= 850 camera Splinter discussion = == Monday 19th May == Splinter Session on New 850 Micron Detector Presentation by Dan Bintley: * [[attachment:Bintley-Dan-Spliter-session-EAO-Futures-2019.pdf]] === Notes on the Discussion === SCUBA-2 has been operating for 8 years. It works extremely well and is doing great science but it hasn’t met the original goals of mapping speeds/sensitivity (yield of arrays affected this, non-uniformities - originally planned capabilities were significantly reduced) So, with a new instrument 10 years after SCUBA-2 was built - we can correct those issues and build something that maps an order of magnitude faster - allowing scientists to achieve far more depth and cover larger areas We need a collaboration to accomplish this - there many technical issues ahead, but we have experience in tackling similar issues and there were many lessons learned from SCUBA-2. We will be working with NIST - where we plan to purchase the detectors. === Concept === See linked talk at the top of the page for images and schematics. Simplest design - smallest cryostat this is compatible with the instrument requirements - half the challenge is finding the better detector array than we have in SCUBA-2 Biggest array in that cryostat (SCUBA-2 doesn’t quite use full FOV) Need to control excess emission and stray light Currently, we see that SCUBA-2 is receiving more power than it should when observing. The excess emission is caused by stray light. It adds to photon noise and limits our ability to detect faint structure The pixels themselves can be polarisation sensitive. We can improve the polarisation sensitivity by a factor of 2. Then, taking into account the order of magnitude increased mapping speed, we can improve the polarisation observations by a factor of 20 We plan to keep SCUBA-2 operational during the whole installation and commissioning process Circular 12 arcmin FOV (SCUBA-2 has effective FOV of ~6-7arcmin) Focal Plane Temperature <200mK - Adiabatic Demagnitisation Refridgerator (ADR) versus dilution fridge (what we currently use) to achieve this. ADR needs to be regenerated daily. Subaru is using an ADR+KID array, so we know it can work. Rotating half wave plate for polarimetry still necessary even though pixels can measure 2 polarisations. We need to decide if ambient temperature is ok (outside the window) or if cooling it to liquid nitrogen or 4 K temperature is necessary! This will be a feed horn array as opposed to bare pixels - MKIDS need a way to get the radiation onto the sensitive parts of the array - a feed horn is a common solution to that An ADR was suggested when we wanted the instrument situated in the cabin because we were worried about the tipping of the dilution fridge. Now, the instrument will likely go on the Nasymuth platform. If we don’t have to consider tipping, a dilution fridge is the better option. Current cool-down time from 4K -> Base temp = 6 hours (75% chance) (25% chance we need to adjust mixture and we get back on sky the second night after power cut). Full warm-up to ambient - down to base temperature = 3 weeks. So we try to catch it at 4K as best as possible. Availability of Helium is an increasing worry in Hawai`i - just to get regular deliveries is becoming more and more difficult. We have had seals that failed - Flexi lines that cracked - we’ve lost all Helium 3 several times Newer systems shouldn’t need Helium 3 refills at all - they are dry systems with pulse tubes (No Flexi lines, no pumps - closed system) TOLTEC has MKID arrays at wavelengths all longer than 850 microns - successfully produced. Shorter wavelengths (250, 350) have also been produced. So, if shorter wavelengths and longer wavelengths work, we are confident it will work at 850 microns. Head mechanical designer at AAO has worked with SCUBA-2 and JCMT on several projects, so they are intimately aware with the system - knows optical and cryogenic issues (See Slides) Question: Why are we using the PT410 as opposed to the 415 or 420? Dan Bintley: No reason - we may of course go with a bigger pulse tube - above is purely a concept. The Original optical design was too simplistic - we thought we could include a pickoff for new instrument and SCUBA-2 Would still work with a bigger 12’ full field of view! The reality is more complex, we would need a series of more mirrors - not final design - need to marry this to the mechanical design and the instrument itself - still work in progress We cannot allow stray light. We can’t make those compromises again as they hindered SCUBA2’s performance. We want to avoid excess power loading if at all possible. How much do extra mirrors contribute to excess power emission? SPT-POL doesn’t have a half-wave plate. Do we need one? Rotating wave plate