First light QHY-8/MPCC/GSO-12

[kinetic]

Wow these files are big compared to DSI II files! :)

This is first light tonight using an MPCC on the 12"
Newt and the QHY-8/ homemade GEM.
I now get almost full frame illumination (I think)
and it shows the extent of the typical fast Newt
coma.
The colour balance is wrong here but I did do a sky
background auto-adjust. No sharpening has been applied here,
seeing is average to bad and the gully breezes are gale force.
Excuses :)
I have a collimation issue too!

Edit: Question for the QHY-8 gurus:
Do you guys capture in RAW FITs with these OSCs and process
as Raws minus Raw darks etc
OR do you capture in RGB FITS and RGB Darks?
The processing regime last night took AGES to do just 35 frames.
I did this:
RAW capture ---> square debayer minus debayered darks--->stacked.

Comments and suggestions welcome.

Steve

[telecasterguru]

Steve,

Interesting first light. How about a pic of your mount? How long are these exposures?

Frank

[kinetic]

10 sec x 35 frames / 10 darks / no flats Frank.
A few running gear pics at post #73,#95 here (http://www.iceinspace.com.au/forum/showthread.php?t=43091):

Steve

[pmrid]

Hi Steve. there is a lot to look forward to in that combination of scope/mount/camera. I think I would put the coma issue at the top of my list of things to fix. I'm more than surprised the MPCC didn't take most of it out. So I have to ask a question about how you have attached it to the camera. Have you maintained the 55mm separation required between MPCC and CCD? That's a must.
Peter

[gregbradley]

QHY8 is a DSLR sized chip and if I recall correctly is the Sony 6.3mp chip first used in the Nikon D70.

So it is what is termed in the CCD world as large format.

This then requires coma correction from your Newt in a circle wide enough to fully illuminate your chip with a fully corrected image. This is referred to as the corrected circle or corrected field. It has a dimension like 36mm etc.


The MPCC per the Baader website will correct even 35mm film size (44 x 44mm I think it is) which is bigger than your chip of 21 x 14mm.

I looked up the MPCC specifications and as pointed out you need to have it set 55mm from the back metal flange of the MPCC
to the top plane of the chip. This is called metal back distance (the last metal edge of the corrector to the surface of the CCD chip).
All correctors/flatteners have a metal back distance. You then need to have the correct length adapter to make sure your MPCC and chip are set at that distance. If one is not available standard you'll have to get one made up. Hopefully QHY sells accessories or Baader does.
There is mention on the Baader site about a T2 system.

Here is the link to Baader's writeup:

http://www.scopecity.com/detail.cfm?ProductID=5515

The unit does not operate properly at the wrong spacing.

Greg.

[kinetic]

Thanks Peter and Greg for the advice.
Yep I had read about the 55mm distance but I just threw it straight
on to get an idea of the 'uncorrected coma'/ rough FOV of my 12".
I knew it was going to have some but I didn't quite think this much :)

If I am correct I need to knock up a quick spacer today on the lathe
with the spacing being 35 mm as shown below in red mm.
The QHY-8 has a CCD/film plane depth of 20mm from the face
of the 2" nosepiece.
I hope I have this right.
I have 2 options for threads....42mm male one end, 48mm female other.
OR
42mm male one end, 42mm female the other (with spacer/lock ring
removed on the MPCC)

Steve

[pmrid]

Someone in these you forums (I can't remember who I'm afraid) had a spacer made up that had a threaded male/female section in the middle, with a locking ring so that you could fine-tune the separation if you wished. Seemed like a good idea and since you're heading to the lathe anyway, this might be a good plan.
Peter

[kinetic]

Yes thanks Peter, I was messing with alu stock just a few mins ago
thinking the same...but it has to also slide down the 2" focuser too
so low profile lock ring of some sort....or recessed grub screws.

cheers,:thumbsup:

Steve

[pmrid]

Yeah. When you're cutting the femail thread, just cut it 5-10mm deeper and cut that extra off - that'll be the locking ring - making the male thread a but longer too of course. That'll keep the OD to the right size.
Peter

[bratislav]

Steve,
your debayering seems wrong to me. With QHY8 Eta should be dominated by RED channel, not blue. Make sure you find the 'offset' information in whatever software use for debayering and set it correctly for QHY-8.

