Alternatives to Drift Alignment?

Hi,

I have a polar scope on backorder from Losmandy, and from others'
experience, it may give alignment good enough for 5-10 minute
exposures, or it may not; meanwhile the weather is bad, so I think :-)

Has anyone ever tried the following alternatives to drift alignment?
(I have a digital level accurate to 1/10 degree, and I know my dec to
1/10 degree, both of which would figure below.)

1. Square the axes of your scope with the ota pointed north.
Take an eyepiece who's fov is 5-10 sec wider than Polaris'
path around the pole. Defocus Polaris until it touches the side
of the fov, then noting the asymmetry as you move the ota from the
left side of the mount to the right, adjust your alignment.

-- or if you can't even see Polaris --

2. Initialize a set of DSCs thru the "orthogonality" portion of
setup. Then center a bright star, say Deneb, in a reticle eyepiece.
Put the encoders in RA/DEC mode and zero them. Then using a calculated
offset to, say, Aldebaran, position the ota to where Aldebaran should
be. Adjust your alignment to center it. (Are repetitions necessary?)

Drift alignment has two words in it which add up to taking wayyyy too
much time: "wait" and "estimate". There has got to be a better way.
Right? :-)

Thanks

On Sat, 29 Nov 2003 21:29:33 -0800, sdh wrote:

snip

Has anyone ever tried the following alternatives to drift alignment?
(I have a digital level accurate to 1/10 degree, and I know my dec to
1/10 degree, both of which would figure below.)

The level only gives you an indication of the angle of some physical
piece of your mount (or OTA), _not_ the actual light path. Thus, your
dec setting will probably err by a greater margin than you suppose.

1. Square the axes of your scope with the ota pointed north.
Take an eyepiece who's fov is 5-10 sec wider than Polaris'
path around the pole. Defocus Polaris until it touches the side
of the fov, then noting the asymmetry as you move the ota from the
left side of the mount to the right, adjust your alignment.

This can work fairly well, but you'll never achieve the accuracy found
with drift alignment.

-- or if you can't even see Polaris --

2. Initialize a set of DSCs thru the "orthogonality" portion of
setup. Then center a bright star, say Deneb, in a reticle eyepiece.
Put the encoders in RA/DEC mode and zero them. Then using a calculated
offset to, say, Aldebaran, position the ota to where Aldebaran should
be. Adjust your alignment to center it. (Are repetitions necessary?)

Fuggedaboutit! DSCs are inherently inaccurate, to the degree
established by the encoder resolution.

Drift alignment has two words in it which add up to taking wayyyy too
much time: "wait" and "estimate". There has got to be a better way.
Right? :-)

Nope. Drift alignment is the ne plus ultra method, and once learned
can be accomplished in 10-15 minutes. BTW, remember to do your
collimation before aligning...

Wayne Hoffman
33° 49" 17' N 117° 56" 41' W
"Don't Look Down"

http://home.pacbell.net/w6wlr/

On Sat, 29 Nov 2003 21:29:33 -0800, sdh wrote:

I have a polar scope on backorder from Losmandy, and from others'
experience, it may give alignment good enough for 5-10 minute
exposures, or it may not; meanwhile the weather is bad, so I think :-)

Has anyone ever tried the following alternatives to drift alignment?
(I have a digital level accurate to 1/10 degree, and I know my dec to
1/10 degree, both of which would figure below.)

I sometimes use a digital level to get things set up during the day.
Realistically, though, figure that the best altitude you will probably be able
to achieve is +/- 0.5 degrees. And that doesn't help with azimuth. Even with a
very good compass, getting within a degree of true north is difficult.


2. Initialize a set of DSCs thru the "orthogonality" portion of
setup. Then center a bright star, say Deneb, in a reticle eyepiece.
Put the encoders in RA/DEC mode and zero them. Then using a calculated
offset to, say, Aldebaran, position the ota to where Aldebaran should
be. Adjust your alignment to center it. (Are repetitions necessary?)

Well, I have some DSCs with 20 arcsecond accuracy. Might do it with that. Not
with your typical 4K or 8K encoders, though.


Drift alignment has two words in it which add up to taking wayyyy too
much time: "wait" and "estimate". There has got to be a better way.
Right? :-)

There are software tools that will work as well as drift alignment (TPoint,
MaxPoint). But they are only faster if you are using a camera and some
scripting. And they aren't cheap.

_________________________________________________

Chris L Peterson
Cloudbait Observatory
http://www.cloudbait.com

Nope! That stuff sounds too damn complicated...even more so than a drift
alignment. But I routinely use a polar alignment scope which is good enough
for long CCD exposures.

