University of Florida Infrared Astrophysics Group - PowerPoint PPT Presentation

1 / 46
About This Presentation
Title:

University of Florida Infrared Astrophysics Group

Description:

The Flamingos-1 Science detector sees a form of cross talk that has not been ... It was found that the Flamingos-1 detector has a well depth of about 250e3 ... – PowerPoint PPT presentation

Number of Views:41
Avg rating:3.0/5.0
Slides: 47
Provided by: Robert1090
Category:

less

Transcript and Presenter's Notes

Title: University of Florida Infrared Astrophysics Group


1
Flamingos-2Electronics Design Overview
2
Electronics Design Presentation Summary
  • Section 1 Detector Status
  • Section 2
  • Electronics Requirements
  • Section 3
  • System Overview Design Implementation

3
Section 1Detector Availability and Delivery
  • Detectors available presently at UF
  • Flamingos 1 Bare Mux
  • Flamingos 1 Engineering Detector
  • Science Detector Delivery
  • Procured and delivered by Gemini
  • Needed by Feb of 2004

4
Detector Anomalies Issues
  • Cross Talk
  • The Flamingos-1 Science detector sees a form of
    cross talk that has not been seen by others.
    This takes the form of a residual image to the
    left and right of the channel with the real
    image. It needs to be noted that this was NOT
    seen in the engineering array and by all other
    testing indications, it is NOT in the external
    analog electronics.

5
Detector Anomalies Issues
  • Well DepthIt was found that the Flamingos-1
    detector has a well depth of about 250e3
    electrons not the 100e3 as noted in the spec
    sheets. This is not to far fetched as the
    integrating capacitance has a posted variability
    of about a factor of 2.This does effect the
    noise and preamp requirements. For F-1 we mapped
    275,000e- to 55,000 counts. That gives a
    conversion gain of about 5 e-/ADU.If we were to
    map the linear ( lt1.0 ) range of about 100,000
    that yields morel like 1.8e- / ADU. This is much
    more compatible with the noise expected from the
    system of about 2.0 ADU.It is probable that a
    scientifically driven choice will have to be made
    regarding what is determined to be the useful
    well depth. ( Linear vs. Saturated )

6
Section 2 Electronics Design Requirements
  • ( Summary )
  • Read out HAWAII-2 Array at maximum pixel rate.(
    Flamingos-1 data shows this to be about 1.3 Sec /
    frame )
  • Clock digitize the array in such a manner so as
    not to significantly degrade the intrinsic
    detector read noise.
  • Transport the co-added and buffered data to the
    ICS

7
Data System Noise Contribution
8
Required Array Readout Rates
  • The array is limited to about 1.3 seconds / Frame
    in the 32 output mode while using an external
    3216 analog mux.\
  • This gives a total pixel rate of about 3.2 Mpix /
    sec
  • The electronics system is capable of between 16
    and 48 Mpix / sec depending on the ADC converters
    used. The FLAMINGOS-2 system will use 1.0 Mhz
    converters and thus yield about 16.0 Mpix / sec.
    Clearly exceeding the array requirements.

9
Section 3Electronics System Overview
10
Main Component Break Down
  • Array Sequencer Processor ( ASP) Rack Components
  • Instrument Control System ( ICS ) Rack
    Components
  • Temperature / Vacuum Monitor Control
  • Motor Control

11
Overall Electronics Block Diagram
EPICS CannelAccess LAN
DHS LAN
12
Dewar Mounted Electronic Modules
  • Bias Generator / Clock Driver
  • Provides array DC voltages
  • Sets clock rail voltages.
  • Converts TTL inputs to clock level output
  • Fanout Board
  • No active electronics, only passive filtering
  • Preamps
  • 32 channels of low noise amplification.
  • Mux down to 16 channels for the electronic system
    inputs.
  • Digital to Analog Converter ( DAC ) controlled
    offsets.

13
Bias Clock Module Features
  • Programmable Clock Bias Voltage Levels ( 8 bit
    ).
  • Programmed via ASP ( Array Sequencer Preprocessor
    ) CPU commands
  • Voltage levels and Clock Phases are TTL Inputs.
  • All Inputs are Opto-Isolated to eliminate ground
    loops.
  • Module has dedicated power supply also to
    eliminate ground loop paths.
  • Will use replication of existing Flamingos-1 Bias
    board.

14
Dewar Modules ( Bias / Clock Diagram )
15
In Dewar Electronics Diagram
16
Preamp Description
  • Two Stage Preamp with 3216 Mux
  • Gain Bandwidth Product (GBW) available in current
    configuration 170 Mhz.
  • Pixel Rates expected ( ( (2048)2 / 1.3 sec/16
    channels ) gt 202 kHz )
  • Band Width ( BW ) needed at pixel rates rates
    expected for 16 bit settling 460 kHz.
  • Hawaii-2 Preamp Gain required 5
  • Gain Bandwidth needed is only 2.3 Mhz
  • Programmable offsets are provided. - Probably
    set once and forget.

