# What accuracy is needed for polar alignment when shooting time lapse of celestial objects?

I'm planning to shoot the solar eclipse coming up in August, and intend to use an astrophotography rotating mount to allow me to keep the camera aimed at the sun for the whole duration.

Needless to say, I will not be able to align it with Polaris optically, since this will be during the day. So I intend to do so by using a compass to align my "local right ascension" to due south, and then use a plumb line and a mark on the tripod head to set my "declination" equal to 90 degrees minus my current latitude. And on the first of these two steps, I am aware of the 15 degrees magnetic declination present where I'm shooting: Salem, Oregon.

I'm expecting that I'll probably have an error of several degrees in each of these adjustments. Assuming that I'm aligned on a axis that is 5 degrees "off true" that will lead to the sun having some rotation.

Without any sort of rotating mount, the sun moves at 15 degrees per hour, i.e. 15 minutes of arc per minute. With the setup I'm describing, including the error, what rate of angle per minute would I see?

More to the point, will it be significant if my longest exposure is an anticipated 8 seconds for the corona, during totality? I don't mind too much if the sun drifts a little between shots during the partial phase, it's the ten minute time period that surrounds the total phase that I'm concerned about.

--Edit--

I'll be shooting with a 1000mm focal length on a full frame Nikon, which according to this web site: Field of View Calculator gives me 2.06 degrees horizontal by 1.38 degrees vertical.

For what it's worth, this gives me an image of the size I want, I've tested the lens / camera combination aimed at the moon, which is pretty much the same size as the sun when viewed from the surface of the earth. It doesn't completely fill the viewfinder, but then I don't want that, since I'm expecting to catch a solar radius or two of the corona during the period of totality.

• You've left out a critical piece of information. What is the angle of view of your lens/camera combination? Or another way of putting it, what lens focal length and what size sensor (full frame, aps-c, 1/2.3", etc.) will you be using?
– scottbb
Jun 5, 2017 at 18:10
• If you have the chance, align the telescope mount in the night before (on Polaris), then don't move it. I did that on a partial eclipse and the sun stayed in the viewfinder all the time. Jun 5, 2017 at 19:24
• You don't have to wait until the August eclipse to test either method. You can shoot the sun any day using the same methods. Just as with shooting the eclipse (other than during totality), be sure to use the proper filter to protect your lens, camera, and most importantly your eyes. You can also shoot the moon (during the day or night depending on its phase) to test either alignment method as well. Jun 5, 2017 at 21:29
• @MichaelClark Indeed I can. And taking it a step further, it's very convenient that the current moon (as of June 5th, 2017) is waxing gibbous, and can be used for testing now, prior to the arrival of the solar filter. Jun 5, 2017 at 21:31
• If your primary goal is to capture the corona during totality, you may want to rethink your angle of view. Although it varies from one eclipse to the next, and even from one minute to the next during an eclipse, the corona can be much larger than two solar radii beyond the eclipsed solar disc (which is two radii wide). Jun 5, 2017 at 21:33

A misalignment of 1'' leads to a drift of 0.23''/min when the telescope is pointing to the Zenit. See here (in german)

Have a look at these formulas here

I've marked @Grimaldi's answer as the answer to my question, but I still want to explicitly quote the section of the referenced article that has the exact answer I want:

In other words, we would have to achieve an 8 arc minute (or better) accuracy on both axes to achieve an 11.25’ overall maximum polar alignment error. Plugging this value into Equation (4) reveals that we would need our drift rate to be no more than 1.72 arc seconds per minute in both axes to achieve this alignment tolerance.

From this, I read that an error of 8 arc minutes of alignment error leads to a drift of 1.72 arc seconds per minute.

In my case, assume an alignment error of 5 degrees, worst case. That's 37.5 times greater, therefore it should lead to a drift rate of 64.5 arc seconds / minute or just over one arc minute per minute. In practice, I'm hoping to do better: within 2 degrees if I can, which drops that to 25.8 arc seconds / minute.

That will work well enough for my needs. If I adjust alignment 10 minutes before totality, then have two minutes of totality and then another 10 after, that's a total of 22 minutes (time). Over this duration I'd experience 23.65 minutes of arc drift, worst case. With a FOV that's 1.38 degrees vertical, my drift in 22 minutes works out to about 0.28 times the FOV.

Hmmm,

What about using a smart phone with a good GPS app mounted on your instrument? That way, using the app, you get time, location, and all the GPS goodies with one step, and that should help your setup easily.

Many of the higher end alt-azimuth GoTo telescopes have GPS capability, and the more expensive equatorial GoTo's do it too. An alt-asm wedge plate that some scopes have gets them closer to equatorial, but there are geometry issues with some alt-azimuth mounts as the altitude and azimuth approach 90 degrees.

In your situation, I'd do a night time test run and see how well things line up, then use the experience to do it in daylight.