Al


"sdh" wrote in message
...
Hi,

I have a polar scope on backorder from Losmandy, and from others'
experience, it may give alignment good enough for 5-10 minute
exposures, or it may not; meanwhile the weather is bad, so I think :-)

Has anyone ever tried the following alternatives to drift alignment?
(I have a digital level accurate to 1/10 degree, and I know my dec to
1/10 degree, both of which would figure below.)

1. Square the axes of your scope with the ota pointed north.
Take an eyepiece who's fov is 5-10 sec wider than Polaris'
path around the pole. Defocus Polaris until it touches the side
of the fov, then noting the asymmetry as you move the ota from the
left side of the mount to the right, adjust your alignment.

-- or if you can't even see Polaris --

2. Initialize a set of DSCs thru the "orthogonality" portion of
setup. Then center a bright star, say Deneb, in a reticle eyepiece.
Put the encoders in RA/DEC mode and zero them. Then using a calculated
offset to, say, Aldebaran, position the ota to where Aldebaran should
be. Adjust your alignment to center it. (Are repetitions necessary?)

Drift alignment has two words in it which add up to taking wayyyy too
much time: "wait" and "estimate". There has got to be a better way.
Right? :-)

Thanks

sdh wrote:
Drift alignment has two words in it which add up to taking wayyyy too
much time: "wait" and "estimate". There has got to be a better way.
Right? :-)

Well, a while ago, I posted the following query:

People ask about drift alignment reasonably often, and usually,
I see it presented as a sequence of magical steps--that is, without
rhyme or reason. Simple, and reasonably easy to remember, but still
magical.

In particular, if the star drifts north or south, you're supposed to
adjust east or west in azimuth (or is it west or east? I find that
part hard to remember). But how far? I never see this addressed. I
can only assume that when you first do a drift alignment, you just
guess, and that over time, you get a feel for how far to adjust. But
that's just an assumption. What do people actually do?

The reason I ask is that this afternoon, in a bit of down time, I worked
out a formula for how far to adjust. Being only a visual observer (as
opposed to imager), I don't often do drift alignment; I only recall
doing it a few times, to get the idea of the thing, so I don't really
know if people would use a formula at all. Would that be something that
people would find useful or interesting? The formula is not terribly
complex.

Well, I got no responses to that. Probably, the fact that the post was
tacked on as an "obligatory astro content" (ObAstro) to an off-topic
thread had something to do with that. Anyway, enough time has passed
that I've completely forgotten the formula. Let's see if we can't work
it out.

Here's the basic idea. Provided you pick a star in the right place
(where the celestial equator is rising for altitude adjustment, where
the equator is transiting for azimuth adjustment), the mount track and
the star track intersect by an angle equal to the error in alignment.
That is, if you're off by one degree in altitude, then the path of a
star rising on the equator in the east and the path followed by the
tracking mount (initially pointing to that star as it rises) intersect
in an angle of one degree.

If you were to wait six hours until that star transited the meridian,
the star would at that point be one degree from the center of the field
of view. Adjust the altitude of the mount until the star is recentered,
and the mount is properly aligned in altitude. Of course, six hours is
an awfully long time to wait, even for people used to drift alignment.

However, suppose that the star were initially centered in an eyepiece.
In t minutes of time, the star moves roughly t/4 degrees of arc, and
if the error in alignment is e degrees, the star will be off from the
center by (t/4)*(e*pi/180) degrees, where pi = 3.14159+, of course.
Expressed in arcminutes, we get a deviation of t*e*pi/12.

Now let's say the true field of view is v arcminutes. That means the
star is initially v/2 arcminutes from the edge. The time it would take
for the star to drift to the edge of the field of view is therefore
given by

t*e*pi/12 = v/2

Or, if we multiply by 12/(e*pi), we get

t = 6*v/(e*pi)

For instance, with a 6 mm Radian on my C5+ (effective focal length
about 1350 mm), the true field of view is about 15 arcminutes. If the
error is as little as one degree, the time it takes for the star to
drift all the way out to the edge is then given by

t = 6*15/3.14 = 29 minutes approximately

Conversely, if it takes t minutes of time to drift to the edge, the
estimated mount error is e = 6*v/(t*pi). If, for example, it takes
10 minutes to do so (again with a 15 arcminute TFOV), the mount error
should be about

e = 6*15/(10*3.14) = 3 degrees approximately

One problem with this is that it still takes a long time. It can be
improved if one estimates the time it takes to drift halfway out to the
edge, and adjusts the formulae accordingly. This is harder than it
sounds (in my experience, people tend to think the star is halfway out
to the edge when in fact it's only about 35 to 40 percent out), but
with a reticle eyepiece, it can be done with some precision.

