Tuesday, July 16, 2013

DSS and the Pentax K10D

In response to a question that appeared on the yahoo group for Deep Sky Stacker, I pulled up some old data to check against how the software rendered the information.

A user was comparing how their version of DSS was leaving some jaggies on debayering and showing some residual effects on the stacks.  The example they used featured M13 at ISO 200 and 3 minute exposure duration.


I dug into my archive and found some data close to what they were using.  I found some source files that were 10 minutes at ISO 200 of M17 the Swan Nebula.  Granted, this is not a globular cluster and it's a longer duration.  Still, it should be good enough for testing as there are quite a few stars remaining in the field.

I threw the original files into DSS 3.3.3 beta 51.  No darks, no flats, no bias.  Just lights.

Here's what the output and intermediary files look like in PixInsight:




The upper left image is the stack of 6 light frames.
The upper right image is the registered and "calibrated" image.
The lower image is the "calibrated" image.

All are subject to a basic Auto STF of the same amount.  All are seen at 200% zoom to show pixel level detail.

Compare to the user's example photos:
Raw in Photoshop
DSS Registered
DSS Calibrated
 
DSS Stacked
DCRAW Stacked






Now that I think about this, I believe I've seen this error before.  I think it comes from cosmetic correction being set too aggressively.  It gives these ringed stars that look a little odd.  I'll see if I can find a way to re-create the effect.

Monday, July 8, 2013

More debugging and M16

Still working on the prize of a missing toenail which keeps me from GSSP.

Last two nights the marine layer has rolled in at 1 AM, stopping any further imaging effort.  At least I can add a few more darks to the library.

Interestingly, I've seen that the RA drift continues.  The mount runs slow - about 9 arcseconds per hour of drift.  This is the same pre and post meridian flip -- in the same direction.  No real variation due to gravity or the guiding settings.

It's down to hardware again.  I am swapping the RA cable with another one that I have in the hopes that it might be a resistance, crosstalk, or communication issue with the Gemini or the encoder.  If this does not save it, then I'll start learning the Gemini 1.  I'll swap the computer and build a quick model to try to make it work.

Here's what the M16 looks like now with 27 subs:

Friday, July 5, 2013

M16 in process

As a consolation prize for not going to GSSP because of an injury, I've worked on some practice shots over the last two nights.  The weather is clear and the moon is absent, allowing a small chance to get some decent data.  It's not as good as if I got out of town, but with enough subs, the signal to noise ratio will improve.

Two nights and 11 reasonably good exposures shows promise.


This is a screengrab from PixInsight, just after a quick crop and Auto STF.  Note the relative smoothness of the lighter areas and the noise in the background areas nearby.  This is a 200% view.

Ambient temperatures were pretty warm this night - lows were only in the low 70s.  EXIF temps on the lights read from 16-18 C.  The TEC was set to -10C.  It's likely that the camera was really near 10C.  The local dewpoint was 54 F and the camera body was wet, giving my conclusion.

I did repeat the Marek-style polar alignment again before starting.  There was a need to touch up in AZ and ALT.  Did these adjustments with the scope facing West.  Also used a 3 minute duration exposure to elongate the star trails.

Based on the calculations on this spreadsheet, so long as I have a tight, single pixel line, I should have no problems with drift over the course of the night from field rotation.

Today, I checked the bolts on the mount and a few were slightly less tight than I'd prefer.  Probably due to the heat cycle of the day.  Tightened everything down.

Before starting another run, I did the V check again.  No need to adjust in ALT, and there was a need to adjust in AZ!  Probably due to the tightening of the bolts.

I'm going to run the session with no PEC to see if I still see RA drift.

Thursday, July 4, 2013

Getting back outside

bounty by S Migol
bounty, a photo by S Migol on Flickr.

I've got the Greens
(not blues)

Beets and cucumbers from the garden.

This is what I've been doing recently, getting plants to grow in the back yard.

The image to the right shows what I pulled yesterday.
 The image to the left shows the kale that we'd been using earlier in the year.  It needed to rest and recover before it went to seed. 
 This is a long view down the garden space showing the raised beds that we're using.  In the scene is the beds of kale and carrots, behind are beets and leeks, and behind that are tomatoes.

...and then there's Bad News


I injured my foot on Monday and thus will not be attending the GSSP star party.  Was able to sell my tickets so it's not a big loss.

Instead, I've set up the scope in the backyard and am enjoying the chance to get some more sky time.

I tried a different method of polar alignment last night.
  1.  Alignmaster was OK but it still suffers quite a bit from backlash introducing errors.  I think that it is the quickest to get aligned if backlash is minimized.  I set TVC to 45 for testing.  Advantage of the Alignmaster is that it can give some control over where the ending star is and thus AZ vs ALT adjustments can be set.  I didn't do this last night but might try tonight.
  2. Gemini's PAC software also works well -- and seems to understand the way the backlash works.  Disadvantage is that it will do a deflection in a part of the sky where either AZ or ALT adjustments mean little apparent change.
  3.  Did the drift alignment with the DSLR like Marek does - 10 seconds to do a spot star, then one minute in RA at guide rate and then another minute back.  Gives a nice V shaped line.  This allows a clear, measurable reduction in the drift.  Takes about 6 minutes per adjustment.  Getting close enough already is important for this technique as it would be a pain to have to keep adjusting when very far off.
The DSLR method may be the fastest after all.  With the PHD showing a bright star, it makes it easy to measure the drift changes.  No backlash issues, either.

Last night, after all these efforts, I was able to see that for ALT, the drift was pretty good - minimal change in the 2 minute exposure (taken at 9 DEC).  For AZ, there was about a pixel (or 1.9 arcsecs) of drift in 2 minutes.  Maybe more.  Thus, in 20 minutes, there was noticeable rotation in the exposure.