Debayering a DSLR's Bayer matrix.

[pixueto]

Hi all,

I have some data now from the Mono modded 450D to compare with the Mono 350D, these two images were made on the same night and although the 450 is not cooled yet, I think is shows very good performance, gladly it shows very good sensitivity, more then my initial test were showing, so perhaps even with smaller pixels, the better processor and 14 bits are pulling that faint details on pair with the bigger pixels of the 350D, the image as a bit less signal but also a bit less total integration (7x7 min. vs 8x7 min. of the 350 image).

So bottom line is, the 450D WILL MAKE ONE HELL OF AN ASTRO CAMERA ONCE IT'S COOLED

So here is the 350D image (8x7 min. @ 800 ISO COOLED:

http://www.flickr.co...in/photostream/

And the 450D image (not cooled camera):

http://www.flickr.co...in/photostream/

Don't forget to right click and choose original size.

Cheers,

Luís

P.S. Way to go Gina!!!!!

That's brilliant Luis. So 7 minute subs is the way to go then? I went for 20 minute subs with my monochrome canon 1000D and the noise was horrendous. I did that because, with the 7nm H-alpha filter, the histogram was completely to the left but then when processed with IRIS, the stretching seemed to go well.

Now, if the image with the 450D is uncooled and the total integration time is 49 minutes, it seems to me that cooling isn't that critical after all, is it? I mean, very good results are being achieved at ambient temperature. Luis, can you tell us which control frames have you used in those photos?

Thanks

[Gina]

Hi all,

I have some data now from the Mono modded 450D to compare with the Mono 350D, these two images were made on the same night and although the 450 is not cooled yet, I think is shows very good performance, gladly it shows very good sensitivity, more then my initial test were showing, so perhaps even with smaller pixels, the better processor and 14 bits are pulling that faint details on pair with the bigger pixels of the 350D, the image as a bit less signal but also a bit less total integration (7x7 min. vs 8x7 min. of the 350 image).

So bottom line is, the 450D WILL MAKE ONE HELL OF AN ASTRO CAMERA ONCE IT'S COOLED

So here is the 350D image (8x7 min. @ 800 ISO COOLED:

http://www.flickr.co...in/photostream/

And the 450D image (not cooled camera):

http://www.flickr.co...in/photostream/

Don't forget to right click and choose original size.

Cheers,

Luís

P.S. Way to go Gina!!!!!

Excellent - great stuff Luis

As for cooling, I think the 450D is similar to the 1100D so I would like to refer you to this thread - http://stargazerslounge.com/topic/167076-the-noise-produced-by-a-canon-1100d-at-various-iso-settings-and-temperatures/

[Luis Campos]

That's brilliant Luis. So 7 minute subs is the way to go then? I went for 20 minute subs with my monochrome canon 1000D and the noise was horrendous. I did that because, with the 7nm H-alpha filter, the histogram was completely to the left but then when processed with IRIS, the stretching seemed to go well.

Now, if the image with the 450D is uncooled and the total integration time is 49 minutes, it seems to me that cooling isn't that critical after all, is it? I mean, very good results are being achieved at ambient temperature. Luis, can you tell us which control frames have you used in those photos?

Thanks

Thanks buddy

I made 7 min. here but this is only with a broadband LP filter (Baader moon & sky glow filter) and don't forget I use a 200mm f/4 newtonian and this pulls in a LOT of photons

But hay, I have some images made with the same exposure time on my backyard using the 1.25" Ha 12 Nm filter that show tremendous detail (for an urban location) on the bubble nebula, latter I'll process that image and show what's possible even with littlçe time on it...this is the beauty of cooling, you don't need a lot of frames to get a final smooth image

For this image in particular I followed as usual the full callibration routine under Iris, BIAS, DARKS and FLATS with median sum combine (2 iterations) and just one iteration of deep space noise reduction under (Photoshop Astronomy tools actions).

Greets,

[Luis Campos]

Excellent - great stuff Luis

As for cooling, I think the 450D is similar to the 1100D so I would like to refer you to this thread - http://stargazerslou...d-temperatures/

Thank you Gina

I have my COLOR 450D beeing TEC cooled, I still have to finish the system to get the MONO 450D also TEC cooled, here's what I'm doing, not sure if sharing the images with this link works but here goes:

https://skydrive.live.com/?cid=afd47809627dc94e#cid=AFD47809627DC94E&id=AFD47809627DC94E!120

Cheers,

[Zakalwe]

At last I've managed to win an auction

An auction sniper is your friend. Goofbid always works for me

http://www.goofbid.com/

I checked the sensor frame and cold finger and it was barely warm to the touch.

This photo shows the sensor assembly and cold finger placed on a 5mm thick aluminium plate. The heat treatment was on the cover glass at the top and towards the left. I started with the top right corner (as in the pic) and then from the corner towards the left. The unit was arranged to show the change in the bond to it's best - in the sunshine.

