Debayering a DSLR's Bayer matrix.

[Luis Campos]

Hi,

I made a coulpe videos showing the CFA removal process, here goes:

Cheers,

Luís

[pixueto]

Thank Luis.

I'm using a wooden paint brush. I don't seen to be able to make the first break into the CFA. I know you used a needle but I'm concerned about leaving scratches on the sensor. Any suggestions? Tips? Maybe you could show us a video when you started removing it.

Thanks

Edited July 26, 2013 by pixueto

[alistairsam]

Thank Luis.

I'm using a wooden paint brush. I don't seen to be able to make the first break into the CFA. I know you used a needle but I'm concerned about leaving scratches on the sensor. Any suggestions? Tips? Maybe you could show us a video when you started removing it.

Thanks

Hi,

it needs a bit of pressure but not too much of course.

if the tip is coarse enough it should bite into the cfa layer.

i noticed the 1000D was a lot harder than the 350D.

I just got delivery of my 600D. man is it a nice camera. feeling nervous about debayering it. replacement sensor costs almost as much as the camera.

but it has to be done.

i'll be doing it over the weekend and will post results.

wednesday should be clear out here, so hopefully the 350D and the 600D (if it goes well), should see first light.

Luis, since you got the sensor glass from your damaged 350d off in one piece, I'd suggest sticking that back on to reduce your condensation issues.

the risk with leaving the sensor open is that the moisture could cause a short circuit on those golden wires. so not advisable running it if there's going to be moisture.

I would suggest keeping subs to 3 or 4 minutes and using darks rather than cooling it. I'm thinking of building a custom body if I'm going to do the cooling, but that's a long way off.

I've also been thinking of adding an ED80 alongside my 10inchF4 on the NEQ6. that way I could use the QHY8L on the ED80 and capture RGB and simultaneously capture mono "L" with the B&W 350D/600D on the 10inchF4. that should reduce imaging time for LRGB. but differential flexure could be an issue. we'll see.

btw, you have some nice round stars and good tracking.

Cheers

Alistair

[Luis Campos]

Thank Luis.

I'm using a wooden paint brush. I don't seen to be able to make the first break into the CFA. I know you used a needle but I'm concerned about leaving scratches on the sensor. Any suggestions? Tips? Maybe you could show us a video when you started removing it.

Thanks

Hi,

I used the tip of a hipodermic needle lying at an angle, but with allmost zero pressure, if you shake you will sratch the sensor, as alistair said is better to start with the wooden tool try to cut it in a sharp angle, do several attemps and you will eventually be abble to crack open the CFA.

Unfortunelly I only remembered shooting the video half way through...sorry

Cheers,

Luís

[Luis Campos]

Hi,

it needs a bit of pressure but not too much of course.

if the tip is coarse enough it should bite into the cfa layer.

i noticed the 1000D was a lot harder than the 350D.

I just got delivery of my 600D. man is it a nice camera. feeling nervous about debayering it. replacement sensor costs almost as much as the camera.

but it has to be done.

i'll be doing it over the weekend and will post results.

wednesday should be clear out here, so hopefully the 350D and the 600D (if it goes well), should see first light.

Luis, since you got the sensor glass from your damaged 350d off in one piece, I'd suggest sticking that back on to reduce your condensation issues.

the risk with leaving the sensor open is that the moisture could cause a short circuit on those golden wires. so not advisable running it if there's going to be moisture.

I would suggest keeping subs to 3 or 4 minutes and using darks rather than cooling it. I'm thinking of building a custom body if I'm going to do the cooling, but that's a long way off.

I've also been thinking of adding an ED80 alongside my 10inchF4 on the NEQ6. that way I could use the QHY8L on the ED80 and capture RGB and simultaneously capture mono "L" with the B&W 350D/600D on the 10inchF4. that should reduce imaging time for LRGB. but differential flexure could be an issue. we'll see.

btw, you have some nice round stars and good tracking.

Cheers

Alistair

Hi Alistair,

On my practice sensor I also broke it but only in half, it wasn't all shattered as this one but useless anyway...

On the working one I sealed it with a thin line of silicone all around the filter case, but obviouslly there was air trapped inside and this little amount of air contains some H2O

I will disassemble the thing again seal the sensor with hot glue and then purge the filter case with Nitrogen, I think this will solve this problem

I'm itching for your "first light" fingers crossed here buddy!