[tlgerdes]

Marc Aragnou (Multiweb) posted pics of his adjustable spacer around August last year. I am sure he can send them to you if you PM him.

[Hagar]

G'Day Steve, Just had a look at the image. A couple of things may be worth thinking about.
1. Send me a PM with your address and phone no. I will give you a call.. I think I still have a custom spacer for the MPCC and QHY8 and I will send it over to you.
2. Debayering, I'm not sure what you used for capture or deybayering your image but it does look somewhat blue. If you are using Nebulosity the debayer protocol used for the QHY8 is incorrect. You have to go to options and check the manual debayer override to set the correct sequence. The correct sequence is x = 1 , y = 1 with a pixel size of 7.8um.

3. Search these forums for custom MPCC adapter as Marc (multiweb) arranged to get a custom adapter made for the MPCC which was adjustable and from memory he was very happy with the end result and I recall him posting a drawing of the adapter which you could probably make yourself.

I always capture in undebayered Raw images in all capture programs.

Nice image just the same.

[multiweb]

Cool shot Steve. I think your MPCC spacing is incorrect and you debayered to the wrong matrix too. QHY8 is RGGB.

Calibration is before debayering. So you shoot your lights in raw FITS 16bits (approx 11MB each). You shoot your bias frames in raw fits and shoot your bias in raw fits too. You do your calibration on your lights with the bias and flats as they are. (with bayer matrix).

Then you debayer to Red, Green & Blue in separate directories. Pick the best FWHM frame in each channel and do your alignment/normalisation/date rejection per channel. WHen you have your masters R, G & B you recombine to color then save as TIFF 16bits scaled. Go to PS and season to taste.

PS: DO NOT align any light/flat or bias with the bayer matrix in. Only align debayered subs.

[kinetic]

Thanks for the input fellas....I shoulda checked in for replies early
afternoon...cos I've just made one Marc, Doug, Pete and Trevor! :)

I made it adjustable so that it covers approx 52-57mm.
Locked with a recessed pair of grub screws that won't foul with
the 2" Drawtube. Thanks Dougie for the offer though mate:thumbsup:
I will have to make a few of these separate distance spacers for any 2"
eyepieces I ever buy, not that I do much visual.
The only one I have is the GSO 32mm, might be interesting to try
the MPCC on that. I do remember my first view of Eta Car through that
when this mirror was brand new, and the coma was evident 7/8ths
out to the edge of the FOV.

Thanks Doug and Marc re the debayering....yep I just did RGB..that would
explain quite a bit :)
I'll sit down and read your tips guys and try and digest the method
you use. Who would have thought the OSC would be such a learning
curve when going from DSI II mono :). Not that I've been very successful
with that either :)

cheers guys.

Steve

[mldee]

Steve,

Quite a lot of QHY8 +MPCC's getting around now, would you be interested in quoting a price to make your adaptor?

Mike

[pmrid]

Make that 2:
By the way - that is a nice, polished (i.e. highly reflective) inner surface on the adapter. You migh want to think iabout cutting in a nice fine pattern of parallel ridges and follow with a coat of matt black blackboard paint or equivalent. The QHY8 has enough reflexions already!!

Peter

[kinetic]

Mike. Peter,

In all honesty, without checking, these things probably go for under
$50 posted, or I wouldn't pay any more, put it that way.
It would cost $20 bucks driving around finding suitable alum stock
to make a few and that's before I added my machining time.
If I were you I would buy one, it would be in the post 2 days later
and probably better tolerance and finish than mine :)
Have I talked you out of it yet :)

As for reflective innards, all you can see in the pics are the M42 thread
surfaces and the inner and outer inteference fit sleeve.
The bore is a vee shaped tool run on slow thread cutting down the guts
as I do all my adaptors. Flat black paint and texta to touch up.
Sweet as.

Steve

[DavidU]

Nice work there Steve, I think that setup will be something special when you nail it all.

[kinetic]

Tonight's effort,
the 55mm spacer now in use. Coma is minimal but I still have
the elongated stars I've had all along radiating in a position angle
sort of 10 o'clock. This is NOT a tracking issue. It's on a live 1 sec shot
or a 10sec shot. Maybe the CCD is slanted slightly? A tip from Theo on it
here:http://qhyccd.com/ccdbbs/index.php?PHPSESSID=5f99f78f470308a 7637a0fbc61bb16ea&topic=1885.0

That and a weird sideways blooming effect on Eta
and a few brighter stars. Debayer artifact?
Seeing was not brilliant either tonight but I think I nailed the colour
albeit a bit unsaturated and lacking blue.
The debayer was done as per Doug and Marc's suggestion 7.8u sq
pixels with a 1x1 ratio.

