(Mostly) GUPPI/PUPPI/ASP observing instructions, with GBT and AO project-specific information
P-ALFA signup page is here.
The GUPPI Astrid guide is here. Older webpages: setup instructions are here and data-taking instructions are here. Both of these are linked from this page. These are GBT pages maintained by Scott and Paul -- contact them with questions!
Initial PUPPI instructions are here, thanks to Paul.
Generic ASP/GASP operating instructions can be found here.
Rob's instructions for writing 0737-3039 GASP data to disk.
A quick version of the operating instructions for the now-defunct BCPM is here. Note that this page contains useful IF/LO convertor-rack frequency information that also applies to GASP.
It is assumed here that all observers understand how to operate the GBT using Astrid, and how to use vnc. For an example of astrid+spigot (now also not used!), see this email from Scott Ransom here.
Current channel offset configurations for ASP, updated May 28, 2008. These change
every time a data server has problems and we need to swap EDT cables around.
This is another useful thing to check in troubleshooting.
ASP, 2-ds mode
chan_offset 0: ds2
chan_offset 8: ds1
Currently dead nodes, as of July 22, 2010:
ASP: many. ASP is slowly becoming unusable.
Setup instructions for specific GBT and AO projects. These are intended as reminders for observers with a little experience; if anything looks overly mysterious, please contact Ingrid for ASP clarification and Paul or Scott for GUPPI/PUPPI clarification.
- P2702, 0337 observing with PUPPI (and ASP)
- Log into CIMA, listing project p2702 (lower-case p). That will tell CIMA to write log files to /share/obs4/usr/pulsar/p2702/. Choose pulsar mode.
- Once in CIMA, hit "Load Setup." This should provide a list of possible setups found in /share/obs4/usr/pulsar/p2702. The only one there right now is puppi_1410_newcima.conf. Choose that and hit "Load and Apply". It will probably take a minute or so to put the right receiver in place and get the IFLO set up.
- Hit "Pointing Control," make sure you're in the "project catalogs" set, and choose 0337.cat. Select 0337 (it's the only source in there!) and hit "Apply".
- While you're slewing, hit "Pulsar observing" and change "Winking cal" to OFF. Then hit "Power control." In the widget that pops up, hit "Adjust all". At this point, run guppi_adc_hist on puppi. (And rms on ASP if you're using it.) The puppi levels should be essentially OK, but if they're not, you'll have to adjust the puppi attenuators. Right now Paul and Scott are the only people who understand how to do that.
- When you're on-source, hit "Utilities" in the main CIMA menu. In the "Send to VX-works" box, type "setcal lcorcal" and hit enter. (The words will not disappear and you will have to delete them manually.) Then type "cal25 on" in the same box and hit enter.
- Now that the cal is on, take the cal scan with puppi (and asp, if using).
- When the cal scan is complete, go back to the Utilities window, type "cal25 off" and hit enter. That will turn the cal signal off.
- Take a real data scan until the end of the observing time.
- GBT07C-084, 1906+0746
- Spigot plus GASP at L-band only.
- Astrid blocks: use the ones in 07C-084. These are clearly labeled to remind you that all observations happen at L-band.
- Use 1400_B_multi_telescope_setup but do not use the the rest of the "multi" setup blocks as they are buggy. (If by mistake you don't use 1400_B_multi_telescope_setup but use 1400_B_telescope_setup insted, you will need to run (bash; source /home/gbt/gbt.bash; config_tool < ~istairs/gbtobs/lband_lin.py) after the initial setup; astrid will prompt you for this.) Edit the spigot_setup and spigot_takedata blocks to set the correct desired data directory and observation length. Follow instructions for when to check on GASP rms, cals on/off, etc.
- GASP config files are in gasp0:/home/gasp/config/GBT07C-084. Many files are experimental and do not actually work. Use these: config.1937.1400.xml.2ds (for the test pulsar 1937+21) and config.xml.1906.1400.2ds. Edit these for cal on/off, observation length, and any dead nodes or unresponsive CGSR machines. Dec. 11, 2007: yes, these are the correct config files to use!
- If at all possible, get continuum-source off/on cal scans first (gasp is more important than spigot) using config.xml.1442_off.1400 and and config.xml.1442_on.1400. Astrid blocks G through J will walk you through this, going to 1442 off-source first, then on-source.
- GBT07A-087, NANOGrav long-term timing
- Check here for instructions.
- GBT10B-033, 0737-3039A/B
- GUPPI and GASP, at either 820 or L-band.
- Astrid blocks are in GBT10B-033.
- Before observing, check the disk space on beef and decide where you are going to put the data. Edit the 1400_A or 820_A block as appropriate to contain either /data1 or /data2. If you do not do this, your data may go to the wrong disk and there may not be enough space!
