Debayering a DSLR's Bayer matrix.

[Gina]

I have now taken the imaging unit out of the newly arrived duff 1100D and put it in my test camera. It works fine so I shall be attacking the cover glass on the sensor before long.

Meanwhile, with a clear sky forecast for tonight I think I shall see if I can set up the test camera for imaging probably on the MN190. I don't have a suitable mounting arranged for lens use as yet.

Edited September 10, 2013 by Gina

[Prador]

Very interesting thread and I wish you the very best of luck in your efforts.

I was wondering if anybody knows what type of material the CFA is actually made of. A polymer or a dioxide of some sort...etc?

Dave

[JohnC64]

congratulations Gina - knew that you would get there - its been a roller coster ride following this thread

Edited September 10, 2013 by JohnC64

[Gina]

Thank you Dave No idea what the CFA is made of but it's mighty tough.

Been attacking the cover glass of the new sensor with hot air. It's coming off in pieces Sensor still working OK

[Gina]

congratulations Gina - knew that you would get there - its been a roller coster ride following this thread

Thank you John (the meerkat)

[Gina]

The clear skies haven't materialised tonight so first light for the debayered 1100D will have to wait, I'm afraid. I need to sort out the cooling too really before it can have a proper test - at least stick a CPU cooler on the back with a 12v supply Anyway, now I have a mono 1100D I must get back to designing the housing for the cooled parts with suitable sealing and moisture removal. A filter wheel within the outer enclosure is on my mind too.

[JamesF]

The clear skies haven't materialised tonight so first light for the debayered 1100D will have to wait, I'm afraid.

Same story here. We've been promised loads of clear sky in the last 24 hours, very little of which has actually materialised

James

[alistairsam]

another product to consider if anyone's contemplating filter wheels with a coma corrector. this gives a T2 thread with about 1mm increase as opposed to the 10mm of the normal T2 rings.

https://www.teleskop-express.de/shop/product_info.php/language/en/info/p2463_Adapter--T2-Innengewinde-auf-EOS-Bajonett---0mm-Baulaenge.html

Are you planning on cleaning the surface any more Gina? with the area in the middle, could that be the abrasive taking away a bit more than the surroundings?

not going right to the edges?

Alistair

[Gina]

another product to consider if anyone's contemplating filter wheels with a coma corrector. this gives a T2 thread with about 1mm increase as opposed to the 10mm of the normal T2 rings.

https://www.teleskop...-Baulaenge.html

That's useful

Are you planning on cleaning the surface any more Gina? with the area in the middle, could that be the abrasive taking away a bit more than the surroundings?

not going right to the edges?

Alistair

I think I'll stop while it's still working Although I believe I should be able to go right up to the edges of the imaging area, I'm not going to risk it. I have a working mono sensor albeit slightly smaller than APS-C in size - at least at room temperature. I want to be sure that it will still work at -15C and for long exposures. An earlier sensor worked at room temperature but stopped working below something like +20C. I believe that to be due to the epoxy resin which this last one doesn't have. So ATM I'm planning to go no further on the present one as far as CFA removal is concerned. I have started on the next 1100D sensor with the cover glass removal and if all goes well with that I may go further with the CFA removal.

Yes, I'm sure the slightly less sensitive area in the middle is where I've gone deeper into the sensor surface. I need to develop a technique for getting a more even surface. Luckily, with this sensor the damage seems minor and should be compensated with flats. At least with polishing only I have avoided deep scratches which previously caused dead rows and columns and progressively worse damage towards one end.

[Gina]

More cover glass removal with hot air - it's cracking off in bits. Now just tested the sensor... "Houston - we have a problem!"

[Luis Campos]

More cover glass removal with hot air - it's cracking off in bits. Now just tested the sensor... "Houston - we have a problem!"

That's why I dont use heat, looks like it's very easy to over do it...sorry Gina

[Gina]

Now I'm going to put the working iming assembly back in this camera, put it all back together, bung a CPU cooler on the back (on cold finger) and set it up on my MN190 in the obsy.

[Gina]

That's why I dont use heat, looks like it's very easy to over do it...sorry Gina

Very true! Thank you Luis

I think it's better with a cold finger to take some of the heat away. I found the sensor kept cooler. I probably won't do any more 1100Ds - the cover galss bond is well nigh impossible and I've got one working mono sensor to use in a Peltier TEC cooled and re-engineered camera.

I have a couple of older cameras coming that I expect to have a go at but mostly I think I'll use my time to get one good 1100D going. I also have some more I want to do on my widefield triple rig. So the debayering will be taking second place.

[alistairsam]

I wonder what the success rate for jtw is with the 1100 and 600 given that he uses the butane torch

[Prador]

Is there any mileage in going the other way? Would exposing it to below freezing conditions make the bonding material brittle?

It may have been suggested already but it would be buried in the 61 other pages

D

[Gina]

Is there any mileage in going the other way? Would exposing it to below freezing conditions make the bonding material brittle?

It may have been suggested already but it would be buried in the 61 other pages

I have a feeling it has been mentioned - can't remember... I think it's the thermal gradient that breaks the bond so I guess it would work. But how woiuld you do it - what would you use? Liquid nitrogen?

