Debayering a DSLR's Bayer matrix.

[Gina]

I thought of gluing on something solid too - something that would withstand the heat used to break the bond. Seems there was a huge difference between the two 1100D sensors. The glass on the first one came off quite easily with a small amount of heat applied whereas the second simply wouldn't budge. I might have applied more heat than I thought.

Yes, I cleaned the 1st sensor with polish and IPA and wiping off with cotton buds. I didn't use distilled water but I think IPA would be sufficient.

[Gina]

Oh dear

I feel for you having such bad luck after all the effort you've put in its a lot of dead sensors and expense so far on this thread, its definately given me second thoughts about giving this a go with my next ebay camera, I might just good old fashioned mod it!

Thanks Chris But they do say "no pain - no gain" But the general rule is - don't take on any of these mods if you can't afford a failure. For me this current exercise is pretty much academic as I already have mono astro CCD cameras and don't need DSLRs for imaging though if I can manage to produce something that works quite well I could make up a separate widefield imaging system and have two independent imaging rigs - that's for the future though.

Having said that it certainly is frustrating when after having got as far as debayering a sensor and getting a mono image the next one fails at the very first step

Edited August 21, 2013 by Gina

[Cath]

Taking a hammer to an imaging sensor was novel

[Gina]

People take hammers to all sorts of things though not always with the intention of doing good Computers in particular have often been under threat

I did wonder about seeing if I could find a glue that was stronger than that used to bond the cover glass onto the sensor and see if I could yank it off

[Gina]

Sensor no.2 needn't be a total loss - I can use it to experiment on for removing the CFA though I'll only be able check it from the front. I have removed that jagged bit of glass by getting the craft knife under the end and breaking it off. It was amazingly strong - I had to lever it off, pointing it away from me.

No more metal scrapers - I'll see if I can find some wood that will do or use the paint abrasive. I guess ordinary household abrasive cleaners are too coarse.

[JohnC64]

No more metal scrapers - I'll see if I can find some wood that will do

- Time to go down the local shop and buy a Magnum

[JamesF]

- Time to go down the local shop and buy a Magnum

I thought a hammer and chisel was a bit heavy-handed, but shooting it is just going too far...

James

[Andrew*]

TNT is the way to go.

[Gina]

Brilliant James - love it

[Gina]

TNT is the way to go.

Is that to blow the glass off or the CFA? Reminds me of the van in The Italian Job with Michael Caine - "You're only supposed to blow the blumming doors off" (The board software has changed a work in the quote.) Edited August 21, 2013 by Gina

[Andrew*]

Didn't get the Magnum joke first time!

Is that to blow the glass off or the CFA? Reminds me of the van in The Italian Job with Michael Caine - "You're only supposed to blow the blumming doors off" (The board software has changed a work in the quote.)

Exactly! The glass, just run it round the edge!

[Gina]

I've checked the thickness of the cover glass that came off the 1100D sensors - 0.5mm.

[Chris]

I've checked the thickness of the cover glass that came off the 1100D sensors - 0.5mm.

0.5mm sounds quite thin, Do you have a diamond ring so you can try scoring the glass to remove the next one

Edited August 21, 2013 by starfox

[Gina]

0.5mm sounds quite thin, Do you have a diamond ring so you can try scoring the glass to remove the next one

No, I don't have a diamond ring I'm afraid - never been that rich or had that rich a boyfriend Have to say, I would rather not use heat on the glass - there seem quite a small margin for error. Maybe I'll get a glass cutter.

[Gina]

I've ordered a diamond glass cutter - should arrive Friday - http://www.ebay.co.uk/itm/DIAMOND-GLASS-CUTTER-SPECIAL-OFFER-/270333293581?pt=UK_Hand_Tools_Equipment&hash=item3ef11eb40d

[Chris]

No, I don't have a diamond ring I'm afraid - never been that rich or had that rich a boyfriend Have to say, I would rather not use heat on the glass - there seem quite a small margin for error. Maybe I'll get a glass cutter.

Maybe glue a retangular bar to the glass (as previously suggested) and use it as a fence to score around the outside with one of these:

http://www.clasohlso...CFfIPtAod6CoA0A

They do cheaper ones with tungsten carbide for about a fiver.

Then the metal rectanglar bar could be used to lever the glass once scored.

Edited August 21, 2013 by starfox

[Chris]

ah beat me to it

[alistairsam]

what was jtw's method again to get the glass off the 600d, not sure if they debayered a 1100d..

the 600d was hard to get off and I had to break it

rottweiler ..?

[Gina]

You certainly need that "dogged determination" in this business

[Gina]

I've practiced carefully applying epoxy resin (Araldite Standard) to the gold wires and up to and a little over the edge of the sensor pretty much covering the blue border. It's very wakward stuff to work with and tends to go anywhere but where you want it. The result is far from perfect but at least I've got some more practice ready for the next go. It certainly seems that the consensus of opinion is that applying epoxy resin does no harm and, of course, protects the delicate wires and other parts while the CFA is removed.

[russellhq]

Gina, do you think a metal rectangle, slightly smaller than the area of the glass could act as a heat stabiliser? You could apply your hot air to the heat sink rather than the glass which might give you more controlled deposition of heat?

[Gina]

Gina, do you think a metal rectangle, slightly smaller than the area of the glass could act as a heat stabiliser? You could apply your hot air to the heat sink rather than the glass which might give you more controlled deposition of heat?

Not sure about that. I think it might be difficult to get a good heat transfer to the glass though thermal compound might work. The other problem might be getting the metal to cool quickly enough so that the sensor doesn't get too hot. I think a small area of concentrated heat would be better and not pass on so much heat to the sensor.

I don't think I'm happy with any heating method though. On current form it works with some sensors and not others and the difference between not applying enough heat to break the bond and too much for the sensor is rather small. And I think that even if the sensor appears unharmed by the heat it could have shortened it's life. The life of semiconductors is shortened by heat and particularly by heat cycling. So bringing in my previous knowledge of semiconductors and my experience of these cover glasses and sensors in particular, I think I'll abandon the heat method. I believe the initial apparent success of the first 1100D sensor gave me a false sense of security.

Edited August 21, 2013 by Gina

[astromerlin]

Dont shoot me, a mad idea perhaps. Depending on the thickness of the bonding material, how about trying a hot wire to slice the bond from the glass?

[Gina]

Dont shoot me, a mad idea perhaps. Depending on the thickness of the bonding material, how about trying a hot wire to slice the bond from the glass?

Great bit of lateral thinking No, I won't shoot you Never thought of that. I think it would have to be very fine wire though - I think the glass and sensor frame might take too much heat away. A heated thin blade might work.

[Zakalwe]

Dont shoot me, a mad idea perhaps. Depending on the thickness of the bonding material, how about trying a hot wire to slice the bond from the glass?

Or bend the tip of a fine pointed soldering iron to 90 degrees. You could then use this to melt the glue under the glass?