This wiki page is a guide to use the Differential Image Motion Sensor Using Multiapertures (DIMSUM) instrumentation that was installed to Aux Tel between December 6th and 16th. I am currently writing this as a wiki page to be able to get feedback during the process of documentation. Once everything is finalized I can create a pdf file if requested.
...
- Check if the MacMini is on. It should always be on, but someone may have unplugged it. If it is off, either unplugged or just not lighting up, turn it on.
- Take the lens cap of the camera. Make sure
- Turn the camera on. When the camera is on, there should be options visible on the small screen on top of the camera (not the flash trigger).
- Turn the flash trigger on. That is the object that is attached on top of the camera. It has godox written in front of it. Turn on the bottom switch. The top should stay off.
- After this step you are done with this part of the setup. Go to the pegboard part of the setup.
- Turn the flash receiver on. It should be attached to the floor under the pegboard. Make sure it is on group 01 and channel A
- Turn the power to the flash lamp on. That is done by switching on the green extension chord.
- This concludes the hardware part of the preparation. Further steps are software related.
ssh into the MacMini using
Code Block ssh stubbslab@auxtel-macmini03.cp.lsst.org
This works if you are on the summit. You will need vpn access if you are not on the summit. It may be good to open up two windows because two separate processes will run.
- 106McKay!Ask someone for the password because this page is public.
Once logged in create a directory for the observing run. You can name the file as you wish but it would be easier to track if naming would be in the form of mm-dd-yy_observing_run, thus if someone were to take data on January 17th 2023, the naming would be 23-01-17_observing_run. The command to create the file in the external disk would be:
Code Block mkdir /Volumes/LaCie/yy-mm-dd_observing_run
Once the directory in the external disk is created go to code directory using the command:
Code Block cd /Users/stubbslab/documents/strobed_dimm/code
Step 12 has been updated. Old instructions are available at the bottom of this point.
Set up the storage managing code. The images taken by the camera are initially sent to the computers local disk, then when a certain number of pictures accumulate, they are transferred to the external disk. 50 images is an ideal number to buffer. The storage management is automatically done by the script python_bash_updated.py. Once you are in the code directory run:
Code Block python python_bash_updated.py ./ /Volumes/LaCie/yy-mm-dd_observing_run/ 50 CR2
This is the way it is used for the context of DIMSUM, but the meaning of arguments are as follows:
path/to/source (where the files are internally saved), path/to/target (the directory of the external disk), number of files to transfer after, file type.
Old description for step 12
Set up the storage managing code. The images taken by the camera are initially sent to the computers local disk, then when a certain number of pictures accumulate, they are transferred to the external disk. 50 images is an ideal number to buffer. The storage management is automatically done by the script python bash. Once you are in the code directory run:Code Block python python_bash.py ./ /Volumes/LaCie/yy-mm-dd_observing_run/ 50 CR2
This is the way it is used for the context of DIMSUM, but the meaning of arguments are as follows:
path/to/source (where the files are internally saved), path/to/target (the directory of the external disk), file type, number of files to transfer after.- This code is an infinite while loop, when the command above is run, you will need a new window to run the image taking commands. Thus once step 11 is completed you need to opene a new terminal window. You may have to log in again.
Once logged in, go to code directory again using:
Code Block cd /Users/stubbslab/documents/strobed_dimm/code
Run the command:
Code Block ./CommunicateTrial.sh
This will only take one burst, if you are taking data all night long you may want to modify the command using repeat N. A single burst approximately takes 11 seconds. For a 2 hour run repeat 900 would be enough. Repeat 900 command approximately takes 160 minutes to run. A single run approximately takes 11 seconds.
Code Block repeat 900 ./CommunicateTrial.sh
- Step 15 is updated. Old Version at the bottom of the new version. New version uses the script python_bash_updated.py and does not result in a problem of a single run with more than 10000 images.
Old Instructions for step 15
Run the command:Code Block ./CommunicateTrial.sh
This will only take one burst, if you are taking data all night long you may want to modify the command using repeat N. A single burst approximately takes 11 seconds. For a 2 hour run repeat 900 would be enough. The 900 is a maximum for now because it may result in more than 10000 photos. If that is the case, the camera may overwrite the previous files. Thus for a run longer than 2 hours, first run the command:
Code Block repeat 900 ./CommunicateTrial.sh
When this is finished:
stop the python_bash script running on the other window (just hit control c).
Transfer the remaining files to the external disk. There will be remaining files because the script transfers them 50 by 50. You can easily transfer all the remaining files by:
Code Block python python_bash.py ./ /Volumes/LaCie/yy-mm-dd_observing_run/ 1 CR2
The steps below this is if the run is still going on after the repeat 900 command finishes executing. If the AuxTel run finishes before the 900 repeat finishes, just end the process using control c and transfer the remaining CR2 files to external disk using the command above.
Create another folder in the external disk named mm-dd-yy_observing_run_2:
Code Block mkdir /Volumes/LaCie/yy-mm-dd_observing_run_2
In the window that is used to run the script python_bash, run python_bash slightly changing the parameters:
Code Block python python_bash.py ./ /Volumes/LaCie/yy-mm-dd_observing_run_2/ 50 CR2
- Repeat step 14 until end if more data is needed. If not, stop both of the scripts when the run ends.
