Carolyn McCoey (Univ. of Waterloo) Emmanuel Caux, IRAP (ex-cesr) Consortium

Transcription

1 NHSC HIFI DP workshop Caltech, 7-9 February 2011 The HIFI Instrument Status, AOTs, and Calibrations Pat Morris Steve Lord, Adwin Boogert, Colin Borys (NHSC) Carolyn McCoey (Univ. of Waterloo) Emmanuel Caux, IRAP (ex-cesr) Thanks Frank Helmich, HIFI PI, and the HIFI Consortium - page 1

2 Welcome NHSC team Science Adwin Boogert Herschel DP Training Champions Group Colin Borys NHSC rep Herschel DP Mgt Group Steve Lord Pat Morris Carolyn McCoey, U. Waterloo Herschel DP Training Champions Group Emmanuel Caux, IRAP HIFI Co-PI HIFI rep Herschel e DP User Group HIFI Users Group NHSC rep Herschel DP User Group Applications Joan Xie Annie Hoac - page 2

3 Outline Part I: Expectations & Goals, Preliminaries, Data Access Part II: Instrument Status and Health Part III: Calibrations Part IV: Cross-Calibrations (HIFI vs APEX, SPIRE) Part V: AOT Status - page 3

4 Expectations & Goals Learn HIPE Become familiar with the GUI concepts and usage, how to execute pipelines and explore your data Learn where to find templates data reduction scripts and how to use them... and import and execute custom scripts Learn how to read/ingest your data in HIPE Learn how to process HIFI data to Level 2+ products Learn about resources and support NHSC can provide Processing science in HIPE with CASSIS (joint session) Sessions will use public data for demo and practice purposes. - page 4

5 Preliminaries Do you have the right version of HIPE installed? Where are the data? Should oudi have aethem already? What if I don t? How do I ensure I m ready to go? Where are the scripts? What are tutorials? The next two presentations will answer these questions. We have allowed for some time to be used for testing your setup (before/during coffee) - page 5

6 Data access this week ESA Mission Ops Center NHSC/HIFI has reprocessed data in HIPE 5.1, and provided in local stores HIFI ICC DB HIPE HSA ESAC, Madrid NHSC/HIFI DB NHCS - page 6

7 Data access at t > this week ESA Mission Ops Center HIPE ESAC, Madrid HSA The HSA is the only repository of observations and calibrations that can be accessed by the observing community. home Help from NHSC and HSC always available. - page 7

8 Part II Instrument health and status - page 8

9 HIFI Single Event Upsets An SEU in August 2009 attributed to CR on a sensitive memory chip in the LO control unit caused premature power-off of the LOU, killing HIFI s Prime side (DC/DC converter failure) HIFI runs on its REDUNDANT side since February 2010, with new intensive fault protection S/W. SEU occurring statistically every 12 days Recovery now well established and rapid. One full power cycle (last resort) has been required, when HIFI stopped reacting to commands. ~2 ODs lost to LCU SEUs 5 ODs lost to ICU SEUs but PACS rescheduled so no observatory time lost. - page 9

10 LSU OXCO Anomaly HIFI s 10 MHz signal (for locking LO chains, LSU, WBS and HRS) is generated by a crystal. Crystal is heated to 100º C for best performances The heater circuit is not working well. Temperatures began spiking in June 2010, investigation concluded that the circuit is reacting to LSU temperature control switching to an on/off mode. The LSU and HIFI are operating within design specifications, and the crystal temperature is within the nominal tolerances so science observations are not affected. - page 10

11 - page 11

12 LSU OXCO scenarios We cannot change anything in software or by commanding. We have to wait and see what happens, three scenarios are possible : Situation remains stable Currently science data are not affected Keep operating HIFI as is Heater circuit dies, but crystal keeps working Drift of HIFI frequency enc expected that needs to be characterized, ed especially at the jump (may then stabilize) Dedicated characterization + observing known lines every OD Heater circuit dies, takes crystal with it End of HIFI - page 12

13 WBS-V comb fade WBS-V comb has a failed component, already discovered in 2007 The WBS-V could not be removed and repaired. Comb power continued to drop Steep decay between OD 320 and OD 420, lighter since No remaining attenuation in subband 4 (2-3 db in others) Comb line intensity dropped to < 40 counts, much lower than specs. S/N still > 400 in sub-band 4 - page 13 Possible reduction of accuracy in determining comb line positions, reduced frequency calibration accuracy expected

14 HRS-WBS cross calibration Measure pair of spectra with WBS and HRS at any input (preferentially from internal 100 K hot load) Compute mutual correlation as function of relative shift HRS can be positioned to cover almost the full WBS scale Some On-Board S/W bugs to use this being worked out. Accuracy of about 100kHz (best case 30kHz) No issues with WBS-H so far. - page 14

15 Reality Check The HIFI team has been working through so-called Failure Mode Effects and Critical Analysis (FMECA) since assembly and test phases. A lot of things are related to problems anticipated at launch Secondary are all the power supplies, which are supposed to be robust, however, not guaranteed Tens of items exist that say switch to redundant that option was required already during PV, thus is no longer available and thus we operate with numerous possible single point failures The FMECA is on functional level Hundreds of components can cause a functional problem. Chance that HIFI survives the remaining mission is not close to 100% anymore - page 15

16 Part III Calibrations - page 16

17 LO Spurs and Spurious Response Impurity features in HIFI are manifest as ~25MHz, Gaussian-shaped features in hot, cold, on, and off data. Hot and Cold Loads from an observation in band 1a Both WBS and HRS are affected. Because they look a lot like astronomical lines, they can fool users. The repeatable ones are tabulated in PHS (HSpot) and pipeline tables. In the pipeline, the hot/cold loads are searched in all science observations, and spurs are flagged. - page 17