I've been removing the rest of the glass from the sensor I've just destroyed - it took a lot of heat, far more than the sensor would have stood. Also far more than the one I was successful with. I think the bond on these varies from sample to sample.

I now have a sample sensor mechanical model that I can use to design the mechanics, sealing etc. of the super-cooled camera. I'm planning to include just the image side of the sensor plus the shutter and main frame in the dry enclosure and to protect the imaging board components with either potting resin or silicone grease or something of that sort. I'm hoping I can thermally insulate the rest of the camera components from the cold so that moisture doesn't condense on the electronics though a coat of water proofing wouldn't go amiss. I'll cover all this in another thread.

Edited September 11, 2013 by Gina

[nytecam]

I've been reading this thread occasionally for nearly a year but the end seems no nearer. Could someone summarise its success or otherwise here and elsewhere perhaps with some sample mono pics that demo its validity. TIA

[pixueto]

I've been reading this thread occasionally for nearly a year but the end seems no nearer. Could someone summarise its success or otherwise here and elsewhere perhaps with some sample mono pics that demo its validity. TIA

This is my modest effort targeting the knot in the Heart nebula (inserted an internet colour picture of the nebula for comparison). Processed with IRIS (with the help of Luis). It's only three subs of 20 minutes each, 7 dark frames, 7 flat frames and 7 offset frames during a warm summer night. Next time I'll use a lot of shorter 8 minute subs (as suggested by Alistair), hopefully during a colder night and see if I can get rid of the thermal noise.

I've used a monochrome Canon 1000D with a 2" H-alpha filter (7nm). During the monochrome conversion, I scratched the sensor badly and unevenly and yet, no defects can be seen once the control frames have been included so the mod is very feasible indeed -in my opinion.

Hope that helps.

[pixueto]

I have a feeling it has been mentioned - can't remember... I think it's the thermal gradient that breaks the bond so I guess it would work. But how woiuld you do it - what would you use? Liquid nitrogen?

I've been removing the rest of the glass from the sensor I've just destroyed - it took a lot of heat, far more than the sensor would have stood. Also far more than the one I was successful with. I think the bond on these varies from sample to sample.

I now have a sample sensor mechanical model that I can use to design the mechanics, sealing etc. of the super-cooled camera. I'm planning to include just the image side of the sensor plus the shutter and main frame in the dry enclosure and to protect the imaging board components with either potting resin or silicone grease or something of that sort. I'm hoping I can thermally insulate the rest of the camera components from the cold so that moisture doesn't condense on the electronics though a coat of water proofing wouldn't go amiss. I'll cover all this in another thread.

Gina, I think we were all doing very well until it has been decided to use heat and epoxy during the process. Unfortunately we all went backwards when that happened. I think best thing we can do is to find ways of breaking the glass covering the sensor without damaging it or else, go for the 1000D or 450D where the glass can indeed be removed in one piece.

Edited September 11, 2013 by pixueto

[Gina]

Gina, I think we were all doing very well until it has been decided to use heat and epoxy during the process. Unfortunately we all went backwards when that happened. I think best thing we can do is to find ways of breaking the glass covering the sensor without damaging it or else, go for the 1000D or 450D the glass can indeed be removed in one piece.

Yes, it would seem so. I have one or two duff 1000Ds in my sights and planning on working with a 1000D to see how well I get on with that. I may go on to apply cooling as I am going to do with my one successful debayering attempt on 1100D sensors (one success and many failures). I was collecting up my DSLRs this afternoon and was rather surprised/appalled at how many 1100Ds I have

[alistairsam]

Gina, I think we were all doing very well until it has been decided to use heat and epoxy during the process. Unfortunately we all went backwards when that happened. I think best thing we can do is to find ways of breaking the glass covering the sensor without damaging it or else, go for the 1000D or 450D where the glass can indeed be removed in one piece.

I wonder if we can minimize damage to the electronics by desoldering the sensor from the pcb before applying the heat.

that way its only the silicon we're heating

not sure if the damage due to heat is to the sensor or the electronics.

given that the general wisdom is not to keep the soldering tip too long on any component, it'd be prudent to remove that risk

Alistair

[Gina]

I wonder if we can minimize damage to the electronics by desoldering the sensor from the pcb before applying the heat.

that way its only the silicon we're heating

not sure if the damage due to heat is to the sensor or the electronics.

given that the general wisdom is not to keep the soldering tip too long on any component, it'd be prudent to remove that risk

Alistair

In view of the number and spacing of the sensor connections I would think desoldering would be hazardous in itself. In assembly, these connections (and the others) will have been wave soldered - a process where the hot solder spends a very little time in any one place. Also, the sort of damage I had with the last 1100D sensor seemed like parts of the sensor array were damaged - faulty rows and columns as well as a large solid area.

[alistairsam]

how about placing a peltier on top of the glass to freeze the bond?

bad idea? cant see any potential damage...

[Gina]

I guess you could get the cold side really cold and then apply it to the glass for a second or two. Might try that - I have an 1100D sensor with some glass left on it.

Edited September 12, 2013 by Gina

[Luis Campos]

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!!!!!

Edited September 12, 2013 by Luis Campos