Flats - what are others doing?

[rcheshire]

A quick question about producing flats - I understand their purpose etc...

...briefly, what are others doing to produce flats? Light boxes, T-Shirts or other methods. Is one better than the other and do you take a fresh set each session or keep a library. Appreciate your thoughts...:question:

[cybereye]

I second Rowland's question!! :)

[pvelez]

Rowland

I used to image with a moded Canon 350D. I didn't bother with flats - darks were far more important.

I traded up to a SBIG ST8300 and have been able to take much longer exposures. My previous OTA was a schmidt-newtonian and there was a fair bit of vignetting so I had to move on to flats.

I picked up an EL screen - it was reasonably even but a bit bright. I also found that it was a bit too bright for some filters - especially when binning.

I picked up a lightbox from Exsfo - Peter kindly added a pot so I could vary the brightness. It was still a struggle to get the brightness down - with a bright source, the exposures are short and this carries with it a few issues.

So lately, if I have the opportunity I take twilight flats. That way I can take longer exposures and so eliminate any shutter artefacts.

What suits you depends on your set-up. If you have a DSLR, a lightbox is definitely the go. You can take the flats before or after your imaging run - I often take them while I pack up everything else.

To be most effective, you will need to take them on the night - while your optics and imager are in the same configuration. Vignetting won't change much from night to night but if your flats are to remove dust donuts, the camera and optics must be lined up in the same way as when you image. Otherwise you will find that you are removing donuts that aren't there - I know this from bitter experience.

Hope that helps

Pete

[multiweb]

Flats are easy to do. There are a number of methods. I use an EL panel for convenience. They need to be taken each session with the same imaging train you took your light frames with. (i.e. same filters, same focus, same camera orientation). You'll have to work out the exposure time that works best with you scope/camera combo and experiment. The improvement flats do when calibrating raw files is well worth the effort.

[batema]

When I used my DSLR i had a light box and set my Canon 400d to AV and plonked the light box on top and shot of 20 flats. Seemed to be OK. Then came the QHY-9 and I have no idea what I am doing. Two nights ago I was shooting through my W/O 70 and took by flats by pointing the scope at my other laptop and using VIRTUAL LIGHT BOX (google it) but for me to get the histogram between 1/3 and half way up from the left hand side I was doing 8 second shots which seems rediculous. If any of this helps I hope so and if any one can tell me if 8 seconds is rediculous that would be good too.

Mark

[rcheshire]

Many thanks so far - a little bit of research on EL panels found these - :question:

Simple little device - fits over dewshield.
http://dasdaveadshead.blogspot.com/2009/12/flat-light-with-electroluminescent.html

Cuttable EL sheet.
http://www.elec2go.com.au/category7_1.htm

[TheDecepticon]

I bought a light box from Exfso also. Great bloke, great light boxes.

I used to have a QHY8, so taking flats was easy by just plonking it on the top of the scope and setting the exposures to around 0.07sec and then taking 50 of those.

With the QHY9M and filters that I have now, it has a minimum shutter speed of around 0.5 sec, otherwise the shutter can be seen in the flats if the exposure is shorter. To alleviate this, I installed a pot I had laying around and dimmed it down to very low and take the flats over 2 - 3 seconds.

The results are the same, just the brightness and the exposure length is different.

For me, the light box has always worked the best, the results are very repeatable and you don't have to deal with changing light conditions when trying to take sky flats. There is software out there that can help with sky flats by adjusting the exposures to achieve the correct ADU count per sub frame.

You want to acheive a nice flat frame of up to half the ADU count of your camera, but you will need to experiment with this. The QHY8 produced the best flats at around 10,000 -11,000 and the QHY9M gives better flats at 20,000 - 30,000, with my set up.

There is an article here on the site about flats if you haven't read it.

http://www.iceinspace.com.au/63-211-0-0-1-0.html

Flats really can be one of those mystery things with astro photography, but once you "get it", you'll wonder what all the fuss was about. You also need to find out what works for you.:hi:

[Astrobserver99]

With my DSLR, I rarely used flats, except where there was likely to be a lot of vignetting...dust motes rarely seemed to be a problem. With the QHY8, dust motes show up much more regularly. A home made light box with an inline dimmer switch purchased off ebay works best for me. I take the flats directly after imaging by putting the light box on top of the scope and adjusting the dimmer to 1/3. Usually take exposures of 2-3s for a background value of 9000.

