pixueto

First attempt at Mars with the C9.25 and a Powermate 2.5. I think I can process it further but some artifacts are preventing me from that.

I think the best advice is to show the type of images that can regularly be achieved with a DSLR and a dedicated astronomical camera so those who are new to the hobby can decide if they want to pay for the extra expense of the latter option. This is what I've done with different modified DSLR cameras over the years: canon 1000D, 450D,  600D, 100D. Nothing special, really -in fact, some can be greatly improved by a more experience hand (vignetting, noise, etc) but it gives an idea of the kind of stuff one can achieve with £200-£300 invested in a DSLR as opposed to ££££s put into a dedicated astro camera.

Hi all,

Just a few pictures I managed to get of the Aurora in Iceland last week.

Gina, who knows? Maybe the fact that in the 1100D the CFA is more firmly attached, could be an advantage in debayering using the dremel and felt tips method. Maybe a perfect finish can be achieved. However, the fact that it's a more expensive camera and the difficulty removing the cover glass really puts me off.

I agree that the blue edge of the sensor seems to be the part to avoid at all costs.  My experience too has been that many of my sensors have worked until I've tried to get that last bit of image round the edge clear of CFA and just strayed over the border.

So from my experience, the causes of failure have been :-

1. Breaking the gold wires
2. Straying into the blue border
3. Going too deep into the imaging area

I shall definitely be working on the 450D sensor and will try to confirm that this sensor can be successfully converted to mono in the central area - leaving a good margin round the edges.  Unfortunately, epoxy resin is far too soft to afford any protection to the sensor border and there is no benefit in covering this with resin.  In fact it would be detrimental in that you couldn't see where the critical border starts.

Actually, once the epoxy sets, if you apply frog tape over that epoxy protecting the blue edges, you will have a very good protection against the dremel and felt tips. You can apply two or three thin stripes of frog tape for extra protection and then you can run the dremel an felt over that area to debayer the sensor entirely. This is what I learnt with my last sensor. So getting 100% of the sensor debayered wouldn't be a problem, I think (that's why I say that if the 450D can take the dremel and felt tips, then the mono mod can be done). The reason why I advise not to go anywhere near the edges now is to make sure we avoid a delicate area and we can determine first if the 450D can take the dremel treatment. Then we can move forward and apply the method I am suggesting to debayer the entire sensor.

Again, the epoxy stage isn't really a problem. As I said, if you use the sensor swab (with microfiber on it) you can place the resin wherever you want as you will have plenty of control in the process. In my sensors, I know the blue edge starts right where the epoxy is (as you can see, the resin can be placed where you want with absolute precision using this method). You want epoxy on the gold connectors so you can put frog tape on top of it without risking damage. Maybe, you could avoid using the resin on the other edges and directly apply 4, 5, 6 stripes of frog tape for protection but I haven't tried this yet.

The annoying thing is that with all the work I put into this mod, I can't say if the 450D can take a severe abrasion method like the rotary tool and felt yet, just because I unfortunately managed to touch the blue edge. I can tell you, though that the 1000D can take almost anything; as long as you use the dremel at the lowest speed using only felt tips and you don't touch the blue edges. If someone devises a way of removing the anti-reflecting coatingS or part of it without any gradient so you can achieve perfectnflat frames, I would be very interested to know how (in the 1000D).

Notice that these sensors don't have only one layer of anti reflective coating; they have several.

Hi Gina, the epoxy (Maplin's one) won't do any harm. The 1000D sensor can take almost everything as long as you don't touch the coloured edges were there aren't any connectors. Problem is that I really want a perfectly uniform sensor and I couldn't achieve that. Once I removed the CFA, I even tried toluene, N-Methyl-2-pyrrolidone and other substances that after some research I've found they are used in the sensor industry but they won't touch the AR coating. I did have success with the 450D as the dremel and felt tips remove the CFA and part of the coating uniformly. Unfortunately, no 450D sensor was left undamaged after the procedure. The annoying thing is that in all my 450D sensors, I managed to touch the coloured sides so the doubt remains if this was indeed the cause of failure.  A good test would be to debayer with the demel only 50% of the sensor in the central area (and leave a perfectly uniform finish with no coating gradients). In this way, without risking getting any close to the edges, sensor failure will show that this method can't be applied in the 450D. However, if the sensor works fine after this, there you go, it can be done with a perfect finish. At the moment I don't reallyknow the answer to this and I won't try it until I can get another 450D sensor at a reasonable price. It would be nice if someone in this thread tries this and see if we can move things forward.

Dear all,

Sorry if I didn't post here lately but, after trying a few times this mod, I got a bit tired. I have quite a lot of experience at the moment with this extreme mod so let me summarise my findings.

