Debayering a DSLR's Bayer matrix.

[Pompey Monkey]

It's top of the list for me   And in fact, the only cause of dead DSLRs.

I've been dipping into this thread since my love of astronomy has been rekindled by modern AP. I am seriously tempted to kill debayer the sensor in my 550D (if the Ebayer ever sends it) too!

Edit: Ebayer/debayer? Coincidence? lol.

Edited July 5, 2015 by Pompey Monkey

[Atreta]

this may be a daft suggestion, but has anyone tried putting the wooden scraper inside a spring loaded biro pen or something similar? that way it would be much harder to put too much pressure on the cmos surface... just a thought.

i thought about something like a micro lathe that would automatically scrape from a fixed(programmed) starting and end position.

don't know if there is such a thing though.

Frank

[flolic]

Not lathe but CNC mill (or router).

[Herra Kuulapaa]

I doubt that a CNC machining could be directly used in color filter removal. Tolerances are so small (and/or the cost of such precision machine would be sky high).

However, use of the CNC like machine in slowly polishing the surface would be a whole another idea.

[flolic]

I agree, that CNC machine Z axis should have tolerances and precision in a sub micron range.

I don't have that precision on my machine, but it should be quite capable in surface polishing if I ever produce precision spring loaded tool...

[Atreta]

I doubt that a CNC machining could be directly used in color filter removal. Tolerances are so small (and/or the cost of such precision machine would be sky high).

However, use of the CNC like machine in slowly polishing the surface would be a whole another idea.

that was what i meant.

[RobRob]

And another black 350D, this time the glass didn't want to come off at all. It scattered each time I tried to lift it. Eventually I was able to remove most of it (some glass left on the edge). This time I stayed away from any edge during the scraping. After scraping, cleaning and testing it it turned out to be broken as well. My theory is some glass particles on the sensor must have cut to deep grooves during the scraping...

[Dave_D]

I doubt that a CNC machining could be directly used in color filter removal. Tolerances are so small (and/or the cost of such precision machine would be sky high).

However, use of the CNC like machine in slowly polishing the surface would be a whole another idea.

getting the sensor absolutely square would be a pain though. wonder if a 'wedge of cheese' profile sensor would still work

[RobRob]

Tried to remove the microlenses and CFA layer on an additional 350D sensor using soda blasting, and... black again. The process was really easy and fast, I created a small mask to ensure only the image area would be cleaned but nevertheless apparently more then needed got removed in the process.

I'll investigate the sensor later today using the microscope to see what has happened (any gold wires might have been disconnected due to the airflow???)

[RobRob]

Here a photo of the blasted sensor showing the soda has removed too much of the blue edges although these were covered by the mask. All the gold wires seem to be still attached.

[russellhq]

The sensor looks very clean, shame it didn't work. Maybe it's worth taking apart and re-assembling again if you haven't already done that.

If you are going to try again, maybe try only a small area in the centre of the sensor. That should confirm one way or another if the soda is too abrasive I should think.

[Herra Kuulapaa]

I bought Canon 1200D few days ago. It didn't stay long in stock though.

Some people may find following flowers interesting.

In color:

In mono with IR, color and UV:

In 950nm IR:

I'm still missing pure UV pass filter, they're so damn pricey

[Stub Mandrel]

Try some buttercups, you may need to stop down but the UV pattern should be quite striking.

[Xplode]

Not a DIY job, but it's still a debayered DSLR so i hope i won't get thrown out of this thread

Last year i had an idea about buying a 6D and get it debayered, around new year there were a special offer for the Canon 6D (N) (N means without wifi and GPS) with a price of 20% under the regular model and on top of that a drawing board that was worth a good amount.

The special offer made the camera cheap compared to a regular one so i decided to go ahead and buy one to get it modified. The drawing board was sold for ~20% of the cost of the camera.

I sent the camera off to Brent Oliver (Hypercams and Mods) in the US that has debayered Canon cameras with great success.

Unfortunately the conversion took a very long time because he recently had a baby (and he also has a kid from before) and there was also some problems with Baader filters that weren't perfect (they had streaks that he couldn't remove by cleaning) so it took almost 7 months from i sent the camera to i got it back, today i finally got it! 

This is probably the worlds first 6Dm (monochrome)

Brend did a very good job and  autofocus works perfectly and there's very little left on the edges so it can be used for regular photography perfectly fine.

