It certainly appears that removing the AA filter (doesn't have to be replaced - in normal mods) or replacing with a UV-IR cut for non spectroscopic work would be a way to go.
I need to find some astronomical/ spectroscopic examples of Moire fringing/ antAlias effects before I start to panic.
Yes, it is one of the reasons I've been using fuji cameras, the later ones have no AA filter and the difference is noticeably sharper images than my Nikon which has the same size sensor, same MP and same lens. It is one of the reasons I have not been a fan of DSLRs for astronomical use, the AA filter is great for stopping moire on patterned shirts in portraits, but not what you want for a 'scientific' instrument.
anyone know if the bayer CFA on cmos sensors would be similar to CCD's on the nikon's?
with Sony's, is there software to use them for long exposure?
Greg mentioned a few that have low noise and high QE. might be worth trying it as well.
I debayered a 1000d spare sensor last night. was harder than the 350d, looks like the layer is a bit thinner.
just waiting for the 1000D body to arrive and I can then test with my scope.
I've removed the front glass on the 350D as well, so its now full spectrum. I will add a clear glass on the sensor face as well.
I might convert a few cameras to B&W, add the Baader UV/IR and put them up if anyone's interested. The normal lenses will work.
waiting for clouds to clear... so can only test indoors.
these are jpg's straight off the 350d with a custom white balance.
I did a comparison with a 300mm lens on a lamp post with an unmodded 1000d and the b&w 350D and noticed the difference in the small coils. not sure if the comparison is realistic, but what else do you do when its cloudy
Some of the results from a debayed DSLRs look very very promising and for a camera that is one it's end of life useful then this looks like a good proposition.
Except... the damn glass over the canon sensor! I hear Nikon's are much easier to debayer (glass removed easily) and are even gaining support for DSLR computer control, like with sequence generator pro!
I put a mono sensor in my SPC900NC and the changes were astounding so ramp this up to a DLSR...I am drooling at the prospect...then I remember the damn canon front glass over the sensor!
which glass were you referring to? the one that's stuck to the sensor itself?
It actually pops off with gentle pressure all around. you'll see it becoming milky white as you're lifting it, but the key is not to lift it too much else it'll crack.
i managed to remove the glass intact from the 1000d sensor this way.
the 350D is similar although I didn't have the patience so just broke it off.
the 1100D is the one that's reportedly extremely well stuck on. the SGL user Gina tried every method possible to get it off but failed.
one clue on which ones are possible could be from the maxmax website and the models they convert. you'd notice they dont have the 1100D there.
these are the ones they do, and I don't see any reason why we can't mod these ourselves.
one point to note is that this process removes the microlenses. so to see that the results are better even without the microlenses shows that its got good promise.
I'm just waiting for a break in the clouds and will try some "L" subs and a few Ha subs.
my next venture will be on a more sensitive one and then cool it.
most important bit is not to go anywhere near those darn golden wires in the sides. they're so thin you can hardly see them and they come off at the slightest contact. that would render the sensor useless.
which glass were you referring to? the one that's stuck to the sensor itself?
It actually pops off with gentle pressure all around. ......CheersAlistair
Yep that be the one. I read that the USA guys were not doing this mod due to this glass being a "pain". I believe they use a ceramic glue to set the glass cover. Which cannot be undone with chemicals.
I'm not worried about the microlenses nor scraping the sensor just getting the glass off and of course not damming the peripheral contact fingers.
You gents have given me inspiration I may indeed attempt this as my 450D is a little long in the tooth and am looking for "yet another project"!.
I'd be interested in a converted camera just for B&W traditional photography use.
Apparently the Nikons are easier to remove the layer, (the Canon 10D for example is apparently nearly impossible) and with BackyardNik being released this year, it could make a great astro combo.
From what I've tried the canon 350D and 1000D are easy.
others have modded the 450D. I think the 20D is hard as well.
with the 1100D and some other models that have the glass very strongly glued on, best bet would be to stick some tape and break it.
you just need to make sure nothing drops on the edges with the thin wires.
since we have software for the canons, i'm staying with modding canons.
I've done a 350d and a 1000D, and will then look at something more sensitive with less noise.
all depends on how the 1000d performs.
Cheers
Alistair
Last edited by alistairsam; 20-07-2013 at 01:22 PM.
I'll post some flats later, it won't be pretty cause I need 500x to see the cfa bits left and I do still have some specks left. but its not too bad either.
as for debayering, the trick is not to debayer. nothing to be done in the camera.
so just treat the files as mono.
in DSS, i believe there is an option not to debayer when opening raw files, in maxim, you can open them as is and stack and process, just don't debayer.
even jpg's straight off the camera can be used, just don't debayer.
edit: i'm not sure how to open them in DSS without demosaicing. I'll test and see.
at this stage, i've only tested in maxim and DPP.
I just bought a 600D. looking forward to debayering that. however, the small pixels and loss of microlenses may not make it that sensitive, but will test and see.
if only these clouds would clear.
Someone mentioned a Pentax K5 as a candidate for mono-mods. There is a Pentax model called the K5-IIs. The "s" I think is meant to stand for sharp because it has no anti-aliasing filter. I don't know anything about the bonding between the bayer and the detector surface.