Edit: 2nd result: 2 iterations of the same curve applied.

Steve

[multiweb]

Way better! :thumbsup: Yeah given your exposure time this is a squaring issue. Your CCD plane is tilted. Should be very easy to fix.

[kinetic]

I think it is actually the drawtube. The cheap GSO ring clamp pushes the
2" barrel off axis a bit.
Just by eye it looks at least 0.2mm, more like even a whole millimetre.
The qhy-forum post I mentioned below says this is unacceptable for
a fast optics Newt like this. This might be a great part of the problem
right there.

Also, Marc, how exactly do you do a RGGB debayer. Are the settings
correct that I have entered here in the manual text box in Neb?

This is what I get when I select CIM debayer and 7.8 sq pixels 1x1:

Steve

[Gama]

Its drawtube/fociser flop.
If the CCD was out slightly, you will find only one side out a little. The whole image is off, meaning focuser/drawtube flop. You need to stregthen it somehow.
The later model QHY-8 cameras have tilt adjustments (3 plastic spirniged screws).

Theo.

[multiweb]

Sounds about right. Tilt is a killer on a field. Sometime I get it right. I found to use extension tubes and keep the focuser barrel in as much as possible helps. I also use GSO 10:1 focusers.

Sorry I use CCDStack not Neb for processing but from your colors you're using the right pattern.
RG
GB

two greens, one red and one blue in that order. Two Gs are recombined in one pixel.

[kinetic]

Thanks Mark , Theo,

I think I've removed most of the off axis tilt tonight because the
focus seems a lot crisper.
The only misalignment now might only be collimation, seen here
with this crop of NGC2822 and Miaplacidus.

Edit: Eta Carina from tonight added too.

Mia was almost centre to the FOV so the ghost shows the extent of the
collimation.
I'm glad that blooming type artifact is gone too.
This camera is a bloody beauty. The dynamic range is lovely under a
big light bucket. This is just a 10s shot series and stars have saturated.

Steve

[Hagar]

G'Day Steve, These settings are correct for nebulosity. I also found just doing a RGB manipulation is all you should need for OSC imaging. I didn't see any real advantage in splitting the image into 3 colours, manipulating the recombining.
Looking better Steve.

[multiweb]

I got into the habit of splitting channels because the very first times I used CCDStack I used to get color ringing around my stars. That was due to undersampling and doing the data rejection on color subs. After a chat with Stan Moore he advised me to process and stack every channel independently and in my case it did smooth out star edges. The problem was gone. I never had the problem with the C11 though because my image scale is 0.4 asp. Overtime I 've also found that processing separate channels allowed me to be more selective with noise reduction because I always seemed to attract more noise in green and blue pixels when the reds have always been relatively clean.

[kinetic]

I captured a few more favourites last night , albeit with only
5 sec sets to check the FOV.

Processing is such a chore now with these huge files :)
20 frames takes about half an hour to an hour to demosaic/stack/ tinker with.

Steve

[DavidU]

5 sec subs !! LOL awesome.
Your setup is going to be amazing.:thumbsup:

[kinetic]

Still experimenting with gain and offset levels.
From this AM...

Steve

[kinetic]

Tonight I'm going to try binning for the first time on the QHY-8,
weather permitting. I have never owned a camera capable of
binning so I'm flying blind tonight....as I do most nights :)

If any QHY-8 gurus could help me...
I plan to:

Do a deep set of something like M83...by deep I mean maybe over
30 sec! WOW...unusual territory for Steve :)
For the lum/raw frame set.

Then do a binned 2x2 set for RGB data.

I assume the 2x2 set are done at 30sec exposures also? or
am I way off here?:shrug:

Also, so far I have used Gain=zero and offset at 126 (for 5-10sec sets)
Is this around the mark for a good offset or logic tells me that the offset
will be exposure length dependent......go longer and the histo creeps up
to the right therefore offset can be less?

