STO-2 JPL/UofA on 05/20/2014

Transcription

1 STO-2 JPL/UofA on 05/20/2014 Date Description Author 5/21/ st complete draft C. Kulesa The overall discussion followed the following outline: 1. Local Oscillator implementation 2. Mixer implementation 3. Overall optical architecture 4. Schedule and milestones This document attempts to capture the sense of the meeting and provide a structure for developments in the immediate future. 1. Local Oscillators The JPL LO design has progressed impressively in the direction established for the GUSSTO Phase A design. This design will also be the implementation for STO-2. The prototype 1.9 THz [CII] LO block diagram and solid model are shown below. The hardware prototype and early results for output power follow. These results were performed with mostly fixed bias on the multipliers (some adjustment on the 220 doublers if memory serves). By eliminating the restriction of fixed bias, much greater (frequency) agility will result. It is expected that the [NII] chain could reach to CO for spectral line pointing and the [CII] chain to the 2.06 THz [OI] line for comparison with the 4.74 THz line with adjustable bias. Neither is a requirement for the STO-2 mission, provided that continuum pointing and focusing on planets is possible for flight. A top-level discussion of the biasing requirements accordingly followed. At the present time, all biases are explicitly provided by lab supplies. It is expected that many multipliers could be self biased for flight, but this has not yet been demonstrated. Because the STO-2 electronics subsystem must be designed now, we will baseline a design that will allow all multipliers to be explicitly and flexibly biased, with sensing of the biases for flight. However, the eventual goal is not to use more biases than

2 needed to complete the science mission. We set the 'ideal' goal of only needing to dynamically adjust the bias on the final 0.6 or 1.9 THz triplers, TBD, to be achieved on a best-effort basis. The expected drive voltages for the two chains are listed below. 1.9 THz LO Component Qty DC power (W) PhaseMatrix Synthesizer (27-40 GHz) Microwave Dynamics Ka amplifiers RF power Bias voltage range Bias current dbm 12V 1.25A W +9V -8V 1.2A GHz triplers mw 15 to 35V sinks 10 ma GHz doublers mw 0 to -15V sinks 5 ma GHz triplers mw 0 to 12V sinks 3 ma GHz triplers 4 0 >10 uw -3 to 1V sinks 1 ma 1.5 THz LO Component Qty DC power (W) Microlambda Synthesizer (12-14 GHz) Millitech AMC GHz RF power Bias voltage range dbm +15V +5V Bias current 1A dbm 8-12V 1A Quinstar GHz dbm 8V? 2.2A? GHz doublers mw 0 to -15V sinks 5 ma 490 GHz triplers mw 0 to 15V sinks 3 ma 1460 GHz triplers 4 0 >10 uw -3 to 1V sinks 1 ma In fixed-bias mode, a prototype [CII] LO chain is already capable of pumping 3 of 4 [CII] mixers in an STO-2 flight array (at or greater than 10 uw per LO beam). More power per pixel will allow for a thinner beamsplitter and lower sky losses; more consistency will make it easier to balance the LO power over the entire array with a common beamsplitter.

