SLAC-2017:opticalcouplingandreadoutsystem

From CMB-S4 wiki
Revision as of 19:03, 28 February 2017 by Jeff (Talk | contribs)

(diff) ← Older revision | Latest revision (diff) | Newer revision → (diff)
Jump to: navigation, search
  • Detectors
    • Instrument Requirements (derived from Science Requirements)
    • specify sensitivity per band so we can figure out what to build
      • band definitions
        • band calibration
        • pair difference relative calibration
        • we should provide guidance on the types of bands to design
        • interacts with calibration plan
      • 1/f requirements (long list)
        • trades against modulators and scan strategy as the requirement is low ell noise
        • shift band: trade studies, split bands vs shift bands
          • how were the NETs put in
    • high priority
    • coupling to optics
    • interface with calibration
      • BICEP group, temperature control the focal plane--> nK stability achievable with NTD + temperature control
      • time constant interacts with scan speed


    • Interfaces to other Technology Choices
      • pixels size
        • interacts with optics and sensitivity requirements
      • loading from telescope and atmosphere
        • interacts with sensitivity requirements
      • bath temperature
        • interacts with cryogenics
    • Paramaters to optimize
      • sensitivity
    • things to develop for S4
      • modular array package, interface to readout
      • testing capabilities
      • specify the key parameters (TC, pass-bands, need to list what needs to be specified)


  • Readout parameters
    • Sensor specification
    • Noise budget (baseline, 1/f)
    • Cross talk
    • Thermal budget (cold components)
    • Power requirements (warm components)
    • range of optical loadings required (30-300 GHz), readout systems can differ due to bias requrmenets
    • ability to test under lab loading (dual TES)
    • DETECor Uniformity in fab
    • polarization angles requirements
    • power dissipation for the readout (thermal budget above)
    • cost needs to be considered
    • risk
    • development priorities
      • readout and testing need to move forward


  • NOTES
    • Simons Observatory Plan
      • universal focal plan module
      • standardize so different arrays of different designs are interchangeable
      • same detector parameters and readout interface
    • SPT3G
      • 1-2 weeks for 5 arrays
      • Testing is part of the S4 development
      • S4 scale fabrication requires developing testing for warm screening esp of film properties
        • this wiggles bands and TC around, (titanium(
      • tested at least 30 out of 60 wafers
      • made two wafers for each deployed wafer
      • build a program to test S4 for scale
        • develop a testing program, options: test all, use S3 cryostats, test in the cryostat
        • batch testing: hybridization could be key to making this go fast
        • also need to measure scatter in parameters
    • NIST
      • two color pixels
      • demonstrated 90/150 , 150/220 and soon 30/40
      • also doing LEKIDs
      • SQUIDS mature, mSQUID maturing, on MUSTANG right now
      • optics: horns and GRINS (gradient Index lenses)
      • testing and assembly is a big deal, need to merge components to simplify the package
        • backside integration could simplify it (baseline for BICEP3 with mSQUIDS)
        • integrated fab could help (KIDS or integrated uMUX)
      • Lessions from recent fab
        • repeatability and uniforming and yield are becoming really good and stable
        • ACT delivered 4 wafers and we could have fielded 4th
        • SPIDER wafers are great
        • tons of process development Gene/Dale/ Shannon did tons of development work to get this stability
        • shared resource risk
        • literally the first four complete wafers fro AdvACT worked, so that is the total made
    • Berkeley/ LBNL
      • fabrication with a commercial foundry (STARcryo didn't work, Hypress did )
      • checked TES bloomer (worked-- define critical functions here)
      • make antenna coupled TES bolometer (worked)
      • did an array (worked)
        • IV curves, and bands work, but the thing to do is to get the right detector parameters for CMB observations. This is the next step.
      • deposit AlMn TES material at Berkeley
        • hypes could do the AlMn, but there is lots of development to transfer that step, so put off. they also bak the XeF2
        • can hypris handle the volume of S4, they have the capacity to run 24/7 and batch processing
        • cost per wafer: tell off line, less than doing it at a university
    • Brad Johnson LEKIDS for ABS
      • dual pol LEkids for ABS---> super simple KIDS, made at Star Cryo
        • JPL (peter day) made them, single aluminum film
      • small cross-pol response
      • 1/f knee comes up at ~10 Hz, bug gently
        • ABS HWP fixes the above two problems
      • MUX factor of 512
      • ABS in NA
    • dual pol MKIDS
      • large CPW MKID in the corners of the pixels
      • microsctip to slotline to CPW was the new development, otherwise based on the ACTPOl design
      • Kent is making them here
      • FIRST array made at Stanford, and tested
      • two neighboring resonators, measured
      • alMn LEKIDS have high Q, magnetic field susceptibility is understood, this is an interesting material for lower frequencies---> eg could go to 30 GHz, with a very low bath temperature
        • lower TC + higher frequency===> multiple cooper pairs per photon, boosts S/N by suppressing recombination noise
      • Dual pol LEKIDS: testing an array with 542 MUX factor, (peter Day at JPL and OPaul Conner at BNL)
      • 1/f noise performance: readout + TLS, 100 MHz has low 1/f: 1/10 of a Hz knees, probably limited by load.
    • PIXEL design
      • relentlessly shrink the footprint of the detectors
      • QUADRILLE could go farther
  • readout
    • uMUX
      • modulate a flux ramp, daisy chaining SQUIDS Together, so you use this to linearize the SQUIDS when all of them in different places, read out as the phase
      • fielded a 64x system in MUSTANG-2
      • 100 KHz ramp, move to 2 MHz
        • using 2 33x uMUX chips
      • achieving GBT background limit
      • showed a first light image
      • sub-Hz 1/f noise has been demonstrated
      • stability of the flux ramp source matters so works, but needs to be optimized
      • includes a LC filter, so 1 bias for an array can work
      • large bandwidth
    • HF DFMUX
      • on wafer DFMUX
      • readout element is just LC resonator
      • LC resonators need to be smaller, they think 100 MHz)
      • fabricating high quality resonators
      • Concern: need to verify good noise performance of TES up at these frequencies
      • capacitors are interdigitated, but parallel plate could make it smaller
      • made a prototype detector readout integrated chip
      • resistance high on the TES to deal with contact resistance
      • relative focal plane usage of the Ls and Cs, so the same
      • reduce R make L smaller and C bigger, cartoon shown for 30 MHz
      • only one more layer, but could be zero with adjustment---> difficult step is making the parallel plate capacitor low loss
    • SLAC RF electronics
      • RF readout electronics
      • ACTA crate
      • Carrier care + AMC (analogy multiplier chain)
      • developing tone tracking firmware, to reduce power, linearize the amplifier, and keeps SQUIDS well biased
      • use DDS and demodulation
      • Carrier card has FPGA
      • ADC DAQ card with a daughter card with RF module
      • reduces the power to the HEMPT by 10-100x lower.
      • 4000 channels per card,
      • 8x cards per crate
      • 8 0-500 MHz channels per card
      • firmare (full time work for three people), professionalize the software
  • KAM readout for SO
    • universal focal plane module
    • DFMUX + uMUX with 650, competitive for cryogenic loading, assuming hard coax-, not the fancy striplings, so could be reduced for uMUX
    • on wafer is a possibility
    • universal focal plan module is a concept interface specification to be specified
    • working to define the schedule and criteria