Sky Flats. (Yes Again, Again)

I agree. Good feature to add. Sky flats are far better than panel flats. The automation of sky flats with CCDAP is extremely good and useful. Would be nice to see this in SGP.

1 Like

+1 - especially for narrowband imaging, sky flats are pretty much the only
way too go

1 Like

+1 also for me

1 Like

+1 - while EL panels may be getting cheaper, larger ones are still expensive and automated flippers are very expensive. This is needed for automating an entire evenings run.

What makes you say that? I’ve been using EL panels for both narrowband and broadband flat fielding for years, and never had issues.

I’ve mounted a big EL panel on our observatory wall. No need for expensive flippers.

Renewing my request (April 2014) for this. I now need 14 flats for each target (7 filters on each side of the pier), and it’s a huge hassle to capture all those sky flats manually. It looks like I can either get a large flat panel or move to a program that automates sky flats. While I really don’t want to leave SGP behind (and climb yet another learning curve), I also don’t want to spend ~$600 on a large flat panel (I image with a C11). Is there any chance that sky flats will be added to SGP 3?

Kevin

Kevin,
I don’t know about sky flats, but if you want to go with a light panel and
make your life easier, check out photoglow.com
http://photoglow.com/products/custom-sizes/ If you email the owner and
explain what you want he will make you up a panel that is evenly
illuminated and you can make a custom size. And it’s way cheaper than
other commercial products. The owner knows what a flat panel is for
astronomical flats and he will create the panel with that in mind. I’ve
been using one for about 6 years now and am happy with it. If you do
decide to get one, make sure you buy an electronic dimmer, not the turn
knob style dimmer. I bought a cheap dimmer on ebay or photoglow has one on
their website. The electronic dimmer makes the brightness illumination
repeatable.

Thanks Joel. I emailed Ed, and I’ll let you know what I hear. But I looked into this a couple years ago, and Ed simply referred me to John Wunderlin at Spike-A, because at that time they were working together on flat panels. We’ll see if I get a different answer now. Back then I decided to stay with sky flats, but I had a different setup then.

Kevin

Sure enough, the response was that I should contact Spike-A.

For me, the options seem pretty straightforward. If anyone knows any pros or cons that I’ve failed to take into account below, please let me know:

  1. Flat panel: $600, plus some user intervention on each series of flats, since I’ll have to move the scope from one side of the pier to the other. Also, I might be getting a 12.5" RCOS on loan soon, and I don’t see any way to position a flat panel on that scope because the secondary holder protrudes well above the cage it’s mounted on. I could rig up a wall mount, I suppose, though it would have to be pretty close to ensure full iliumination, and I’m not sure my obs allows that geometry…

  2. CCDAP: $295 plus $75/year. As near as I can tell - having never used it - the whole flats process would be fully automated. It sounds like it can be pre-programmed to take flats in dawn twilight without user intervention, and I’m guessing I could position the scope (and rotator) however I wanted…

Just to be clear, I’m definitely not complaining about SGP, just trying to figure out the best way forward for me. I’d much prefer to stick with SGP, and if sky flats are going to be added to it, I’ll definitely keep doing my t-shirt flats until that feature is available.

Kevin

That’s really a bummer. I got a 12" square flat panel from Ed for around
100$ and I really like it. Since I now have a 10" scope I was considering
buying a larger one from him, so now I’m quite disappointed too. There
simply has to be less expensive options around than Spike-A…

Alnitak. It’s about negative $100 cheaper. :wink:

Kevin

One way to make a flat system is to make a flat box. What you need is a cube that fits over the OTA This can be made out of white foam board. The four corners next to the OTA have white LEDs facing forwards, away from the OTA and illuminating the far wall of the cube. What this gives is a remarkably uniform flat white source.
It costs very little, some foam board, glue, four white LEDs and a power supply. It would be difficult to spend more than $20.

It’s important to make this a cube where the length and width are the same as the distance from the LEDs to the far surface. I found some details about this on the web but every time I try to search for it find it very difficult to locate, I think it was on a RGO site.

I’ve also had good results by sticking a white panel to the observatory dome then slewing the dome to where the panel is uniformly illuminated by the observatory lights and slewing the scope so it’s looking at it.

There are alternatives to throwing money at a problem.

I have used a variety of flat methods including my own light box - but I always found that sky flats worked best as long as you have nice blue sky to work with.

So I ended up buying a fairly expensive flat box - only to realize that its nice illumination control was based on pulsing the lights rather than actually dimming them. That means that there are artifacts when using the flat box with my asi 1600 at short exposures.