helps mitigate low frequency noise (many more advantages). It may also introduce dynamic range issues. Stepped half wave plates? Can we rotate purely in the software? In heterodyne when you correlate in phase - that’s possible. But how do you do it in continuum? This wouldn’t have the advantages of overcoming low-frequency noise Question: Are we thinking about a cassegrain focus instead of nasymuth? This is better for correcting instrumental polarisation. Dan Bintley: We would need to deal with the tipping - but we have an optical model to put the instrument in the cabin. The disadvantage is that it would take most of the cabin - and we would need to change the structure of the telescope. Without changing that structure - we risk losing the other instruments (like the new receiver, Namakanui). So, practically this would be difficult. We want a strong science cases, however, to drive the technology development - so if some very good cases exist for putting the instrument in the cabin - that’s what we should do! MKIDS can respond to much faster scanning - but we are already at our mechanical scan speed limit! Still, there is some re-optimisation that will need to take place so we may not be able to immediately use the data reduction software with the new system. Factor of 10 increase in speed = FOV/Yield/Sensitivity of the detectors. If you believe the models - it should increase by a factor of 30 - but 10 is probably closer to reality. There is currently no dedicated project manager overseeing everything. --- We need help to produce a document that captures all the pros and cons of each design choice --- === Collaboration=== Note that while Canada is leading a proposal to fund the readout electronics, this doesn’t mean they must be developed entirely in Canada. In fact the funding proposal includes talk about external collaborations. There should be collaborations between different groups across the EA regions. === 450 microns - Necessary? === Is dropping 450 microns a real problem over the next 5 years? We would prefer to have SCUBA-2 in the mix - but would it hurt to lose? Much discussion about the importance or lack of importance of 450 microns. Magnetic fields! Both wavelengths (450 and 850) help better constrain the properties of dust. The Transient observations of protostellar outbursts rely on 450 microns and the great resolution that bridges the gap between ALMA and space telescopes. Extragalactic observations take advantage of the resolution. 450 microns is a critical wavelength for cosmological surveys of high-redshift galaxies. Question: How much can we improve the 450 micron sensitivity? Jess: Probably by a factor of 10. A viable next generation option is 450 microns. The 850 array should be considered a “first generation“. There are practically no single wavelength instruments in existence. Multi-wavelength is important, but it is this complexity that introduced some of the issues with SCUBA-2 IRAM+NIKA2, Toltec observing at longer wavelengths, sot the JCMT has a niche. The LMT may going to 850 microns. Paul Ho: Faster Detectors can make better use of marginal site (in terms of weather)! If we can map 10 times faster, even though there are only a small percentage of nights suitable for 450 micron observing throughout the year, on those good nights we will get much more data than we are now! Large surveys are critical in this era. Why not 2 cheaper instruments - one on each Nasymuth - a 450 micron and an 850 micron, both MKID arrays? This is cheaper than operating SCUBA-2 for when “good weather” nights are happen. 2 instruments - 450 and 850 - may cause more than a year delay and would result in higher costs if we keep SCUBA-2 running during installation and commissioning. Can we just put MKID arrays into SCUBA-2? JCMT one of only places to test 450 microns - but 450 microns should be a focus of the next generation of Submillimetre telescopes. === Cryogenics === ADR (Adiabatic demagnetization refrigerator) vs Dilution refrigerator ADR beneficial if you want it in the cabin (where it would tilt). However, on the nasmyth this is not necessary. Also the magnetic field issue is a problem. Any 'pro's' for an ADR? Cheaper. Anything else? Chase fridge (CARDIFF) - for Muscat. Miniature dilution fridge (with paired continuous sorption fridge at 350mK, and continuous 1K sorption fridge) - sustained 150mk (five to six days). Can survive brownouts... but still need pulse tubes. Cool down time is 48 hours... full 80 hours to get down to 150mK. Uses 9L of He3. Liquid cryogenic sourcing on Hawaii islands is difficult, and increasingly so. This might be an issue if we need cold traps. The new closed systems should not require this however neither was this in the original SCUBA-2 design.