PDA

View Full Version : CCD Camera design


Alain J
February 6th 04, 06:48 AM
Hi

I want to build my own astronomical camera. I don't have decided which
I'll use. I want to put a color CMOS sensor but many web sites relates
that CMOS sensors are not very good (high dark level)
I've seen a color CMOS sensor with 1,3 Mpixel (1280x1024) at a very low
price (about $140).

For best performance, do I use a color sensor or B&W sensor ? In term
of sensitivy, dark noise which is the best (colour or B&W)?

Cordially

Alain

Tony Turner
February 6th 04, 07:32 AM
"snip>
> I want to build my own astronomical camera.
snip
This site is a good read. I've almost finished my camera and the
instructions have generally been clear.

http://www.astrosurf.com/audine/English/index0.htm>

There is also a North American site called Genesis. I've forgotten where it
is but the ever-reliable Google will find it for you.

Helmut G÷bkes
February 6th 04, 09:04 AM
Hi folks,

maybe it is interesting for you that a astronomical
camera called "Platinum" like Audine / genesis is available
as Kit or as complete camera from this site too.
have a look to www.trinityccd.de
Also available are many parts and modules for general
use in CCD-Camera constructions like cooling components,
cool-chambers, peltiers, Cases, Interfaces,
USB-Converters and such nice things every body who wants
to build CCD-Cameras.

helmut

David Nakamoto
February 6th 04, 09:09 AM
Color CCDs achieve their results by placing a matrix of color filters over
the sensor array. This cuts down on the sensitivity of the array,
increasing exposure times 2 to 3 times or more. It also reduces the
resolution, since every three to four pixels are used to recreate a single
color one.

CMOS does have higher noise, and less low light sensitivity.

The best color results are still from using a black and white CCD with a
color filters, taking one exposure per filter.


"Alain J" > wrote in message
...
> Hi
>
> I want to build my own astronomical camera. I don't have decided which
> I'll use. I want to put a color CMOS sensor but many web sites relates
> that CMOS sensors are not very good (high dark level)
> I've seen a color CMOS sensor with 1,3 Mpixel (1280x1024) at a very low
> price (about $140).
>
> For best performance, do I use a color sensor or B&W sensor ? In term
> of sensitivy, dark noise which is the best (colour or B&W)?
>
> Cordially
>
> Alain
>

Davoud
February 6th 04, 01:29 PM
Alain J:
> > I want to build my own astronomical camera.

Tony Turner:
> This site is a good read. I've almost finished my camera and the
> instructions have generally been clear.

> http://www.astrosurf.com/audine/English/index0.htm>

*****

I think that M. Alain J might appreciate this link even more than the
above: <http://astrosurf.com/audine/index_fr.htm>

Davoud

--
usenet *at* davidillig dawt com

SHFORSTER1
February 6th 04, 11:14 PM
Check out the new color cameras coming from SBIG, David.

www.sbig.com

They might make me trade in my ST-10 and filter wheel on the 11 Megapixel color
unit.

STU

David Nakamoto
February 6th 04, 11:38 PM
Admittedly, not looking at the site, but remembering how high priced the
SBIG cameras are, I don't think these will ever be in my future. Also,
doesn't it obey the rule that a single shot color has exposure times two to
three times longer for an individual exposure than the equivalent black and
white?

"SHFORSTER1" > wrote in message
...
> Check out the new color cameras coming from SBIG, David.
>
> www.sbig.com
>
> They might make me trade in my ST-10 and filter wheel on the 11 Megapixel
color
> unit.
>
> STU

SHFORSTER1
February 7th 04, 12:16 AM
The price of the ST-2000 with the autoguider is $2995. Yes , Quantum Efficiency
is down, but this is a tradeoff. If I go to widefield with a color camera, I'll
be spared all the stacking and still get good images. Hydrogen alpha would be
out with these cameras, while pretty pictures would be in.
STU

Tony Turner
February 7th 04, 03:02 AM
"Davoud" > wrote in message
...
> Alain J:
> > > I want to build my own astronomical camera.
>
> Tony Turner:
> > This site is a good read. I've almost finished my camera and the
> > instructions have generally been clear.
>
> > http://www.astrosurf.com/audine/English/index0.htm>
>
> *****
>
> I think that M. Alain J might appreciate this link even more than the
> above: <http://astrosurf.com/audine/index_fr.htm>
>
> Davoud
Tres bon! Merci.

Leonard
February 10th 04, 05:32 AM
There is another big drawback to using a "color" detector.
You are stuck with the filters the manufacturer placed on the front of
the chip. While fine for the daytime snapshot they do not have the
high efficiency of or color balance available in astronomical filters.
You could not do a Ha or OIII shot very well (yes you could put one of
those filters in front of a color detector but I believe the results
would be less than desired.