Any advice most welcome.

cheers,

Steve

[mill]

The offset is a set and leave setting, same for gain.
You could take the color set at a shorter exposure time with 2*2 binning.
When i do 600 Sec L exposures at 1*1 binning then i take the RGB's at 2*2 binning and 300 Sec.
In the end, the RGB is only there to get the colors in the picture and just have to be noise free.
This is with the QHY9, so in your case it might be different and someone owning a QHY8 might say something different.

[kinetic]

Thanks Mill for the tips!

Steve

[kinetic]

From tonight.
NGC3324 was the target.
Inspired by Danielsun's great result in the Deep Sky section (http://www.iceinspace.com.au/forum/showthread.php?t=56421)
here is a quick 17 frames at 15sec.
No Darks, flats or bias frames.....some hot pixel artifacts
showing in processing.

Steve

[kinetic]

Another curve tried on this result.

Steve

[AlexN]

Traveling along really nicely mate..

As for your comments earlier regarding taking lum at 1x1 and RGB at 2x2 with the QHY8... If you bin the QHY8 2x2, the resulting image is monochromatic, not RGB... It is unfortunate, as it would make a good camera fantastic... 2x2 colour binning is supported by some of the newer sony sensors, but not the ICX-453 in the QHY8. Your best bet is to capture as much data 1x1 as possible.. capturing a lot of 2x2 data might work for brightening the image, but it will not be a true luminance layer in that you would be getting your best resolution and detail from the 1x1 colour image, and the 2x2 monochrome lum would just be used to brighten up the image...

[kinetic]

I'm not quite sure which lobe of the NGC3372 complex
this one from tonight is. I looked at Mill's great widefield
shot in DS to try and pinpoint it but I'm about to give it away
for the night...work tomorrow :(
Interesting little string of stars highlighted in the neg version.
Grabbed a few favourites again tonight....
Apologies for tacking these on to the 'first light' thread but
it's a better way for me at least to check my progression.

Steve

[kinetic]

And M42 preliminary result from last night.

Steve

[kinetic]

This one is similar curves but different noise reduction.
More frames needed. A different curve to show Trap detail.
I discovered last night I have a drive PEC index error and
the PEC was engaging at a different point on the curve
to where it should. This means all of my recent results have
had PEC actually working against me rather than correcting
or zeroing most of it out.
This is not all bad news. It means the pics can only get better
once it's fixed. The culprit is the toothed belt gear on the end
of the worm. It has somehow slipped and turned approx 2hrs
on my 'clockface' index wheel. Loose grubscrew!

Steve

[kinetic]

Tonight I grabbed another set of the 3372 complex
and it stitched together quite OK with my 18/19Feb set.
This is a good sign that the MPCC is keeping things
under control as far as coma and field flatness goes I suppose.
You can see the join, not a lot of effort done here
to hide it.
These cameras are such a pleasure to use.

Steve

[Octane]

Steve,

That's looking marvelous.

Do you use Photoshop? If so, use the File -> Automate -> Photo Merge command. It should seamlessly blend the images together for you.

Also, I'm so glad to read the middle line in your signature. That is my very thought -- which I've made known on here on numerous occasions.

Well done!

H

[kinetic]

Thanks H for the tip. I don't do mosaics that often so when I do
I usually can't recall the app I used to do them :)
I know once I used a SONY stitcher my brother had with his DSLR
package and it worked wonders.


Amen , brother :)

[kinetic]

Seeing was very good this morning but what ruined it
was a persistent axis problem with the GSO focuser
and Baader MPCC/ QHY8 combo.
At such a short F ratio either collimation and / or off
axis tilt in the image train is still causing this trailing of stars.
By rotating the entire QHY-8 I can minimise it but not totally
eliminate it. The result here shows the tilt at about the halfway
point of the severity.
If rotation changes the level of this tilt, does that mean it's
NOT collimation?

Steve

[kinetic]

This image is from the 24th when there was a first
quarter moon. Still playing with misalignment issues
with the QHY-8 and MPCC in the GSO focuser.
I have run a heap of results through CCDInspector and
they all seem to agree with some misalignment and
definitely bad collimation.
Collimation is good to the eye but that's not good enough for
fast Newts. Apparently the tolerance on-axis for my F5 is
less than 1 mm! Not good timing, so many unbelievable 2944s
posted lately:)

Steve