Argonne-2018:TechDevWorkingGroup

From CMB-S4 wiki
Jump to: navigation, search

Session notes for Technology Development Working Group Update and the Path Forward

Ahmed: TD Input to CDT

Want some pre-project spending to reduce risk, maybe improve cost, schedule, and science return.

At end of last workshop we identified a need for an R&D plan. Decided to start a community process. CDT also needed an R&D plan, but on a short timescale. Ahmed & Benson provided a plan to the CDT, based on a high-level risk registry. Used CDT concept, risk register similar to LSST. ~40 items in the register. Top risks are focal plane performance and detector array production. Developed a generic plan for burning down risk. CDT report included the broad conclusions.

Q. Bischoff. Is the risk burn down plan available? A. Yes on the CDT wiki.

Suzuki: TD Community Overview and Organization

Technology book after Harvard meeting. Homework to evaluate which tech developments are important for CMB-S4. Charge to community groups was to evaluate tech dev based on impact on cost, schedule and science return. Topics split into 4 groups. Common evaluation metric. Reference design and schedule from CDT, or assumptions by TD group. Subgroups identified the most important developments: production rate, systematics, timely refinement of sims. TD evaluation table will be used to evaluate options for the reference design.

Q. Kovac. What do the developments mean? Systematic isn't really a development. A. These are really technical issues, rather than developments.

Summary from TD Sub-Groups

Borill: Data Management

Identified key DM challenges. Prioritized based on importance to current project phase and likelihood that the required technology will not have been developed by others in time. Current DM risks are concentrated in statistics and systematics. During project definition, refining details of the experiment and extending the breadth and depth of science analyses are the highest priority. Most of the work for the CDT was in map domain, but some time domain work will be needed to support instrument definition. Need to respond to HPC developments. Highest priorities: observation modeling, time domain synthesis, framework scale & quality.

Q. Kovac. What is environmental modeling. A. Atmosphere, ground. Q. Kovac. What is unique about the highest priorities? A. Data volume, multi survey, multi site. Q. Zeesh. Any timing associated with the highest priorities? A. Not yet.

Cecil: Detectors and Readout

Review of CDT detector and readout specs. Developed a straw person evaluation tool in a Google sheet. Max number of pixels/wafer from AdvACT. Wafers/camera from BICEP. MUX factors from fMUX. Identified current status of technology, and where it needs to be for CMB-S4. Ranking based on weighted average for cost, schedule, statistics, systematics; also highest in each

Top rankings:

Detectors:TES fab at CMB-S4 scale; MKID spectral sensitivity; feed horn designs; AR coating of lensless; production rate for optical coupling

Readout: uMUX design optimization; FDM noise, parasitics, SQUID optimization; FDM resonator size, bandwidth, mux factor

Integration & testing: modular design; get characterization; MKID on-sky demo

Q. Ruhl. Looks like some lower freq bands might need trichroic pixels. A. 20GHz detector numbers don't fit nicely into dichroic pixels. Not proposing to combine, e.g., a 40GHz band with a pair of mm bands.

Salatino: Cryostat and Cold Optics

Starting point is CDT stew man. 100mK dilution fridges. freq range up to 270GHz. 0.5m and 6m telescopes. Polarization modulators TBD. 20 optics tubes per large tel. 1 tube per small tel. 75 tubes total for CMB-S4. Many lenses, filters etc. ~250kW for 20 pulse tubes.

TD topics: cryogenics; cryostat; optics. All divided into sub topics.

Top priorities: sidelobe systematics; characterization of cold optical properties of materials.

Next priorities: production rates of AR coatings, metal-mesh filters, lenses; HWP rotation cold; AR coating performance; magnetic shielding.

A few topics still missing. Need to identify cost impacts.

Comment from Zeesh. Sideline systematics will show up in many areas, not just cryostats.

Comment from Devlin. If we double capacity at Cardiff, it takes 3ys to make the filters we need. Cost is $5k/filter.

Other comment. Large aperture cryostat may need 5-7 pulse tubes.

Benson: Telescope and Site

CDT identified ground pickup as highest risk.

No significant TD required for telescope and site, but we should do optical and ground pickup sims and measurements.

TD categories:

Ground pickup: shielding (model this); boresight rotation (estimate cost and schedule impact, need to decide if it is in the reference design)

Telescope: mirror gaps; support structure and thermal stability

Site: power; data transfer; better estimates of observing efficiency

Q. Charles. Atmospheric noise vs. freq differences between sites is not well known. Observations over next few years will answer this.

Comment from Clem. Modeling pickup is a good goal, but it has not been successful to date. A. Can do some physical optics calculations. Maybe piecemeal modeling to support other estimates. Difficult to do a calculation that will give an accurate estimate of pickup.

Comment from Natalie. Regarding site, would be helpful to understand site performance earlier rather than later because it impacts ops cost.

Comment from Charles. Physical optics is not good enough for pickup calculations because it does not do edge effects properly.

Comment from Kovac. Important to distinguish between using optics modeling to compare designs vs. estimate actual pickup. May be possible to validate models based on Stage-3 measurements. Big job. Need to figure out how to get the required inputs.

Benson: Charge to parallel sessions

Determine as much of the reference design as possible.

Start with major components/systems and work towards the details. Identify viable options we should consider.

Use Google sheet template, but add more detail as needed. Goal is to list one technology for the reference design.

Q. Stages. How do we include options? A. Add rows to the spreadsheet.

Q. Ruhl. What if we cannot reach consensus because of process, not lack of information? A. Need to make choices so the Decadal plan is sound. Keep talking until you reach consensus. We will have to choose when we write the reference design.

Q. Natalie. What about deciding where the telescopes go? A. It will be covered in reference design, but ICCC did not want to include it for this workshop. Natalie. Reusing existing telescopes affects cost, so we need to say something in the reference design.

Comment from Keith Thompson. Many viewgraphs say the reference design is not final. Keith wants us to say that the reference design will not have greater weight when we choose the final design. Charles. Remember not to decide over and over. Do the best you can at every stage. Brad. DOE process does include design alternatives.