Yes, the great thing about this is having to cameras one mono and other color and then merge the results, no need for color filter wheels

Cheers,

Luís

[pixueto]

Hi,

it needs a bit of pressure but not too much of course.

if the tip is coarse enough it should bite into the cfa layer.

i noticed the 1000D was a lot harder than the 350D.

I just got delivery of my 600D. man is it a nice camera. feeling nervous about debayering it. replacement sensor costs almost as much as the camera.

but it has to be done.

i'll be doing it over the weekend and will post results.

wednesday should be clear out here, so hopefully the 350D and the 600D (if it goes well), should see first light.

Luis, since you got the sensor glass from your damaged 350d off in one piece, I'd suggest sticking that back on to reduce your condensation issues.

the risk with leaving the sensor open is that the moisture could cause a short circuit on those golden wires. so not advisable running it if there's going to be moisture.

I would suggest keeping subs to 3 or 4 minutes and using darks rather than cooling it. I'm thinking of building a custom body if I'm going to do the cooling, but that's a long way off.

I've also been thinking of adding an ED80 alongside my 10inchF4 on the NEQ6. that way I could use the QHY8L on the ED80 and capture RGB and simultaneously capture mono "L" with the B&W 350D/600D on the 10inchF4. that should reduce imaging time for LRGB. but differential flexure could be an issue. we'll see.

btw, you have some nice round stars and good tracking.

Cheers

Alistair

Thank you Alistair. I'm finding it very difficult to start breaking the bayer array. If I'm not mistaken, this is the first a 450D sensor that is being mono converted, right? Maybe this sensor is harder than the 350D or the 1000D? Most likely it's me not doing it right, though.

[alistairsam]

Thank you Alistair. I'm finding it very difficult to start breaking the bayer array. If I'm not mistaken, this is the first a 450D sensor that is being mono converted, right? Maybe this sensor is harder than the 350D or the 1000D? Most likely it's me not doing it right, though.

you could try breaking through it with a plastic tool as it'll be slightly harder than wood. but watch out for scratching it as its your working sensor.

other option is to break through the blue area on the sides that don't have the golden wires. the cfa extends all the way there.

i don't think the 450d would be very different from the 350d, so it might just be your pressure or tool. try sharpening it in different ways. as in not too pointed but sharp

cheers

Alistair

[pixueto]

Well, I did just that. I used a needle to break the bayer matrix in the blue area (which is not used for imaging) and continued from there with the wooden tool. I have removed all the microlenses and the CFA as far as I can see. Now, I'm still waiting for the microscope but to me the surface surely should now be full of scratches, shouldn't it?

Edited July 26, 2013 by pixueto

[Luis Campos]

Well, I did just that. I used a needle to break the bayer matrix in the blue area (which is not used for imaging) and continued from there with the wooden tool. I have removed all the microlenses and the CFA as far as I can see. Now, I'm still waiting for the microscope but to me the surface surely should now be full of scratches, shouldn't it?

Those are great news, you can use an eye piece to check for the surface, try an inverted 20 or 25mm eyepice and place it near your eye and near the sensor, focus by aproaching the sensor, you should get a highly magnified image, i did that, works good

I'm very interested in 450D adventure, that is a great deep sky camera out of the box, very low noise, 14 bit, live view, etc...in mono that should be a serious machine

Cloudy here now and looks like the weekend is not good...Murphy's law at it's best

Please keep us updated!

Cheers,

[pixueto]

Thanks Luis, Thanks Alister. I think I still have a lot of tiny bits of the bayer array left. The eyepiece helps Luis but not a lot. When I look at the golden surface at an angle under certain light there seem to be scratches. I can't really tell without magnification. Well if that's the case, they are all over the place! I wonder if they will show up in the microscope or in a flat frame. What about you guys? Can you also see that in your sensors? Your flat frames and photos are perfect, though. No sign of any damage (apart fron that little area in the right bottom corner)?

[pixueto]

Well guys, the microscope finally arrived and I got a chance o having a proper look at the surface. I'm afraid it's badly scratched. Well, that was the practice sensor. If I'm going to continue with this now it's going to cost me some money!

I put just enough pressure so I could remove the bayer array which makes me wonder about the wooden tool. I used a small paintbrush.

Luis, in your videos I can see a bit what you've used. Could you please explain what is it and where you got it? I would like to have a another go but this time only using the same tools than you so I can eliminate variables and explain the failure: is it me? the sensor? the wooden tool?

