OSELOTS Troubleshooting

On this page, we document errors/problems we've encountered working with OSELOTS.

We encourage other users to write down/document other issues as they come up, and the solutions they find.

PROBLEM: When acquiring images, code fails with an error like:

Traceback (most recent call last):
    File "/home/sashab/Documents/sashas_python_scripts/pixis/ConvertPIXISRawToFits.py", line 143, in <module>
        convertRawToFits(source_file, target_file, source_dir = source_dir, target_dir = target_dir, target_suffix = target_suffix, header_elems_to_add = additional_header_elems); 
    File "/home/sashab/Documents/sashas_python_scripts/pixis/ConvertPIXISRawToFits.py", line 49, in convertRawToFits
        flat_img_arrays = [struct.unpack(img_fmt, data_set) for data_set in separated_data]
    File "/home/sashab/Documents/sashas_python_scripts/pixis/ConvertPIXISRawToFits.py", line 49, in <listcomp>
        flat_img_arrays = [struct.unpack(img_fmt, data_set) for data_set in separated_data]
struct.error: unpack requires a buffer of 2097152 bytes

OTHER SIGNS OF ERROR:

Identified instrument is a Demo:
- (Something like): PIXIS: 100B (SN:12345:Demo) [E2V 1340 x 100 (CCD 36)(B)]
Temperature set point is -40 degrees C:
- (Something Like): Temperature is -40 degrees C

SOURCE OF PROBLEM:

The computer cannot connect to the camera, and is therefore generating data off of the 'Demo' PIXIS instrument. This artificial image does not conform to the format expected by the code that converts the .raw image to a .fits image, and the code fails with the above error.

POSSIBLE SOLUTION(S):

Make sure that camera AC-DC converter (black, labeled PIXIS Camera) is turned on (switch is flipped to | rather than 0).
- Power strip has a green light to indicate it is turned on, though it is covered by black aluminum tape. You can peel back the tape to check, if you like.
Make sure USB connecting computer to camera is plugged in at both ends (perhaps give the ends an extra push to make sure they are connected).
Make sure camera power cord is connected to the camera.

PROBLEM: When attempting to move stage, system throws an error like:

Traceback (most recent call last):
    File "/home/sashab/Documents/sashas_python_scripts/pixis/moveStageToFocusPosition.py", line 22, in <module>
        zaber_stage.home( )
    File "/home/sashab/anaconda3/lib/python3.7/site-packages/zaber_motion/binary/device.py", line 323, in home
        call("binary/device/home", request, response)
    File "/home/sashab/anaconda3/lib/python3.7/site-packages/zaber_motion/call.py", line 46, in call
        process_response(response_buffers, response_data)
    File "/home/sashab/anaconda3/lib/python3.7/site-packages/zaber_motion/call.py", line 122, in process_response
        raise convert_exception(response_proto.error_type, response_proto.error_message, response_buffers[1])
zaber_motion.exceptions.binary_command_failed_exception.BinaryCommandFailedException: BinaryCommandFailedException: Command 1 with data 0 resulted in error code 1.

OTHER SIGNS OF ERROR:

Zaber translation stage does not move, so focus does not change even when different focus values are given (with the -f command line option)

SOURCE OF PROBLEM:

The computer is having some issue passing commands to the translation stage. It can connect to the stage, but the stage will not accept commands.

POSSIBLE SOLUTION(S):

Either the connection between the stage and the computer needs to be reset or the stage is otherwise occupied

To reset the stage connection, try the following (in sequence):

Disconnect and reconnect the USB, pausing for about 10 seconds before reconnecting
1. Grey USB cord, at time of this writing
2. Did this solve the problem? If so, stop here. If not, next step.
Disconnect and reconnect the stage power, pausing for about 10 seconds before reconnecting
1. Either look for AC-DC converter labeled 'zaber' on the powerstrip
2. Or open up OSELOTS and unplug from stage gearbox (center of spectrograph base)
3. Did this solve the problem? If so, stop here. If not, next step.
Reboot control computer.

If none of the above work, the stage may be under some other motion control. Most likely, the manual knob to move the stage is not perfectly centered

Open OSELOTS and locate the knob on the zaber translation stage gearbox
Turn the knob to move the stage to approximately halfway along its control axis
Center the knob until the stage is still
1. There is a slight catch at the knob center position, but it is very subtle
2. The stage can appear still, but in fact still be slightly engaged. Watch the stage for at least 30 seconds to make sure it is truly still
Now that stage is not moving due to its internal motion control, it should be controllable over USB once again.

PROBLEM: Translation stage slipping

OTHER SIGNS OF ERROR:

Two images taken with same focus position do not show same focus results
Two images with similar but not identical focus positions show focus results that appear identical
If the stage is watched while moving, wormgear rotates, but stage does not move smoothly or does not move at all
- Occurs most obviously when the stage is moving AWAY from home, rather than towards home

SOURCE OF PROBLEM:

The internal gears of the translation stage have lost good contact. This often happens after OSELOTS has been transported (vibrations loosening the screw connections)

POSSIBLE SOLUTION(S):

Thankfully, a typically easy fix.

