Hubble telescope: how to effectively get 12 hour exposure times?

Ground-based telescopes can use film to make 12 hour exposures or even
longer so the integrating power of photographic film is an immense
advantage. The CCDs on the Hubble probably have a response time of a
fraction of a second. Even if the elements contain capacitors, their
time constant can't be more than a minute.
It's another instance of the gain-bandwidth constraint: the CCD's can
take 10 pictures a second but it does them no good to keep listening
for 12 hours. So, despite its favorable environment, it seems to me
that the Hubble is otherwise at a severe disadvantage.
It seems feasible to have electronics that would scan, every few
seconds, the elements which would be fitted with a bit of capacity
and emptying them and accumulating the result to get a de facto
integration.
Or is that what they're doing now?
John Polasek

"John Polasek" wrote in message
...
Ground-based telescopes can use film to make 12 hour exposures or even
longer so the integrating power of photographic film is an immense
advantage. The CCDs on the Hubble probably have a response time of a
fraction of a second. Even if the elements contain capacitors, their
time constant can't be more than a minute.
It's another instance of the gain-bandwidth constraint: the CCD's can
take 10 pictures a second but it does them no good to keep listening
for 12 hours. So, despite its favorable environment, it seems to me
that the Hubble is otherwise at a severe disadvantage.
It seems feasible to have electronics that would scan, every few
seconds, the elements which would be fitted with a bit of capacity
and emptying them and accumulating the result to get a de facto
integration.
Or is that what they're doing now?
John Polasek

Your assumptions are not born out by the facts. Point a web
cam at a light source and you'll soon seen over-exposure,
especially if you point it at your own monitor.
If a star is variable (and let's face it, those that are not don't
interest anyone - the bigger the variation the better - "Oh LOOK,
that's a nova! I must report that!") and you integrate the light then
you've lost part of the variation. The only reason for staring at a
source for 12 hours is to get enough light to detect it is even there.
Photographic film has no special advantages over CCDs, the finer
the granularity the longer the exposure time needed. So called
"fast" film is grainier. The multi-megapixel cameras available
today enable one to zoom in on part of an image, but my monitor
resolution gets no better for an 8 megapixel image when they
can only display 2,592,000 of them (2 x 1440 x 900) - GeForce
512 Mbyte graphics card.

John Polasek wrote in
:

Ground-based telescopes can use film to make 12 hour exposures or even
longer so the integrating power of photographic film is an immense
advantage. The CCDs on the Hubble probably have a response time of a
fraction of a second. Even if the elements contain capacitors, their
time constant can't be more than a minute.
It's another instance of the gain-bandwidth constraint: the CCD's can
take 10 pictures a second but it does them no good to keep listening
for 12 hours. So, despite its favorable environment, it seems to me
that the Hubble is otherwise at a severe disadvantage.
It seems feasible to have electronics that would scan, every few
seconds, the elements which would be fitted with a bit of capacity
and emptying them and accumulating the result to get a de facto
integration.
Or is that what they're doing now?
John Polasek

CCD's will integrate for as long as desired. The limitations are
elsewhere, most notably thermal noise, hence why infrared sensors
are cooled with liquid helium or better.

It's almost literally like a bucket out in the rain. So long as
the lid is off, it will collect rain. If it takes 12 hours to
collect a cup of water or 12 seconds it matters not.

The Hubble Deep Field images have a total integration time of
1 million seconds - over 277 hours. Mind you it wasn't done all
at once, but there's no reason why it couldn't have.

Brian
--
http://www.skywise711.com - Lasers, Seismology, Astronomy, Skepticism
Seismic FAQ: http://www.skywise711.com/SeismicFAQ/SeismicFAQ.html
Quake "predictions": http://www.skywise711.com/quakes/EQDB/index.html
Sed quis custodiet ipsos Custodes?

"John Polasek" wrote in message
...
Ground-based telescopes can use film to make 12 hour exposures or even
longer so the integrating power of photographic film is an immense
advantage. The CCDs on the Hubble probably have a response time of a
fraction of a second. Even if the elements contain capacitors, their
time constant can't be more than a minute.
It's another instance of the gain-bandwidth constraint: the CCD's can
take 10 pictures a second but it does them no good to keep listening
for 12 hours. So, despite its favorable environment, it seems to me
that the Hubble is otherwise at a severe disadvantage.
It seems feasible to have electronics that would scan, every few
seconds, the elements which would be fitted with a bit of capacity
and emptying them and accumulating the result to get a de facto
integration.
Or is that what they're doing now?
John Polasek

It is my understanding CCD's at maximum sensitivity provide almost a "click"
for every photon that hits them. You can turn down the sensitivity to get
less background noise (equivalent to reducing film ISO). The output of each
element of the CCD goes into an adder to produce a total brightness. Hubble
has to point at each location for hours to get enough photons to record an
image - it may only get a few photons per hour from that exact spot in the
"sky".

John Polasek wrote:
Ground-based telescopes can use film to make 12 hour exposures or even
longer so the integrating power of photographic film is an immense
advantage. The CCDs on the Hubble probably have a response time of a

You could integrate a similar time with a CCD, if you cooled it to low
enough temperatures (liquid nitrogen) to nearly eliminate thermal noise.
The problem then is the accumulation of "cosmic-ray" hits, or blemishes due
to natural radioactivity of the environment. The quantum efficiency of CCDs
far outweighs that of even the fastest film (typically, 75% vs 2%), so a CCD
will capture faint stars in a fraction of the time needed for film to do
this. The added advantage of detector linearity is important, as is the
"dynamic range" or ability to measure bright and faint objects on the same
frame.

fraction of a second. Even if the elements contain capacitors, their
time constant can't be more than a minute.

Not sure what you mean. If the shutter is open, the CCD will accumulate
photons.

It's another instance of the gain-bandwidth constraint: the CCD's can
take 10 pictures a second but it does them no good to keep listening
for 12 hours. So, despite its favorable environment, it seems to me
that the Hubble is otherwise at a severe disadvantage.

I've never heard an astronomer say this. Due to the design of the Hubble,
it isn't set up for really short exposures.

It seems feasible to have electronics that would scan, every few
seconds, the elements which would be fitted with a bit of capacity
and emptying them and accumulating the result to get a de facto
integration.
Or is that what they're doing now?

This can be done, but during readout noise is added. So if you do 10
shorter exposures and do readouts and co-add the data, the noise level will
be higher for the same total signal. And during readout, the shutter is
closed, so you stop integrating.

There are special ways of using CCDs with short exposures and continuous
readout. Video cameras for example.

John Polasek

I checked the calendar, and it's Christmas, not April 1st.

--
Mike Dworetsky

(Remove pants sp*mbl*ck to reply)