Transcript Presentation for Electronic system review

CMS ME CSC HV system
Alex Madorsky
University of Florida
Cathode Strip Chambers
Main purpose of the CMS EMU CSC HV
system:
 Provide
High Voltage for CMS Endcap
Muon Cathode Strip Chambers (CSC)
CSC features that affect HV system
design:
HV segments – high tolerance to
HV failures
 Same working voltage with small
variations from segment to segment
 Problematic segment can be fixed by:
 Reducing voltage
 Disconnecting from HV
 Needs precise consumption current
measurement for each segment
 Small
One HV segment
November 2003, CERN
2
Alex Madorsky
Voltage and current parameters
Cosmic Ray Count Rate, 4/6
0.2
0.18
Count Rate (kHz)
0.16
Ar+CO2+CF4
0.14
0.12
40+50+10
0.1
0.08
0.06
0.04
0.02
0
3
3.1
3.2
3.3
3.4
3.5
3.6
3.7
3.8
3.9
4
HV (kV)
Voltage:
•The operational point 3.6 kV (full efficiency)
•The end of plateau is at 3.9 kV
Current:
•Current per channel averaged over the full Encap Muon System: ~0.7 uA/segment
•Maximum expected current per segment: 2uA
•Needs to be monitored on each segment with good precision, to detect possible troubles.
November 2003, CERN
3
Alex Madorsky
UF/PNPI design
UF/PNPI
HV system design:
 3.5
years of development
 3 prototypes + pre-production prototype produced
 Prototypes passed all tests
November 2003, CERN
4
Alex Madorsky
Target specifications (1)
System structure
See Figure 1
Mainframe Power supply functions
1. “Master” High Voltage (Vmax) generation and regulation
2. Distribution boards control
3. Distribution boards low voltage power (under discussion)
4. User interface
5. CMS SCADA interface
126 + 13 spares 36-channel boards
144 + 15 spares 30-channel boards
Number of distribution boards
Number of output channels of the
distribution board
36 or 30
Distribution board output organization
Output HV connectors
One connector LEMO REDEL SLA.H51 for 30 HV outputs or two LEMO REDEL SLA.H51
for 18+18 HV outputs
Pin assignment defined by customer, connectors with the HV wires (1 m) attached are
supplied by customer. There are 30 or 18 wires for HV outputs in each connector, 3 ground
wires, 2 interlock and 2 reserved wires.
4000 V
Maximum output voltage, Vmax
Voltage regulation individually for each
output, software programmable
Vmax – 500 V to Vmax, with the possibility to turn off
Voltage regulation resolution, individually
for each output, software programmable
Less or equal to 50V
November 2003, CERN
5
Alex Madorsky
Target specifications (2)
Channel to channel output voltage
difference
Voltage output
Capacitive load for each output
20 V max
Maximum output current, Imax
100 A
Individual output turn-off (trip) speed
Programmable, from  1 s
Trip level, software programmable
1 to 100 A
Trip level setting resolution
1 A
Hardwired trip level (erroneous software
protection)
100 A
Maximum total output current of the board
(sum of all outputs)
40 A * number of outputs
Ripple and noise
Common mode ripple, measured on 1
KOhm resistor
10 mV p-p maximum, bandwidth 100 Hz – 20 MHz
50 mV p-p maximum, bandwidth 100 Hz – 20 MHz
Mutual influence of channels
No trips because of other channel(s) tripping
Voltage measurement, individually for
each output
Current measurement, individually for each
output
Channel to channel measured current
difference
Via software, resolution 10V, 0 to Vmax
November 2003, CERN
Floating (HV return wire insulated from the system ground)
0 - 60 nF
Via software, resolution 100 nA or better for currents 0 - 1A, 10% or better for currents
>1A to Imax
10% of measured value max.
6
Alex Madorsky
Target specifications (3)
Current measurement period for each
output in the entire system
10 sec or less
Rate of voltage change, software
programmable
2 to 50 V/s
Output control via software
Low voltage power input and control
Status: OK, tripped, interlock status, overload, on/off, ramp up, ramp down, current
limit/measurement
Required, with an option to disable, software programmable. The chambers are equipped with
the interlock switch, and HV cable has two interlock wires.