1. First, all you have to do with the epoxy is to keep it at bay with a sensor swab (with microfiber) until it starts setting and stops flowing onto the sensor. This will happen in about 20-30 minutes. It doesn't matter if the sensor has some epoxy residue; you can clean it with isopropanol once the epoxy is fresh but not moving (or even not bother to clean it at all as it will easily go away once you star debayering the CFA). The epoxy part is not a critical step in this mod at all.

2. The golden connectors are delicate but not as delicate as most people here think. I can tell you that they have taken a lot of punishment from my methods and they still work. Yes, try to avoid touching them but, although delicate, they aren't as troublesome as everybody thinks here.

3. The part that REALLY is very delicate, is the coloured edges on the sides where there aren't golden connectors. This is a REALLY DELICATE area and if you remove the CFA a tiny bit from there, chances are that your sensor will be damaged. Thus, all the work should be put into protecting those coloured sides. The golden connectors are delicate but much, much less critical than that area.

4. There are two sensors who are easy to work with: these are  the canon 1000D and 450D. Also, the cover glass is very easy to remove in one piece in those without applying any heat.

5. I have managed to debayer the sensor in the 1000D with the dremel and felt tips until I nearly got to the circuitry and the sensor still works. I can tell you that this sensor is really robust. The problem I have encountered here is that I found it impossible to debayer the sensor in a completely uniform way, which has always been my goal.

6. With the sensor in the 450D, I can debayer it leaving a completely homogeneous finish without any gradients left in the anti-reflecting coating so, although this sensor seems nearly identical to that of the 1000D, this leads me to believe that their construction varies in some way). The problem is that I haven't managed to debayer a 450D sensor without damaging it. I think this sensor is intrinsically weaker than that of the 1000D but I might be wrong as in my 2-3 failures with this sensor, I managed to get to the coloured sides by mistake which I know they are very delicate. I am not going to sacrifice another 450D for a while so it would be interesting if someone can proof that. The idea is to try debayering uniformly a 450D sensor only in the central area with a dremel and felt tips. If you manage to debayer only the central area in an homogeneous fashion without damaging the sensor, then this mod is perfectly achievable.

7. If you are using the dremel, only the felt tips are adequate. Any other: silicone, rubber, etc will damage the sensor. Even the felt will damage the sensor in the 350D model as its construction is very different from the others.

As I said, I am not going to spend more time and effort on this for a while (or maybe not) but it would be nice if someone could check if the 450D sensor can take this method.

By the way, it helps applying masking tape on top of the epoxy even when this is completely dry as the dremel and felt can remove a chunk of epoxy (this is what happened with my last 450D that got debayered a bit on one of the coloured sides). Green frog tape applied on top the epoxy once it's dry will effectively protect your golden connectors and coloured sides from the dremel and felt tips.

Well, I hope that helps. A few pounds were spent reaching these conclusions so it would be nice if someone can put them to good use and contributes to the thread to find a method for those brave enough to try it in the astrophotography community. The thing that I like about the dremel and felt approach is that it leaves no CFA residue and it can potentially achieve perfect flat frames. Nobody has achieved this so far; not even those companies selling monochrome converted cameras.

Also, I wouldn't worry about any other model apart from the 450D and the 1000D. The 1100D seems a really difficult one to crack (and this mod is a tremendous challenge even with a 100D or 450D). Other models: 500D, 550D, 600D, etc I don't think they are worth trying as failure would be really expensive and the nature of their sensors makes the loss of the microlenses a real inconvenience. The 1000D, for instance, clearly has a gain in sensitivity with the monochrome mod even though the microlenses are lost. I doubt the same can be said about any other models  (except perhaps the 450D -but I didn't manage to debayer one without damaging it to test this).

Hi guys

Got my debayered 1100d today - have posted a quick daytime comparison in the 'cameras' section if anyone is interested.

Cheers

Louise

Would it be possible to have a look at a flat frame?

The way to go is to apply the resin with a syringe and keep it at bay with a sensor swab (with microfiber) until the epoxy stops flowing into the sensor. The resin will stop moving in 10-15 minutes after it has been applied). Then clean the sensor with isopropanol. Any residue on the sensor will go  away during the debayering procedure.

Based on my experience comparing photos from debayered and Baader modified sensors and also taking into account what Luis has said, I find the results from RawDigger a bit confusing. Is it possible that this software has been designed to work with a Bayer matrix and it isn't showing correct results for debayered sensors? Just a thought.

Thanks for that and welcome back Luis   You're certainly right about the 1100D - not just the CFA but the cover glass is super-stuck on - with a vengeance!

