Transcript Design of the HV box

The Babar IFR High Voltage
System
Naples, 12/15/1997
Electronic review
Marcello Piccolo
Napoli Dec., 1997
1
Requirements
 Rpc’s require about 8 KV to operate (with
the babar gas mixture)
 No dramatic differences in knee voltages
have been experienced in testing the 800
modules needed for the IFR.

 Design a system based on relatively
small number of supplies and a complete
monitoring system.
Marcello Piccolo
Napoli Dec., 1997
2
Knee Voltage Distribution
(Barrel Rpc’s)
50
45
40
35
30
25
20
15
10
5
9000
8750
8500
8250
8000
7750
7500
7250
7000
6750
6500
6250
6000
0
HV for 90% efficiency (Volt)
Knee voltage distribution barrel chambers
Marcello Piccolo
Napoli Dec., 1997
3
Features
 The resistivity of the RPC electrodes allow the
compete decoupling of the H.V. From the pulse
read-out .
 Ground loops can be minimized and , provided
that power supplies for the front end are on the
detector, no connection at the ground level exists
between the detector and the LEACH.
 Monitoring individual current on the ground leg.
 Easy decoupling of different modules
Marcello Piccolo
Napoli Dec., 1997
4
Power Supply Selection
 At 8 KV operating voltage the RPC dark current
has been measured 5 ma/m2 at 18oC.
 The total current for the IFR would be @10 ma.
 Sy-127 from CAEN has been selected as H.V
power supplies.
 This device consists of a mainframe which does
H.V and current setting , ramping, computer
interfacing for up to 10 H.V. Pods.
 Pods are specifically designed for different types
of detector: The RPC model is the A330P/N.
 One pod has two independent outputs that can
deliver 10kv@1ma.
 The nominal design capability of the system has
been set at 36 ma,…. But everything can be easily
upgraded to 80 ma.
Marcello Piccolo
Napoli Dec., 1997
5
Design Philosophy
 Given the relatively small variance of the knee
voltage one is led to design a system with a
relatively small number of independent power
supplies, and split a single (pod) output into
many H.V. Feed for the RPC’s.
 In doing so one has to abide few rules:
 A single RPC connection must not bring down
the pod output even if is shorted out.
 Some provision should be made to remotely
disconnect sick//broken modules.
 An early warning system should be part of the
H.V. Distribution so that potential malfunctions
could be spotted and cured as soon as they
develop.
Marcello Piccolo
Napoli Dec., 1997
6
The Schematic Connection to
the RPC’ Modules
At the detector level there is a second connection between
the H.V. ground and the detector ground (the black
wire that carries a 10KW resistor in series
Marcello Piccolo
Napoli Dec., 1997
7
Design of the H.V. box
 In order to comply with the design rules
mentioned above one has to implement:
 Current limiting on every output channel
 Fuse type of device (remote disconnect) in case
of serious fault
 Extensive monitoring of each individual counter

Bounds on the series resistor and fuse
Single current monitoring to be implemented
Marcello Piccolo
Napoli Dec., 1997
8
Design of the H.V. box (cont.)
 A safe design ( which will handle @ 10 KV )
must rely on the principle:
small distances  small voltage differences
H.V. and ground from the power supplies separated
in the box and kept at 10 cm. distance :
distribution boards do not have more that few
hundred V. differences between components.
Use H.V. wires and connector at the joining of the
wires: H.V. feed to the RPC have to use single
wires as the two connection are separated in space
(on opposite corners of the detectors).
Dimension the H.V. series resistor so that the extra
current load due to a short afterward does not
knock down the relative pod output.
Typical current output 400 mA , so choose 20 MW,
( 400 mA @ 8 KV)
Marcello Piccolo
Napoli Dec., 1997
9
The H.V. box
Marcello Piccolo
Napoli Dec., 1997
10
Hardware components
 A330/P (the H.V. pods) use KINGS type
connectors (Reynolds parts # 1064-1)
H.V. coaxial cable ( Reynolds type C parts #
167 2699) does the long haul from the
LEACH to the detector in the standard wireways.
A total of 80 cables will be pulled
(within January)
 H.V. boxes on the detector will be housed
in Al containers which are mechanically
interlocked with the H.V. coax feed.
 Output of the H.V. box utilize bipolar
connectors from Alden (parts #
G400/G401 ) rated @ 10 KV.
 Connection to the detector done with
H.V. wire from Alden (parts # 8014-22)
 Detector connection with adhesive
copper tape glued onto Bakelite with
conductive epoxy.
Marcello Piccolo
Napoli Dec., 1997
11
Hardware components
(H.V. boxes)
 Two single layer boards:
 H.V. board contains a 20 MW alumina resistor
rated @ 5 W with a 2.2 MW , .125 W resistor in
series, which acts as a fuse at @ 300 mA .
 Ground boards contains the LEACH-Detector
decoupling resistors , the current monitor resistors
and the L-shaped connector to output the current
monitoring signal.
 H.V. boards housed in metal container (6 of
them). Half a barrel sextant or 1/3 of each half
end-caps is serviced by one container.
 Three H.V. cables feed one container.
 Nom. voltage and 2 extra voltages.
 H.V. wires exit the container through small
vertical slits, to prevent accidental disconnection
and exposure of the H.V. pins.
 Containers lined with Palusol ( fire barrier
material)
 One Klixon @ 85 oC interlocks the CAEN
supplies
 One + 1/5 60 cond. Flat cable suitably split inside
the container carry away monitor signals.
Marcello Piccolo
Napoli Dec., 1997
12
Cabling plant
The total number of long haul cables needed is 72 as
the total number of Al container is 24. We plan to
install 80 H.V. cables from LEACH to Detector.
This number matches the total number of pods
(40) that we could install.
We plan to have four mainframes in LEACH: one for
the barrel, one for each end-caps and a spare to
feed sick/odd modules.
The Cylindrical RPC could be powered by extra pods
located in one of the four mainframe or have a
dedicated mainframe.
Long haul cables will be available within January.
The (already installed) RPC’s connections will need
few splicing (about 20) as the H.V. boxes have
been grouped in metal containers.
Assuming that all the RPC’s could be powered at the
nominal voltage with one single cable per
container, the current capability would be 36 mA.
Marcello Piccolo
Napoli Dec., 1997
13
Installation and scheduling
 Power supplies and pods:
 purchased 75%; at SLAC 50%
 balance within first quarter 1998
 H.V. boxes:
 95% built; 30% at SLAC
 5% special boxes for cyl. RPC to be built.
 H.V. containers:
 Designed in Genoa (design approved)
 24 to build: 12 in production at Frascati, 12 in
Genoa and/or outside company.
 Ready to install by February
 Cables
 Coax barrel : cut and ready
 Coax end-caps: length not defined yet; material on
hand at SLAC stores.
 Installation during January
 Monitoring hardware
 Cables and connectors purchased
 Installation January / February
Marcello Piccolo
Napoli Dec., 1997
14
Conclusions
 The components of the H.V. system of the IFR
have been thoroughly tested in the quality control
runs done during the RPC’s production year.
 The system seems to be quite simple and
reasonably robust
 The bulk of the custom made hardware turned out
to perform satisfactorily.
 Installation of the various barrel components
should be completed by the end of February.
 End-caps stuff is a bit more uncertain, as
especially in the back, mechanical placement is
not that easy. We expect, however, that the entire
H.V. system for the IFR to be operational by the
end of March.
Marcello Piccolo
Napoli Dec., 1997
15