This is a typical situation with artificial flats - they are too bright for cameras that have shutters and need long exposures - and they are too dim when used with narrow band filters.

And with progressive shutter cameras the short exposures don’t work because there is varying illumination across the sensor.

So - I haven’t used my expensive light box much with my asi1600 - and instead I use sky flats - and sit there with sgp manually figuring out the correct exposure for each filter - every time I take flats.

I find sky flats to be the best for a good flat that calibrates well - it costs nothing - and the only issue in taking them is getting the exposure right - which is relatively trivial to automate. It also involves no tracking and storing of “state” - i.e. the exposure needed in each filter. Every time you take flats you just figure out - at the time - what exposure you need for each filter - and you take them. If the lum exposure is 0.001s and you have a shutterless camera - no problem. If it’s 10s with Ha - no problem.

Anyway - I support the need for on the spot calculation of flat exposure times. You should just say - I need 15 flats each for Lum, Ha, Oiii - please take them now. And it would figure out the exposure needed at the moment for each filter and take them. Like I do now - manually.

Frank

1 Like

Isn’t the problem with sky flats that it’s like nailing jelly to the wall. The light level is continually changing so you don’t know from moment to moment what exposure you need. People are good at coping with this sort of thing but computers aren’t.

I would also very much like to have a sky flat function.

Fortunately, the flats do not need to have exactely the same average value: anything from about 40% to 60% of the saturation will do. Pixinsight is doing a very good job at scaling them. That makes an algorithm which continuously (i. e. from image to image) adapts the exposure relatively easy to implement: accept the actual light if it is at 40% to 60% and adjust the next exposure aiming for 50%.

Kind regards,
Horia

1 Like

You can actually make your own flat panel as well. Electro-luminescent panels are not that spendy and couple them with a frame and some milk plexi or well sanded translucent brown they work quite well. I now use an Alnitak because it was easier to automate with SGP but the home made one served me well for years.

1 Like

of all people, neil degrasse tyson wrote a paper analyzing where in the sky to point to get the most gradient free twilight flat… and that paper gives an equation that tells you what your exposure should be as the sun goes down (or comes up). this is what ACP implements and it works really well. basically it slews to the right part of the sky (dawn or dusk) and then takes test exposures of the minimum (dusk) or maximum (dawn) flats until the ADU range at the center is within a specified target range. at that point it starts increasing (or decreasing) the exposure as given by the equation. it all works really well.

rob

2 Likes

Unfortunately an illuminated panel on the wall of my obs isn’t an option for my scope, which is a pre-Edge C11 = mirror flop. I’ve tried it, and pointing horizontal gives nasty circular artifacts that are very difficult to handle.

OTOH, looking further into CCDAP reveals I’ll also need to buy (and learn) FocusMax, so now we’re talking $445 plus $75 per year, so the cost issue is starting to disappear. And FocusMax’s technique - slewing to a bright star to focus, then slewing back to target - also may not be the best for a moving mirror SCT like mine.

Has anyone ever looked into off-the-shelf flat LEDs like this one? I have no idea if the illumination is even enough for flats, and you’d certainly need to reduce the brightness somehow for LRGB flats.

Kevin

How about:
http://glowhut.com/electroluminescent-panel-square-12-12-inch-24-24-inch/
+

but would need to make your own frame …

DaveNL

1 Like

If the camera doesn’t have a shutter than you can take flats in bright daylight in a uniform section of blue sky - in which case the exposure isn’t changing much at all and the flats are captured fairly quickly. That is how I normally operate with the asi1600.

Even if the exposure changes a bit it doesn’t really matter because the flats are usually normalized before stacking. You just don’t want them too bright to introduce nonlinearities, or too faint to reduce the total photon count in each pixel.

In my case I don’t adjust exposure at all while capturing the flats in a given filter. But I do have to tune in the exposure for each filter in each session since the sky brightness varies from session to session. This would be much simpler if it were automated. A more sophisticated routine could look at each flat exposure as it comes in and slightly adjust the exposure to keep them exactly on target - but I don’t think that added complexity is needed.

For ccd’s with shutters you need longer exposures so you can’t do daylight and you need to do twilight flats. And since the exposures are longer and the sky brightness is changing - that is a much more challenging situation - and it would benefit from automatically adjusting exposure after each flat. But for shutterless cameras and daytime flats this isn’t really needed.

Frank

1 Like