It can't be harder to remove than the glass in the 350D! I nearly fried the sensor with the blowtorch and yet, I  couldn't remove it in one piece!

Hi guys!

Been awhile

By now I've debayered quite a few sensors (350D, 450D, 1000D, 1100D) and what I founs is this, the older cameras with bigger pixels do have a gain is sensitivity due to the large pixel size, the loss of microlenses is offset by the removal of the CFA and by quite a bit specially on the 350D, the 1000D is also very noticeabble as Pixueto has pointed out, it is correct that the light meter does overexpose the images on a debayered sensor, so this means that there is a gain in sensitivity, also I have some images done with half sensor debayered that prove this

For the "newer" cameras there is a very slight loss in sensitivity unfortunately, but it's not much but for me exposing maybe 10% more and get a full resolution image in narrowband full of detail is well worth the slight sensitivity loss

BTW, the 1100D sensor is BY FAR the most difficult one to debayer, the CFA is so stuck to the surface that is really complicated to remove!

Best,

Thanks Luis. have you removed the anti-reflective coating in those debayered sensors? I suppose flat frames can sort out imperfections but I'm  obsessed with having a nice perfect golden sensor.

BTW, you wouldn't have a tutorial for a deep cooled 1000D, would you? I haven't found anything online for this model.

You are probably experiencing differences in brightness because you are not imaging the same scene with the same settings and sensor.

My previous post demonstrated the sensitivity loss on the 1100D to be around 50% on a pixel by pixel basis. In this case, I used a camera with a sensor that had half of the CFA removed, this guaranteed that the same settings and light conditions were seen by the debayered and original pixels. I also used narrowband filters to ensure a straight comparison could be made between the pixels with CFA and those without. The RAW images files are linked to at the bottom of the post if you wish to inspect them.

http://stargazerslounge.com/topic/166334-debayering-a-dslrs-bayer-matrix/page-81#entry2308747

All the parameters are the same, apart from the sensor and I don't expect this to be a factor that influences the result that much. When I first debayered my 1000D, I couldn't use the auto settings as all the photos would invariably come up overexposed; I had to go to manual and lower the ISO or speed to get the right composition so, with the same settings than the stock camera, the photos definitely turn out overexposed. 

Now, doesn't that mean that the sensor is more sensitive to gathering light? If not, what's going on here? I'm trying to understand this.

We knew the removal of the microlenses would bring about a loss in sensitivity but we also knew that the removal of the CFA would increase it. This is more in line with what I would expect. Nevertheless, the tests with the narrowband filters in the 1100D show a different story and I don't really expect the results for the 1100D to be very different from other cameras, really.

I have a 7nm H alpha filter and I'll be able to replicate those with the 1000D but I need a stock1000D for that. If I manage to get perfect flat frames by removing the entire anti-reflective coating all over the sensor then I might get another 1000D. The idea is to do H alpha with the debayered camera and 'colour' the picture with the stock 1000D.

Guys, the settings are the same: ISO 200 Speed 1/125 F5.6 Same lens

I'm not contesting that there is a loss of sensitivity brought about by the removal of the microlenses. I just want to make sense of all this.

And now the raw files right from IRISwithout any adjustment whatsoever (resized to 50% and converted to jpeg):

First jpeg files right from the camera: grayscale Baader modified canon 450D and debayered 1000D with custom WB to get a grayscale picture for comparison:

Dear all,

I've been pretty busy these days with the monochrome mod. This is what I've found and where I am at the moment:

-I have debayered a Canon 1000D with the rotary tool method and the felt tips without leaving any CFA leftover. Now I am in the process of achieving a perfect flat frame throughout the sensor. I have achieved this in the central area of the sensor  (I didn't go any further yet as I wanted to avoid damaging the delicate edges. I wanted to check if the sensor would take it). So, it is then possible to achieve a perfectly uniform debayered sensor. The trouble is finding the right tool to go near the edges to remove the final anti-reflective coating -I have tried the dental silicone tips that Rinaldo suggested but they are too hard and will damage the surface (I tested this in a duff sensor). The rubber tips won't do either. It seems to me that the felt is the only option. I need now to find a high precision felt tip. The CFA in the 450D sensor is very similar to the 1000D so these results can be applied to that camera.

The canon 350D is a completely different type of beast. The CFA is much, much weaker and the rotary tool will damage the sensor very easily. I think the solvent approach might work with the 350D. Taking into account how much different the 350D is from the 1000D and 450D model, it is possible that the CFA on the 1100D sensor is very different too but I haven't tried with a 1100D yet.

