DES Collaboration Meeting, Chicago, December 11-13, 2006 T. Shaw1 DES Collaboration Meeting Front...
-
Upload
lesley-hunt -
Category
Documents
-
view
213 -
download
0
Transcript of DES Collaboration Meeting, Chicago, December 11-13, 2006 T. Shaw1 DES Collaboration Meeting Front...
DES Collaboration Meeting, Chicago, December 11-13, 2006 T. Shaw
1
DES Collaboration Meeting
Front End Electronics StatusT. Shaw, D. Huffman, M. Kozlovsky, J. Olsen, W. Stuermer (FNAL)
T. Moore, V. Simaitis (UIUC)
M. Barceló, L. Cardiel (IFAE)
J. Castilla, J. DeVicente, G. Martinez (CIEMAT)
P. Moore, R. Schmidt (NOAO)
DES Collaboration Meeting, Chicago, December 11-13, 2006 T. Shaw
2
DES FE Project Definition
We must read out a large focal plane of CCDs
62 Science CCDs
8 Alignment CCDs
4 Guide CCDs
The Science and Alignment CCDs are read out at the same rate with the same clocking patterns
250k-pixels/s @ 10e- rms noise
The Guide CCDs will be readout at a faster rate; will not interfere with required “quiet” time of Science and Alignment CCDs
DES Collaboration Meeting, Chicago, December 11-13, 2006 T. Shaw
3
Monsoon Readout System is the chosen platform for test and production electronics
Monsoon– Developed by NOAO
– Crate based system (Eurocard format)
– cPCI backplane used (no cPCI protocol)
– Open Design• Schematics available• Firmware available• Layout files available
– Boards exist• Master Control Board• Clock and Bias Board• 8 Channel Acquisition Board
– Large software framework exists
http://www.noao.edu/ets/monsoon/
DES Collaboration Meeting, Chicago, December 11-13, 2006 T. Shaw
4
Monsoon Crate
MCBCBB8-ChCCDAcq.
DES Collaboration Meeting, Chicago, December 11-13, 2006 T. Shaw
5
History of Monsoon at FNAL
Mid 2004DES group looks into the Leach and Monsoon Systems as possible readout
solutions. Leach is an established proprietary readout system for CCDs. Monsoon offered some advantages – higher density (8 channel card vs 2 channel card of Leach); telemetry readout (readback of clock rail voltages and bias voltages); designed for large mosaics; open architecture; easy to customize. (Lots of manpower already expended by NOAO on Monsoon ~12 engineering years/~5 programmer years/~5 board layout years.)
University of Illinois Urbana-Champaign began to purchase the initial Monsoon test stands.
March 2005FE workshop at FNAL; decision to continue to pursue Monsoon as readout platform.
April 2005 Receive MCB, CBB and a blank prototype 8 ch CCD Acq Board to be hand
assembled.Never able to achieve low noise.
Nov 2005Delivery of 2 complete Monsoon systems (MCB, CBB and production 8 ch CCD Acq
Board).Fairly quickly began running with <10e- noise at ~12us/pixel.
DES Collaboration Meeting, Chicago, December 11-13, 2006 T. Shaw
6
History of Monsoon at FNAL
Feb 2006Delivery of two additional complete Monsoon Systems. One was shipped to
Barcelona
Mar 2006With “short” cable run, we are able to achieve ~6.5e- noise at ~6us/pixel
April 2006 With a pre-amp located outside of dewar wall, we are able to achieve ~6.5e- noise at
~6us/pixel with a long cable
July 2006Delivery of fourth Monsoon system for FNAL
August 2006FNAL Director’s review recommends use of JFET source follower close to CCD
video output
October 2006Begin instrumenting the Multi-CCD test Dewar (MCD)
December 2006Simultaneous readout of two CCDs in MCD.
DES Collaboration Meeting, Chicago, December 11-13, 2006 T. Shaw
7
Proposed Changes to Monsoon
• To use the Monsoon System for Production DES Electronics, we will customize the boards for our specific application:– Build a 12 channel CCD Acq Module which would provide
customized bias voltages, RTD readout, and Vsubstrate ramp control.
– Build a customized Clock Board to better address our clocking needs; greater fanout ability to drive more clocks.
– Slight modification to the Master Control Board to allow for switch to S-Link as the fiber optic link; also, look into multi-crate synchronization
DES Collaboration Meeting, Chicago, December 11-13, 2006 T. Shaw
8
12-Channel BoardWhy do we need to change the Monsoon Acquisition Board?
• System density– Limited space in proximity to the telescope. Increasing the
Video Channel count to 12 to better match the system.
• Improved Control of CCD Biasing.– Better trimming of HV Biases– Control of Substrate voltage rise/fall time (Ramp).
• Adding Temperature read back for RTD sensors.
DES Collaboration Meeting, Chicago, December 11-13, 2006 T. Shaw
9
CCD Acquisition Card Description
• 12 Video Signal Channels AC Coupled 18-bit ADC (same as Monsoon!) – Dynamic range and Sensitivity 4V pk/pk 15.6uV/ADC– Dual Slope Correlated Double Sampling (CDS) with DC Restoration– Offset controlled to maximize dynamic range– Maximum 250 kpixel/sec– Noise <3 ADU rms
• 48 High Voltage Biasing Signals (Telemetry)– 12 bit DAC adjustable between fixed range.– 12 bit ADC read back on each Bias channel– Outputs can be Enabled/Disabled via software control
• Micro-Sequencer Pattern Memory– Drives the Analog Front End circuits when triggered across the Back Plane.
