Skip to main content

W30 reset

We recently released a significant update to the backend systems for the AAVSO website. While most of the bugs introduced by this update have been fixed, there may still be problems we haven't fixed. If you run into a problem, please email webmaster@aavso.org
GFB's picture
GFB
Offline
Joined: 2010-07-08

I found a script with some code that was causing the first image in a series to not be centered correctly.  The script has been replaced.  Somewhere along the way, maybe because of the script, the pointing model got polluted with pointing errors.  I've replaced the model and pointing has returned to normal as of last night.  Looks like a good run.

 

Bill Goff

Re: W30 reset
mrv's picture
mrv
Offline
Joined: 2010-12-21

Bill,


I had noticed that with images from a week ago, the first frame of each target was off center.  However, images from the last two nights are perfect.  Thanks for fixing the script.

Bob

W30 reset
GFB's picture
GFB
Offline
Joined: 2010-07-08

Thanks Bob,

I'm glad the images are good and you can make use of them.

 

Bill Goff

Re: W30
mrv's picture
mrv
Offline
Joined: 2010-12-21

Bill,

I was wondering how often master darks are created for W30?  I've noticed that many of my images from the last two weeks have a significant number of residual light and dark pixels in them.  If these pixels are near the variable or comp star, it can cause problems for the photometry.


Bob

W30 darks...
GFB's picture
GFB
Offline
Joined: 2010-07-08

Hi Bob,

It's ironic that you mention this, I had noticed it with my personal scope on some short exposures lately.

 

We did a new set of darks June 9th.  The nighttime tempertures were getting high enough that the camera wouldn't cool to -20 reliably so we had to drop back to -10.  We shot a battery of darks at various times, 1, 3, 10, 30, 60, 120, 240 and 300 seconds.  The calibration is done by a pipeline on the Hq server so I'm not sure how that is set up; how the dark master are chosen for the processing.  It might be that if the columns are appearing on one set of frames at one exposure time and not another, the pipeline might not be selecting the correct dark master.   Did they appear at one exposure time and not others?  You may not have a way of knowing if the frames you are seeing are all the same exposure.

 

I'll hope that Matt or Arne will see this and comment if a change is needed.

 

Bill

Re: W30 darks
mrv's picture
mrv
Offline
Joined: 2010-12-21

Bill,

I've seen the light and dark pixels on all frames.  Exposures range from 120 to 180 seconds.  Might be a dark frame scaling issue, or the wrong dark being applied, or perhaps its time for some new darks.  When I would use my ST-7XME, I found that I could use my master darks for about two weeks.  Beyond this point, residual light and dark pixels would start to show up in my calibrated images.

Bob

W30 darks...
GFB's picture
GFB
Offline
Joined: 2010-07-08

Hi Bob,

I wanted to check with Matt first about running darks to make sure non-science frames are handled by the pipeline.  I know he's made several adjustments to it lately.  I'm going to run a new batch of them this evening.

 

Bill Goff

dark frames
HQA's picture
HQA
Offline
Joined: 2010-05-10

The AAVSOnet pipeline scales the mean dark level for the appropriate exposure time.  We interpolate between several dark frame exposure lengths; in Bob's case, between darks at 100sec and 300 sec. What often happens is that the hot pixels are not completely eliminated by this procedure, as a very small temperature difference between the darks and the science frame will manifest itself as a small residual value for hot pixels (which are more sensitive to temperature).  Some software programs scale the dark frame so as to subtract the hot pixels, rather than a purely exposure time scaling, and they will look visually better.  Some programs also truncate negative pixels, which are possible after a dark subtraction on a long exposure image, so again their images will look visually better.

Bill, go ahead and take new darks, but wait for a lousy night (or full moon).  New darks are not worth losing science imaging time.  If these work better for Bob, then we should perhaps take darks on a more regular basis.

Arne

dark frames...
GFB's picture
GFB
Offline
Joined: 2010-07-08

Hi Arne,

Ok, thanks for the clarification.  Matt and I decided to go ahead with new ones which I did two nights ago.  He had to alter his scripting some on the Hq end as file names are just a little different as they are saved.

 

Bill Goff 

Re: W30 darks
mrv's picture
mrv
Offline
Joined: 2010-12-21

Bill & Matt,

Thanks for the new darks and the extra processing.  I see that my images from earlier in the week were re-calibrated.  I will remeasure these images.  I did a quick comparison of an image set from 8/6.  The new images have somewhat fewer residual dark and light pixels.  The resulting measurement of the variable also has a somewhat lower error as well.  However, there are still thousands of residual light and dark pixels present in these images.

Arne,

I have a suggestion for further reducing or eliminating these light and dark pixels.  You mentioned that 100sec and 300 sec darks are used for W30.  My exposures for CI Aql and X Ser are 120 and 180 seconds, respectively.  What if we make them all 100 seconds to match one of the darks?  We could increase the number of exposures for each target to get the same equivalent exposure time as we have now.  I'm suggesting 100 seconds because I'm not sure if W30 can track well enough for 300 seconds (many of my 180 second exposures have been slightly trailed).

