Debayering a DSLR's Bayer matrix.

[fimpen]

I've been searching the net and this seems to be the only place that DIY debayering actually has been tried. Kudos to you for that! I really hope this tread will continue as I am very eager to know how to do it. i have an old 450D waiting to be butchered. So 12dstring, how are things going?

[Photosbykev]

JTW Astronomy have just announced they can debayer DSLR sensors for 200 euros. Sounds very interesting

[12dstring]

Sorry guys, I've been super busy over the past couple of months so unfortunately haven't made much progress, but I'll get back to it as soon as I can.

Can you post a picture of your paintbrush handle tool? Also what power magnification do you think would be good for this operation?

Hi Tristan, I've attached a picture of the scraper tool. It's simply the end of a paintbrush sharpened with a knife. The plastic is hard enough that the sharp edge holds well and can easily scrape away the layer (unlike other plastic things I tried, or toothpicks), but not hard enough to scratch the underlying surface (as the metal scraper did).

Most of the time I was using a 4x objective, switching to 10x sometimes for a close look, with a 17mm Baader Hyperion eyepiece - they make great wide microscope eyepieces.

I'm still looking out for the perfect smaller tool to get to the edges, I don't want to ruin it having made it this far in the search for perfection.

Also debating whether to turn it into a dedicated astro cam and rehouse/cool/amp-off, or keep it as a normal camera - could be pretty cool for daytime IR photography.

[grzegorz]

Hello everybody,

Amazing topic! Really great.

Being inspired by the idea, I tried to use 'chemical approach'. I took ccd sensor from damaged Canon A75 camera.

First I ground the edges of covering glass with Dremel and little diamond bit, and then removed the glass.

As a solvent I used mixture of dichloroethane (C2H4Cl2) and trichloroethylene (C2HCl3) (3:1 proportion).

All layers of Bayer's array (together with microlenses) came off rather easily, not very fast, but after 10 to 15 minutes I did manage to remove them with cotton swab moistened with the mentioned mixture.

There are some problems around the edges - as the swab is thick and I didn't want to damage tiny wires, but I think this could be overcome with some finer tool.

Anyway I'm going to apply this procedure to another, working camera (also with small sensor unfortunately).

regards

grzegorz

[GlassWalker]

Interesting stuff... in a quick look dichloroethane can be got from ebay, but it does sound a bit of a nasty chemical to have around! Still, this talk is making me want to have a go at this myself with whatever I have around at home.

[billhinge]

I just converted a Lumix GF1 to full spectrum at the second attempt but in the process fried the first GF1

So I decided to dissect the sensor and debayer it

The glass is held on by a ring of rubber/silicon glue around the edge of the sensor plate, this comes off with a razor blade under the glass gently cutting the glue

I tried several solvents for removing the bayer layer from alcohol to toilet cleaner

Finally I tried medicinal compound and it comes off in seconds, perhaps too quick as I'm not sure how far down I went but I was rubbing reasonably hard - don't try this at home

I was able to get to the gold layer easy with gentle rubbing

see pic which shows the various layers, doesn't show up well in the pic but there is a red, green, gold and silver set of layers

[billhinge]

In case anyone wants to try this, autosol removes the CFA very easily, but don't rub too hard.

There is also a gold and silver autosol which is the finest and is designed for cleaning gold & silver including thin plate

I also found colgate toothpaste works very gently ;-)

[nightvision]

It seems many approaches sort of work and there are also structural differences between webcam sensors and DSLR sensors so I guess it remains an inexact (and potentially expensive) science for now, it's great that we have so many pioneers.

Edited April 26, 2013 by nightvision

[billhinge]

just checked toothpaste again, very gentle on a cotton bud, doesn't even seem to scratch the gold layer at least to the naked eye

plus it smells minty and should wash off in distilled water

[nightvision]

Makes me wonder what it does to your teeth!

[Gina]

It wears them away.

