Документ взят из кэша поисковой машины. Адрес оригинального документа : http://www.apo.nmsu.edu/Telescopes/SDSS/mirror_report/node4.html
Дата изменения: Thu Nov 25 00:45:52 1999
Дата индексирования: Sun Apr 10 00:20:54 2016
Кодировка:
Поисковые слова: р п р п р п р п р п п р п п р п п р п п р п п р п п р п п р п п р п п р п п р п

TPM log files

There are TPM log files for every night of the dark run except for the nights of 08 October and 14 October 1999. In both cases it appears that the usual location for the creation of the TPM log files (sdsshost:/mcptpm) had filled up, and so there was not enough room for the files to be created. This is especially unfortunate for the night of 14 October 1999, as it was on this night that the severe problems with both mirrors were reported. In any event, for the other 13 nights of the run, we have about 50 quantities logged every 0.05 seconds, including the MIG readings, the ALT, AZ, and ROT positions of the telescope, the velocities of those axes, and some weather parameters. The TPM generates roughly 1 GB per 12 hour session. There is only one piece of code that which worked and allowed one to translate the binary data of the TPM log into a usable ASCII file: ``log2asc''. We tried to use the GUI ``TPMrt'' but had no success in getting this code to plot any of the TPM data. Unfortunately ``log2asc'' will translate all of the data in a given log; there is no ability to select the fields of interest. There is also no way of specifying a time interval, so that the entire log needs to be converted. In any event, we found that we could convert the log into ASCII and then pipe the ASCII to an awk script to obtain the fields of interest, and selectively despose of data points (i.e. it was not neccessary to use every data point taken each 0.05 seconds to obtain plots of the positions of the axes or the MIGS). We used every 100th point or sampling every 0.5 seconds for the generation of the plots. Even with these caveats, we found that to took at about 2 hours wall-clock time to process 1 hour's worth of TPM log.

From the night logs, the critical nights were 10 October 1999 (MJD 51462) where we had an ALT runaway, and the nights of 14 and 15 October 1999 (MJD 51466, 51467) where problems were reported with the secondary. Kleinman therefore looked at the recorded MIG readings in the existing TPM logs during those nights. He also extracted MIG readings from 08 October 1999 (MJD51460) as a test of the system. The goal was to extract MIG readings from all nights, but due to the time consuming extraction process, did not go beyond these three nights. The figures below are plots of the MIG data with the y-axis in TPM MIG units (real MIG units*1000) and the x-axis in seconds since midnight. The temporal resolution is about 5s.

If we assume that secondary MIG values of 100 are not problems (this value was hit on the nights of 08 October and 15 October 1999), there do not seem to be any problems in the secondary MIG values as recorded in the TPM logs.

• For 10 October 1999: the night of the ALT runaway: which occurred around 05:10. Figure 1 shows no MIG irregularities around that time. The large jump occurring at 21000s ($\sim$ 05:50) was due to setting the focus anew on a new field.

  
  

   

  Figure: Plot of MIG values vs elapsed time from midnight 10 October 1999. The plot starts recording data from 04:35:22.0 on that day.

  
  
  
• 15 October 1999: This is actually the second night of the reported secondary communication problems. This night's problems were less severe than those of the previous night, but unfortunately there are no TPM logs for that night. Referring to Figure 2, the large jump near 75000s ($\sim$ 20:50) is due to homing the secondary. The next little spike at ($\sim$ 82000s ($\sim$ 22:45) is due to an incorrect focus adjustment where the value was given in absolute terms when a relative adjustment was expected. This was quickly fixed as evidenced by the narrowness of the spike. Nothing else unusual here, just the normal focus changes during the night.

  
  

   

  Figure: Plot of MIG values vs elapsed time from midnight 15 October 1999. The plot starts recording data from 20:13:37.0 on that day.

  
  
  
• 08 October 1999: This was the first night after the power outage. The plotted MIG in Figure 3 do not cover much timebase. The raise to values near 100 around 21:45 appears to be due to a homing of the secondary, but nothing of this is noted in the log, but the on-duty Observers cannot positively recall performing this. However, with everything else going awry that night, it's not surprizing.

  
  

   

  Figure: Plot of MIG values vs elapsed time from midnight 08 October 1999.The plot starts recording data from 21:26:52.0 on that day.

  
  
  

  
  

   

  Figure: Plot of Commanded focus position vs elapsed time from midnight 15 October 1999. The plot starts recording data from 19:26:40.0 on that day and should be compared to Figure 2.

  
  
  

We also looked at the ALT positions as reported by the MCP on the night of 10 October 1999 (MJD 51462) as it was on this night that an ALT runaway occurred, to the limit of 90.1$^{\circ}$. In this case, the thought was that the sudden stop at the zenith limit may have caused a breakage or malfunction in the secondary mirror support syetem, leading to a stressing of the mirror and hence to the eventual fractures found. Hayes used the TPM log of this night to see if the velocity of the runaway was potentially harmful of the telescope and the optics. This does not appear to be the case as the telescope obtains sustained speeds of $2.5 \times 10^{5}$ in MCP encoder reading, and does this on a number of occasions during the same night. It would appear that the motion of the telescope and its speed when the zenith limit was hit was not out of the ordinary, and so should not have caused damage to the secondary support system. We plot the TPM logging of the reported MCP position with time in Figure 4, and the reported MCP velocity for the same times in Figure 5.

 

Figure: Plot of MCP position vs elapsed time from midnight 11 October 1999. The plot starts at time 04:35:22.0 on that day.

 

Figure: Plot of MCP velocity vs elapsed time from midnight 11 October 1999. The plot starts at time 04:35:22.0 on that day, as in the previous figure.