From mainframe or external, 9 boards per 1 crate
High voltage power input
From mainframe, one SHV connector
HV, LV and control connectors’ positions
See Figure 2
LED indication
HV on/off, LV on, trip, interlock open
Ambient magnetic field for distribution
boards
0.3 Tesla constant field, B-field map is available
Radiation hardness of the distribution
boards
2*1011 neutrons/cm2 and 0.5 krad of ionizing particles
Slow control
Connection to CMS SCADA required
Construction of distribution board
6U or 7U Eurostandard board, up to 700 mm long, 9 or 8 boards per 19” crate (see Figure 2)
Protection loop (interlock)
November 2003, CERN
7
Alex Madorsky
Target specifications (4)
Mainframe Power sources
(270 outputs)
One HV conductor per
distribution board (270) plus
control and Low Voltage
power
Distribution Boards
• System structure defined by us
One conductor
per segment (8856)
• Master HV sources and control
computers in Control Room
Chambers
• Voltage regulation and
monitoring, current
measurement by Distribution
boards near disks
~100m
~12 m average
Figure 1
November 2003, CERN
8
Alex Madorsky
Target specifications (5)
36-channel distribution board
Two types of distribution boards:
• 36 channels (two small chambers)
• 30 channels (one large chamber)
6U-7U
Output connector defined by us.
Output connectors
SLA.H51, for two
small chambers,
shown with the
mating cable
connector
Input connectors: HV, control.
HV connector is SHV.
Front panel
81 mm
30-channel distribution board
6U7U
Output connector
SLA.H51, for one
large chamber,
shown with the
mating cable
connector
Input connectors: HV, control
Front panel
~ 700 mm max
Figure 2
November 2003, CERN
9
Alex Madorsky
UF/PNPI HV system architecture
Counting Room
Detector Area
Card 72
Remote
Distribution
Card Type 1
Card 9
Primary HV Supply
Card 1
Card 1
Master
Distribution
Card
Long
Distance
HV
Cables
Multiwire
HV
cables,
~ 100
long
100 m,
onemper
18
SHV Connectors
distribution boards
on both ends
Remote
Distribution
Card Type 2
Multiwire HV Cable
~ 12 m long
LEMO REDEL Connectors
on both ends
 •Primary
Three Main
Units: Primary
Master Distribution Card, Remote Distribution Card
HV power
supplies:HV
offSupply,
the shelf
 One Primary HV Supply per up to 9 Master Distribution Cards
board:
One output
distribution
board. Outputs
Regulates voltage 0-4KV (VMAX), measures current on each
 •Master
Master
Distribution
Card:per
1 Input,
8 Independent
output.
 One Master Distribution Card per 8 Remote Distribution Cards
 •Remote
RemoteDistribution
Distributionboard:
Card Type
1: 1one
Input,
30or
Independent
Outputs (36 outputs max). Regulates voltage 1KV down
powers
large
two small chambers
(One
Card
per
one
ME
23/2
Chamber)
from VMAX, measures current on each output. Each output can be disconnected from HV if necessary.
 Remote Distribution Card Type 2: 1 Input, 36 Independent Outputs
(One Card per two ME1 Chambers)
November 2003, CERN
10
Alex Madorsky
Control interface
HOST PROCESSOR UNIT
VGA
USB-GP-IB
HOST PROCESSOR
ETHERNET
SCADA INTERFACE
(DIM SERVER)
PRIMARY HV
POWER SUPPLY
PCI BUS
SERIAL BUS 6
72 HV CABLES
(ONE PER
REMOTE CARD)
9
9
REMOTE DISTR.
CARD 1
REMOTE DISTRIBUTION CRATE 8
REMOTE DISTR.
CARD 9
REMOTE DISTRIBUTION CRATE 2
REMOTE DISTR.
CARD 1
REMOTE DISTR.
CARD 9
REMOTE DISTRIBUTION CRATE 1
MASTER DISTR.
CARD 9
HOST CARD
6
UP TO 144 MULTIWIRE HV CABLES IN TOTAL UP TO 2582 LEADS
November 2003, CERN
11
Alex Madorsky
REMOTE DISTR.
CARD 9
9
REMOTE DISTR.