- Astrid blocks are labeled starting with 820_ or 1400_ depending on the desired frequency. Follow them in order (Except for the the 1929 track; see below!).
- Astrid will prompt you to run guppi_adc_hist to check the guppi convertor-rack attenuators (most likely in modules 4 and 8, but check the bottom of the gbt_status screen to see which backend is using which modules). These attenuators should get automatically set by astrid now, though, so you shouldn't have to adjust them. You should adjust the GASP CR attens on CM 11 and 15 using rms as usual. Then, during the test scan on the test pulsar, astrid will prompt you to using guppi_monitor to check on the guppi scaling. If you have to change guppi.scale, you'll need to abort, edit for guppi.scale and then re-run the B block. Note that the convertor rack attenuators should not need changing the second time through; if they do, something is wrong, so call Scott or Paul if you can't figure it out.
- GASP config files are in gasp0:/home/gasp/config/GBT10B-033 and subdirectories. For 820 MHz, try running with config.xml.0737A.820.realtime.less_nodes. If this overflows too much, then move to the one without .less_nodes. Remember to cgsr_send_all before each folding scan, once the config.xml and polyco.dat files are in place.
- Use 0740-28 as a test pulsar (make sure it appears on stat_asp and with pav -FY on the guppi output file), then slew to 0737 and calibrate. At 820 MHz, use the config file in gasp0:/home/gasp/config/GBT10B-033/820/config.xml.0737A.realtime.cal. At 1400, use gasp0:/home/gasp/config/GBT10B-033/1400/config.xml.0737A.1400.3ds.realtime.new.cal .
- Then edit the astrid D_takedata block to set the scan length, leaving time at the end to do cal scans on 1442off/on if desired (a good idea if you have the time and most important on gasp!).
If you try to edit the D_takedata block and it says it's invalid: it is NOT invalid; you just don't have the background information stored in the astrid memory. Solution: save it anyway. Switch to block A in the edit tab (for the correct frequency) and hit "Validate". Do the same for block B. That will put all the source coordinates and guppi scale factors in memory. Then go back to block D, validate, and then save agin. It should be fine this time.
- Our standard procedure is now to fold pulsar A in real time on gasp and write data to disk only during eclipses. To see when eclipses are coming up, log into gasp (I suggest on another desktop than the one where you'll be running the gasp programs), type "dopoly_plotorbit.pl 820 55150" (replacing the frequency and MJD with current values!) and then run "plotorbit -p polyco.dat", which is Michael's display program. You should make sure to take gasp data to disk when A is between orbital phases ~80 and 100 as indicated by this program. NB it is a bad idea to rely on the tempo mean-anomaly orbital phases!!
- For real-time data acquisition, you'll need polycos for A. To make these, cd to ~/runtime on gasp and do "makepoly 0737-3039A 822 64 1 8" (substitute 1402 and 128 if at L-band). This will create the standard ~/runtime/polyco.dat; make sure to copy it to polyco.dat.0737multi.54620 (or other mjd at end).
- For eclipses disk-writes, Rob's full set of instructions can be found here, but they include everything: cals, etc. You've already calibrated, so you don't need all that rigmarole, but read over it anyway. The main things you need are "record_setup_data (or
record_setup_data2 obs freq", start_record, start_ds and go. Deciding what "obs" should be, and whether to use data or data2, are a little tricky, and I will do my best to communicate that information to the observer beforehand. If you haven't been told what to do, use node_disks to see whether /data or /data2 has more space. Then use something like foreach_node "ll /data/gasp/GBT10B-033/obs?" to determine which obs directories are empty. Select your obs number (for record_setup_data) to match one of the empty directories, or you will overwrite data and cause much grief!
- For the 1929 track, the astrid block order (all at 820 MHz) is: 820_A_setup, 820_B_1541_test (not 0740!) (take gasp data using 1541 config file), 820_E_slew_cal_1442off, 820_F_slew_cal_1442on, 820_G_slew_cal_1929, then 820_H_takedata_1929 (edit the scan length as needed). Guppi disk settings, attenuation prompts, and gasp need for cgsr_send_all as the same as during the regular 0737 run. Note that the cal block is set up to take cal scans both on and off the 1929 position; please take gasp scans at both locations.
- GBT06B-028, 1620-26
- BCPM plus GASP, L-band.
- Setup: Astrid blocks are in GBT06B-028. Use 1400_config to configure, 1400_slew to slew, and 1400_calscan to toggle the cals.
- Important: On cleo receiver screen, check the RF filters when astrid asks you to, checking that they are the 1.3-1.45 GHz filters.
- GASP config files are in gasp0:/home/gasp/config/GBT06B-028. The correct file to use for L-band observations is config.xml.1620.1400 . Edit this for cal on/off, observation length, and any dead nodes or unresponsive CGSR machines.
