Spectrograph Detector

1
Spectrograph Detector Hardware
2
Developments Since CoDR

• Collaborating with IfA on NIFS detector subsystem
• Rockwell PACE technology HAWAII-2 HgCdTe ordered
• Designed detector mount and cryostat wiring
• Bypassing detector internal output amplifier and
  using external output amplifier to eliminate
  amplifier glow
• Purchased components for detector temperature
  control and demonstrated mK stability
• Determined final configuration and ordered SDSU-2
  detector controller
• Using Quad Channel Coadder Video Board instead of
  two Dual Channel Video Boards for IfA
  compatibility
• Using detector reference to improve bias
  stability
• Plan to implement digital filtering to overcome
  mux. glow

3
IfA/RSAA Collaboration

• Collaborate on the design and characterization
• IfA world leaders
• Similar low dark current and noise to IfA AEOS
• NIFS design adapted from AEOS
• April 2002 - Science detector delivered to IfA
  for characterization
• End July 2002 - delivered to RSAA for integration

4
Spectrograph Detector
5
Spectrograph Detector

• PACE Technology HAWAII-2 HgCdTe ordered
• Measure QE at 65K
• Lower dark current (0.01e/s/pix) at 65K than 77K
• Mount temperature sensors on Eng. and Science
  Detector
• Need information on sensor wiring
• No mention of AR coating
• Still developmental
• Reduces mux glow, fringing and ghosting off front
  of detector
• Possible upgrade should we pursue this option

6
Detector Housing

• Stray light shield
• Electrically isolated to reduce noise pickup
• Temperature controlled 60K 1K

7
Detector Mounting Board

• Two-layer, Teflon flexi-rigid board
• Space constraint and reduce solder joints and
  connectors
• Cooling block 65K 1mK
• Inner 223 non-signal pins soldered to cooling
  copper block cool detector
• Cooling block large enough such that 2mW over 10s
  raises temperature lt 1mK

8
Detector Mounting Board Layout

• Four mil tracks with minimum four mil spacing
• Technology being trialed on our DBS CCD system

9
External Output Amplifier

• Eliminate amplifier glow
• Use J270 JFET source follower
• For acceptable drift, J270s mounted in detector
  cooling block to stabilize temperature ( 1K)
• Microphonic noise pickup from cyrocoolers
  possible in high impedance path from output pin
  to gate of J270 - Use rigid wiring
• Reference improve bias stability
• using similar circuit
• thermally couple and match reference and signal
  J270s

Reference
10
Flex Circuits

• Two Teflon flex circuits
• Clocks
• Outputs and biases
• Grounds
• Cold strapped by
• Clamp to CWS at radiation shield
• Clamp to CWS at Skirt
• Micro D connectors to Detector Housing
• Four mil tracks with minimum four mil spacing

11
Detector Temperature Control
Temperature Test Cryostat
DC IOC Controller
Lake Shore Temperature Controller
12
Detector Temperature Control

• Achieve low drift ? mK stability
• Purchased
  Lake Shore Model 340
• Cernox CX-1080-LR sensor
• Vishay RTO TO-220 packaged resistor
• Demonstrated mK stability in a test Cryostat over
  days
• Response to disturbances
• cold strap
• detector dissipation

13
Detector Wiring Discussion

• Is our thermal design sufficient to obtain mK
  stability of detector?
• Demonstrated components can achieve mK stability
• Controlling the temperature of detector assembly
  environment
• Wiring is sufficiently cold clamped and of high
  thermal impedance
• Cooling block is in as close thermal contact as
  we can achieved
• Have we done enough to eliminate microphonics as
  a problem?
• Using rigid wiring in sensitive areas
• Added additional vibrational isolation to
  Cryocooler mount

14
Spectrograph Detector Controller
15
Design Overview
16
Bias Stability

• High level of bias stability is essential
• John Barton measured HAWAII-1 signal drift
  1.5e/mK
• gt control detector temperature to mK level
• Rockwell has provided a reference output on the
  HAWAII-2 for drift compensation

17
Reference Circuit

• Enabled on 1025th horizontal clock
• Signal and reference processing circuits are made
  as similar as possible
• Same amplifier circuit with matched FETs J270s in
  close thermal contact
• Feed into same video channel by utilizing IfA
  Video Switch Board.
• Take 32 samples, average and subtract as with CCD
  overscan

Signal
18
SDSU-2 Controller Stability

• Limited by
• video ADC ADS-937 ? 10 µV/K
• bias/clock/video offset DAC DAC8420 ? 20 µV/K
• Spec. is drift lt read noise 5e ? 15-30
  µV at det. gain 3.0-6.0 µV/e
• Achieved by controller stability lt 2C over
  longest exposure
• Relax further if reference pixel works
• Plan to water-cooled the controller ? stability ?
  water temperature
• Looked at WFI on 40inch for example of
  water-cooled controller
• lt 2C over two days
• OK if Gemini water chiller lt 2K/hr

