PHD Dark frames?

[whzzz28]

I took out my QHY5 and was going to do some drift alignment via camera to save me having to sit on wet grass the other night, and i am new to using PHD (and drift alignment). I spent a few hours playing before some clouds rolled in (Oh Brisbane).
I watched a few tutorial movies but none seemed to cover the dark frame ability of the software, nor is there much (anything?) on the net about PHD dark frames.

Correct me if i am wrong but my thought with regards to a dark frame was that the camera takes a few frames (with the cover on) to work out problematic pixels and then subtract it from future captures to remove noise/bright pixels. It does this by taking the photos in a pure dark environment (ie scope cover on).

What i find is after i do this, if i try and do a loop to just "see" what the camera is seeing, i end up with a white screen. My first thoughts were: Gain/gamma on either ASCOM or PHD was too high. But even if i set gain in ASCOM to 0% and PHD gamma to 0% the screen is still pure white. This is looking up at a dark sky.

I put the cover back on the scope and i end up with this:
http://core-au.net/astro/1.png
I would be expecting a pure black picture.

If i don't use dark frames, this is the view through the scope of my sky (was cloudless when i took this, yes gamma is on 0): http://core-au.net/astro/2.png

I spent a while trying to calibrate on a number of these (what i thought were stars) only to get "this star didn't move"... being my first time trying to use PHD it was a bit of a '...what?' situation.

Tonight i decided to play around a bit inside.
I ended up with this:
http://core-au.net/astro/3.png

This is with a dark frame, on-screen gamma at 100%, ASCOM at 50% gain.
Dark frame was taken by pointing the camera at a light then blocking the camera with some cardboard when looping. The white dots are noise/bright pixels/whatever (if you look at 2.png you can see them on this picture as well, and i probably tried to use some of these pixels for tracking which wont work).

Is my understanding of a dark frame wrong?
Is there something i am doing wrong in PHD? I wasn't able to find any stars to track for the hour i spent trying, although i believe this is due to my on-screen gamma being down very low and that i would of tried a number of these white dots which were in fact dots from the camera, not stars. Default in ASCOM for the camera gain is 50% as well. If i had the gamma at say 50% i might of seen a few other dots to try and calibrate on. Also i was pointing at close to the meridian (south) and it is pretty bare out there, although i should be able to find something.

One other thing i noticed but i am not sure how normal it is but the QHY often has "lines" going through the video. I didn't take a screen shot of it but it is sort of like what is in 1.png but on a black background. Video's i have seen do not have these lines, but then again i am not sure what camera they are using.
I think if i get dark frames solved/working then the lines may not be an issue.

I'll get back out to try again whenever the weather is good enough, but I'd like to have a bit more info this time so i don't get disheartened and sit out there while the mozies bite.

Any thoughts are appreciated.

[mswhin63]

I understand how you feel, my dark frame look the same as your first image but unsure how to get it like you second image. I have noticed though the so called dark images are smooth and are still able to rack well. In some instances I have the time set to 5 or 10 seconds.

Mossy here too, I have just set-up teamviewer on the small Netbook and going to attempt to remote control it tonight.

[Rigel003]

Those are definitely hot pixels you're seeing. Is it possible that your guidescope is substantially out of focus so no stars appear? You could try a test frame on the moon to nail down the focus.

[scagman]

Ni Nathan,

I have been playing with PHD lately (well its been a week or so due to #$%#! clouds) and had the same prob with finding stars too. thought I was really good and had nice sharp focus and no drift. it was a hot pixel. Asked a few questions on IIS and found that it was my focus, or lack of, Once I played around with the focus I was able to get stars and all was good. So I would agree with Graeme, play with your focus. I didnt take any dark frames and all worked fine. (once i stopped triing to use hot pixels).

Once you get it sorted PHD is very nice and like you say sure beats getting down on the wet grass.

