Debayering a DSLR's Bayer matrix.

[Gina]

Very interesting and useful Luis - thank you

When I was using a cooled 1100D as OSC with ISO 3200 I found a noticeable improvement in going down to -10C but the 450D is different and max ISO is 1600 and I can understand that this may mean that cooling to -10C or lower might give little benefit.  I have yet to test the 450D.  I am leaving the CFA removal of my latest 450D sensor until I'm feeling more confident in my abilities.  As things are ATM I cannot afford to but any more cameras for their sensors.

I was interested to read your comment on the high reflectivity of the 450D cover glass Luis, means I'm now thinking twice about replacing it.  If 0C is low enough for cooling, the problems of condensation are much reduced and I think I could get away with one optical window.

[libraryman]

"The cold finger is very easy to do, I do mine with a 2mm aluminium plate and grind the sensor contact area with a Dremmel tool till it fits on the sensor gap (about 1.3mm) but you also have to modify the camera's chasis (nothing special...) I don't have the measurements here now but I'm thinking on making a detailed tutorial on how to do the full cooling mod on the 450D once I have some good free time"

I made my cold finger out of 2mm copper...yes, a little heavier but no doubt a little more efficient! i have not yet completed the cold finger mod still waiting for a bit more info on the connections and trying to decide whether i should put the body into a box or go with the aluminium heat sink on the side of the camera!! decisions..decisions!!

Iv'e attached a couple of photo's for the measurements, the black cover is thin self adhesive Teflon tape to prevent electrical contact with the sensor. the small screws are to fix the cold finger to the sensor frame.

they fit perfectly.

Ray

[Luis Campos]

Very interesting and useful Luis - thank you

When I was using a cooled 1100D as OSC with ISO 3200 I found a noticeable improvement in going down to -10C but the 450D is different and max ISO is 1600 and I can understand that this may mean that cooling to -10C or lower might give little benefit.  I have yet to test the 450D.  I am leaving the CFA removal of my latest 450D sensor until I'm feeling more confident in my abilities.  As things are ATM I cannot afford to but any more cameras for their sensors.

I was interested to read your comment on the high reflectivity of the 450D cover glass Luis, means I'm now thinking twice about replacing it.  If 0C is low enough for cooling, the problems of condensation are much reduced and I think I could get away with one optical window.

You're welcome Gina

Yep, there's really no need to supercool at least with the 450D

The original coverglass is really bad reflections wise, there's no big problem on a original RGB sensor but with a monofied sensor is very different, what happens is that the glass layer under the CFA is highly reflective and when we polish the sensor face to remove the CFA it gets really shiny and light from brighter stars ends up bouncing from the sensor face to the coverglass and big uglly halos show up and it's even worse with the Ha filter, the shiny surface of the Ha filter get this phenomenon even worse!

See here some comparasion images of several cover glasses used, these are all unfiltered test shots, this is a worte case scenario with bright Arcturus:

Original cover glass placed directly on the sensor face, this one creates a very bright cross pattern of multiple reflections:

LPF1 only filter placed on the filter frame, but NO original cover glass, this one is coated but bad quality...still big bad halos and a cross pattern:

And this is the Astronomik clear glass filter (no filtering but very good anti reflection coatings) this one also placed on the filter frame and sensor with NO cover glass:

Hope you find this usefull

Cheers,

Luís

[Luis Campos]

"The cold finger is very easy to do, I do mine with a 2mm aluminium plate and grind the sensor contact area with a Dremmel tool till it fits on the sensor gap (about 1.3mm) but you also have to modify the camera's chasis (nothing special...) I don't have the measurements here now but I'm thinking on making a detailed tutorial on how to do the full cooling mod on the 450D once I have some good free time"

I made my cold finger out of 2mm copper...yes, a little heavier but no doubt a little more efficient! i have not yet completed the cold finger mod still waiting for a bit more info on the connections and trying to decide whether i should put the body into a box or go with the aluminium heat sink on the side of the camera!! decisions..decisions!!

Iv'e attached a couple of photo's for the measurements, the black cover is thin self adhesive Teflon tape to prevent electrical contact with the sensor. the small screws are to fix the cold finger to the sensor frame.

they fit perfectly.

Ray

Yeah copper is much better but around here I can't find any usefull bit for this so went for aluminium, still good cold transfer

Mine is at the side, works great

Cheers,

[libraryman]

Many thanks Luise, i looked at your camera mod pic's on flicker, the side on cooler looks great and obviously word well for you.

Ive already modified the sensor frame for the cold finger so that's my likely route!

