Attending: Brian, Ranajoy, Shaul, Qi, Karl, Francois, Kris, Jeff, Richard, Jacques, Hannes, Tom

Agenda:

• focal plane model and noise (Young)
• Readout

Delayed:

• Study of deployable shields (Qi)

Notes:

Roger's update on Readout:

I've written some code that can accept a somewhat arbitrary number of detector channels. But I need input (from you guys?) on a final detector/color count. This code adds up the wires needed for TDM.

It presumes 64 channel MUX, which is the state of the art recently implemented in ACT. For 4K-UC wiring, there are a pair of wires for each of the 64 row-select switches. Each column has a common TES bias pair, a common SQUID bias pair, and a common SQUID feedback pair. 300-4K wiring includes these wires plus SSA bias pairs.

The code does not allow readout columns to cover more than one tile (which would complications that we should avoid) , but otherwise attempts to make efficient use of the wiring.

I'm making no assumptions about size limits of the warm electronics (MCE) since UBC has shown a willingness to expand their crates if given the proper incentive.

focal plane model and noise (Karl)

• Focal plane version 0, hex packing pixels. close to 15000 px fit.
  • Hannes: hexagons preferred, 130 mm max
  • Using lenslets. Is this reasonable?
  • Hannes/Tom: horns can be made smaller, but can't do 3:1 bandwidth. Current F pixels is just outside what can be done.
    • Could reorder bands in pixel to avoid this.
• Mirror emissivity. 1% was being used. Planck data pre-flight and during suggests < 0.1%
  • Richard: Planck never saw sign of dust on mirrors increasing emissivity. Emissivity of pure aluminum drops with T, levels out around 30 K. But alloys level out sooner. Consider this.
• Change in temperature of components changes noise.
  • 6.5 K stop vs 4 K. NET change depends on band. bands with large spillover suffer the most.
    • Karl calculate CMB map depth for both cases.
  • 45 K primary. NETs worsen at high frequency, more load from mirrors.
  • 50, 100, 150 mK focal plane. Changes NET at lowest frequencies (phonon noise goes as sqrt(T1/T2). Small effect.
  • Plan is using Planck mirror temperatures and emissivities.
• Rough check of alignment sensitivity using codeV.
  • need to align to 0.2 mm level for 900 GHz.
  • Richard: this is less strict requirement than Planck, as expected for this type of system. Should test mirror motion in its plane and normal to its surface. Volunteers to do more thorough analysis over next few weeks, if can find the time.
  • Kris: Paper on Planck mirrors emerging now, Tauber et. al. expect in few weeks.

Readout

• Roger calculating TDM mux case, x64 (see above) (Jeff/Hannes: no issue up to 100 mux.)
  • Jeff: 64x is being deployed now, 100 is doable in near future
  • Hannes: no TDM. always limited multiplexing. microwave mux for example. Plans to demo 512x in 1 GHz band by end of 2017.
  • Jeff: current microwave mux is high warm power. unclear what the power reduction would be in satellite case. TDM is easy to calculate currently, so easy to cost. (TDM at 15k bolos is 600 W with current unoptimized crates)
  • Shaul: Can we calculate end-to-end microwave mux system today? Hannes: Will calculate.
• Need a leader for FDM. (Franky at UMN can do some of this)