3 Discussion points: Q: What is the operating environment? A: 'ambient' both in lab and in flight. JPL should stipulate environmental requirements: temperature limits for operation and thermal stability requirements. Currently, I am carrying the following very loose criteria. Speak now or deal with it! Thermal characteristic of LO box Operating temperature range Absolute temperature range Desired setpoint Stability Assumed requirement 0 o C to 60 o C -40 o C to 80 o C 30 o C to 45 o C Stdev in bins of 1 min: <1 o C Stdev in bins of 1 hr: <10 o C Startup and shutdown procedures: nothing new here. LO Chain is brought up by biasing from high frequency stages to low, with an initial low gain setting on the W-band power amp. RF is then brought up, and the W-band amp drive is gradually increased while monitoring the multiplier state. LO chain is brought down in the reverse order. Discussion of AGC loop to stabilize mixer current and total power. Initial results shown by Abram Young look very promising, with current and total power stabilization due to sudden impulses, ~20 Hz microphonics, and long-period drift noise demonstrated using the Cologne 1.9 THz test flight HEB unit pumped by the VDI 1.9 THz LO. The mixer current monitor is the input to the PID loop, and the LO power control is provided by modulating the PSAT drive on a 26 GHz Spacek amplifier as input to the VDI multiplier chain. Everyone expressed the need to test stabilization on a JPL LO at some point soon. The intent is to modulate either the secondto-last tripler (630 GHz) or the final tripler (1900 GHz). UofA to bring the next-gen PID prototype to JPL once testing on the Cologne HEB and VDI LO is complete? Separately discussed: 1.9 THz LO in good shape, many parts and devices available, first to ship THz chain not at the same level of readiness.

4 2. HEB Mixers Jon showed the now-standard view of the 1x4 HEB mixer block staring at the so-called 'brass knuckle' beam transformation mirrors THz and 1.9 THz HEB arrays sit side by side in the dewar and share the same polarization. Without a wire grid for polarization diplexing, the feedhorn requirement for low cross-pol is greatly reduced and a simple diagonal horn seems adequate. Discussion between Jon and Groppi regarding adoption of a spline horn to further improve the beam shape is ongoing. Similar to the development of the LOs, the [CII] mixer arrays are in a more advanced state of development than the 1.5 THz arrays. Jon has the LNAs directly stood-off from the mixer arrays via G10 or comparable thermal isolator. This is to provide mechanical support to eliminate shearing of IFs as we experienced in STO-1 and to provide an LNA temperature of ~20K. Looks good, but I'm not sure we can tolerate any direct conductive coupling to 4K. I hope that a dewar design that allows safe opening and disassembly will eliminate the risk of coax breakage. The expected LNA temperature from the UofA side is ~40K if memory serves. This imposes a minor penalty in T IF. Jon emphasized the need to have direct copper straps from the 1x4 mixer arrays to the 4K CWS, not simply relying on the one-piece brass knuckle mount. Mounting holes and details are not yet in the current 1x4 mixer block design but this is planned. Current plan is to extend the 1x4 blocks slightly to provide enough meat to thermally anchor the blocks to the brass knuckles and the copper straps. Jon has sourced adequate UT-020-SS-SS coax from Micro-Coax for STO-2. This all-stainess coax has higher RF losses but much lower thermal losses than the STO-1 coax of the same diameter (no Ag flashing on the center conductor). Per foot RF losses look to be 5.5 db at 1 GHz and 12.3 db at 5 GHz. UofA needs to specify length of coax and we need to converge on a plan to safely mount the LNAs. Based on the lengths of the required coax, we can decide whether we should use the UT034-SS coax (w/ag flashing) from STO-1 if length >= 3 OR Jon's UT020-SS-SS coax (if length can be <3 ). 3. STO-2 Overall Optical Architecture Abe's 2 window port, 1 Si lens design remains the baseline as it is the only optical design in an advanced state of readiness and there are no showstoppers yet. Concerns remaining: The wide fan-out of the beams creates a very curved focal plane. This can be accommodated in the mirror design for the brass knuckles, however it generally means that the optical presciption (at least tilt) for each horn is different. This really shouldn't be a problem, AFAICT. Discussion of the 'silicon lens as dewar window'. Think we ought to be OK with this. Must prepare more than 1 window. Broad parylene coatings to be centered at 1.7 THz and 4.7 THz. LO relay optics now becomes the tall pole to tackle. Mounting orientation of LOs, fold mirrors, OAP mirror of minimum off-axis angle to feed both beams into dewar, entrance port diameter, location of silicon lens to match beam f# before diplexing at the beamsplitter plane. Requirement to get the spot diagrams under control in the final phase of lens and mirror design. Tolerancing discussions akin to those in STO-1. How to maintain alignment of the LO or sky windows and lenses with the 4K insert? Ball claims that the He can will not move at all but needs to be measured. Setting requirements and assessing design options require a minimal tolerancing analysis. i.e. how much can the innards shift w.r.t. the silicon lens and not disturb