Good luck and clear skies

[mill]

The only thing you forgot to do and i can't belief others didn't see this :P
Only the last picture shows the slider moved to the left completely.
I should be about a quarter from the left.
My guide camera gamma is normally on 95%.
The slider doesn't change the camera gamma at all.
slide the bottom slider to the left till you can see the picture appear, this is in the main PhD window.
This slider is there so humans can see stars and pick one for guiding :)

[mswhin63]

Last night I also found the sweet spot to be 1/4 from the left as well. Also I do not have a black background it can discriminate stars well at that point.

It is interesting that PHD come with darks like it does but EQAlign dark-frame blacken the background very well and I only see the stars.
I have not successfully got EQAlign to guide well for me as yet but just the dark frame feature on EQAlign is so much better.

[whzzz28]

Quickly took the scope out tonight, didn't mount it though as i didn't want to cart the mount out for 15mins of testing - and there is a bit of cloud out.
Hopefully tomorrow the sky will be a bit better and since it's the weekend i'll be able to spend awhile with it playing.

Still can't figure out how exactly dark frames in PHD are meant to be used.

Tried to get focus and see if i could find any stars but i didn't have much luck. The only thing i was able to find was Jupiter (nice and bright).
The hot pixels are very annoying (and for a new camera, i think it does have a fair few hot pixels).
I watch videos like: http://www.youtube.com/watch?v=XPvHQhdLJ4o
and
http://www.youtube.com/watch?v=jSGmqMTjX04

There display is very dark with obvious stars in view. IF i have my on-screen gama (slider in PHD) set to 1/4 from the left i get a very bright screen that you probably won't be able to discern any stars. If i go down the other end to 3/4 from the left it is dark but all i can see are hot pixels. Sure the scope probably was out of focus but i was expecting to see at least some blobs of out-of-focus stars.

Oh and another question; what do people use for camera gain when selecting the camera?
ASCOM defaults to 50% gain.

Thanks.

[mill]

I have a QHY5 and it is set to 95% gain in PHD.
You don't need dark frames at all, i just use the 2X2 noise reduction in the settings and that works very well.

[DJT]

hate to say this but let everything run in PHD at default and whack a regular eye piece in your guidescope just to make sure there are stars in your field of view. What were you imaging or is this a general issue?

I was convinced I had everything aligned yesterday,finderscope, guide scop, main scope etc, (after all, theres not much else to do when La Nina is in town) when chasing tarantula nebula only to find I could not get a sniff of a star to guide with.. Put in an eye piece, adjusted the guodescope till there was one in the FOV and there you go..spot on guiding.

PS, havnt a clue what darks are for in PHD..

[mithrandir]

So you can dark subtract the hot pixels and not pick one of them as a guide star. Hot pixels give you beautifully straight lines in the graphs but terrible guiding.

Of course since I use a guide camera with a Sony chip I don't have any hot pixels. :D

[whzzz28]

More of a general issue. Had a great fine night on Friday so went back out.
Was able to actually see stars through the camera and got it guiding fine, although even with gain at 100% (camera gain, not PHD gamma) the stars were faint, but it was guiding fine.

I tried to drift align with PHD but after maybe 10seconds or so PHD would lose the star so the graphing wasn't very accurate.

I thought i had aligned it fairly well, so i tried to do an align with the synscan hand controller (3 stars). First star was Rigel. Told it to go-to and then had a look through the eyepiece. Wasn't even close to Rigel. As i don't have a guidescope (will rectify that soon) on either of the scopes i spent awhile moving around until i found it.
Had difficulty finding the next two stars as well, and after an hour gave up for the night.
This weekend is looking to be good weather wise so i will try again with a guidescope.

[mswhin63]

I made some changes based on discussions here and was also able to get great dark frame result, clear stars and clear guiding. I believe I am still going to have problems with my webcam for some areas of the skies but in most I get good results which will keep me going for a while.

Thanks all for the information.