[billhinge]

I decided to try autosol and old toothbrush to go a bit deeper - you could probably use the gold polish or toothpaste but wouldn't be as quick

appears to be a layer of glass under the layers

this thread seems to imply that the CFA is printed on to glass and the actual sensor is underneath. Also says that the companies that do debayering, even their products leave streaks at $2000 a pop! The scratches on the glass were from my earlier attempts with a sharp tweezer point

http://www.cloudynights.com/ubbthreads/showflat.php/Cat/0/Number/5606114/page/0/view/collapsed/sb/5/o/all/fpart/5/vc/1

[12dstring]

I gave toothpaste a try when playing around with the broken sensor. It didn't seem to help an aweful lot - although I was going more for scraping away the CFA layer rather than polishing it down.

It did get rather messy though. The bright microscope light dries it out pretty quickly which didn't make it easy to clean off.

[billhinge]

Hi Dave

Perhaps its the different methods of different manufacturers? Autosol is easy but quick, toothpaste works for me but is much slower.

I'm thinking I may get another cheap GF1 in a month or two and go for a full debayer, I'm confident I can get down to the bottom layer without damage, the GF1 seems easier than the canon to open up

Can I ask, even by your scraping method that the sensor is fully working and there is no issue reading the raw file , if so thats great since it means the sensor is tougher than we think

[12dstring]

It's easy to imagine that different manufacturers, and even models would have different types of layers in the sensor.

Interestingly in the CN thread you linked to Jim said that Nikons were easier to work with than Canons. Seeing how we're into relatively uncharted territory each new model could throw up suprises.

My first sensor died when I broke two of the tiny gold connecting wires when removing the cover glass, so I used that to practice the technique. With the second attempt I've had no problems. Electrically it's all fine, I can't see any problems with the sensor at all at the pixel level using DCRAW to convert from RAW.

The layer below the CFA is pretty tough, I suspect it's a glass 'planarizing layer'. That said it's possible to scratch it with the metal pick I first used, and this will leave a permanent mark - although hasn't appeared to damage the photodiodes underneath at all.

[billhinge]

Cool, glad yours works. I gives me an incentive to have a go when I get a new s/h one (I've spent my budget this month). There don't seem to be any exposed wires on mine, just autosoling the rest of the the cfa off to the edges, nice and shiny now. looks like a very good diffraction grating for a diy spectroscope ;-)

[billhinge]

Decided to pull the ribbon cable off, it is indeed a layer of glass

One thing puzzles me though, won't removing the CFA break a circuit somewhere, I know that some circuit failures are non fatal. Any thoughts?

[12dstring]

Be very glad you don't have the tiny connecting wires, they're the only thing that makes me nervous!

The circuitry is on the bottom layer of the sensor, where the photodiodes are. The CFA is just an array of glass filters on top of that, it's not connected electrically in any way, and the camera electronics will unaffected.

[billhinge]

cool, I'll order some gold & silver autosol from amazon http://www.amazon.co...ds=autosol gold

https://autosol.com/...a-crystal-face/

(ps did some research, solvent for removing dry autosol is stoddard solvent also called white spirit)

Edited April 26, 2013 by billhinge

[billhinge]

I decided to take a screwdriver to the bottom glass layer, unfortunately I don't have a microscope but under an eyepiece it looks like the glass is the sensor top and the wires are in the glass, there isn't much under the glass

this seems to be confirmed by the article I found here http://www.cameratechnica.com/2011/06/23/technology-demystified-backside-illuminated-sensors/

it is glass since it shatters etc, but it seems to have a diffractive quality so I'm guessing its the millions of tiny wires inside?

[tonedeaf]

sorry to highjack the thread so late...

...@ all the scrapers and modders here...hats off!

i thought recently about something like that (sleepless nights haunt me ) but this is a thing i have to try...maybe on a cheapo 350D

rgds

[Gina]

Just don't try it with the 1100D - the sensor cover glass is welded on

[tonedeaf]

thanks gina, i appreciate your efforts with the "sticky" glas cover

as i have seen the 350D seems to be easier...

maybe i´ll do a cooling mod too...

mmh...peltier etc...don´t know if a temperature regulation would be needed...

don´t want to dissmantle my 550D-allround workhorse

[Gina]

No, I wouldn't recommend messing with your best camera.

[billhinge]

Think I killed my spare camera removing the CFA , I'll have to wait another month or two till a cheap one comes up on ebay

Camera switches on, no error message, menus work and playback but no live view or autofocus

Shutter works and saves raw file but it just contains noise when stretched

anyway, off to do full spectrum conversion on my other camera, my quartz filter arrived this morning