Of course if you have already selected a specific chip there may not
be a color filter option for it. You can only buy / get what they
make.

In general CMOS sensors do not offer as low as noise as CCDs.
Howerver the new Canon digital SLR cameras have CMOS sensors and dark
noise far below what my CCD based camera would produce (at the same
sensor temperature).

I wish you the best of luck on your endeavors.
Please publish the design if you get it to go.

Leonard

On Fri, 06 Feb 2004 09:09:18 GMT, "David Nakamoto"
> wrote:

>Color CCDs achieve their results by placing a matrix of color filters over
>the sensor array. This cuts down on the sensitivity of the array,
>increasing exposure times 2 to 3 times or more. It also reduces the
>resolution, since every three to four pixels are used to recreate a single
>color one.
>
>CMOS does have higher noise, and less low light sensitivity.
>
>The best color results are still from using a black and white CCD with a
>color filters, taking one exposure per filter.
>
>
>"Alain J" > wrote in message
...
>> Hi
>>
>> I want to build my own astronomical camera. I don't have decided which
>> I'll use. I want to put a color CMOS sensor but many web sites relates
>> that CMOS sensors are not very good (high dark level)
>> I've seen a color CMOS sensor with 1,3 Mpixel (1280x1024) at a very low
>> price (about $140).
>>
>> For best performance, do I use a color sensor or B&W sensor ? In term
>> of sensitivy, dark noise which is the best (colour or B&W)?
>>
>> Cordially
>>
>> Alain
>>
>

Alan Chen
February 13th 04, 12:49 AM
Hi Leonard, Stu,

With all due respect, narrow band imaging can be done with color detectors.
There is an interpolative process to yield the final image, but the results
are not as bad as one might think, especially for imagers with small pixels.

If you get a chance, here are links to my only 2 H-alpha images with a
one-shot.

http://www.heavenlyview.com/b33halphasxvh9c.htm

http://www.heavenlyview.com/ic405sxvh9c.htm

I don't have an OIII filter, but there's no reason to suspect this would be
much different. Depending on the filter cross-over, this could generate
signal in the blue or green filtered pixels, or both.

Maybe these are "less than desired" to some, but I'm quite happy with the
outcome. Judge for yourself.

Best regards,

Alan
http://www.heavenlyview.com/

"Leonard" > wrote in message
...
> There is another big drawback to using a "color" detector.
> You are stuck with the filters the manufacturer placed on the front of
> the chip. While fine for the daytime snapshot they do not have the
> high efficiency of or color balance available in astronomical filters.
> You could not do a Ha or OIII shot very well (yes you could put one of
> those filters in front of a color detector but I believe the results
> would be less than desired.
>
> Of course if you have already selected a specific chip there may not
> be a color filter option for it. You can only buy / get what they
> make.
>
> In general CMOS sensors do not offer as low as noise as CCDs.
> Howerver the new Canon digital SLR cameras have CMOS sensors and dark
> noise far below what my CCD based camera would produce (at the same
> sensor temperature).
>
> I wish you the best of luck on your endeavors.
> Please publish the design if you get it to go.
>
> Leonard
>
> On Fri, 06 Feb 2004 09:09:18 GMT, "David Nakamoto"
> > wrote:
>
> >Color CCDs achieve their results by placing a matrix of color filters ove
r
> >the sensor array. This cuts down on the sensitivity of the array,
> >increasing exposure times 2 to 3 times or more. It also reduces the
> >resolution, since every three to four pixels are used to recreate a
single
> >color one.
> >
> >CMOS does have higher noise, and less low light sensitivity.
> >
> >The best color results are still from using a black and white CCD with a
> >color filters, taking one exposure per filter.
> >
> >
> >"Alain J" > wrote in message
> ...
> >> Hi
> >>
> >> I want to build my own astronomical camera. I don't have decided which
> >> I'll use. I want to put a color CMOS sensor but many web sites relates
> >> that CMOS sensors are not very good (high dark level)
> >> I've seen a color CMOS sensor with 1,3 Mpixel (1280x1024) at a very low
> >> price (about $140).
> >>
> >> For best performance, do I use a color sensor or B&W sensor ? In term
> >> of sensitivy, dark noise which is the best (colour or B&W)?
> >>
> >> Cordially
> >>
> >> Alain
> >>
> >
>
>

Leonard
February 14th 04, 09:13 PM
They are nice shots.
I didn't say you could not.

But assuming a RGBG Bayer color filter on the front of the chip,
it throws away 3/4 of the light / signal.
Same depth in 1/4 the time or 4x the resolution sounds really good in
comparison.