17
Dewar Modules ( Preamp )
Analog Outputs from Array
Analog Outputs to ASP Rack
Second Stage Preamp
First Stage Preamp
16
16
21 Mux
TTL Offset Control from ASP
Preamp Power
Offset DAC
Opto - Isolators
16
Analog Ground Plane
18
Overall Electronics Block Diagram
EPICS CannelAccess LAN
DHS LAN
19
ASP Rack Contents
  • Array Signal Processor Main MCE-4 Contents
  • VME CPU ( MC 68020 based )
  • Pattern Generator Board ( PGB )
  • ADC / Coadder Board(s) (ADC )
  • Optical Fiber and Aux. I/O ( UFB )
  • Preamp / Bias Main Analog Linear Supplies

20
ASP Physical Layout
VME CPU
21
Pattern Generator Board
  • Identical to existing OSCIR, Flamingos TRECS
    units.
  • Simple RAM based clocking patterns.
  • Clock Phases available 24
  • Internal clocks 8
  • Clocking resolutions to 50 ns at highest frame
    rates.
  • Will drive arrays to 250ns / pixel.
  • Pattern Ram depth up to 512K words.
  • All logic is Xilinx FPGA based.

22
PGB Block Diagram
23
ASP Physical Layout
24
ADC / Coadder Boards
  • 1.0, 2.0 or 3.0 MHz, 16 bit Converters
  • Coadders
  • Word depth to 32 bits per pixel
  • One Coadder per channel.
  • On Board Data Buffers isolate average data rates
    from peak image data rates.
  • Analog section Opto-Isolated from Digital
    portions to eliminate ground loops through the
    digital system.
  • ADC Boards Individually Analog Powered to
    eliminate ground loops through the analog power
    supply commons.

25
ADC Board Block Diagram
26
ASP Physical Layout
27
Upper Fiber Board
  • Data input path is 128 bits wide
  • Relatively slow 10 MHz synchronous data clock
  • Serves as mother board for Gatir Remote Camera
    Interface (GRCI) Fiber Module
  • Three opto isolated RS-232 ports for aux..
    Control functions.
  • Not used with Flamingos - 2
  • Opto-isolated TTL I/O port ( 8 bit in / 8 bit out
    )
  • Not used with Flamingos - 2

28
Upper Fiber Board Block Diagram
Input Data Registers
Science Data Bus 128 bit
Science Data32 bit
32
Mux. Control
Handshake Control
Science Data Command ControlFiber Link
Clock Drivers
Clocks10 Mc.20 Mc.
VME Bus
Backplane
29
GRCI Module Description Features
  • Semi-Custom module designed by EDT( Engineering
    Design Team ) .
  • Supports direct 32 bit parallel input.
  • Supports data rates as high as 25 M samples/sec
    (100MB/sec)
  • Direct technical support and warrantee supplied
    by original manufacturer at no additional cost.
  • Custom features are mechanical only. Electronic
    features are identical with EDTs standard
    product, leveraging experience and increasing
    reliability .

30
Grounding Issues Topics
  • Entire structure will be safety Earth grounded
    for safety.
  • The analog data system single point ground will
    be at the preamp ground plane in that module.
    All other portions of the data system are
    referenced through this point.
  • All analog modules are powered with separate
    power supplies.
  • Bias Pod, Preamp Pod each ADC Board
  • Clock Bias TTL input lines are opto-isolated at
    the input of the Clock / BIAS module.
  • The analog outputs of the preamps are isolated
    from the remainder of the data system by
    opto-isolators after the ADCs.
  • Each ADC board is also powered with a dedicated
    supply.

31
Grounding Diagram
32
Science Data Path
Fiber Aux.. TTL I/O Board
ICS Rack
Fiber Link
Rackmount Spark-2
Temperature Monitor Vacuum Monitor
DHS LAN Connection
Motor Drivers Supply
PPC board may replace main CPU in slot 1
33
Temperature Monitor Control
  • The cold bench will be thermally controlled at
    two points for thermal stability and monitored at
    several other points as needed.
  • A Lakeshore 332 will be used for control as it
    has two control channels. One _at_50 Watts, one _at_10
    Watts.
  • A Lakeshore 218 will be used for all other
    temperature monitoring.

34
Temperature Monitor Points
35
Pressure Monitor(s)
  • A Pfeiffer TPG262 controller and two Pfeiffer
    PKR251 sensors will be used with
    Flamingos-2.These will monitor pressure in both
    the camera as well as the MOS dewars.
  • The 262 allows for up to two sensor inputs and
    is RS-232 interfaced.