I'll post some flats later, it won't be pretty cause I need 500x to see the cfa bits left and I do still have some specks left. but its not too bad either.
as for debayering, the trick is not to debayer. nothing to be done in the camera.
so just treat the files as mono.
I opened it in photoshop and they seemed to come in okay, can see the areas where there are still blobs of bayer layer, but they opened up ok. Really shallow depth of field makes it hard to judge the performance. I might have to have a go at this. The D90 sensor is relatively cheap http://www.ebay.com.au/itm/Nikon-D90...item5650b780ab
Looking forward to the flats.
Last night was relatively clear, so had a play with the mono 350D. first light actually.
Here are links to some unprocessed subs from the camera. https://www.dropbox.com/sh/zq4yhkjf01i72f6/DCKL0AR99R - files are still uploading.
Low resolution jpg's attached.
60s x 1 sub - iso800
180s x 1 sub - iso400
I'm pretty happy with the outcome given that this was taken through thin clouds, awful seeing, urban skies.
subs are unguided as I couldn't get the oag working with the dslr spacing.
will try again tonight.
I haven't tried stacking or processing these yet. will try later on.
from my perspective, the results show very good potential, so I'll be pursuing with a 600D.
another amateur in Portugal also managed to convert his 350D to mono by removing the CFA layer using my method.
he's had better weather than us, so here is a comparison of a mono 350D Ha image with a QHY9 Mono image taken by marc from IIS.(with permission)
the 350D is uncooled, 12bits, 12nm ha filter, 5x5min Sub, 8inch F4
QHY9 mono is 16bit, cooled, 7nm Ha filter, 15x5min subs, 90mm APO
Obviously the qhy9 image is a lot smoother but I was surprised to see that there is no significant loss of detail in the dslr image and it compares very well with the qhy9 given that the cost of the 350D was $150, and the qhy9 is $2300 or so.
This makes narrowband very affordable in my opinion.
I had to scale up the 350D image as the imaging scales are quite different with the F4 and the F7.
Seems like you stirred quite the interest! It's a very novel idea and also very cool. The comparison shows just how similar they two are, but this is of course bearing in mind a 90mm refractor vs a 203mm f/4 newt - I suspect the newt might always win This sort of set up seems very well suited to fast newts, since the DSLRs obviously dont have peltier cooling for long subs.
I had no idea the D3200 had 60% QE, if the bayer matrix is easy to remove it might be worthwhile pursuing that path rather than the canons? I'd be very tempted to. Question then becomes how to use NB filters; you would have to get a M42 extension tube from the T-ring (Atleast 40mm to push the FW away from the DSLR where the prism sits), and if you are using a OAG that equates to almost 90mm backfocus! The steep f/4 light cone may result in a smaller-than-the-sensor image on the guidecam since the pick-off mirror is almost the same size as the guidecam sensor (I'm not sure if this is going to happen but food for thought).
I wish you many clear nights, can't wait to see the outcome of this project!
Objective was not to compare an Apo and a newt but to see just how the low qe chip stacks up to a 16 bit ccd cause there are several factors against the mono dslr like qe, dynamic range, read noise, smaller pixel being a CMOS as opposed to a larger well in a ccd, loss of microlenses etc.
The subtle details have always been so enticing with a ccd compared to a dslr. So it was interesting to see how the 4 pixels reading lum contribute to details and gradations.
This hobby being so expensive, it's worth looking at how to get decent results without breaking the bank and its just plain exciting to do something radical and see it pay off.
Also with our weather and the light pollution, narrow band is mostly what you can do from your backyard and this puts it well within reach.
As for the oag, it's such a pita. I just couldn't get it focused. Have to try again.
I was planning on ditching the mpcc and getting the RCC that gives 90mm back focus.
It would cut the light cone of an f4 but should be fine as the aps-c size is small enough for vignetting not to be too bad.
Then a filter wheel and an oag can be used.
Cooling is my next step, and isn't too much of a hassle other than condensation control.
Jtw astronomy sell cooler kits for 199euro that has been known to be fitted by a school kid. I'm planning on a sealed housing like central ds. Rowland's work here is very informative as Well.
Yup I definitely agree, a little bit of ingenuity and hard work will go a long way in this hobby. I also know the objective was not to compare, but I was bascially saying it's not entirely apples to apples as the f/4 obviously collects lots more light! I would really like to see how the DSLR fairs on your 10" f/4, that would be very easy to compare to other 10" f/4 images with a CCD.
As for the OAG, I'm surprised you couldnt? Did you not have enough in or back focus? I'd imagine its back focus, if you like you are more than welcome to use my extension tube on my TS9OAG and see if that helps You shouldnt have any issues with focus on the guidecam once you can move it back far enough, but yes I think the MPCC doesnt have enough backfocus for your set up, especially a DSLR + FW.
Well done and thanks for sharing.
.
You shouldnt have any issues with focus on the guidecam once you can move it back far enough, but yes I think the MPCC doesnt have enough backfocus for your set up, especially a DSLR + FW.