I notice you mention Sigma reject.
I know Russell Crowman. He really is a nice guy.
Here is his shot.
It is less than 10% longer in exposure time on a 4" f5 scope.
Granted it is a different camera (Kodak CCD) but it is deeper and
higher resolution (click on image to see the full resolution details).
http://www.rc-astro.com/nebulae/ic434_region_ha.htm

It is quite a bit better than my shot.
Mine was shot unfiltered line of site to 3 streetlights.
CB245, 27 x 15s (6.75 minutes), cooling off, ambient ~ +35 deg F, with
darks, no flats, no defect map, 8" f5.3 Newtonian
http://photos.yahoo.com/z_leonardh

Leonard

On Fri, 13 Feb 2004 00:49:48 GMT, "Alan Chen" >
wrote:

>Hi Leonard, Stu,
>
>With all due respect, narrow band imaging can be done with color detectors.
>There is an interpolative process to yield the final image, but the results
>are not as bad as one might think, especially for imagers with small pixels.
>
>If you get a chance, here are links to my only 2 H-alpha images with a
>one-shot.
>
>http://www.heavenlyview.com/b33halphasxvh9c.htm
>
>http://www.heavenlyview.com/ic405sxvh9c.htm
>
>I don't have an OIII filter, but there's no reason to suspect this would be
>much different. Depending on the filter cross-over, this could generate
>signal in the blue or green filtered pixels, or both.
>
>Maybe these are "less than desired" to some, but I'm quite happy with the
>outcome. Judge for yourself.
>
>Best regards,
>
>Alan
>http://www.heavenlyview.com/
>
>"Leonard" > wrote in message
...
>> There is another big drawback to using a "color" detector.
>> You are stuck with the filters the manufacturer placed on the front of
>> the chip. While fine for the daytime snapshot they do not have the
>> high efficiency of or color balance available in astronomical filters.
>> You could not do a Ha or OIII shot very well (yes you could put one of
>> those filters in front of a color detector but I believe the results
>> would be less than desired.
>>
>> Of course if you have already selected a specific chip there may not
>> be a color filter option for it. You can only buy / get what they
>> make.
>>
>> In general CMOS sensors do not offer as low as noise as CCDs.
>> Howerver the new Canon digital SLR cameras have CMOS sensors and dark
>> noise far below what my CCD based camera would produce (at the same
>> sensor temperature).
>>
>> I wish you the best of luck on your endeavors.
>> Please publish the design if you get it to go.
>>
>> Leonard
>>
>> On Fri, 06 Feb 2004 09:09:18 GMT, "David Nakamoto"
>> > wrote:
>>
>> >Color CCDs achieve their results by placing a matrix of color filters ove
>r
>> >the sensor array. This cuts down on the sensitivity of the array,
>> >increasing exposure times 2 to 3 times or more. It also reduces the
>> >resolution, since every three to four pixels are used to recreate a
>single
>> >color one.
>> >
>> >CMOS does have higher noise, and less low light sensitivity.
>> >
>> >The best color results are still from using a black and white CCD with a
>> >color filters, taking one exposure per filter.
>> >
>> >
>> >"Alain J" > wrote in message
>> ...
>> >> Hi
>> >>
>> >> I want to build my own astronomical camera. I don't have decided which
>> >> I'll use. I want to put a color CMOS sensor but many web sites relates
>> >> that CMOS sensors are not very good (high dark level)
>> >> I've seen a color CMOS sensor with 1,3 Mpixel (1280x1024) at a very low
>> >> price (about $140).
>> >>
>> >> For best performance, do I use a color sensor or B&W sensor ? In term
>> >> of sensitivy, dark noise which is the best (colour or B&W)?
>> >>
>> >> Cordially
>> >>
>> >> Alain
>> >>
>> >
>>
>>
>

Alan Chen
February 15th 04, 12:27 AM
Thanks Leonard,

I was just questioning the phrase "less than desired", since I'm quite happy
with this result. As it turns out, when I increased the contrast and
resampled down to match sizes, the 2 images are not all that dissimilar. If
you are interested, I can send my resampled shot for you to compare. It
really doesn't give much away at all at such a small scale. In fact, mine
is deeper in stars (due to use of a wider 13nm H-a bandwidth), but Russell's
has better contrasts of detail, due to the small 3nm bandwidth. The
passband has much to do with the outcome of H-alpha images. I do agree that
the resolution cannot be equal all else constant, due to the interpolative
routines required to fill in the data, but the depth is similar for the
given times.