Ray

[alistairsam]

I came across a forum where someone swapped the sensor of an olympus e-300 with a mono because it uses the kodak kaf8300 colour sensor which is pin compatible with the mono. the mono sensors can be purchased for $200 odd, so might be an option for some. 

just thought I'd share this info. 

e-300 looks like an old model though. 

the 450D is proving tough to remove the cfa. anyone used the glass polish on a sensor successfully? does it actually dig into the sensor layer? 

Cheers

Alistair

[conehead]

Hi,

i read the hole thread before i started to debayer a sensor from my 40d. i bough one for a few buck to start with it. i also wan to share my experience with you:

first i removed the front glas, i heat it up with a heatgun and a small extension to 60°celsius. then i lift up the glas with a scalpel starting at the Corner:

then a start to scratch of the lenses and the Matrix with a wooden tool:

finished removing the Matrix:

to protect the bonding wires with epoxy i used the 2 component uhu endfest . i take a small drop and let it fallen down to the pit. it is not perfect, but the protection is superb:

starting to pollish the ar coating away:

and here a streched flat without a lens and some paper in front of:

and now the Tools i used for it:

meguiars PlastiX for polish:

http://www.meguiarsdirect.com/product/4596

a pice of Wood grindet to a sharp top from a cleaning rod like this:

http://www.leder-lauber.de/index.php?id=1

a dremel felt tip, but i polished without the dremel only by Hand!!!:

http://www.hornbach.de/data/shop/D04/001/780/493/579/55/DV_220x214_7947100_01_4c_DE_20131204174638.jpg

markus

[Gina]

Brilliant!  You've done very well indeed

[RAC]

Looking very good conehead. Wow you read the hole thread!

[Herra Kuulapaa]

I have one Nikon D5100 body lying around and the plan is to convert it monochrome ie removing bayer matrix.

Before actually starting do it I’d like to ask if anyone has experience on this specific sensor model (also D7000 sensor) and are there any special advices for it?

I read the whole thread through, relatively fast though and noticed that one shouldn’t go too close the edges because of some invisible wiring or something. Most of the examples were however Canon sensors so I’m wondering how this applies to Nikon ones?

Reference picture:

https://d3nevzfk7ii3be.cloudfront.net/igi/WLYSvKbyGSpyIxlv

[Herra Kuulapaa]

I did some tests with slow, semi fast and fast curing household epoxy resins and conclusion seems clear.

Faster the resin, more shrinkage and heat the curing causes (at least with tested substances).

Slow curing Araldite was least shrinking although some minor shrinkage was visible. There war clear shrinkage with Araldite rapid. I’d estimate that because of the low magnitude of shrinkage the slow curing Araldite should be ok and not tear any gold wires off the sensor when curing.

Method was to fill up straight wall .22 long rifle cartridge casing with resin, even the resin amount to casing and observe how much resin surface was lowered during curing.

[D4N]

I don't know about your specific chip but I did notice when I took mine out to remove the IR filter that there are no exposed wires like there are on other chips. I would guess that there will be a lot of variety in chip designs, the problem is that means you can't even be sure that the bayer matrix will actually come off without damaging the sensor unless someone else has managed it on the same sensor.

Sent from my iPad using Tapatalk

[Gina]

I'll try and find the epoxy I used.  ATM I'm totally distracted in sorting out my main imaging rig and problems with power supplies and have not been working on this project for quite some time now.  I haven't forgotten about it and plan to get back to it once I'm up and running on my imaging systems using Atik cameras.  For me this project is mostly academic I'm afraid but there is the prospect of producing a half-frame size imager that's good enough for worthwhile imaging in parallel with my high spec Atik setup.

[Unassorted]

Well, after a good day or two of reading I finally reached the end. I've taken apart my 1000D and started the process of removing the cfa. Got most of it removed and went to take a test shot to make sure everything was still working correctly. Image came back as being completely black. It would appear that I hit one of the golden fibers while cleaning it and I have a new sensor on the way! Here is what my sensor did look like before attempting to take a photo and I'm still unsure on if I went too far down or not far enough.

[Gina]

I think I can see at least one, maybe two or three, broken gold wires on the right.  It's so very easy to catch these however much care you take as my graveyard of dead sensors will testify

[Thalestris24]

I can see lots on the right (top and bottom)! Might some low-melting point gluesticks dribbled on be used to protect them? Just a random idea...

Louise

[Dave Lloyd]

I'm sorry, this is far too scary for me!

[Nitrous]

I'll try and find the epoxy I used.  ATM I'm totally distracted in sorting out my main imaging rig and problems with power supplies and have not been working on this project for quite some time now.  I haven't forgotten about it and plan to get back to it once I'm up and running on my imaging systems using Atik cameras.  For me this project is mostly academic I'm afraid but there is the prospect of producing a half-frame size imager that's good enough for worthwhile imaging in parallel with my high spec Atik setup.