[Luis Campos]

Well guys, the microscope finally arrived and I got a chance o having a proper look at the surface. I'm afraid it's badly scratched. Well, that was the practice sensor. If I'm going to continue with this now it's going to cost me some money!

I put just enough pressure so I could remove the bayer array which makes me wonder about the wooden tool. I used a small paintbrush.

Luis, in your videos I can see a bit what you've used. Could you please explain what is it and where you got it? I would like to have a another go but this time only using the same tools than you so I can eliminate variables and explain the failure: is it me? the sensor? the wooden tool?

Hi,

Blimey...I'm sorry to hear that

That's strange, I also used a small paint brush handle, nothing special, I guess it's about the pressure applied on the surface, it has to be enough to remove the CFA but not enough to damage the underlayer, my practice sensor was too badlly scratched, so I'm sure you wil succeed in another try, but hey, why don't you try with that sensor and see how it goes? Maybe flats will take care of the scratches? Yep, the 450D sensors are quite pricey around 120 to 160 eur. on Ebay that's allmost the price of a complete body :/

On my sensor there was a very shiny surface once it was done, no visiblle scratches but I think that area you say was when I started scraping, I had to make some extra force to get it going at the begining.

Some more shots of the sensor for you to check, it may help:

Best wishes,

Luís

[alistairsam]

hi

I'm almost finished removing the cfa from a 600d sensor and let me tell you its not easy. its similar to the 1000d and by the sounds of it, similr to the 450 as well in that its a lot harder to remove than the 350d. the glass just would not budge so I had to break it with tape on top and that wasnt easy either

I'm not very sure if I've wrecked the sensor but I'll know tomorrow when I assemble it and test.

I dont think the scratches show up in images.

and to add to it the 600d sensor is spring loaded so has to be precisely aligned back.

we might have to stick with the older models with larger pixels as that could influence the thickness of the cfa layer

cheers Alistair

[alistairsam]

I meant to say the cfa layer itself is a lot harder to scrape off for the 600d

[Luis Campos]

hi

I'm almost finished removing the cfa from a 600d sensor and let me tell you its not easy. its similar to the 1000d and by the sounds of it, similr to the 450 as well in that its a lot harder to remove than the 350d. the glass just would not budge so I had to break it with tape on top and that wasnt easy either

I'm not very sure if I've wrecked the sensor but I'll know tomorrow when I assemble it and test.

I dont think the scratches show up in images.

and to add to it the 600d sensor is spring loaded so has to be precisely aligned back.

we might have to stick with the older models with larger pixels as that could influence the thickness of the cfa layer

cheers Alistair

Hope everything goes well buddy, fingers crossed here.

If you 600D works well it will be a HUGE achievement, that's for sure

Cheers,

[flolic]

Hi all, this is my first post on this forum

I just sucessfully removed cfa from 450D sensor

I used proven method, sharpened wooden paintbrush handle. It was hard to start scraping, but I started with needle and from then it was relatively easy.

I removed glass window from sensor with ease and in one piece. I used hot air rework station with 5mm nozzle, with temperature set at 250C. Just circle

around the edge of glass and when hot enough bonding glue will turn white. From then you just need to pick up glass with the tweezers. It's so easy.

I still need to clean up the edges and some dust.

Full size image on the next link. This is with original IR/AA filters in place, so it cean be even sharper when I remove them. Only histogram adjust, no sharpening.

https://dl.dropboxus...7/slike/1_1.jpg

Filip

[pixueto]

Hi all, this is my first post on this forum

I just sucessfully removed cfa from 450D sensor

I used proven method, sharpened wooden paintbrush handle. It was hard to start scraping, but I started with needle and from then it was relatively easy.

I removed glass window from sensor with ease and in one piece. I used hot air rework station with 5mm nozzle, with temperature set at 250C. Just circle

around the edge of glass and when hot enough bonding glue will turn white. From then you just need to pick up glass with the tweezers. It's so easy.

I still need to clean up the edges and some dust.

Full size image on the next link. This is with original IR/AA filters in place, so it cean be even sharper when I remove them. Only histogram adjust, no sharpening.

https://dl.dropboxus...7/slike/1_1.jpg

Filip

Do you think you could post a flat frame? I managed to get my 450D sensor full of scratches! Any tips, suggestions?

I wonder if that method for removing the cover glass could be used with a 1100D?