Bob

W30 darks...
GFB's picture
GFB
Offline
Joined: 2010-07-08

Hi Bob,

I'm not sure where to go next with this, I guess the warmer operating temperture might be have some effect so that the frames aren't as uniform even with the new darks.  They says there's a doubling of noise for each 7 degrees of change and we're working at 10 degrees warmer right now, which effects both darks and science frames.  W30 has a water pump I could work with, but I'd found with other SBIG cameras that it only gains about 5 degrees and the inconvenice goes way up.  Arne, if you'd like me to work with that I'll give it a try.  I did not check this pump at all, the biggest problem is 'head height', the pump has to have enough grunt to raise the water column from floor level to the camera height.  I don't know how this one was used before.

 

Tracking does seem not completely reliable.  I've seen some series at 120 or 180 seconds where 3 of 4 look fine. What happened to the fourth?  Just can't say.  The drives on the Meade mount are very touchy as I learned on the declination.  I haven't touched the RA as there's so much more to disassemble to get to it.  That might be a winter, bad weather project.  We might consider taking some 240 or 300 second frames and see what we get.  I haven't seen any that long yet.

 

Bill Goff

Re: W30 darks
mrv's picture
mrv
Offline
Joined: 2010-12-21

Bill,

I wouldn't bother with the water cooling.  Too little gain for the effort.  Lets try matching the exposure time of my science frames to whatever exposure time you're using for darks (ie. 100 seconds).

Bob

Re: W30 darks
mrv's picture
mrv
Offline
Joined: 2010-12-21

Arne,

Thanks for changing my exposure times for X Ser and CI Aql to 100 seconds.  Unfortunately, the residual light and dark pixels remain in the newest images (8/11).  This suggests that the cause of the problem may be somewhere in the calibration pipeline.  Perhaps someone at HQ could look into this and see if there is a bug or some other problem with the dark frame subtraction for W30.


I did receive some images from OC61 a few days ago and they look fine.  No significant numbers of residual light or dark pixels.  So this problem seems to be only with W30.

Bob

W30 darks
HQA's picture
HQA
Offline
Joined: 2010-05-10

Hi Bob,

I went through the processing steps manually, subtracting the master bias and the bias-subtracted master 100sec dark from one of the X Ser V frames, then flatfielding it.  I get the identical results that the pipeline obtained, so there is no problem with the pipeline.  What I do see is that this particular chip is pretty noisy, with a significant population of pixels that are 5 sigma from the mean dark level.  I'm currently checking with another site that has an STL-1001E, and seeing what their chip looks like.  I'll get back after I know more.  However, in the meantime, the reductions are correct and you just have to deal with more hot/cold pixels than with other cameras, and that is compounded by the fact that we are running the camera at -10C for the summer.

Arne

Re: W30 darks
mrv's picture
mrv
Offline
Joined: 2010-12-21

Arne,


Thanks for checking the processing pipeline for W30. I agree with your assessment that the chip in the W30 camera has rather noisy dark current.  In addition to the more obvious light and dark pixels,  I can see a faint mottled pattern that lies between the light and dark pixels and is identical from frame to frame in a set of images.

I have some ways of working around this problem,  but it takes more time per image set.

Bob

W30 darks
GFB's picture
GFB
Offline
Joined: 2010-07-08

Arne and Bob,

Thanks for checking this.  I wonder now if it had been noticed in any of the -20 frames from the spring.

 

Bill

W30 darks
HQA's picture
HQA
Offline
Joined: 2010-05-10

Hi Bill,

I had checked earlier W30 data (including the -20), as well as when the camera was on Sonoita's C14.  It always has been noisy, and the bias/dark look pretty ratty to me in comparison to other Kodak chips.  The Clay Center has a similar STL-1001E, and I'm having them run some biases and darks for me with their camera to see if we get something similar, or whether this is just a poor sensor.

Arne

Re: W30 darks
mrv's picture
mrv
Offline
Joined: 2010-12-21

Arne,

Is the STL-1001E currently used on W30 the same camera that was used on K35 last year?  Images I received last year from K35 had the same type of noise. 

Bob

W30 == SRO35 == K35
HQA's picture
HQA
Offline
Joined: 2010-05-10

Hi Bob,

The STL-1001E camera on W30 is the same camera that was on K35 and SRO35.  As far as I can tell, it always was noisy.  That is why I'm checking with another site with a different STL-1001E to see how it performed, to understand whether this is a a generic feature of the KAF-1001E or whether ours is particularly bad.

We have a couple of other cameras that could be of interest, but replacing cameras is lower priority right now than getting additional sites commissioned.

Arne

AAVSO 49 Bay State Rd. Cambridge, MA 02138 aavso@aavso.org 617-354-0484