36
Motion Control
  • Flamingos-2 will use the same IMS-483 indexers
    that were used with TRECS.
  • The TRECS motor indexer chassis has all the
    indexers on a daisy chain. While this is
    functional, it was a problem at times in
    diagnosing any problems that may arise on the
    communication channel(s).
  • Therefore, on FLAMINGOS-2, the indexers will NOT
    be daisy chained via RS-485. Instead, each
    indexer will be individually controlled from the
    Pearl-e Terminal Server via RS-232.

37
Bar Code Reader
  • The MOS plates for FLAMINGOS will be bar coded.
  • At this time a reader has not been selected.
  • Issues
  • Mounting of a reader within the confines of the
    MOS dewar.
  • Possible Cryogenic operation or survival.
  • Bar Tag adhesion at cryogenic temperatures.
  • Tagging of slot positions as well as MOS plates.
  • Tag Generation

38
Grounding Issues Topics
  • Entire structure will be safety Earth grounded
    for safety.
  • The analog data system single point ground will
    be at the preamp ground plane in that module.
    All other portions of the data system are
    referenced through this point.
  • All analog modules are powered with separate
    power supplies.
  • Bias Pod, Preamp Pod each ADC Board
  • Clock Bias TTL input lines are opto-isolated at
    the input of the Clock / BIAS module.
  • The analog outputs of the preamps are isolated
    from the remainder of the data system by
    opto-isolators after the ADCs.
  • Each ADC board is also powered with a dedicated
    supply.

39
Top Level System Cable DiagramVISIO -
Flamingos-2 Preliminary
40
Individual Cable Diagram Example
41
Areas of Risk
  • 1) System Read Noise
  • 2) Excess Low Frequency Noise
  • 3) Bar Code Reader issues
  • 4) Timely Construction
  • 5) Parts Procurement Spares

42
Areas of Risk
  • 1) System Read Noise
  • F-1 had read noises in excess of those found in
    other instruments.
  • This was due primarily to the scaling of the
    system to 275,000 e- instead of 100,000 as most
    others are.
  • Resolved by Rescaling to the 100K linear portion
    of the integration curve.
  • Risk LOW

43
Areas of Risk
  • 3) Bar Code Reader
  • These have been enumerated in a previous slide.
  • Resolution
  • See what others have done.
  • Buy one and play with it. ( cheap 600 )
  • Do it early!
  • Risk if addressed soon Low if we
    wait Medium

44
Areas of Risk
  • 2) Excess Low Frequency Noise
  • F-1 had 50-60 Hz pickup wich varied from site to
    site.
  • This was comprised oc two components. A) long
    single ended cables from the preamp to the ASP
    rack and b) long single ended cables from the
    array to the preamp inside the dewar.
  • Resolution a) F-2 will be using differential
    signals from the preamp to the ADCs in the ASP
    rack. This was shown to lower the 50-60 hz noise
    in TRECS both in the lab and on site. It should
    be noted that the TRECS noise levels showed no
    change when installed on the telescope.
  • Resolution b) F-2 will have a MUCH shorter path
    from the array to the preamps. In F-1 the length
    was about 1 to 1.5 meters while the length for
    F-2 will be about .3 meters.
  • Risk Low

45
Areas of Risk
  • 4) Timely Construction Delivery July 2004 is
    the schedule for most of the electronics SW IT
    to begin. There are two main issues here a)
    MCE-4 construction b) System cable
    construction.
  • Resolution
  • a) MCE-4 is a copy of at least three other
    systems of identical design. This should be a
    strait construction project with good
    manufacturing documentationRisk Low
  • b) If the cable definitions are ready to go in
    January 2004, then construction of all needed
    cables should not be a problem for two FTEs
    devoted to the task of construction and final
    documentation.RISK Low

46
Areas of Risk
  • 5) Parts Procurement Spares
  • Parts are becoming harder and harder to locate.
  • Currently identified problem parts include the
    Xilinx chips, and memory simm modules.
  • Resolution
  • An advanced purchase of these critical components
    has been made and have been received.
  • If spares are desired, they should be considered
    in the VERY near future.
  • Risks Program Delivery Low -gt Medium
    Spares Medium -gt High

47
Appendix Item ListingsSupport Materials supplied
on CDor at CDR
  • A CDR Board Level Schematics
  • B CD Individual Cable Diagram(s) example
  • C CD Overall Cable Map
  • D CD Various Excel Preamp Calculations
  • E CDR Recommended Spares Listing
  • F CD Lakeshore and Pfeiffer Manuals
Write a Comment
User Comments (0)
About PowerShow.com