Transfer the remaining files to the external disk. There will be remaining files because the script transfers them 50 by 50. You can easily transfer all the remaining files by:
Code Block python python_bash_updated.py ./ /Volumes/LaCie/yy-mm-dd_observing_run/ 1 CR2
- Once the run is finished and all of the scripts are stopped. Go back to AuxTel and:
- Turn off the power to flashlight
- Turn off the flash trigger receiver.
- Turn off the flash trigger.
- After everything is done and you have
...
- Check that all the steps through 1 to 7 are correctly done. That is making sure the flash trigger is on, flash receiver is on, the power to flash lamp is on.
- If the receiver or the trigger cannot be turned on or do not show anything when they are turned off and on, replace their batteries. Spare batteries were left on the table close to staircase.
...
- (batteries were replaced with AC adapter in Dec 2023)
- Sometimes, if the camera was operated via EOS utility on the computer, it can be left in Live Mode. That needs to be off (affects mirror lock) in order for flash to trigger.
Flashes are flashing but they are clearly out of sync with the camera. Pictures taken ended up being all black.
...
The SMA threads are 1/4-36. Digikey sells them, part number A104989-ND. Bought 100 of them.
Asked Helen to place this order since credit card failed.
Feb 14 2024 Pachon, Stubbs
Changed optical fiber configuration to have Xenon source feed attenuator module to 600- micron 1:4 fiber then 4 of 200 micron 1:4 bundles. This gives us 16 sources 200 microns in diameter.
I did find one fiber that was flopping/loose in the pegboard, that was presumably the one that showed anomalous behavior in analysis done Feb 2024 by Ali K.
Checked using headlamp that all fibers look comparable.
When running remotely on data collection using gphoto-2 to interrogate aperture worked fine, after power cycling the camera.
But the
./CommunicateTrial.sh
script failed.
...
1600 Boston time:
Kicked off a test collection with file xfer to LaCie.
Verified with image 5993 that it's working well!
Started a long term run
Feb 16 10:38 local Pachon time.
The cores of the images from Feb 14-morning of Feb 16 are saturated.
I installed an aluminum foil attenuator in one of the filter holders, replaced the Kapton tape that was in there.
This should reduce light level by at least a factor of two.
We can adjust f-number remotely to get desired light levels. Note high f-number makes us less sensitive to thermally driven focus changes.
We also have less aperture averaging and a smaller beam cone. Currently operating at f/20. With 220mm focal length our input aperture is 220/20 ~ 11mm.
I think this is better than operating at low f-number and cutting light level at the source.
Tasks:
OK to let it stop running over the weekend.
review performance and adjust f-number as needed to attain desired light level.
Imaged obtained after 1030 pachon time on Feb 16 should be OK to look at.
Image 6636 is most recent on disk, now. Wait for 6637 and verify light level. 6642 is a good test- grabbing that one.
Also now that we know how to extract dome shutter status from EFD (See image quality dashboard) we can better correlate that with measurements.
Well, that definitely cut light out! Peak is now around 600 ADU. So we should open up f-number more. Each f-stop is factor of two in flux.
So going from f/20 to f/10 should be 4x the light. Going from f/20 to f/5 should be 16x the light. That would give us around 9500 peak.
1214 Pachon time.
I used EOS utility to change lens to f/5, after halting current acquisition. We are getting perilously close to filling SD card on camera but I don't have time to change it and re-align.
Image 6694 should be at f/5.
As I think about this, I believe I changed camera focus and that's why things are so mushy now. Need to go up and do an autofocus with dome light cracked. That's also why flux peak is so low.
Rats.
1218 stopped data collection, going up to focus. Image 6708 should be good test frame.
Nope, no change.
Did a focus on camera as follows:
Set to "Av" for auto exposure. Set lens to AF
Partial button press to get a beep
change back to 'M' and MF.
lens is set to f/5. test image is 6710.
PSF looks good but it's misaligned. Drop back on zoom setting to around 150
redid focus as above, slid attenuator out to get more flux.
test image is 6711
OK this looks acceptable. Going with this config.
Seems fewer than 16 spots but outta time.
started another auto-run at image 6713.
christopherstubbs@Christophers-MacBook-Pro DIMSUM % dcraw -i -v CU2A6711.CR2
--------------
Filename: CU2A6711.CR2
Timestamp: Fri Feb 16 15:47:19 2024
Camera: Canon EOS 5D Mark IV
ISO speed: 100
Shutter: 1/10.4 sec
Aperture: f/5.0
Focal length: 188.0 mm
Embedded ICC profile: no
Number of raw images: 1
Thumb size: 6720 x 4480
Full size: 6880 x 4544
Image size: 6744 x 4502
Output size: 6744 x 4502
Raw colors: 3
Filter pattern: RG/GB
Daylight multipliers: 2.074708 0.932504 1.176114
Camera multipliers: 2040.000000 1024.000000 1456.000000 1024.000000
christopherstubbs@Christophers-MacBook-Pro DIMSUM %
-----------------
