Tuesday, October 25, 2016

Estimating gain of my camera

I wanted to try out using the subframeselector script to assign weights to all frames. For that, I needed the gain of my camera. I can't find the data sheet from FLI anymore, but Richard told me a quick way to estimate the gain good enough:
  1. Take two flats.
  2. Raise the level of one of the flats slightly.
    I am doing that in Pixinsight by using the Pixelmath expression $T[0]+0.2.
  3. Subtract the other flat from this first flat
    Again, in Pixelmath: Ha_01270-Ha_01269. (these are the names of the flat fit files).
  4. Measure the standard deviation in the middle.
    In Pixinsight, I create a preview in the middle of the new image and with the Statistics process measure 5.1259e-003 = 328.06 DNA (*64000)
  5. Divide by sqrt(2) = 231.97 - this is the noise (or to be exact the quadrature sum of shot noise and read noise)
  6. Now subtract a bias frame from one of the two flats
  7. Use the same region (I drag and drop the preview that I created in step #4) and measure the average value: 6.1413e-001 = 39304.32 DNA (this is the signal)
  8. Now, gain = signal / (noise^2)
    gain = 39304.32 / 231.97^2 = 0.7304

Sunday, October 23, 2016

IC 1318 - Narrowband enhanced

I wanted to see if and how my image of IC 1318 could be improved with Narrowband (Ha and OIII) data:

Which is a big improvement compared to the original. When zoomed in, the improvement in color and detail are quite amazing:




I took 440 minutes of HA data and 720 minutes of OIII from our backyard. I then used a tutorial from LightVortexAstronomy to fold these in.

Fixing my vertical stripes - Fixing the darks! (not the individual frames)

My first approach was to use the measured differences to correct each light frame that I took. But that would have meant to do this over and over again every time I have new data...

Instead I decided to adjust my dark frame (subtract the difference in the respective columns). Then I would have to do it only once. Of course, I would only correct the dark frames that are used for lights. The much shorter dark frames that I use for calibrating flat frames I would leave as-is.

Using Pixelmath in Pixinsight, I could easily correct each column:

Doing this for both columns resulted in this dark frame:
Compare this to the original dark frame:

One can see that the columns are still there, but much more subdued.

With this dark, I calibrated and stacked my Ha frames again:

No dark columns - compare with the original stack:

I also checked other filters and they got corrected well too. So, from now on I'll use the corrected dark frames for calibrating and everything is fine!!!

Fixing my vertical stripes - Measuring the Difference

Now, where I knew how to fix my vertical columns, I needed to a) find a way to measure the difference in the two columns, and b) find an efficient way to apply these corrections.

1. Measuring the difference
I first toyed with the idea of writing a simple Python script to do this, but I can never figure out how to install a Python distro on my Windows machine and then install Python libraries on top of it - they always seem to come with their own Python environment ...
So, I thought about Pixinsight:

Determining the to-be-corrected columns:
Zooming into the image showed the bright columns in the darks really well:

Hovering over the column shows the coordinates at the bottom. X=4633 is the one on the right, X=1285 the one on the left.

Next, I needed to measure the average level 3 columns to the left and right of these columns and of the columns itself in an image that has as little as possible detail in those columns.

Luckily my images of the Bubble Nebula that I recently took are a good fit:
Above is the dark with the two bright columns - at the bottom is the Bubble Nebula image that shows no details there.
First, I calibrated my 10 Ha frames. Then I loaded the first one and used Pixinsights Previews to define an area of 3 columns to the left of the problematic column:

I then created the other two previews:

Now, I could use the Statistics process to get the average level of these three previews:

I manually entered the "mean" values into a spreadsheet:

When I was done with one image, I opened the next one and dragged the previews from the previous to the new image (to avoid having to define them by hand over and over again):

Pretty cumbersome, but at the end I had all the values (3 readouts x 10 frames x 2 columns):

Taking the averages of all differences, I ended up with the following values:

Column X=1285 Difference=0.0003619335296 (~DN=23)
Column X=4633 Difference=0.0003021386857 (~DN=19)

Now, I needed to find an effective way to apply this.

Saturday, October 8, 2016

Still pier collisions!

When I imaged the Bubble Nebula, I still experienced pier collisions. The filter wheel banged into the AP control box:
Filter Wheel colliding with the control box
I tried to put the control box in other positions, but unfortunately, it has to stay in the back. The solution was an extension cable from Astro-Physics:
CABGTO24 - Extension Cable for the AP Control Box
With that cable, I can mount the control box much lower and it can't interfere with the scope/filter wheel anymore:

Moved the AP Control Box lower, using the extension cable

Bright, vertical columns in Darks

I took new darks and also bias frames, and they show bright vertical columns:

Master Bias (1000 frames)
10 min Master Dark (51 frames)
But these don't show up in the flats or lights:
Master Flat (20 frames)
10 min Light Frame

And when I know calibrate the light frames, they show a dark column in the place where the bright column was in the master dark/bias:
Calibrated Light frame
I retook my master frames (bias, darks, flats), calibrated with Pixinsight or CCDStack - always with the same result ...

--- Update ---

I asked Richard, Tim, Jim and Richard about it. Apparently this is a rare but known problem with KAI/KAF sensors. Tim measured the pedestals and said that I need to add 24 DN to my lights in column 1285. With that, I get this as my calibrated light frame:
Calibrated Light frame with correction in column 1285
No more dark column on the left!!! Now I have to figure out the pedestal in the other column and then an efficient way to apply these corrections to my light frames.

Saturday, October 1, 2016

My backlash issue in DEC

As I wrote before, I observed a pretty severe backlash in both axis. When I asked about this in astro-phyisics forum, I got a talking to from Roland that this is a red herring. Well, I still tried to fix it.

I used the Guiding Assistent from PHD2 which can also measure DEC backlash:

And PHD2 measured 1008ms backlash compensation!!!

After I entered this in the backlash compensation settings, I had a couple of "overshootings" where PHD2 correct in DEC so much that the star went to the other side. I asked my friend Tim Kahn about it who told me to only use 60% of the correction to avoid this. I did that, and since then have MUCH better DEC corrections.

I still wonder if I never noticed that before (would be strange) or if something really changed in the mount...

But the result is that I can now track REALLY well. I get 0.5 arcsec accuracy on pretty much all nights now.