• On-Board Temperature Sensor (10 bits)• 6 Channels for RTD temperature sensors on CCD
– 12 bit ADC with Sensor Current Source• Single FPGA for control: JTAG programmable• Board Identification
– Unique Digital Serial Number– Firmware Version– Board Type
DES Collaboration Meeting, Chicago, December 11-13, 2006 T. Shaw
10
CCD 12 Channel Acq. Board
Temp.Sensor
DigitalSerial
Number
ResetInterface
6 ChanRTD
12 bit ADCTelemetryInterface
X48
12 bit DACInterface
X48
18 bit ADCInterface
X12
CDSInterface
X12
Back Plane
RAMMemory
-Sequencer
FPGA
900 pinBGA
JTAG Prog.
Transition Card (12 Video Channel Inputs)
DES Collaboration Meeting, Chicago, December 11-13, 2006 T. Shaw
11
Top View
DES Collaboration Meeting, Chicago, December 11-13, 2006 T. Shaw
12
Bottom View
DES Collaboration Meeting, Chicago, December 11-13, 2006 T. Shaw
13
12-Channel Transition Module
DES Collaboration Meeting, Chicago, December 11-13, 2006 T. Shaw
14
DES New Clock Board
• New Clock Board is being designed by engineers at CIEMAT (Madrid) which will drive many more clock lines.
• The current NOAO board provides 32 clocks, the new board will provide 135, or enough for 9 CCDs.
DES Collaboration Meeting, Chicago, December 11-13, 2006 T. Shaw
15
9 CCDs
CB-36 clk
Backplane
MAIN board
Transition
CRATE
135 clk
VIDEO Board
VIDEO Board
VIDEO Board
DES Clock Board Concept
CCD 3X15 CLKCCD CCD
CCD CCD CCD
CCD CCD CCD
3X15 CLK
3X15 CLK
9 CCDs
CCD 3X15 CLKCCD CCD
CCD CCD CCD
CCD CCD CCD
3X15 CLK
3X15 CLK
CB-36 clkTransition
135 clk
DES Collaboration Meeting, Chicago, December 11-13, 2006 T. Shaw
16
FPGA
36 CLKVoltage swing generation
135 Cable Drivers
Hardware redesign
MONSOON CBB DES CB
sup
ply
Te
lem
etr
y
32 CLK
BIAS CPLD
CPLD
FPGA
• BIAS section removed• 4 clock generation channels added • 135 cable drivers included on board• 2 CPLD removed
su
pp
lyT
ele
me
try
DES Collaboration Meeting, Chicago, December 11-13, 2006 T. Shaw
17
Firmware redesign
• BIAS control removed• Code integration: from 2 CPLD+FPGA to 1 FPGA• New microsequencer development
microsequencer
SEQ BUS (write)
TELEMETRY
MUXFRONTPANEL
DAC/ADCCONFIG
ADCs
DACs
FRONTPANEL
36 Clocks
BackPlane
Interface
PATTERNMEMORY
(DUAL PORT)PROGRAMCOUNTER
GENERATION
CLOCKTRIGGERREGISTER
PIXDATA BUS (read)
FPGA
DES Collaboration Meeting, Chicago, December 11-13, 2006 T. Shaw
18
• Status:
– Schematics and layout ready
– Firmware ready for testing
– Prototype is under production
• Future work:
– CB testbench installation in Madrid (Monsoon based)
– Prototype validation and characterization
Clock board status
DES Collaboration Meeting, Chicago, December 11-13, 2006 T. Shaw
19
Barcelona work/Master Control Board
Waiting for Laia’s slides…
should cover s-link, multi-crate synchronization, transition cards
DES Collaboration Meeting, Chicago, December 11-13, 2006 T. Shaw
20
PF Packaging
UIUC has worked with FNAL mechanical engineers and proposes to instrument the telescope using three FE crates.
These crates would each contain two backplanes; one with 6 slots and one with 4 slots.
A single 4-slot section would provide for readout of up to 9 guide CCDs
The remaining backplane sections would provide readout of up to 72 science and alignment CCDs.
UIUC is also providing engineering support for crate power supplies.
MCB
CLOCK
CCD
Acq
CLOCK
CCD
Acq
CCD
Acq
MCB
CLOCK
CCD
Acq
CCD
Acq
Provides Readout of up to
18 CCDs
Provides Readout of up to
9 CCDs
DES Collaboration Meeting, Chicago, December 11-13, 2006 T. Shaw
21
Vacuum Interface Board(VIB)
Vacuum Interface Board will provide the connection between the Monsoon boards and the CCDs on the focal plane.
Monsoon cables will plug into the VIB outside the vacuum.
Kapton cables will be used to connect between Vacuum Interface Board and the CCDs inside the dewar.
DES Collaboration Meeting, Chicago, December 11-13, 2006 T. Shaw
22
Vacuum Interface Board
Kapton Cable
AlN Board
Camera Vessel Instrumentation
DES Collaboration Meeting, Chicago, December 11-13, 2006 T. Shaw
23
Power Supply Studies
• Ongoing work by engineers at the University of Illinois on testing power supply performance.
• Goal is to prove viability of switching power supplies.• Tests were run with Linear supplies and Vicor supplies
attached to original Monsoon 6-slot DHE.• Remote sensing was removed & local sensing added.• Noise measurements are taken with video inputs connected to
ccd emulator
DES Collaboration Meeting, Chicago, December 11-13, 2006 T. Shaw
24
UofI Monsoon Test StandWiring Scheme
DES Collaboration Meeting, Chicago, December 11-13, 2006 T. Shaw
25
Linear Supply Plot Results
DES Collaboration Meeting, Chicago, December 11-13, 2006 T. Shaw
26
Vicor Supply Plot Results
DES Collaboration Meeting, Chicago, December 11-13, 2006 T. Shaw
27
Conclusions