You may already have this link, but the 1100D sensor cover glass can be removed with heat.

This is a very interesting thread!

Thanks

Doug

[Gina]

You may already have this link, but the 1100D sensor cover glass can be removed with heat.

This is a very interesting thread!

Thanks

Doug

Yes, I have indeed seen that video and tried it without success - the sensor was dud afterwards   I also tried this several time in different ways.  I think I succeeded once but managed to destroy the sensor at a later stage.  It's all written up on here somewhere

I think it quite likely that I might have another go at 1100D sensors some time in the future

[spokeshave]

Greetings all:

I have been curious about hte possibility of debayering a DLSR sensor for a while now and finally came across this thread. It seems that the entire world-wide knowledge base on the subject is contained here!

I have read all 93 pages of this discussion and am amazed at the determination that some of you have exhibited. Kudos to everyone for being nearly the sole pioneers in the world in this endeavor.

After reading of the successes and learning opportunities, I have given the idea of debayering a lot of thought. It seems like chemical methods have been largely unsuccessful at least on the Canon sensors. Physical scraping or abrasion seems to be the best approach so far. As I understand it, the microlens layer is easy to remove, ut the CFA layer can be fairly tenacious. Given the relative successes of several scraping and abrasion methods that leave the layer underneath intact, it seems like a logical conclusion that  the CFA layer is a good bit softer than the layers underneath. This leads me to consider an approach that I use in my other hobby - auto restoration. The analogy is the removal of automotive paint without harming the metal, fiberglass, plastic or carbon fiber underneath. The parallels are undeniable. One method is chemical stripping. Another is scraping and abrasion (sanding) but by far the most effective method is media blasting. I'm left wondering how well media blasting would work for debayering. It would not requiring physical touching of the sensitive photosite layer and if the correct media and media grit is selected, I can imagine that it would quickly, easily and effectively remove the entire CFA layer while leaving the layer underneath completely unharmed.

I would like to test this hypothesis and will eventually, but I thought I would share the idea here first, since right now my time is very limited and it may be a few weeks before I can get a chance to try it out. I would expect that an "air eraser", which is essentially an air brush designed for precision media blasting of small parts, would be the ideal tool. They are inexpensive and there are many grits and compounds available.

I would expect that the work flow or media blasting would be similar to the more traditional approaches. I think the gold wires would need to be encased in epoxy to protect them from both the media and the air currents. I suppose one could then experiment with physically removing the microlens layer or trying to "erase" it with the air eraser. Given the descriptions of it being "gel-like" it may be more difficult to remove with an air eraser. The CFA layer on the other hand appears to be an ideal candidate for blasting. Since many of you have dead sensors laying around, you would have some good candidates for experimentation.

Anyway, I think it is an approach that may have merit and I will experiment with it as time permits. Meanwhile, if any of you have an opportunity to try it out, please let me know how it goes.

Tim

[Unassorted]

I think I can see at least one, maybe two or three, broken gold wires on the right.  It's so very easy to catch these however much care you take as my graveyard of dead sensors will testify

Yea =( I knew going in with that sensor that I would have an 85% chance of it being dead after I even touched it but my new sensor should be in today and I already have a much better set up and lighting so that I won't hit any of the golden wires.

[Gina]

Good luck with your second sensor

[Unassorted]

Thanks Gina!

Ill update either tonight or tomorrow on the progress and or success/failure!

[Unassorted]

So a bit of an update!

New sensor came in. I was a bit more careful this time around and the results are pretty appealing.

Here is the camera with the glass 100% off with no breaking in it - http://i.imgur.com/MKKrRu3.jpg

Starting the removal - http://i.imgur.com/Mb9P23M.jpg

Alot more done and this was the end of the day for me - http://i.imgur.com/8JzGIBL.jpg

Flat image(yea, nice ol scratch in the middle and some tiny ones on the botton but flats galore should help) - http://i.imgur.com/Xanb6dh.jpg

My hand! - http://i.imgur.com/qpi5BO2.jpg

So far this has turned out quite well. Going to tackle more of the removal and then polish it up when I have some free time. Hopefully i can make the scratch less prevalent after removing the rest of the CFA. Currently in my forecast I am seeing Friday and Saturday as nights that will be somewhat free of clouds so I plan on getting first light with the mono mod then! 

[Herra Kuulapaa]

The Nikon D3100 sensor is almost ready but I'm wondering one thing of it.

That reddish/orange color can be seen on the sides but otherwise the whole sensor is greenish. (that different tone area inside green tone is just a lamp reflection).

I have worked the green area quite much but still it doesn't change. Should I keep going on the green, what do you think?

It's dirty I know, still in process...