I have made some testshots and picked the best one that was of my cat sitting outside.

I opened in RAWTherapee and saved without debayering.

Resized to 25% and not cropped.

RAW file on Dropbox

Edited July 14, 2015 by Xplode

[RobRob]

@Xplode: Now that must be an awesome camera to use for narrowband!!! What's left is to put some cooling facilities on it and you are good to go!!!

[Xplode]

Narrowband and Luminance is what it will be used for.

I have thought about sending it off to CentralDS in South Korea to get cooling on it, but not sure if it's worth it as it's pretty cold in Norway anyway during astronomy season.

[Tr2sa]

 Removing CCD/CMOS cover glass - diamond burr works well for milling through the glass. Does work well also for the ceramic casing. Mild ultrasonic cleaning is needed to clean glass dust off and one has to watch, where the dust accumulates - it can break bonding wires if compacted by the burr. Cutting without coolant with slow feed is perfectly ok.

 Some processes seem to use reflow to give final form to microlenses, so rising chip temperature to 120...170c should make mechanical removal way easier without compromising chip bond to casing. It is a theory, I have not tested it myself on Canon sensors.

http://www.google.com.ar/patents/US7388270

 About microlenses - highly speculative, but maybe it is possible to recover some efficency by recreating lens array. Alignment (in XY and Z) will be hardest to solve probably. The process described below is too coarse but maybe there are other cheap non-lithographic shortcuts to create array.

http://cameramaker.se/microlenses.htm

[Herra Kuulapaa]

Here's a monochrome D5100 sun test, with proper filtering of course. I don't have a comparison, but I have a feeling that the mono clearly out performs a regular dslr.

Larger image: http://www.kuulapaa.com/Solar/1872015_AG.jpg

Edited July 19, 2015 by Herra Kuulapaa

[Wayne_Li]

Did anyone tried this on D200? I am thinking to convert my old D200...

[Herra Kuulapaa]

Some older Nikon sensors have more easily releasing resin than the modern ones, but I don't really know about the D200. Based on the fact that the D200 may not be your active camera, I say: Go ahead!

BTW, I just finished the first proper tests of monochrome D600 and it really rocks!!

Full frame Monochrome Nikon D600 with true dark current mode and 12nm Ha, 8x11min

Edited August 9, 2015 by Herra Kuulapaa

[Xplode]

Great image Herra!

I'm doing my first real test of my monochrome 6D just now, 10min on the heart nebula with 7nm Ha filter with my 100Q. It's a little to warm so pics have more noise than i would have liked...not sure how it will end up after 2-3 hours imaging time.

[Herra Kuulapaa]

Issue with the Nikon D600 and D800 sensors (and later models) is that the cover glass is bonded with resin/adhesive, which has a release point above the heat kill temperature of the sensor. This means that you are practically heating glass hotter than what sensor photosites can take if you are planning to use heat to release the cover glass.

But now there is an alternative method, sadly it's not diy, but at least there is one: I've got some good results with micro milling. Procedure is rather simple; sensor glass is milled off with a microtool from three sides and then glass fixed to a holder to prevent it falling on sensor surface. After that the final side is cut off. Problem is that those machines and cutting tools are insanely expensive and I don't have a production scale access to them. Sensor glass is actually eating the tools and not the other way around

Edited August 29, 2015 by Herra Kuulapaa

[Herra Kuulapaa]

Ok, I'm happy to inform you guys that I have a fully working and exceptionally flat field full frame Nikon D800am (astromono;) on my desk. Flat field because the sensor is absolutely flawless.

And in additionally there is the same type D600am, yep I'm using that "am" on both of them to create confusion to model D810a

Process was very difficult and I struggled with the cover glass removal even though the micro CNC was used. I found out that that there really is no guarantee with that method either and it resulted one dead D800 sensor But the other one is success

Left reads D800 and right one D600

[Fred_76]

Congratulations Herra !!!

What's next ? Will you debayer a Sony A7s ???

[Herra Kuulapaa]

I don't know if I can justify the cost to myself. This starts to be a bit too expensive hobby even though I like to experiment

I haven't really tested the D800am yet so that comes next.

D600am is a solid performed though:

http://stargazerslounge.com/topic/250891-nikon-d600-monochrome-images/

Edited September 29, 2015 by Herra Kuulapaa