Calibration question... calling all gurus

[bianry]

Hi.
This is most likely a FAQ but I can not find the answer. So please, someone....
A short recap: I do not have a scope so I use the iTelescope.net rent-a-scope system. This image of NGC5033 was taken with this equipment. http://www.itelescope.net/telescope-t21

I attach a screenshot with four images. The master flat (made by BPP-script), an integrated image made with the masters from iTelescope but calibrated with PI, an integrated image made from frames calibrated by iTelescope and a frame integrated from frames calibrated with PI-made masters.
From upper left:
1. Flat frame
2. Calibrated with PI with iTelescope masters
3. Calibrated by iTelescope
4. Calibrated with PI, with masters made with PI

To my untrained eye it looks like the calibration is off. In upper right it looks like the vignetting from the master flat is still present. Lower right is flatter but there are streaks in picture. Same streaks can be found on the master flat. Lower left picture looks flattest except the left and upper bright stars where the sky is brighter.

I can of course try to correct this with DBE but I feel I have not made an optimal calibration yet. And it bugs me.

All files can be found at the link below.
https://www.dropbox.com/sh/u4y8lzb01lxo1qw/TxOWnfBSZ0

I promise a beer to anyone who can shed some light on this. I have sent a request to PI support but so far no reply.

Best regards

Mats

[jkmorse]

Mats,

How did you process the Flats?  Harry and others have pointed out that using Darks to process flats may create problems and it is best to just use bias frames instead.  I have found this to be true since I was suffering the same problem you are experiencing but once I switched over to processing my flats with just a master bias, the problem went away.

Worth a try,

Jim

[bianry]

I fed everything in to the BPP script and pressed Run. I disabled the Optimize checkbox so I would not get the 'Warning: No correlation....' error.

I will try to do it without the dark. But the darks contain a 'lot' of  hot pixels that will not be removed then. But I supposes Cosmetic Correction script can take care of that.

regards

Mats

[bianry]

Mats,

How did you process the Flats?  Harry and others have pointed out that using Darks to process flats may create problems and it is best to just use bias frames instead.  I have found this to be true since I was suffering the same problem you are experiencing but once I switched over to processing my flats with just a master bias, the problem went away.

Worth a try,

Jim

Nope, I tried a run without darks and got the same result. There are dark streaks in the integrated picture.

r

mats

[Phil Leigh]

OK so the light frames are faulty. They suffer from bad dark streaking that NO amount of calibration can fix.
If you just look carefully at a single light frame you can see that the data is bad.
Likewise, the bias and dark frames have strange artefacts at the left hand edge which I would not want to see from my camera.. and the flats are a complete mess.

Something is badly wrong with the raw data.

Attached is an upper right quadrant zoom into an untouched flat from your dropbox. STF has been applied.

[bianry]

Thank you Phil.
Here is the beer as promised
http://beeroverip.org/pilsner-urquell/

It never occurred to me the enlarge the flats like that. And I would not even have known if the pattern was supposed to be there or not.

I have contacted iTelescope on this issue. Will report back.

Best regards

Mats

[Phil Leigh]

Thanks for the beer but I'm puzzled... do you have the bias, flat and dark masters produced by Itelescopes?

[bianry]

Thanks for the beer but I'm puzzled... do you have the bias, flat and dark masters produced by Itelescopes?

I added them to the Dropbox folder.

2014-01-02-T21-Master-Flat-Bin1-Luminance
2014-02-03-T21-Master-Bias-Bin1
2014-02-03-T21-Master-Dark-Bin1-180

r

Mats

[bianry]

I checked an other flat from an other telescope/camera and it shows the exact same pattern. This must be taken with something covering the lens that gives this pattern.
I suppose this pattern screws up the calibration?

r

mats

[Phil Leigh]

Thanks for the beer but I'm puzzled... do you have the bias, flat and dark masters produced by Itelescopes?

I added them to the Dropbox folder.

2014-01-02-T21-Master-Flat-Bin1-Luminance
2014-02-03-T21-Master-Bias-Bin1
2014-02-03-T21-Master-Dark-Bin1-180

r

Mats

OK - thanks...

[bianry]

I understand there are two type of flats. Dome flats and sky flats. Dome flats is taken with something covering the lens (a t-shirt seems to be classical)
Does this not by default introduce a pattern on the flat? If enlarged enough? Could some kind soul show me a sample of a good flat. Maybe even let me look at one in PI.

Fascinating this new learning is....

r

Mats

[bitli]

No, the flat are taken totally out of focus (your tee-shirt is somewhat nearer than the nearest star), so there should be no pattern from the the tee-shirt, dome, etc....
Sorry I cannot look at the files from where I am now.
-- bitli

[Phil Leigh]

I checked an other flat from an other telescope/camera and it shows the exact same pattern. This must be taken with something covering the lens that gives this pattern.
I suppose this pattern screws up the calibration?

r

mats

Well, I've never seen a flat quite like that! When I first looked at it I thought the sensor was covered in thousands of dust motes...

I'd like to hear some other folks opinions. I don't think that flat  would work properly - and I don't think the dark patches on the lights are good either...

Seems very odd - I thought iTelescopes were supposed to keep the imaging systems in great condition?

[bianry]

No, the flat are taken totally out of focus (your tee-shirt is somewhat nearer than the nearest star), so there should be no pattern from the the tee-shirt, dome, etc....
Sorry I cannot look at the files from where I am now.
-- bitli

This page also talks about Dome flats
http://www.astro-wise.org/portal/howtos/man_howto_flat/man_howto_flat.shtml

I think these flats are made like this.
Still waiting for a reply from iTelescope. Night in Australia now

r

Mats

[chris.bailey]

There is something amiss with the master calibration files. I have just tried manually calibrating a single light frame with a single flat, a single bias and and a single dark and it all works. When you introduce the masters it all goes pear shaped. If I look at the FITS header of the calibration files they seem to have a Pedestal value of -100, whereas the lights don't, and neither do the single calibration files. Trying to calibrate files with different Pedestal regimes is likely to be the culprit.

They may not look pretty but the flats do remove the streaking in the lights.

Hope that helps.