18 Spur clean-up at instrument level Many of the most troublesome impurities (e.g. near key lines) have been cleaned up, by (time/labor intensive) tests and changes to multiplier settings and bias voltages. H Band 1a spur eliminated (557 GHz H2O line region) System temperature increased at upper 1.5 GHz of band Sensitivity in band 1a is better than that in band 1b for LO frequencies up to GHz. V V - page 18

19 Band 5b purification LO-band 5B had several LO spikes at the same time Result is that the IF consists of several frequency bands superposed on each other. Calibration of all frequencies within the IF was unknown Purification efforts took many weeks. AORs in 5b were released in December Band 5B was not alone in impurities but in the diplexer bands the diplexer can moved and the signal evaluated at each setting to give purest signal more efficiently (in observing terms). Purification in bands 3b/7b done this way (Band 5 has no diplexing). Purification is a major triumph for modeling of the LO chains! - page 19

20 Sensitivities Sensitivities are driven mainly by System Temperatures and Aperture (or beam) Efficiencies. Tsys remains essentially as measured just before launch --- some small changes at specific frequencies to mitigate impurities and unstable tunings. - page 20

21 Beam Calibrations H/V beam sizes and co-alignment Measured mainly from Saturn, Uranus, Mars Beam offsets are taken into account in the HIFI pipeline since page 21

22 Efficiencies Beam efficiencies measured on Mars + Overall ~10% better than pre-launch estimates o Data taken before Nov 2010 should have somewhat better noise performance (HSpot 5.2 uses updated effciencies). Band 5a/b are an exception, ~10% lower. - page 22

23 Error Budget Error source B1/B2 B3/B4 B5 B6/B7 Sideband ratio <10% <30% <10% <30% Hot Load Coupling <1% < 1% < 1% < 1% Cold Load Coupling <1% < 1% < 1% < 1% Hot Load Temperature < 1% < 1% < 1% < 1% Cold Load Temperature <1% <1% <1% <1% Beam efficiency 5-10% 5-10% 5-10% 5-10% Pointing (APE+co-alignement) < 5% < 5% < 5% < 5% Standing Wave 2-4%??? Total (quadratic sum) <16 % <32 % <15% <32% - page 23

24 Part IV Cross Calibrations - page 24

25 HIFI / Ground Comparison made of HIFISTARS spectra with APEX spectra (CHAMP+ SRON/NOVA mixers in MPIfR instrument) Scaling consistent with different beam sizes - page 25

26 HIFI - SPIRE From HIFISTARS program Continuum less reliable for HIFI at higher frequencies (drift) Observations not optimized for continuum measurements HIFI-V HIFI-H Good agreement in continuum at low frequencies CO 10-9 in CRL618 - page 26

27 Part V AOT Status (Not an Intro to the AOTs) - page 27

28 HIFI AOTs in OT1 (mid-november 2010) Relative distribution of SScan and Map times is ~reversed from GT/OK KPs. - page 28

29 - page 29 Band requests in accepted OT1 programs

30 General Status All HIFI Observing Modes released and being scheduled. Certain restrictions and recommendations remain in place: FSwitchwith Spectral Scans is forbidden in Bands 6 and 7. LO power is difficult to stabilize between ON and OFF (throw). If FastDBS can t be used, LChop is the alternative. Not quite see next slide! FSwitch with OTF maps is forbidden in Bands 6 and 7, except over a range of frequencies around 1897 GHz (C+ tuning). Fast(chop)DBS is recommended over DBS in Bands 6 and 7. All FSwitch and LChop with all modes should be taken with sky reference, esp in the diplexer Bands 3,4,6,7. Data which were taken inconsistent with these recommendations (mainly AOT tests and some early PSP/SDP data) to be dealt with in IA. - page 30

31 WBS-V leveling/saturations Some systematic issues Spectral Scans Certain LO tunings in Spectral Scans lead to strong IF instabilities, as a result of saturating the WBS-V on the hot load. First noticed over ~ GHz; investigation reveals other isolated ranges over most all bands. 4a and 6a most affected (SScans there on hold), 3a/b least affected. Several problems lead to this behavior e.g. initial power levels deduced from the cold loads. will be corrected by increasing the WBS attenuator tuning grid and update to power level tables where not sufficiently covered AOT validation (remember HIFI went on a fast-track track release after loss of Prime side). Testing in progress. - page 31

32 Some Systematic Issues Spectral Maps OTF mapping modes were the last ones to release, could not receive detailed attention during HIFI s time-out. Several changes were necessary: Sky sampling definitions (Nyquist vs half-beam ) Density: Maps taken < OD 280 uses Nyquist = HPWB/2. (half-beam), > 280 Nyquist ~ /2D. / DBS Raster affected also. Compensation for a zig-zag effect - page 32 Slight timing mis-correlations due to command timing in the uplink system, scan speed granularity. Overall not major problems but could affect what the user expected vs sees in the data. OTF only.

33 More details on Observing Mode performances Are in the HIFI AOT Release Notes at HSC: Update to the AOT Release Notes for OTF modes is pending (next week!) The site also contains informational notes about Beam efficiency calibrations, and effects in HSpot and the pipeline. H/V beam pointing in the pipeline. More info on the data by mode will be covered by Carolyn (Level 1 and 2 pipeline overview), and in the mode-specific presentations and demos. Further resources (pages, online and video tutorials, science tools) will be explained during the workshop, and summarized at the end to take all inputs into account. - page 33

34 - page 34 Questions?