CARD 1
HOST CARD
5
SERIAL BUS 5
HOST CARD
2
SERIAL BUS 2
SERIAL BUS 1
HOST CARD
1
MASTER DISTR.
CARD 1
MASTER DISTRIBUTION CRATE
US CMS Review
on June 24th 2003 in UF
UF/PNPI system selected over CAEN
Reasons:
 Price
 Design features:
Conducted
Simple and robust design
No programmable logic in radiation – no SEU
November 2003, CERN
12
Alex Madorsky
UF/ PNPI
CMS EMU CSC HV System Main Design Features
Main technical approaches are shown
HV
regulator
Current sensor
Fuse control
Digital control interface
Mechanical design
November 2003, CERN
13
Alex Madorsky
HV regulator (distribution board)
HV IN
HV OUT
Q1
C2
R1
HV MONITOR
+
C1
R2
HV CONTROL INPUT
U1A
+
-
R3
U2A
CONTINUOS CLOCK
Q2
• Output voltage controlled by linear regulator (Q1)
• Regulates down to –1000V from input voltage
• Voltage measured by divider R1-R2 and U1A opamp.
• Regulator feedback via U2A
• Q2 and C1 provide HV decoupling
November 2003, CERN
14
Alex Madorsky
Current sensor
I
U=IRs
R2
R3
Rs
Cv=KU
Ug
D1
C1
C2
CHARGE SENSITIVE AMP.
R4
Q=UgCv
U3A
Uout
+
R5
Cf
Uout=QCf=UgCvKuIRsCf=KI
• Current measured across Rs
• Varicap D1 is used as voltage-sensitive element
• Input pulse is applied via C1
• U3A is a charge-sensitive amplifier
November 2003, CERN
15
Alex Madorsky
Fuse control
HV IN
MASTER CARD
4KV REGULATOR
POSITIVE
REMOTE CARD
HV CABLE
HV CA B LE
1KV REGULATOR
FUSE
100mA
HV RELAY
DIODE
RES IS T OR
Q3
LV POWER SUPPLY
NEGATIVE
GND
100K
FUSE CONTROL
COUNTING ROOM
DETECTOR AREA
• Situation requiring permanent disconnect is extremely rare (never happened on FAST sites)
• Fuse is used to disconnect channel from HV permanently
• To blow fuse:
• Low negative voltage applied to channel input
• Switch Q3 shorted
• Fuse can be quickly replaced during short access
November 2003, CERN
16
Alex Madorsky
Control interface
LOGIC
FROM/TO 36 HV CARDS
36 CHANNEL HV MONITOR
36 CHANNEL I MONITOR
36
CHANNEL HV CONTROL
ANALOG
MULTIPLEXER
SERIAL DAC
ANALOG
MULTIPLEXER
SERIAL ADC
CLK
N MODULE
COUNTER
RST
ANALOG
MULTIPLEXER
SERIAL ADC
CLK
ADDRESS
COUNTER
RST
CLK R/W
DATA OUT
DATA OUT
12 LINE HVM SERIAL BUS
• Differential signal transmission (RS-485)
• Optically insulated
• Built completely on discrete logic
November 2003, CERN
17
Alex Madorsky
DATA IN
Control software
Based
on PVSS and DIM server
Initial version of DIM server and PVSS shell works
Written with excellent assistance of Valery Sytnik (UC
Riverside)
Targeted for full DCS compatibility
Work in progress
November 2003, CERN
18
Alex Madorsky
Mechanical construction
• Final mechanical construction
• Simple and rugged design
• PCB is optimized for automatic assembly
November 2003, CERN
19
Alex Madorsky
Distribution Rack
Fan unit & heat exchanger
Need from CMS:
Distribution crate
Distribution boards
1.
Racks
2.
Fan units & heat exchangers
3.
Strain reliefs
4.
Space in front and behind the
racks
5.