- BCPM gets run by hand for this project : 1 MHz channels/72 usec sampling.
- GBT07B-034, 1756-2251 and 1802-2124
- GASP only, L-band.
- Setup: use astrid, project GBT07B-034. There are various blocks for 1756 and 1802 at the bottom of the block list, starting with "1756_" or "1802_". For setup, run "1756_1802_1400_config".
- Important: On cleo receiver screen, check the RF filters when astrid tells you to, making sure they are the 1.3-1.45 GHz filters.
- Slew using either 1756_slew_to_1756 or 1802_slew_to_1802.
- GASP config files are in gasp0:/home/gasp/config/GBT07B-034. The correct files to use for L-band observations are config.xml.1756.1400.3ds and config.xml.1802.3ds . Edit these for cal on/off, observation length, and any dead nodes or unresponsive CGSR machines.
- There is a 1756_1802_calscan block which will turn the cals on and off for you, so that you can then run a gasp cal scan. For the actual data scan, you don't run an astrid block; all
you need to do is start up gasp.
- GBT07B-034, 1740-3052
- BCPM plus GASP, L-band, 3 (or at least 2) different frequencies.
- Setups have now also migrated to Astrid, under GBT07B-034. These setup blocks only configure, point, and toggle cals; you still need to run both gasp and the bcpm by hand. 1400 and 1190 MHz are the two most important frequencies; do 1660 only if you are able to get a full half-hour at each of the other two frequencies without problems.
- Continuum cals off/on 1442+10: only at 1400 MHz, and only if you have at least 2 hours including setup time. April 2007: skip doing these.
- GASP config files are in gasp0:/home/gasp/config/GBT07B-034. The correct files are: config.xml.B1442.1400.off.2ds , config.xml.B1442.1400.2ds , config.xml.1740.1400.L , config.xml.1740.1190 and config.xml.1740.1660. Edit these for cal on/off, observation length, and any dead nodes or unresponsive CGSR machines.
- BCPM gets run by hand for this project: 1 MHz/72 usec at 1400 MHz; 0.5 MHz/72 usec at 1190 and 1660 MHz.
- Arecibo P2286, long-term timing
- Check here for instructions.
- Arecibo P2284, B1534+12
- Setup: cima setups in /share/obs4/usr/pulsar/p2284/P2284set_430.gui (ASP-only, though wapp screen needed to activate cals) and P2284set_1400.gui (ASP plus wapp1; should migrate to using 3 wapps).
- Source catalog: flux.cat for 1442+10, psr.cat for 1534+12
- Make wapp polycos "at home", ASP polycos on aspmaster.
- Plan: continuum cals (1 deg N, then on-source) off/on 1442 first at 430, then at 1400, then 1534 at 1400, then 1534 at 430 MHz.
- ASP config files in aspmaster:/home/asp/config/P2284 . Good files are: config.xml.B1442+10_430_off , config.xml.B1442+10_430 , config.xml.B1442+10_1400_off , config.xml.B1442+10_1400 , config.xml.1534+12_1400.new , config.xml.1534+12_430_new . Edit these for cranky or dead slave nodes.
- Arecibo P2180, 1906+0746
- Setup: cima setup in /share/obs4/usr/pulsar/p2180/1906+07_Lwide_WAPP_asp.gui . This will put the L-band bandpass right-side-up going to ASP, unlike in most ASP observations.
- Source catalog: /share/obs4/usr/pulsar/p2180/P2180.cat . Position is 19:06:48.652 +07:46:29.29 . Do not use a mouldy old position from the P2099 directory!
- This will use 3 wapps in parallel with ASP.
- Make wapp polycos "at home", ASP polycos on aspmaster.
- ASP setup is a little special because we write one channel to disk while folding 7 in realtime, using a special "Joeri" version of the ASP code. Look in aspmaster:/home/asp/config/P2180 and subdirectories: two config files are required ("with" and "without") and ASP needs to be started in a manner similar to that in the "commands" script. Check with Joeri before any observation under this project number to make sure the node numbers are all up-to-date.
- Arecibo P2203, "Parkes binaries"
- Setup: cima setups in /share/obs4/usr/pulsar/p2203/p2203setup.gui . Probably need to change the directory for the wapp polyco files from p2017 to p2203.
- Source catalog: P2019.cat for most, psr.cat for 1953+29.
- Make wapp polycos "at home", ASP polycos on aspmaster.
- Plan: 1853+1303, 1910+1256, 1905+0400, then big slew to 1953+29. 1-min cal scan before each data scan, aiming for ~30 minutes per source. Cut the time for 1905 and/or 1953 if necessary.
- ASP config files in aspmaster:/home/asp/config/P2203 . Names are self-explanatory. Edit these for cranky or dead slave nodes.