Maxim won't align images.

[Tandum]

First off, here's a link to a series of maxim tutorials on utube (http://www.youtube.com/watch?v=3i-WZ8IPxvc&feature=related)if anyone is interested. I haven't watched them all so can not comment on quality but something is better than nothing ;)

Now, I've found my last couple of targets are suffering from bad flats and I don't know why.

I capture everything in Maxim and use the stacking tool to stick it all together. I'm talking narrowband images here, that's what I normally shoot from home. I have a dark library of 30minute darks and bias files at -25C for summer on the disk and shoot new flats when I change scopes or clean something. Maxim scales the darks to whatever exposure time I use.

Lately I've found that my flats alter the image files in a way that stops the maxim alignment process from working correctly. I always do an Astrometric alignment and have never ever had a problem with it until now. With a tip from the maxim group, to figure out what was happening, I got one image on screen, pinpointed it and it showed 2234 stars in the image and it resolved. I then clicked calibrate and pinpointed again. It now said 9 stars in the image and would not resolve. On screen it all looked ok. The flat looked ok as did the numbers, a bell peaking at about 20K, The calibrated image looked ok to the eye, yet it would not resolve.

Tonight has been clear for a few hours and as I changed scopes a few days ago, I shot off a few Ha tarantula subs before the clouds came back and shot off a set of flats for all filters and binning modes with this setup to make sure it's all going. The subs won't align using the flats I took.

This happened to me with thor's helmet last week and I had to reduced the bell curve on the flats by 50% to get it to calibrate, but I don't understand why this is happening all of a sudden. I tried jumping back several versions of maxim but got the same result so it's in my method not the software. I've ended up stacking the files uncalibrated and then calibrate the result?

I've attached some files so you can see they look normal. I've done a ddp on them all. Anyone got any clue?

Here's a master flat, stacked with no cal and stacked with a post cal.

[RickS]

Very weird, Robin. I used to do astrometric alignments in Maxim and never had any problems like that. These days I use PI which works brilliantly, of course :D

If you can put them up on your web site temporarily I'd be happy to have a play with the master calibration files and a light frame to see if I can figure out what is going on. I seem to spend a lot of time poking around in FITS files these days.

Cheers,
Rick.

[Karls48]

I had same problem about year ago after taking set of new flats. Never figured out as what happen. PinPoint would not solve with only few stars available from hundreds on the image. Fixed it by deleting all calibration files - Bias, Darks and Flats and started calibration process from the scratch. I'm not saying that this is best solution, but it worked for me.

[RickS]

Robin,

Have you tried doing Pinpoint solves on all the image files and saving them before you calibrate? IIRC, Astrometric alignment won't do a new plate solve if you've previously done one (and the FITS headers have been added to reflect this). This might be a workaround at least. You can tell Maxim to automatically do a plate solve for each image as it is acquired.

Cheers,
Rick.

[Tandum]

Here's some files (http://starshed.celticcomputers.com.au/temp/FitFiles.zip)if you wanna look Rick. There's 5 files totaling about 65meg in that zip.

I'm wondering if the bias point has shifted in the camera. I might need to redo the cal files as Karl suggests. The master bias file shows a low end of 2155, the darks show 782 while the lights show a bottom end of 1750. Sounds a bit odd. Definitely have to wait till tonight for a look, it's really bloody hot here today.

[RickS]

Got 'em, Robin. I'll have a poke around tonight.

[RickS]

Robin,

I was able to reproduce what you were seeing with Maxim. If I don't calibrate the subs then they plate solve fine. After calibration, fewer image stars are found and the plate solve fails.

If I run the full version of Visual Pinpoint then it will solve both uncalibrated and calibrated files, but it does find fewer image stars in the calibrated files (e.g. 10485 uncalibrated, 8510 after callibration). The solve claims the FWHM in the uncalibrated file is 5.14 arcsec vs 4.88 arcsec in the calibrated file.

I can't see anything obviously wrong with the calibrated files. Most of the hot pixels are gone as is the vignetting, but that's expected :) The overscan area looks pretty funky after calibration but cropping it off didn't make any difference.

That's as far as I've got so far. Something appears to be upsetting Pinpoint's star detection algorithm but I don't have any clues what it is yet.

I'll have a bit more of a play if I get a chance... Did you try new calibration files yet?

Cheers,
Rick.

[Tandum]

Cheers Rick, It's been a busy day, didn't get the camera cooling till after 8pm, when I got back to it just after 9pm it was sitting on -24C @ 98% cooling :D It normally sits around 75% so it's frigging hot here and I don't want to leave the slot open to let cool air in, it sort of looks like rain :sadeyes:

I've knocked it back to -20C @ 78% cooling and will capture a set of cal files there I guess. I believe we are in for another week of temps in the high 30s low 40s ...

I note the bottom end of these 30min darks are very close to the bottom end of the bias shots this time round :shrug:

[Tandum]

Just to finish this off, it has finally cooled off enough to get a fresh set of cal files @ -25C and all these alignment problems have gone away.
The old set where under 3 months old and I haven't changed anything here so something must have changed in the camera.
Hopefully for the better :question:

[RickS]

Robin,

The bias values in the old master bias are significantly higher than in the new master bias (see mean and median values below). Also, the old master bias is quantized (see histograms).

I don't have an explanation for the difference. If you still have the original bias frames then it might be worth taking a look at them to see if there is anything odd about them. I wonder if your original master bias got overwritten somehow and that's when your problems started happening?

Cheers,
Rick.

[Bassnut]

What do the flats look like now?. The old ones look pretty extreme.

[Tandum]

The timestamps on the masters are all the same Rick. Who knows what happened but it's fixed now :)

That's just the stretch Fred, the histogram for them look OK.