Low Voltage power for
distribution boards
HV and control cables patch
panel
Output HV cables to chambers
November 2003, CERN
20
Alex Madorsky
Distribution Racks
Disk 1(Station 1)
Disk 2 (Stations 2 and 3)
Disk 3 (Station 4)
Position in Rack
Rack 1
Rack 1
TOP
Crate 1: 936
Crate 2: 936
Crate 3: 936
Crate 4: 936
Rack 1 (right half
of the disk)
Crate 1: 930
Crate 2: 930
Crate 3: 930
Crate 4: 930
Crate 5: 936
BOTTOM
Rack 2 (left half
of the disk)
Crate 1: 930
Crate 2: 930
Crate 3: 930
Crate 4: 930
Crate 5: 936
Crate 1: 936
In the table above:
• 9x30 means 9 boards of 30 channels. One board of 30 channels powers one ME23/2 chamber
• 9x36 means 9 boards of 36 channels. One board of 36 channels powers two ME23/1 (or similar)
chambers
• This table shows the HV distribution boards necessary for one Endcap (+ or -).
November 2003, CERN
21
Alex Madorsky
Rack position for YE1 and YE2
YE1 has only one rack
November 2003, CERN
22
Alex Madorsky
Low Voltage Requirements for
Remote Distribution Cards
Parameter
Min
Max
Positive voltage
7V
8V
Negative voltage
-8 V
-7 V
Current on both
channels
Power per
distribution board
Ripple/noise
300 mA
4.2 W
4.8 W
100 mV
• Low voltage power will be provided by CMS AC/DC LV system
November 2003, CERN
23
Alex Madorsky
Cooling
Only remote distribution racks are discussed.
Dissipated heat:
4.8
W maximum per distribution board (about 3-4% of one
chamber LV power)
~216 W per rack maximum (45 boards)
~1335 W for all distribution boards
Cooling of distribution boards:
No
enforced cooling is currently planned
Racks must be open on top and bottom for convection
Need heat exchangers to remove generated heat
May need fans (unlikely, will decide later)
November 2003, CERN
24
Alex Madorsky
Safety
HV Cables
KERPEN
halogen-free cables
Passed CERN flammability test
HV Connectors
LEMO/REDEL,
bought from CERN stock
PCB material
FR-4,
flammability rating 94-V0
Other components
Will
be checked for CERN safety compliance
November 2003, CERN
25
Alex Madorsky
Design status
Boards’ design
complete (electrical and mechanical)
Pre-production prototype constructed in UF, under tests now
Tests of the pre-production prototype:
 Full bench test – OK
 Chamber test on FAST site – OK
 Radiation test – OK
 Magnetic field test – November ’03
Production boards - exact copy of the pre-production
prototype
November 2003, CERN
26
Alex Madorsky
UF-PNPI collaboration
MOU
between UF and PNPI is signed
Arrangement is very similar to chamber production
UF
responsibility:
 Development
and production management
 Pre-production prototype construction and testing
 Test stands construction
 Test procedures verification, instructions
 Off-the-shelf components procurement
 Bare PCBs manufacturing
 Automated SMT assembly
 US labor and components contingency
November 2003, CERN
27
Alex Madorsky
UF-PNPI collaboration
PNPI
responsibility:
 Simple mechanical components manufactured
 Pre-production and production manual assembly
 Pre-production and production testing
 PNPI labor and space contingency
November 2003, CERN
28
Alex Madorsky
Schedule
– November ’03
Board production and SMT assembly start in US – end of
November ’03
Start of pre-production run in PNPI – end of January ’04
Pre-production system test in UF – May ’04
PNPI production readiness review, production start – July ’04
Production finish – June ‘05
ESR
November 2003, CERN
29
Alex Madorsky
Installation and commissioning
Installation:
To
be done by CERN crew & UF/PNPI visitors
Will start as soon as the first shipment arrives to CERN (Oct
04’)
Very uncomplicated
278 distribution boards, 30 crates
HV cables already installed by that time
Commissioning:
power supplies are necessary – at least prototype
Would like to start as early as possible (Oct ‘04)
LV
November 2003, CERN
30
Alex Madorsky
Conclusions
Design
solutions are proved to be working
Pre-production prototype built
Pre-production prototype passed tests
Satisfies CMS EMU CSC HV system specs
Production documentation is being prepared
November 2003, CERN
31
Alex Madorsky
Radiation environment
Expected:
Neutron
Fluence: (1 - 4) x 10^10/sq cm
Total Ionizing Dose: ( 0.07 – 0. 7) kRad
November 2003, CERN
32
Alex Madorsky