From l.rizzi@jach.hawaii.edu Tue Oct 23 15:49:30 2007 Date: Tue, 23 Oct 2007 01:55:35 -1000 (HST) From: Luca Rizzi To: Mike Irwin Subject: draft of notes from my visit Visit to CASU, Cambridge, Oct. 22/23, 2007. - Re-reducing data upon PI request? There is no easy yes/no answer. Depends on whether the request is sensible, meaning it actually improves the results, either because of the complex observation strategy or because of a problem with the standard reduction. Also, if the re-reduction requires a considerable effort from CASU, then the PI should probably consider a co-authorship with one or more of the CASU staff, since CASU members have research time and this would probably be a viable way to use it. In any case, this is on best effort basis. - Sky 1. The sky subtraction performed by the pipeline is not a sky subtraction in the usual IR meaning. It's more of a correction for large scale features. Indeed, the sky frame are normalised to zero before being subtracted. 2. CASU is interested in collaborating with JAC in putting in place the infrastructure to deal with ABBA or On/Off cases. This requires: (a) JAC needs to be able to perform these kind of observations (the SKY eye should not crash the OCS) (b) only a fixed set of observing templates should be allowed, not full freedom (c) a sky frame should not be used with more than one MSB (d) CASU need to look for these frames and reduce them properly, then associate them with the MSB in which they are observed - Access to raw data. It has been verified that raw data are fully online, and that no human interaction is required at CASU to provide them - PI vs. UKIDSS, any priority? None on the CASU side. Nights are reduced exactly in the order in which they are received. Might need further investigation at WFAU. - Internet transfer. Tests have been done showing that only with multiple simultaneous transfers we can achieve a significant speed improvement but at the expense of server load. The problem seems to be related to buffering. Tests would be much easier with a dedicated account for CASU on our machines. - Requests, improvements, suggestion by CASU for JAC 1. Carefully check the number of darks, especially for rarely used observing modes 2. Make sure narrow band flats are taken when needed 3. Increase the dither size of the flat fields to at least 1/4 of the detector 4. CASU notes that refocussing has a very noticeable effect, so do it often 5. Avoid taking flats near bright stars 6. IMPORTANT: we have lost entirely the synchronization of the data sent to CASU. They often receive partial nights, with only a subset of the 4 chips. This is a major reason for delay in the reduction. This needs to be fixed, and did not seem to happen during early phases of WFCAM operations. 7. Marco would like to receive a schematic of the physical position of the temperature sensors, giving a key to identify them in the header keywords - ACTIONS: 1. Provide CASU with a fully operational account on one of JAC machines, to be able to run FTP tests without constant interaction with JAC staff 2. Perform linearity tests. Possibly on a narrow band filter using as constant as possible a light source and alternating with a fixed exposure time to monitor the light source variations and/or the effects of previousl illumination. 3. Provide a formal way to relate to CASU all the events that affect the camera (Change in software, board swaps, filter swap). The contacts should be frequent and regular, especially near an engineering night. CASU might have engineering requests that we could easily carry out. - Other issues: - closing the loop: we have discussed the possibility of using the quality control data provided by CASU to decide whether an MSB should be repeated, or to address issues related to the general health of the data being taken. It is noted that it would be extremely useful, but probably not practical at present due to the delay between observations and reduction. The issue should be explored again as soon as we have the FTP transfer in place. It might be very helpful for support astronomers to assess the quality of UKIDSS data. - logging: Comments put in the log are not usually read at CASU unless a problem develops with the reduction of a specific night. In any case, comments on data reduction are surely not automatically used to steer the behaviour of the pipeline. Special PI requests should be dealt with separately (see before). - new paper on calibration is being prepared. Among other results, it shows evidence of a spatial dependency of the zero points. Overall, the 1/2% photometric accuracy is achieved with 2 mass. - Campaigns proposals: I suggest that campaigns proposals would benefit from having a similar progress page as UKIDSS. - reporting processing or data quality issues discovered whilst running the pipeline. This would benefit from a more regular input from CASU, even if to the effect of nothing found.