Another more serious problem is that its convergence slows down. As
you get closer, it takes longer and longer for drift to reveal the
alignment error. I don't know if there's anything that can be done
about that.

Thoughts, anyone?

Brian Tung
The Astronomy Corner at http://astro.isi.edu/
Unofficial C5+ Home Page at http://astro.isi.edu/c5plus/
The PleiadAtlas Home Page at http://astro.isi.edu/pleiadatlas/
My Own Personal FAQ (SAA) at http://astro.isi.edu/reference/faq.txt

On Sun, 30 Nov 2003 05:54:38 +0000 (UTC), (Brian Tung)
wrote:

snip of thoughtful analysis and lotsa formulae

One problem with this is that it still takes a long time. It can be
improved if one estimates the time it takes to drift halfway out to the
edge, and adjusts the formulae accordingly. This is harder than it
sounds (in my experience, people tend to think the star is halfway out
to the edge when in fact it's only about 35 to 40 percent out), but
with a reticle eyepiece, it can be done with some precision.

Another more serious problem is that its convergence slows down. As
you get closer, it takes longer and longer for drift to reveal the
alignment error. I don't know if there's anything that can be done
about that.

Thoughts, anyone?

Yup - just do it, over and over, and you'll develop a sense of how far
you need to go. I can now routinely get my G11 accurate enough to
hold an object at the center of a ¼ in² CCD chip in a 5000mm system
for more than 30 minutes, and it takes me less than 20 minutes to do
so.

Of course, this works only if you consistently use the same mount...

Wayne Hoffman
33° 49" 17' N 117° 56" 41' W
"Don't Look Down"

http://home.pacbell.net/w6wlr/

I have a polar scope on backorder from Losmandy, and from others'
experience, it may give alignment good enough for 5-10 minute
exposures, or it may not;

Hi:

Yes, in my experience, polar finders can do at least that well with care. I've
even hit the pole dead-on with one once in a while (confirmed via a drift
alignment). It all depends on your requirements. If you're shooting film,
expecially color print film, and can limit yourself to something in the 10-15
minute exposure range, are careful in setting up the polar scope, etc., yes,
you can probably get away with this.

If you're shooting a high resolution film like 2415, or smaller CCD chips at
anything other than very fast focal ratios, no, a-drifting you will go! Forget
DSCs. The accuracy ain't there for this purpose.

I don't understand your reluctance here. Even a novice at it can do a drift
alignment in less than 30 minutes. And there ain't much "estimate". The star's
moving in dec or it ain't. It's just not that hard to do a drift alignment (and
nothing else is as effective, period), and after a few iterationsit will be
second nature. It even gives you something to do while you wait for
astronomical twilight! ;-)


Peace,
Rod Mollise
Author of _Choosing and Using a Schmidt Cassegrain Telescope_
Like SCTs and MCTs?
Check-out sct-user, the mailing list for CAT fanciers!
Goto http://members.aol.com/RMOLLISE/index.html

Brian Tung wrote:

sdh wrote:

Drift alignment has two words in it which add up to taking wayyyy too
much time: "wait" and "estimate". There has got to be a better way.
Right? :-)


Well, a while ago, I posted the following query:


People ask about drift alignment reasonably often, and usually,
I see it presented as a sequence of magical steps--that is, without
rhyme or reason. Simple, and reasonably easy to remember, but still
magical.

In particular, if the star drifts north or south, you're supposed to
adjust east or west in azimuth (or is it west or east? I find that
part hard to remember). But how far? I never see this addressed. I
can only assume that when you first do a drift alignment, you just
guess, and that over time, you get a feel for how far to adjust. But
that's just an assumption. What do people actually do?

The reason I ask is that this afternoon, in a bit of down time, I worked
out a formula for how far to adjust. Being only a visual observer (as
opposed to imager), I don't often do drift alignment; I only recall
doing it a few times, to get the idea of the thing, so I don't really
know if people would use a formula at all. Would that be something that
people would find useful or interesting? The formula is not terribly
complex.


Well, I got no responses to that. Probably, the fact that the post was
tacked on as an "obligatory astro content" (ObAstro) to an off-topic
thread had something to do with that. Anyway, enough time has passed
that I've completely forgotten the formula. Let's see if we can't work
it out.

Here's the basic idea. Provided you pick a star in the right place
(where the celestial equator is rising for altitude adjustment, where
the equator is transiting for azimuth adjustment), the mount track and
the star track intersect by an angle equal to the error in alignment.
That is, if you're off by one degree in altitude, then the path of a
star rising on the equator in the east and the path followed by the
tracking mount (initially pointing to that star as it rises) intersect
in an angle of one degree.