5 the LO coupling by 1 db? A broad discussion of an off-axis Gaussian Beam Telescope followed at the sketch level. In short, a ~300mm x 300mm right angle GBT comprosed of two 90-degree focusing mirrors on top of the dewar injects an f/17 beam to just beyond the brass knuckle, which then images to the feedhorn. Looks quite fine! But I suspect that no one, particularly APL, is willing to take the Ball dewar off-axis without a really compelling reason. Is there one? Assuming that the tolerancing analysis turns out to be acceptable, above Schedule and Milestones The STO-2 master schedule is driven by two fundamental deadlines. 1. PIC meeting in January 2015 in which the flight payloads will be chosen 2. Hang test in Palestine, July The former requires T-V testing of the Ball dewar in as close to a flight configuration as possible by the end of The latter requires full integration of the STO-2 payload in the science flight configuration at APL in May 2015, and demonstrated again in Palestine in July To meet the PIC deadline, the baseline designs that were initially sketched in March 2014 need to be finalized to the extent that we can fabricate all dewar internals starting in June, for cold testing in July, and with available representative components installed in August for preliminary testing. Fun. Thoughts on the following? Items in bold represent deliverables and dates leaving JPL. Deadlines Baseline demonstration Minimal demonstration Delivery: 15 August 2014 Demo by: 10 Sept 2014 Delivery: 15 Nov 2014 Demo by: 10 Jan 2015 Delivery: 15 Feb 2015 APL Ship date: 1 May Ball dewar cold w/ Cryotel CT. 2. Existing lab prototype JPL [CII] LO installed at dewar port? 3. Single pixel JPL [CII] mixer installed? 4. Prototype brass knuckle and silicon lens. 5. STO-1 flight electronics and prototype augmentations installed for mixer stability testing w/pid on dewar w/ cryocooler. 1. Prototype [CII] 1x4 array installed and tested in dewar. 2. Prototype JPL [CII] LO installed. 3. Flight mixer insert, brass knuckles and silicon lens AR-coated and installed. 4. All of the above T-V tested in chamber. 1. Flight [CII] and [NII] arrays in dewar. 2. Flight [NII] and [CII] LOs installed. 3. All dewar internals ready for flight. 4. All instrument components and subsystems tested. Ready for payload integration. 5. Second T-V test scheduled before ship. 1. Ball dewar cold with Cryotel CT 2. VDI [CII] LO installed at LO port 3. Single STO-1 [NII] or test flight [CII] mixer installed 4. Prototype brass knuckle installed. 5. STO-1 flight electronics and prototypes installed for stability testing on dewar w/ cryocooler. No descopes really possible. Prefer no deviations in dewar... But possible to continue to improve and replace LO components or other assemblies outside the dewar.

6 Acceptance testing before shipping is a rather loose notion at the level of an APRA grant, but shall we also try on a best effort basis to do the following before the shipment of final (or near-final) hardware in February 2015? Flight Component Acceptance test Requirement 1.46 and 1.9 THz LO units Power output > 10 uw in each of 4 beams. Beam shape, f# Spectral purity Documentation 1.46 and 1.9 THz mixers Sensitivity Beam pattern Documentation Consistent with 0.3mm waist at 1.9 THz. Single tone, spurs >17 db down (FTS scan) Tuning sheets for primary and secondary lines (1461, 1497, 1900, 2060 GHz). Do not exceed bias parameters for all multipliers that are not already self-biased. < 1500 K DSB Shape/position emerging from test dewar consistent with the optical prescription. Must hit positions of Si lens and secondary! Recommended bias points for operation. Allan variance result (or TP/spectral data available)