[Alchemy]

T shirt, early morn before sunrise........ Has worked for me for years, can't see me changing method now...... Only do it occasionally, particularlarly if I open the ccd so different dust can get in, if it stays closed theoretically ( and in practical use) no change happens.

[rcheshire]

That's a good point. With DSLR's the sensor os frequently exposed to dust, so flats should probably be taken at every shoot.



Now I know what I'm seeing on my raws - the dark spots except for one have moved with use and are probably dust.

[Terry B]

It depends on your purpose as to whether flats are important.
I do photometry so flats are very important regardless of the exposure time. My camera remains permanently on the scope so I don't do flats every time but reuse them.
I take them on overcast days in my observatory with the roof in place, The scope aims at a white board on the wall and hopefully the diffused light on an overcast day gives me a pretty even illumination. For R and I (infrared) filters I still need to put a T shirt over the front of the scope to get the exposure to be long enough.
You should take separate exposures through each filter if you are using filters and at different binning if you need to. I tried not doing this but if you bin the flats in the puter afterwards the gain data is incorrect in the fits header.
Try not to have too short an exposure as the shutter can give uneven exposure. 5 secs seems a good minimum.
If your setup isn't permanent then you will need to do them each time. The T shirt works well but you must make sure that the T shirt is evenly illuminated itself and that it is aiming at an evenly lit surface. The focus position must also be the same as when you are imaging.
Hopefully this is helpful.
Cheers

[rcheshire]

Thanks Terry - very helpful. My setup is temporary, so it's flats each time.

Because I have to time sessions carefully around sleeping time for the shift worker in our home, I need to plan a work flow that optimises every moment on clear nights.

Darks and bias frames are not such a big deal - a library will usually do, I think. Flats require some preparation, and because the camera will possibly be rotated to frame an object, flats should really be taken following a session.

I'm thinking that an EL panel is probably the best solution for me:question:

[higginsdj]

I use sky flats. My observatory uses ACP and ACP has dark and flats acquisition scripts that take all the guess work out - it's automated and the results are just fine from my precision photometric work.

Basically the script defines what types of flats you want (ie filters) and how many flat images for each filter you want. It then points the scope and starts imaging and continues until it gets the flats or it gets too dark. The script decides if the image is suited by looking at the max pixel values and either accepts of rejects the flat and tries again.

Hey, I like to do things the easy way - sleep nights :)

Cheers

[Astrobserver99]

If you are shooting flats, its a good idea to take bias frames at the same time, otherwise during processing, the flats may not be calibrated properly.

[Astrobserver99]

Almost every time I change the field flattener and focal reducer, the front filter comes off also, letting dust in. I may end up adding a locking screw to the front filter to lock it place...

[higginsdj]

Ummm, bias frames affect darks, not flats (unless I've been missing something all these years. I do take bias frames but I thought that was for the purpose of scaled darks)

Cheers

[Astrobserver99]

To get a properly calibrated master flat, both darks and bias frames should be used to calibrate the flat. While dark flats are not absolutely necessary unless there are a lot of hot pixels, bias frames should be used to calibrate the flats as readout noise is usually present. I usually use lights, darks, bias ... no flat darks...and let the softawre do the processing

Read this:

http://pixinsight.com/tutorials/master-frames/en.html

Skip to the master flat section

[higginsdj]

Ah, yes, but the bias is about the quality of the dark used for the flat rather than the flat itself but I see your point. Ideally all darks should have a bias frame but its not absolutely necessary where the dark has the same integration as the image (or flat in this case)

[Astrobserver99]

The bias also improves the quality of the flat (not sure how?), probably by removing readout and electronic noise. Usually if bias are used, dark flats are not.

I always take darks, bias and flats new with each imaging session, as temperature matching of the darks to the lights is also very important. With a library, the temperature at which the darks were taken might not match with the lights. My qhy8 does not need darks though.

[higginsdj]

No, flats have to have darks subtracted, bias alone won't do. If your flats were integrated sub second, then yes, you could probably get away with just a bias. My flats are typically 4 to 10 second integrations and they need a full dark.

If your doing scientific work then you will need darks.

Cheers