If you were to wait six hours until that star transited the meridian,
the star would at that point be one degree from the center of the field
of view. Adjust the altitude of the mount until the star is recentered,
and the mount is properly aligned in altitude. Of course, six hours is
an awfully long time to wait, even for people used to drift alignment.

However, suppose that the star were initially centered in an eyepiece.
In t minutes of time, the star moves roughly t/4 degrees of arc, and
if the error in alignment is e degrees, the star will be off from the
center by (t/4)*(e*pi/180) degrees, where pi = 3.14159+, of course.
Expressed in arcminutes, we get a deviation of t*e*pi/12.

Now let's say the true field of view is v arcminutes. That means the
star is initially v/2 arcminutes from the edge. The time it would take
for the star to drift to the edge of the field of view is therefore
given by

t*e*pi/12 = v/2

Or, if we multiply by 12/(e*pi), we get

t = 6*v/(e*pi)

For instance, with a 6 mm Radian on my C5+ (effective focal length
about 1350 mm), the true field of view is about 15 arcminutes. If the
error is as little as one degree, the time it takes for the star to
drift all the way out to the edge is then given by

t = 6*15/3.14 = 29 minutes approximately

Conversely, if it takes t minutes of time to drift to the edge, the
estimated mount error is e = 6*v/(t*pi). If, for example, it takes
10 minutes to do so (again with a 15 arcminute TFOV), the mount error
should be about

e = 6*15/(10*3.14) = 3 degrees approximately

One problem with this is that it still takes a long time. It can be
improved if one estimates the time it takes to drift halfway out to the
edge, and adjusts the formulae accordingly. This is harder than it
sounds (in my experience, people tend to think the star is halfway out
to the edge when in fact it's only about 35 to 40 percent out), but
with a reticle eyepiece, it can be done with some precision.

Another more serious problem is that its convergence slows down. As
you get closer, it takes longer and longer for drift to reveal the
alignment error. I don't know if there's anything that can be done
about that.

Thoughts, anyone?

Brian Tung
The Astronomy Corner at http://astro.isi.edu/
Unofficial C5+ Home Page at http://astro.isi.edu/c5plus/
The PleiadAtlas Home Page at http://astro.isi.edu/pleiadatlas/
My Own Personal FAQ (SAA) at http://astro.isi.edu/reference/faq.txt

I assume this is Tung-in-cheek? Exact equations for the
interpretation of drift alignment are available in many references.
Go to the bottom of
http://www.astro.ufl.edu/~oliver/ast...asicScopes.htm
to see a represenation. Basically, assume you are looking at the CE
due south. The sky and scope are both moving/tracking at about 15
degrees/hour = 15'/minute. [If the star is not at 0 degrees dec then
multiply by cos(dec).] A tilt of the pole of beta degrees to the
East will cause a drift S given by 15'/min*tan(beta). In your
example if it takes 10 mins to drift 7.5', tan(beta)=(7.5/10)/15=
2.86 degrees which comes close to you approximation.

If you have a CCD camera, drift alignment can go quite rapidly since
a drift of 0.1' (easily detected) in 5 minutes indicates an error
of 5' in alignment which is about as small as one can expect to correct.
--
John Oliver

John Oliver wrote:
I assume this is Tung-in-cheek? Exact equations for the
interpretation of drift alignment are available in many references.

Not exactly...but I admit that after I posted this, I did find a number
of sources on-line that give equivalent formulae.

If you have a CCD camera, drift alignment can go quite rapidly since
a drift of 0.1' (easily detected) in 5 minutes indicates an error
of 5' in alignment which is about as small as one can expect to correct.

Really? You can't reasonably expect to get closer than that? That's
kind of interesting.

Brian Tung
The Astronomy Corner at http://astro.isi.edu/
Unofficial C5+ Home Page at http://astro.isi.edu/c5plus/
The PleiadAtlas Home Page at http://astro.isi.edu/pleiadatlas/
My Own Personal FAQ (SAA) at http://astro.isi.edu/reference/faq.txt

Brian Tung wrote:
John Oliver wrote:

I assume this is Tung-in-cheek? Exact equations for the
interpretation of drift alignment are available in many references.


Not exactly...but I admit that after I posted this, I did find a number
of sources on-line that give equivalent formulae.


If you have a CCD camera, drift alignment can go quite rapidly since
a drift of 0.1' (easily detected) in 5 minutes indicates an error
of 5' in alignment which is about as small as one can expect to correct.


Really? You can't reasonably expect to get closer than that? That's
kind of interesting.

Brian Tung

My assumption was that this was for a portable scope being setup for
a single imaging session. Of course a permanent scope can get
better than that but in that case why worry about how long the
measurement will take?
--
John Oliver