Finally, after what has been discussed here, I am a bit confused. Maybe you can help me with this. It has been said that the loss of the microlenses could account for 30-50% of the sensor sensitivity. Yet, here you have a comparison of pictures taken with my 1000D debayered and a Baader modified 450D. Even if the cameras are different, I would expect the 450D to be superior to the 1000D so I think my point stands. These are the set of photos in twos; comparing both cameras: greyscale camera jpeg from the 450D versus custom WB jpeg from the 1000D. Also raw from 450D versus raw from debayered 1000D. I think they are interesting. How is it possible that the 1000D has lower sensitivity if in all the comparison photos, the result is invariably brighter? By the way, all photos have been taken with the same lens and the same settings: ISO 200, F5:6, same speed (I don't remember now)

Sorry about the debris in the 1000D. I haven't cleaned it properly as this is an ongoing project. As I said; it is only the central area, where the camera has been not onl;y debayered, but with the entire anti-reflecting coating removed completely without any visible damage.

Gina, have you noticed a layer below the bayer matrix? When I've been debayering the 1100D, once I've gotten the green layer off with a scraper, there's a dull gold layer underneath. If I then use a polishing compound, the dull layer is removed and there's a second gold layer.

Surely that's the bottom antireflection coating colour filter, right?

We should be able to measure actual losses due to removal of the microlens if someone can take a test image indoors, remove just the microlens layer but not the cfa which is easy enough and take an image of the same object. Then compare red and blue values.

Light conditions would need to be identical though.

This test would remove the cfa from the equation.

That's a good idea. I would say, even better removing the microlenses only in half of the sensor -just in case the microlenses removal physically damages the CFA somehow. Then proceed to completely debayer that side (with the other side still intact) and see if the results hold true. I might actually do that.

Here's one method :

http://www.google.com/patents/US8530356

Plus side, the gold wires won't react with the nitric acid.

Well, I've read through and the nitric acid is used to make...explosives! I don't feel so brave now about heating a 69% solution to 60C. Interestingly, it can be purchased quite easily:

http://www.ebay.co.uk/itm/Nitric-Acid-69-500ml-Free-Entry-into-Monthly-Prize-Draw-Same-Day-Shipment-/140573783993?pt=UK_BOI_Medical_Lab_Equipment_Lab_Supplies_ET&hash=item20bad9b7b9

Here's one method :

http://www.google.com/patents/US8530356

Plus side, the gold wires won't react with the nitric acid.

Brilliant, I'm reading this patent now. Very interesting. Thanks! Nearly there with this mod.

But will it eat into the underlying silice substrate? That would be a disaster.

Here's one method :

http://www.google.com/patents/US8530356

Plus side, the gold wires won't react with the nitric acid.

But will it eat into the underlying silice substrate? That would be a disaster.

So my original analysis stands then.  Removing the microlenses (and maybe some additional effect) is reducing the pixel sensitivity by a factor of very roughly 3. 

This has always been my understanding also.  All the pixels will now be sensitive to H-alpha instead of one quarter of them.  But if the price to pay is a reduction in sensitivity by a factor of 3 then not much has been gained. 

I'd love to be proved wrong so if someone has some evidence of the improvement.  I would also happily accept that resultswith  the 1100D may not be typical of what can be achieved with debayering other DSLR sensors.

Mark

Forgive me if I am saying something dumb (which it's probably the case) but... isn't the case that the removal of the CFA nearly compensates for the reduction in sensitivity brought about by the microlenses removal?

Found more information about the bottom antireflection coating colour filter:

In general, BARC layer 16 may be formed from any conventional BARC material, including a dyed organic film-forming BARC material or a light-absorbing polymeric film-forming BARC material. In some embodiments, BARC layer 16 preferably is substantially absorptive of radiation in the wavelength range used to pattern color filter array 14 and is substantially transmissive to radiation in the wavelength range to be imaged by image sensor system 10 (e.g., the visible radiation spectrum). BARC layer 16 may be formed from an organic film-forming material or a polymeric film-forming material. In one embodiment, BARC layer 16 is a photoresist-based antireflective coating that is substantially transmissive to radiation in a wavelength range of about 400 nm to about 700 nm (e.g., a Shipley AR2-600 antireflection coating, which is available from Shipley Company, L.L.C. of Marlborough, Mass., U.S.A.). In this embodiment, BARC layer 16 may be applied by a conventional spin-coater operating at 2000 rpm during deposition and at 4790 rpm during spreading; the resulting BARC layer 16 has a thickness of about 60 nm. After deposition, BARC layer 16 is exposed to a 60 second proximity bake at 205° C. on a DNS track.

Any ideas which solvent can be used to remove it?