Could you not run the Peltier to keep the sensor cold. whilst applying heat to break the bond? Having the sensor cold might even help to break the bond more quickly??

[Gina]

An auction sniper is your friend. Goofbid always works for me

http://www.goofbid.com/

Thank you - I thought I remembered reading about something like that a while back - so that's how they do it I'll check that out.

Could you not run the Peltier to keep the sensor cold. whilst applying heat to break the bond? Having the sensor cold might even help to break the bond more quickly??

Brilliant!!! Now why didn't I think of that? And some of the sensors I've had a go at already had cold fingers... Yes, with a sensor already nice and cold it should need far less heat on the glass. I still have a sensor with some of the glass stuck fast so I could try it on a non-working sensor.

[Gina]

Thought I'd just mention that I have finished writing up my own successful debayering attempt on the 1100D sensor - link here - http://stargazerslounge.com/topic/194781-debayering-an-1100d-sensor-removing-colour-filter-array/

Mostly copied from here but with a few extra bits added to improve readability - explain things that were covered in other posts here.

[badgerchap]

Thought I'd just mention that I have finished writing up my own successful debayering attempt on the 1100D sensor - link here - http://stargazerslou...-filter-array/

Mostly copied from here but with a few extra bits added to improve readability - explain things that were covered in other posts here.

Cheers Gina, you're an angel. Now for some reading.....

[Gina]

I had 2 non-working cameras arrive yesterday - a 300D and a 1000D. The 1000D is most interesting from the debayering POV. Easy glass removal and a reasonable one to try cooling - looks like an interesting exercise. As for the 300D I might just see if I can combine parts from the 3 I have and make up one working one and sell that on. Then some of the parts can be sold separately. This is a very old model now and doubt I'll get much interest. Of course, if anyone expressed an interest in a modded and debayered 300D I might well oblige

[foundaplanet]

You may be starting a new business then Gina...

[Gina]

I rather doubt it - but who knows

[russellhq]

Gina, another option to cool the glass would be to use one of those pipe freeze sprays plumbers use:

http://www.diy.com/nav/fix/plumbing/plumbing-supplies/pipe_freezing/Rothenberg-Quick-Freeze-Pipe-Freezer-Spray-11247981?skuId=11758686

[Tf2k]

And another entry for the knowledgebase (this thread ) -

Dichlormethane does a very good job on a dead D40 Sensor, dunno what went wrong with my last Sensors (CFA leftover).

Spray it on, wait one minute and easily clean the sensor scratch free with a Q-Tip.

My only problem is, the sensor died while applying epoxy to protect the wires

[Gina]

My only problem is, the sensor died while applying epoxy to protect the wires

Yes, I had that problem too I'm not using epoxy anymore. The filter frame sans filters works fine.

[pixueto]

And another entry for the knowledgebase (this thread ) -

Dichlormethane does a very good job on a dead D40 Sensor, dunno what went wrong with my last Sensors (CFA leftover).

Spray it on, wait one minute and easily clean the sensor scratch free with a Q-Tip.

My only problem is, the sensor died while applying epoxy to protect the wires

That's very good news indeed!

I have a nikon D80 sensor ready for the Dichlormethane treatment. Have you got any experience with that one?

[pixueto]

Also for the knowledge database . This is what I've learnt from my humble experiments and those of the intrepid pioneers in the forum:

1. The sensor cover glass removal is a pain in every camera model apart from the Canon 1000D and 450D. It can only be removed intact in those two models.

2. The use of epoxy to protect the golden connectors or the heat to remove the glass is indeed a very bad idea and with all likelihood it will destroy or damage the sensor -the epoxy may take a while before the harm can be noticed, though.

3. In the canon models, the CFA in the 350D is the easiest to remove without scratching badly the sensor. Some people like Luis are getting very good at this. However, the wooden tool and the mechanical procedure is needed.

4. Dichlormethane doesn't seem to work with the canon models.

5. The sensor, once debayered, even badly scratched and polished unevenly can work nicely and all defects can be fixed with control frames: flats.

6. Dichlormethane seems to work with the nikon CCD sensors; at least with the 40D. Thus, it's possible to completely debayer a sensor without leaving any scratches or CFA leftovers.

Now, I have a nikon D80 sensor. I'm going to apply Dichlormethane and see what happens. If the CFA can indeed be removed completely without any mechanical means, this may be the choice for a monofied astrocamera for me. The questions will be then, how good this nikon model es for astrophotography and how can the raw files be extracted without any demosicing being aplied by nikon software?

Anybody can shed some light on this? Any experience with the nokon D80 for astrophotography?