19
Sampling Techniques

• Implement three types
• Double Correlated Sample
• Fowler Sampling
• Linear Fitting
• Complicated by multiplexer glow

20
Multiplexer Glow

• Mux. glows every time it is clocked
• Affects multiple sampling - more reads, then read
  noise ? but glow ?
• As number of reads increases, total system noise
  initially decreases, bottoms and then increases
• Extent with HAWAII-2 unknown
• Reduced by
• AR coating array to stop light piping effect
  through substrate
• Skewing the clocks
• Overcome by digital filtering each pixel
• Multiple ADC samples per pixel access.
• Noise improvement ?(n)

Noise
HAWAII-1
Finger et al 2000
Digital Filtering
Total noise
Read noise
Glow
Number of Reads
21
Detector Controller Discussion

• Will water cooling the detector controller obtain
  good thermal stability?
• Need data from Gemini.
• variation of water temperature with dome
  temperature at cassegrain
• variation of NIRI jacket temperature with dome
  temperature
• likely dome temperature variation during longest
  exposure
• Have we done enough to achieve good bias
  stability?
• Designed the detector temperature control system
  to achieve mK stability
• Temperature stabilize the detector controller to
  less than 2K over longest exposure
• Use reference circuit to compensate for drift
• Will multiplexer glow be a problem?
• Affects how well linear fitting will reduce noise
• HAWAII-2 mux. designed to have less glow. Extent
  unknown
• Overcome by implementing digital filter in all
  three sampling techniques

22
Detector Characterization
23
Characterize

• Determine operating point
• good QE, sufficient well depth (50ke), and
  linearity and minimizes read noise, dark current,
  and cosmetic problems
• Bias stability
• Check adequacy of mK detector stability
• Measure detector controller stability
• Determine best way to use reference
• Multiplexer glow
• Determine best mix of digital filtering and
  sampling method to obtain best noise performance
  for different observing situations
• Persistence
• Linearity
• Fringing stability and how to flat field

24
END
25
CdZnTe MBE Technology 5µm Array

• Possible loan of 5µm MBE Array from NGST project
• MBE has advantages over PACE
• lower dark current (0.01 ? 0.001 e/s/pixel)
• eliminates persistence. (0.3-0.5 1st read, none
  thereafter)
• higher uniform QE (60 ? 85) (CdZnTe better
  refractive index match)
• NIFS designed to accommodate 5µm MBE Array
• Packaging different ? redesigned of detector
  mounting

26
(No Transcript)
27
Detector Discussion

• Information required on sensor wiring of
  temperature sensors on Eng. and Science Array
• What is the progress of AR coating and will it be
  considered as an option?

28
Spectrograph Detector Wiring
29
Thermal Shunts

• Notes

30
Changes Since CoDR

• Change of SDSU-2 Video Processor Board
• CoDR proposed two Dual Channel IR Video Board
  (Dual Channel)
• Instead one Quad Channel Coadder Video Board
  (Coadder)
• Reason - compatibility with IfA
• Added IfA Video Switch Board
• Switches between signal and reference and enables
  the same video channel to sample both reference
  and signal.
• SDSU-2 power supply mounted and air-cooled inside
  the DC thermal enclosure instead of mounted and
  water-cooled on the Cryostat

31
Choice of Video Processor Boards

• Dual Channel IR Video Processor Board
• two analog signal video processing chains
• six bias generators
• Quad Channel Coadder Video Processor Board
  (Coadder)
• four analog signal video processing chains
• no bias generators
• image-processing capability, but not useful
• clock board provides biases

32
Comparison of Video Processor Boards

• Identical video processing channels
• Coadder has image processing capability, but not
  enough memory to be useful
• Modest readout rate gt all image processing done
  in external processor
• Coadder complex gt DSP code more complicated than
  necessary
• IfA are developing AEOS, a spectrograph with
  HAWAII-2 , using the Coadder and will use the
  Coadder to characterize the NIFS detector
• More compatibility with IfA gt less changes to
  IfA DSP code
• Dual channel board has six low noise bias
  generators. The coadder will require the biases
  to be provided by the noisier clock board
• Summary

more complicated potentially noisier
Versus
IfA compatibility
33
Bias Stability

• High level of bias stability is essential because
• Long integration times 3600secs
• The way the readout is performed (multiple reads
  over full exposure time)
• Accurately subtract fixed pattern dark current
• Low read noise 5e, bias stability specified as
  read noise
• John Barton measured HAWAII-1 signal drift
  1.5e/mK
• gt control temperature to mK level
• Rockwell has provided a reference output on the
  HAWAII-2 for drift compensation

