Why Are My Stars Trailing? Diagnose It From the Shape

Nobody photographs a star trail on purpose at 300mm. You get home, open the subs, zoom to 100%, and every star is a short comma instead of a point.

The unhelpful advice is “improve your polar alignment”. Sometimes that is the answer. Often it is not, and you can spend three clear nights chasing a problem you do not have. (If you have not processed a session yet at all, the walkthrough is the better place to start.)

The useful thing is that each cause leaves a different signature. Before you change anything, open one bad frame and look at where the stars are trailing and which way.

Read the frame first

Zoom to 100% and check four corners and the centre.

All stars streak the same direction, everywhere in the frame. The mount moved relative to the sky. This is drift, wind, a cable snag, or a sub that was simply too long. Most common by a wide margin.

Stars are points in the centre and streak outward toward the corners, worse the further out you go. That is not tracking. That is optics — coma, or a field flattener at the wrong spacing. The streaks point radially, like a starburst, rather than all lying parallel.

Stars are elongated in one direction on one side of the frame and a different direction on the other. Sensor tilt, or a sagging focuser. Again not tracking.

Stars are points in most frames and trailed in a handful. Something intermittent — a gust, someone walking on a wooden deck, the mount hitting a cable. This is the easiest one to deal with, because you just throw those frames away.

Stars trail as short arcs, curving, worse at the edges, worse the longer the session ran. Field rotation. Skip to that section, because the fix is different from everything else here.

Only the first and the last are worth losing sleep over. Let’s take them in the order actually worth checking.

1. The sub is longer than your mount can hold

Start here, because it costs nothing to test and it is the answer more often than polar alignment is.

Every mount has a period of time over which it tracks well enough. Unguided, on a decent consumer equatorial head, that is often 60 to 120 seconds at short focal lengths and rather less at 600mm. Periodic error in the worm gear is the usual limit, and it is cyclic — it comes and goes on a period of several minutes, which is why some of your frames are fine and some are not.

Take thirty subs at your current length. Count how many have round stars. If it is 60%, halve the exposure and take thirty more. If it goes to 95%, that was your problem.

Shorter subs cost you a little read noise and buy you a much higher keep rate. On a modern low-read-noise CMOS sensor that trade is almost always worth taking. There is a fuller version of this argument in how long your sub-exposures should be.

2. Polar alignment

Drift from polar misalignment has a tell: it is slow, consistent, in one direction, and it gets worse with declination. If your stars trail in the same direction across a whole hour and the streaks lengthen steadily rather than coming and going, this is a fair suspect.

The thing beginners get wrong is thinking of alignment as pass or fail. It is a number. Five arcminutes off is fine for 60-second subs at 250mm and ruinous for 300-second subs at 1000mm. The tolerable error scales with everything else you are doing.

Polar alignment routines built into modern mounts and the plate-solve-based ones in phone apps will get you under two arcminutes in a few minutes of work, which is enough for most beginner setups. Getting below one requires drift alignment and a lot of patience, and it is rarely the highest-value hour of your night.

3. Balance, and the cable you forgot

An unbalanced mount tracks badly in a way that mimics almost everything else. The motor is either fighting the load or being pushed by it, and gear backlash turns into star movement.

Balance in right ascension with the counterweight bar horizontal, then in declination with the scope pointed at the meridian. Slightly east-heavy in RA is the traditional advice and it does help, by keeping the gear teeth loaded on one face.

Then walk the cables. A USB cable with enough tension to matter will pull the scope a few arcseconds every time the mount slews past a certain angle, and it will do it sometimes, which is the worst kind of fault to diagnose. Leave slack. More than looks necessary.

4. Wind, and things that are not wind

Wind trailing is intermittent, direction-random, and correlates with nothing else. If you can hear the trees, this is probably it.

Less obvious: a tripod on a wooden deck picks up every footstep in the house. A tripod on soft ground sinks unevenly over three hours. A tripod on a balcony picks up the building’s own sway. None of these are fixable in processing, and all of them are fixable by moving.

5. Field rotation on an alt-azimuth mount

This one deserves its own section because it affects a lot of people who own smart telescopes and do not know it applies to them.

A Seestar S50, a Dwarf, or a Celestron on an alt-az fork tracks the sky by moving in altitude and azimuth at the same time. That keeps the target centred, but the field rotates around the centre as the night goes on. Stars in the middle of the frame stay round. Stars at the edges swing through an arc.

The rate depends on where the target is. Near the zenith it is fast. Near the horizon it is slower. Near the celestial pole it is nearly nothing.

You cannot fix this with better alignment, because nothing is misaligned. Your options are: shorter subs, shorter sessions, cropping the outer part of the frame, imaging targets lower in the sky, or a wedge that turns the mount equatorial.

Stacking software handles the rotation between frames perfectly well — registration solves for rotation as readily as for shift. What it cannot fix is rotation within a single 30-second exposure, which is where the arcs come from.

6. It is not trailing at all — it is coma

If the streaks point radially outward and the centre is sharp, your stars are not moving. A fast Newtonian without a coma corrector does this, and so does a refractor with the flattener at the wrong backfocus. The standard spacing is 55mm from the flattener’s shoulder to the sensor, and being 2mm out is visible.

Worth ruling out early, because no amount of mount work will touch it.

What processing can and cannot do

It cannot straighten a trailed star. The light was smeared across those pixels during the exposure and the information about where it should have been is gone. Deconvolution sharpens a blur that is symmetric and known; a directional smear from a mount lurch is neither.

What processing can do is notice and discard, and be specific about why. Akastroid scores every frame out of 100 and shows the six measurements behind the score — among them Roundness, which is eccentricity, and Trailing, which counts actual streaks and reports the longest one in pixels. A frame rejected for trailing tells you it had four streaks, longest 31 pixels, rather than just going grey in a list.

That distinction matters here more than anywhere. Eccentric stars and a long satellite streak are different faults with different causes, and a single “bad frame” verdict hides which one you have.

One detail worth knowing: the score is relative to the rest of the session, not to an absolute standard. If every frame in the night is slightly eccentric because your alignment was off, that is not grounds for throwing all of them away — you would be left with nothing. It only counts against a frame when the others did better. Full walkthrough in reading the frame quality table.

The thing to understand is that this is triage, not repair. If every frame is trailed, rejection has nothing to choose between. Fix it at the mount.

The short version

• Look at the frame before changing anything. Direction and position identify the cause.
• Halve your exposure and re-test before touching polar alignment.
• Balance the mount and free the cables. Both are free and both matter.
• Alt-az owners: field rotation is real, it is worst at the zenith, and it is not your fault.
• Radial streaks from a sharp centre are optics, not tracking.
• Software throws bad frames out. It does not straighten them.