Find the smallest hex key in the OSELOTS toolkit.
Locate the small hex-hole on the top of the stage (the threaded part on which the focusing lens is mounted)
Tighten this screw until it is firm (over-tightening should not be necessary)
Repeat stage motion test, and verify that stage moves smoothly to and (especially) away from home.

PROBLEM: Acquired images show very little or no light, particularly noticeable with calibration light sources:

Here is an example of such an image (right panel of Figure below), taken with the HG-2 calibration light source. Notice how the lines in the right panel are very ragged; not the clean image of the fiber tip, as seen through the slit, that we hope for (left panel):

OTHER SIGNS OF ERROR:

Images taken of the sky show little or no light, although all connections are good and the lens is on sky.

If you disconnect the cage system joining the slit and fiber to the grating, it is possible to pear through the slit and look for light from the fiber tip behind. When the fiber is illuminated by a bright enough source (HG-2 light, for example, one can see the light from the fiber tip shining through. UNDER GOOD CONDITIONS, the data should look like this (note the bright yellow-white light shining through the slit):

If the system is not working, the light shining through the slit will be either non-existent, very faint, and/or ragged.

SOURCE OF PROBLEM:

The multimode optical fiber and the 5 um slit have slipped out of alignment. OSELOTS requires manual realignment.

POSSIBLE SOLUTION(S):

Be warned, this is a ~1 hour process. We are working on designing a more user-friendly repair process. At no point in this process should the fiber be removed from Here are the steps (somewhat difficult to describe in text):

Remove the rubber baffle around the cage system connecting slit to central column
1. Unwraps by hand
2. Not shown in photo below
Use a small, imperial hex key to just loosen (DO NOT REMOVE ALL THE WAY) the small screws locking the cage into the central column
1. In the hex key set provided with the OSELOTS toolkit, the key you need is the small, red imperial key
Disconnect the slit holder from the support post, just to the right of the entrance hole
1. We recommend unscrewing the small, horizontal, silver beam from the holder of the optical slit, keeping the larger poster pieces in place
2. You should first loosen the screw locking this silver post in place, and then rotate it
Slide the cage posts holding the slit and fiber tip out the front hole of OSELOTS
1. If all the screws have been appropriately loosened, the whole front of the cage system should cleanly slide out.
Connect a light source (like the HG-2 calibration lamp) to lens-end of the optical fiber
1. See the section on taking wavelength calibration images, starting with "(OPTIONAL, BUT SHOULD DO IF POSSIBLE)", on the Nightly Observing Sequence page for guidance on how to do so
When the slit and fiber are well aligned, you should see bright light shining through the slit from the fiber
To adjust the alignment, you must gently move the slit around in it's holding tube
1. (The slit, unfortunately, does not sit perfectly centered in the tube; it is held in place (imperfectly) with a ThorLabs 1" tube ring)
2. Slide the slit holder off of the cage struts
3. Loosen the ring holding the slit in place
  1. Can be done with a flathead screwdriver or the ThorLabs tube ring adjuster
4. Gently nudge the ring's position perpendicular to the long axis
  1. Best done with a soft, long object with a small tip, such as a toothpick or a cable tie
5. tighten ring, gently
6. slide slit holder back down cage struts until the fiber and slit are flush
  1. Bring them into contact gently
7. Do you now see light from the fiber through the lens (as in the image above) when slit and fiber are flush?
  1. If yes, then you're good
  2. If not, repeat steps c. through g., using the results from the previous adjustment to inform your next adjustment
Once good alignment is retrieved, replace the cage struts and tighten system down
1. i.e. - reverse steps 2. through 4.
Take an image with OSELOTS:
1. As su in the ~/Documents/OSELOTS/ directory, run:
2. ```
bash doPixisImaging.bash -e 5 -o HG2 -p HG2_f23p0 -n 1 -h 0 -f 23.0
```
3. (you might need to cool the camera and find focus home if this is the first image taken today):
  1. ```
  bash doPixisImaging.bash -e 0 -o Cool -n 1 -h 1 
```
If the lines look good, the slit and fiber have been successfully realigned!
But the slit is likely slightly rotated out of position. Are the lines perfectly vertical in the image, or is there a slight tilt?
1. ds9's crosshair is helpful
2. Here is an example (note in the left panel, how the crosshair traces the line, while in the right panel, the line drifts right off of the crosshair):
If the slit requires rotation correction:
1. Use a hex key to loosen the rotation stage for the slit
2. Rotate the slit slightly, making note of the direction
3. Take another image and check the slit orientation
4. Repeat until good vertical slit orientation is achieved.
The slit-fiber coupling error has now been fixed.