Russell's SigmaReject is a great tool. I only recently started to use that.
I've also enjoyed his work over the past year and he is one of the premier
imagers. Your shot only needs more time, but it's tough to overcome light
pollution unfiltered. I'm thinking of going to a 10" f/4.5 Newt, since the
optics on my SCT really aren't all that good along with the mirror shift. I
think you have the right idea with the 8" Newt!

Take care,

Alan
http://www.heavenlyview.com/

"Leonard" > wrote in message
...
> They are nice shots.
> I didn't say you could not.
>
> But assuming a RGBG Bayer color filter on the front of the chip,
> it throws away 3/4 of the light / signal.
> Same depth in 1/4 the time or 4x the resolution sounds really good in
> comparison.
>
> I notice you mention Sigma reject.
> I know Russell Crowman. He really is a nice guy.
> Here is his shot.
> It is less than 10% longer in exposure time on a 4" f5 scope.
> Granted it is a different camera (Kodak CCD) but it is deeper and
> higher resolution (click on image to see the full resolution details).
> http://www.rc-astro.com/nebulae/ic434_region_ha.htm
>
> It is quite a bit better than my shot.
> Mine was shot unfiltered line of site to 3 streetlights.
> CB245, 27 x 15s (6.75 minutes), cooling off, ambient ~ +35 deg F, with
> darks, no flats, no defect map, 8" f5.3 Newtonian
> http://photos.yahoo.com/z_leonardh
>
> Leonard
>
> On Fri, 13 Feb 2004 00:49:48 GMT, "Alan Chen" >
> wrote:
>
> >Hi Leonard, Stu,
> >
> >With all due respect, narrow band imaging can be done with color
detectors.
> >There is an interpolative process to yield the final image, but the
results
> >are not as bad as one might think, especially for imagers with small
pixels.
> >
> >If you get a chance, here are links to my only 2 H-alpha images with a
> >one-shot.
> >
> >http://www.heavenlyview.com/b33halphasxvh9c.htm
> >
> >http://www.heavenlyview.com/ic405sxvh9c.htm
> >
> >I don't have an OIII filter, but there's no reason to suspect this would
be
> >much different. Depending on the filter cross-over, this could generate
> >signal in the blue or green filtered pixels, or both.
> >
> >Maybe these are "less than desired" to some, but I'm quite happy with the
> >outcome. Judge for yourself.
> >
> >Best regards,
> >
> >Alan
> >http://www.heavenlyview.com/
> >
> >"Leonard" > wrote in message
> ...
> >> There is another big drawback to using a "color" detector.
> >> You are stuck with the filters the manufacturer placed on the front of
> >> the chip. While fine for the daytime snapshot they do not have the
> >> high efficiency of or color balance available in astronomical filters.
> >> You could not do a Ha or OIII shot very well (yes you could put one of
> >> those filters in front of a color detector but I believe the results
> >> would be less than desired.
> >>
> >> Of course if you have already selected a specific chip there may not
> >> be a color filter option for it. You can only buy / get what they
> >> make.
> >>
> >> In general CMOS sensors do not offer as low as noise as CCDs.
> >> Howerver the new Canon digital SLR cameras have CMOS sensors and dark
> >> noise far below what my CCD based camera would produce (at the same
> >> sensor temperature).
> >>
> >> I wish you the best of luck on your endeavors.
> >> Please publish the design if you get it to go.
> >>
> >> Leonard
> >>
> >> On Fri, 06 Feb 2004 09:09:18 GMT, "David Nakamoto"
> >> > wrote:
> >>
> >> >Color CCDs achieve their results by placing a matrix of color filters
ove
> >r
> >> >the sensor array. This cuts down on the sensitivity of the array,
> >> >increasing exposure times 2 to 3 times or more. It also reduces the
> >> >resolution, since every three to four pixels are used to recreate a
> >single
> >> >color one.
> >> >
> >> >CMOS does have higher noise, and less low light sensitivity.
> >> >
> >> >The best color results are still from using a black and white CCD with
a
> >> >color filters, taking one exposure per filter.
> >> >
> >> >
> >> >"Alain J" > wrote in message
> >> ...
> >> >> Hi
> >> >>
> >> >> I want to build my own astronomical camera. I don't have decided
which
> >> >> I'll use. I want to put a color CMOS sensor but many web sites
relates
> >> >> that CMOS sensors are not very good (high dark level)
> >> >> I've seen a color CMOS sensor with 1,3 Mpixel (1280x1024) at a very
low
> >> >> price (about $140).
> >> >>
> >> >> For best performance, do I use a color sensor or B&W sensor ? In
term
> >> >> of sensitivy, dark noise which is the best (colour or B&W)?
> >> >>
> >> >> Cordially
> >> >>
> >> >> Alain
> >> >>
> >> >
> >>
> >>
> >
>
>