[alistairsam]

nice work filip.

that hot air is a good idea but I doubt it will work with the 1100d or 600d.

thats because when I removed the glass on the 1000d with a blade, I aplied pressure inward and lifted slightly at the same time and I could see I getting white very easily and quickly. whereas with the 600d, it wouldnt budge. filip, if you can get your hands on a 1100d or 600d sensor, do try with the hot air and let us know.

looks like there are generations where they changed the bonding method.

good to know the wooden tool is working well

cheer

Alistair

[flolic]

Do you think you could post a flat frame? I managed to get my 450D sensor full of scratches! Any tips, suggestions?

Flat frame here (stretched): https://dl.dropboxusercontent.com/u/21387397/slike/flat.jpg

I found 450D sensor very hard to scratch with wooden tool, in fact I did not make any...

That what looks like scratches are in fact leftovers of cfa matrix.

[flolic]

Alistair, I don't think bonding material is much different between generations. I tried this method with heating on 10D, 300D and now 450D sensors, and there is no noticable difference between them.

But you must heat that glass really well and it will just pop off. Don't be afraid, sensor can tolerate more than 200C without any problem

Filip

[pixueto]

Flat frame here (stretched): https://dl.dropboxus.../slike/flat.jpg

I found 450D sensor very hard to scratch with wooden tool, in fact I did not make any...

That what looks like scratches are in fact leftovers of cfa matrix.

Thank you very much. In this picture you can see I scratched the sensor badly. I was struggling to remove the bayer array so I couldn't apply less pressure really. Now, you obviously succeeded with the same sensor. I was using a wooden paint brush, though. Any suggestions, ideas? I'm at a loss here. I would love to convert a camera to mono.

[pixueto]

Hi,

Blimey...I'm sorry to hear that

That's strange, I also used a small paint brush handle, nothing special, I guess it's about the pressure applied on the surface, it has to be enough to remove the CFA but not enough to damage the underlayer, my practice sensor was too badlly scratched, so I'm sure you wil succeed in another try, but hey, why don't you try with that sensor and see how it goes? Maybe flats will take care of the scratches? Yep, the 450D sensors are quite pricey around 120 to 160 eur. on Ebay that's allmost the price of a complete body :/

On my sensor there was a very shiny surface once it was done, no visiblle scratches but I think that area you say was when I started scraping, I had to make some extra force to get it going at the begining.

Some more shots of the sensor for you to check, it may help:

Best wishes,

Luís

Unfortunately, I don't have a working 450D; that was just a sensor from a faulty camera. If I try again, I would go for a cheaper model as that would cost money now. However, I don't see the way forward. I seem to have damaged the sensor where all of you succeeded guys.

[pixueto]

Alistair, I don't think bonding material is much different between generations. I tried this method with heating on 10D, 300D and now 450D sensors, and there is no noticable difference between them.

But you must heat that glass really well and it will just pop off. Don't be afraid, sensor can tolerate more than 200C without any problem

Filip

Most people here think that isn't possible to remove that glass from a 1100D without damaging the sensor.

[Gina]

I haven't found anyone who has successfully removed the cover glass from the 1100D sensor and AFAIK the few firms offering mono converted cameras do not include the 1100D in their range.

Yes, although I've given up, I'm still reading this thread with interest I still have my super-cooled 1100D lying around and if I ever find someone has found a way of removing the cover glass without killing the sensor I might just have another go just for the sake of trying it. An OSC camera is of little use to me as my main interest is NB DSO imaging. For most galaxies a smaller sensor is perfectly adequate.

[alistairsam]

Hi Filip

if the solution is to heat the glass then I might give that a go

i must've bricked my 600d sensor. The camera starts up fine but when I try to take a pic, I get an "err 70"

from looking under the microscope I don't think I've damaged any of the gold wires but I'll have to check again.

i saw a big difference in the bond for the 600d and 1000d. 1000d popped off with just pressure and it turned white as I was lifting it. I took it out in one piece. But the 600d would not.

i had to stick tape, apply pressure sideways to cause a crack, then put a toothpick under the glass on top of the sensor to lift the glass in pieces.

this is pretty crude and I saw glass particles fall on the sensor between the wires so not recommended.

my next method will be to get a diamond tipped glass cutter and score a rectangle on the glass. I think getting it off in one piece is important so not to damage the wires.

but I'll try the heat method . Can you post a pic of your rework station

gina, do you have a thread on the cooling for your 1100d.

If i can achieve similar cooling I'll try a 1100d

alistair