34
Detector Circuit
35
Component Drift

• Dominated by ADS-937 and DAC8420
• Drift lt read noise 5e ? 15-10 µV at detector
  gain of 3.0-6.0 µV/e

 
36
Sampling Techniques

• Implement three types
• Double Correlated Sample
• Signal End read - Start read
• Bright object and background limited
• Fowler Sampling
• Average( n/2 End reads) Average(n/2 Start
  reads)
• Noise improvement ?(n/2)
• Medium length exposures
• Linear Fitting
• Reads at equal periods throughout exposure
• Linearize and least-square fit to obtain photon
  rate
• Noise Improvement ?(n/12)
• Long exposure length

37
DSP code

• Three Motorola 56000 DSPs
• VME interface board
• Timing board
• Coadder video board
• Two code
• VME interface board DSP code
• Use GMOS code as starting point
• Add interrupts to sequence data transfer
• Timing/coadder video board DSP code
• Preliminary command interface document written
• Copy IfA code after Video Switch Board tested
• modify where necessary to suit and to add
  features such as
• Linear fitting
• Digital filtering

38
Grounding and Shielding
39
Grounding and Shielding

• Adhered to good grounding and shielding practices
• Grounding system designed
• Minimized earth loops
• star point
• make sure cryostat earthed to telescope only. no
  earth through components controller, cryocooler
  lines, cables etc
• Separate noisy grounds from low level signal
  grounds
• All cabling use shields
• Keep noisy circuitry and cables away from low
  level signals.
• Only earth shields at one end, preferably at
  source
• Orient cryocooler so that its radiated large
  magnetic field has least affect on the detector
  controller.
• Electrically isolate shields from each other
  especially noisy shields from low level shields
• eg. clocks from biases from output signal

40
Detector Ground

• Star point ground no ground loops
• Used isolated internal shield around detector
• Cryostat grounded once to telescope
• All cables are shielded and shields connected at
  one end only
• Noisy signals/cables are separated from low level
  ones
• Orient cryocooler so that its radiated large
  magnetic field has least affect on the detector
  controller

41
Component Controller Ground
42
Detector Test Facilities
43
Test Facilities

• Test Software
• CICADA RSAA Instrument Control and Data
  Acquisition software
• support SDSU-2 for CCDs
• modify to support IR detectors
• Test Cryostat
• IfA
• Do initial characterization and optimization
  using KSPEC, cross-dispersed spectrograph
• RSAA
• Modify existing 8 IR labs cryostat (CWS at 55K)
• No optics, but neutral density filter for looking
  out
• Manually operated cold shutter for darks
• IR LEDs inside cryostat for flat fields -
  controlled from outside cryostat
• Accommodate final Detector Housing and assembly
  and wiring
• Data Analysis Software
• IRAF

44
Test Cryostat Measurement

• What measurements do we want to do
• read noise, dark current - completely blank off
  with cold stop
• photon conversion gain, full well capacity and
  linearity - illuminate evenly with stable source
  intensity (e.g.IR LEDs inside cryostat)
• quantum efficiency - illuminate with repeatable
  stable source of known intensity (e.g. black
  body) - absolute vs relative
• cross-talk and persistence - test image generator
  such as hot spot and lines
• fringing???? - should we wait until in final
  cryostat

45
Test Cryostat

• IfA
• Initial detector characterization at IfA
• KSPEC, cross-dispersed spectrograph
• RSAA
• Modify existing Irlabs 8 double can cryostat
  (CWS at 55K)
• Place neutral density filter over detector so
  that detector does not saturate when it looks out
• Manually operated cold shutter directly over the
  detector for doing darks
• IR LEDs inside cryostat for flat fields -
  controlled outside cryostat
• Accommodate final Detector Housing, assembly, and
  wiring
• Test image generator
• Hot spot to one quadrant for measuring cross-talk
  and persistence
• Lines for testing spatial characteristics

46
Operating Point

• Vary temperature and detector voltage, measure
  and plot
• Read noise (e)
• Photon conversion gain (e/ADU)
• Dark current (e/s/pixel)
• Full well capacity (e)
• Cosmetics
• Quantum efficiency
• Select suitable operating point with
• good QE
• sufficient well depth, (50ke) and linearity
• minimizes read noise, dark current, and cosmetic
  problems

47
Detector Characterization Discussion

• Can we learn to live with persistence?
• Will fringing be stable and be able to be flat
  fielded?
• Will we achieve low enough read noise, dark
  current at a sufficient well depth and quantum
  efficiency to do good science?

48
Persistence

• Decay of residual images in 10 seconds darks
• Point source of 5 million photons/sec
• After 60 sec, residual 6000e
• After many hours, still much greater than dark
  current signal.

G. Finger ESO
49
HAWAII-1 Fringing

• Constructive and destructive interference effects
  in sapphire substrate - stable with time
• peak-to-valley variation 10