Re: Astro CCDs still dragging their pixel feet

Based on my experience at JPL, it is not that easy to attach a heat sink to a
CCD or CMOS chip. The requirements seem contradictory. The adhesive must be
thermally conductive so the heat from the chip can be drawn away, but
electrically insulated so it won't affect the electrical properties of the chip.
Usually the two properties go hand in hand, thermal conductivity usually
associated with some electrical conductivity. Also, the housing for most CCDs
and CMOS imagers is not a good thermal conductor, so simply attaching the heat
sink, thermal finger, and/or cooler to the back of the IC housing is not
efficient. For normal cameras this type of housing is not a problem, since
cooling is not done because the signal levels (brightness of the scene) is so
high that the noise is negligible. So the solution is to take the chip out of
its housing, or at least take the backing off the housing, and adhere the
cooling device to the CCD or CMOS imager.

The adhesive has to have no thermal expansion in either direction, as this would
place mechanical stress on the chip. This usually means the adhesive being used
is has pretty much permanently attaches the CCD or CMOS imager to whatever it is
that's cooling it, making replacing a damaged chip an impossible affair without
replacing the heat sink or thermal finger it's attached to, a costly affair.

And this is just trying to cool the chip, a necessary requirement since this is
the best way to decrease the amount of noise generated by the chip, a result
mainly of the heat it generates, which it sees as part of the signal coming from
the image.

An added requirement due to the low signal levels is the need for low noise
electronics to read out the signals. If you look at a histogram of the number
of pixels verses the signal from those pixels, on normal images (day or
nighttime) most of the pixels will have a high signal value. This is just the
opposite of deep sky images, or even planet images unless the planet covers more
than half the pixels, where most of the pixels are at the low signal end, often
in the lowest 10% of signal values if you isolate the pixels just containing the
object you wanted to image. Any noise from the readout electronics will add to
the difficulty in trying to get clean clear images of deep sky objects, as you
can see from this discussion, since noise is also in that low 10%.

Compared to what everyday cameras need in order to generate images, astronomical
cameras need higher grade electronics and other things not needed by those
consumer cameras that are still a requirement for getting good images of very
faint objects.

And for those that say "the professionals use this or that," I suppose if we all
individually had tens and hundreds of dollars of spare change, we could also
afford such large cameras, assuming we had enough image size at the focal plane
of our telescopes to be able to take advantage of such large arrays. Such
comments and comparisons with professional equipment are, in this case, merely
tools of a troll, and the poster should be considered as such, or at least
horribly misinformed or educated on the subject.

--- Dave
--
----------------------------------------------------------------------
Pinprick holes in a colorless sky
Let inspired figures of light pass by
The Mighty Light of ten thousand suns
Challenges infinity, and is soon gone

david.nakamoto@xxxxxxxxxxx


<AFM@xxxxxxxxxxxxxxxxxxx> wrote in message
news:1113680250.439023.126960@xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx
> Chris L Peterson wrote:
>> On 15 Apr 2005 18:10:51 -0700, AFM@xxxxxxxxxxxxxxxxxxx wrote:
>>
>> >I admit though this all does cost some money and does not lend
> itself
>> >to inexpensive cameras.
>>
>> Yes, when I said "expensive" I meant it in a relative way. Some folks
>> seem to think that since digital cameras are just a couple of hundred
>> dollars, a cooled camera for $500 should be easy. But considering you
>> need the thermoelectric cooler, some machined parts to get rid of the
>> heat, a sealed chamber with a window, and the control electronics,
>> that's probably close to $100 to the manufacturer, which means $500
> or
>> so retail (that's basic small volume manufacturing rules). And that
> is
>> before the camera itself, and the development time that goes into it.
>> Expecting cooled astrocameras to drop much below a couple of thousand
>> dollars is probably not too realistic. Some garage operations could
> pull
>> it off, but if you want good support, you will pay a little more.
>>
>> And it's crazy, but the B&W sensors required for good imaging cost
> more
>> than the same sensors with integral color filters.
>>
>> _________________________________________________
>>
>> Chris L Peterson
>> Cloudbait Observatory
>> http://www.cloudbait.com
>
> Both FLI and SBIG use off the shelf TEC coolers with $5 heatsinks fan
> combos. Mead could have added it for $100. It wouldn't be much of a
> leap to use the IR filter to seal the chamber. You then end up with a
> $500 camera, almost exactly what SAC is selling it for.
>
> Ian Anderson
> www.customopticalsystems.com
>


.

Relevant Pages

  • Re: Astro CCDs still dragging their pixel feet
    ... pull the chip out and the thermal load dropped by 5 watts if i remember for a 2K*2K 15 um pixel chip. ... For normal cameras this type of housing is not a problem, since cooling is not done because the signal levels is so high that the noise is negligible. ... An added requirement due to the low signal levels is the need for low noise electronics to read out the signals. ... If you look at a histogram of the number of pixels verses the signal from those pixels, on normal images most of the pixels will have a high signal value. ...
    (sci.astro.amateur)
  • Re: Micro ATX / water cooled question
    ... I'll just need to acurately compare the two boards for the locationof the chip. ... and four copper tubes running to a heat sink with a fan about 120MM. ... chips are running cooler than the one I have. ... There are no moving parts on a heatpipe assembly. ...
    (alt.comp.periphs.mainboard.asus)
  • Re: TEC and current driver chips for laser?
    ... heat to be conducted into a copper pour plane. ... some pins of the device before soldering the ground connection from the ... don't involve the chip getting warm to the touch. ... The more flex you design in the faster fatigue can set in. ...
    (sci.electronics.design)
  • Re: Digital Video has its limitations
    ... The HAD CCD's used in cameras like the PD170 for example exceed the capabilities of any 20 year old type. ... The real problem in shrinking down chips is having to use all sorts of enhancement technology to clean up the image which can often add noise ect.. ... Using those older larger chips allow for as little image "doctoring" as possible. ... we have one of the new Sony A1's with a single CMOS chip. ...
    (rec.video.production)
  • Re: Micro ATX / water cooled question
    ... The chip has no fan just an alum. ... and four copper tubes running to a heat sink with a fan about 120MM. ... A sintered tube helps resist gravity, ...
    (alt.comp.periphs.mainboard.asus)
Quantcast