 krad. Determine what is required (time and
resources) to produce a radiation hard control chip
• Test the PRM module independently of other modules to
determine radiation damage effects.
• Use multiple VTMs in parallel to investigate issues of
current sharing and noise.
• Determine magnetic field effects for VTMs and PRMs
Wiener Advantages and
Disadvantages
• Advantages
• Company has a history of cooperating in producing
products for radiation and magnetic field environments
particularly for high energy physics.
• Disadvantages
• Products are not specifically designed for radiation
environments
• For the existing power requirements power density is an
issue and the existing products would have to be
redesigned and/or repackaged
• Small company – has limited staff and resources
Wiener ( future work)
• The MOSFET that is used in their existing
products will be tested for Single Event and
displacement damage effects.
• Testing of the present product line to the
qualifying Total Ionizing Dose
• As part of a backup solution to the MDI power
supply a repackaging of the current Wiener
modules might be done to achieve and
demonstrate the required power density.
• Parallel multiple modules to investigate issues of
current sharing and noise.
AEi Systems Advantages and
Disadvantages
• Advantage
• Experienced in power supply design and circuit analysis. This
approach has the greatest probability of being trouble free.
• Company is very interactive, cooperative and open with information
• Disadvantages
• Expensive Uses radiation hard components in design. Estimate of
cost would be > $4 million for upgraded power supply system.
• Company is and engineering firm, not manufacturing firming. A
manufacturing partner would have to be found (Algen?) which
complicates this approach and increases the risk.
AEi Systems ( future work)
• We have initiated a discussion with AEi on strategies to
develop a supply for the upgrade.
• They are open to finding ways to produce a lower cost
solution using commercial component part substitution
for radiation hard components. An initial suggestion from
them includes the substitution of an IGBT for the
MOSFET. Another possibility would be the use of a pchannel MOSFET rather than an n-channel MOSFET.
• As part of a backup solution to the MDI power supply the
MOSFET and other components selected by them will
be tested for ionizing radiation, Single Event and
displacement damage effects.
Conclusions
•
Power supply has been much harder than originally
anticipated.
• Lessons learned on the present power supply should be
taken to heart and the power updating should be started
now. Any solution should include a Worst Case Circuit
Analysis as performed by AEi Systems. This gives some
confidence that the power supply system would have an
adequate life.
•
Testing of Wiener, Vicor, AEi and perhaps other vendors
should continue.
Conclusions (FY08 Budget)
• Manpower
• EE: 0.25 FTE - 40k
• Tech: 0.25 FTE - 30k
• Material costs
• Lab Instrum. - 12k
• Material/Samples/Modules - 10k
• Radiation Testing
• Beam/Travel - 16k
• Total: 108k