Thanks

[pixueto]

Yes, I had that problem too I'm not using epoxy anymore. The filter frame sans filters works fine.

Indeed, the filter frame naked works nicely protecting the golden connectors when removing the CFA. I've just removed the CFA in the fringes of my 1000D sensor.

[Gina]

The weather forecast said we would get some clear skies last night - HUH!  However, I did manage first light with my mono 1100D on the moon through the clouds   The moon only showed for a few seconds at a time so there wasn't time to get good focus even with the remote focussing.  I was able to use live view with ISO 6400.  This was with the Baader 36mm Ha filter - I was set up for a go on the Eagle Nebula but that was out of the question.  I had the camera on the MN190 so exposures were at f5.3 and 1000mm FL.  I also took some longer exposures when the dark clouds covered the sky to see how longer exposures worked.  These are not quite darks but give an idea   All exposures were at ISO 6400 and captured with EOS Utility.  I couldn't get APT to work   This is the first time on the present laptop - I was using the netbook before.  I had only slight cooling with a CPU cooler on the cold finger but without the fan working.

1. Moon behind the clouds
2. Dark clouds 30 seconds
3. Dark clouds 3 minutes

   

Edited September 15, 2013 by Gina

[Gina]

I am not entirely satisfied with my debayering but at least this sensor is working to some extent.  The longer exposures show that I went too deep in places and upset some of the imaging devices.  I don't yet have any more 1100D sensors to work on but I do have some others - 1000D, 350D and 300D.  If I succeed with a 1000D sensor I might go on to install cooling and see how all that goes.  I know there are several members with 1000D cameras who might well be interested in this.

[alistairsam]

Hi Gina

I'd suggest pursuing a 450D if the 1100D fails. As Luis has shown, that shows the most promise being 14 bit and reasonably high res. 

I've finally removed the glass of my 600D sensor. hooray. 

I just used heat, 320C and about 15secs in one spot till I saw fringing, that then rapidly spreads and I follow the fringing.

but before I even applied heat, i used a knife and scored the sides of the glass all around and I applied the heat at an angle toward the centre of the sensor. that kept the electronics underneath only warm as most of the hot air was deflected off the glass. 

once I finished the heat I used a damp tissue to cool it off. 

I then used the knife and again scored it sideways with an attempt to lift the glass. that worked and it slowly lifted and I got the glass off intact in one piece!!

so time to test the sensor to make sure it wasn't cooked, turn it on, and its dead. nothing on the screen even when you turn on. that sounded more like a connector issue so removed everything, put everything back and it now worked, was able to take test images. 

now for the debayering. i'm keeping the filter frame as Gina has done to protect the wires, but I'm not too sure I'll be able to get to the edges. we'll see. 

i'll post results soon. 

I'll be using my wooden tool under a microscope as I did the 350D and 1000D.

excited to say the least...

Cheers

Alistair

[Gina]

I've started attacking the cover glass bond on my next 1000D sensor.  Just using the tip of a craft knife and quite a lot of pressure   As the white approaches the inside I'll reduce the pressure - only applying enough to produce a bit more white at a time anyway.  I would like to get this off in one piece - that'd be a first for me

Edited September 15, 2013 by Gina

[Gina]

Hi Gina

I'd suggest pursuing a 450D if the 1100D fails. As Luis has shown, that shows the most promise being 14 bit and reasonably high res.

Yes, I might do that Alistair - thanks for the suggestion

I've finally removed the glass of my 600D sensor. hooray. 

I just used heat, 320C and about 15secs in one spot till I saw fringing, that then rapidly spreads and I follow the fringing.

but before I even applied heat, i used a knife and scored the sides of the glass all around and I applied the heat at an angle toward the centre of the sensor. that kept the electronics underneath only warm as most of the hot air was deflected off the glass. 

once I finished the heat I used a damp tissue to cool it off. 

I then used the knife and again scored it sideways with an attempt to lift the glass. that worked and it slowly lifted and I got the glass off intact in one piece!!

so time to test the sensor to make sure it wasn't cooked, turn it on, and its dead. nothing on the screen even when you turn on. that sounded more like a connector issue so removed everything, put everything back and it now worked, was able to take test images. 

now for the debayering. i'm keeping the filter frame as Gina has done to protect the wires, but I'm not too sure I'll be able to get to the edges. we'll see. 

i'll post results soon. 

I'll be using my wooden tool under a microscope as I did the 350D and 1000D.

excited to say the least...

Cheers

Alistair

I'll look forward to that

[Gina]

Oh dear... I wonder how you split up quotes to answer parts individually....... aarrgghh!

[JamesF]

Oh dear... I wonder how you split up quotes to answer parts individually....... aarrgghh!

You can insert the quote tags manually and it appears to handle things properly.  It's a bit of a pain though

James

[Gina]

Nearly off