Autocollimator: The Precision Finish

Full guide. Single-pupil autocollimator. This version explains how the tool works and why each step does what it does, in addition to the procedure. If you just want the steps, use the Core procedure instead.

What this does

The autocollimator is the final collimation step. It does not start the job — it refines and verifies what the laser and Cheshire have already set. It carries a small flat mirror that faces back down the telescope, so the optics are turned back on themselves. Any small alignment error left over is reflected several times over, which magnifies it until you can see and remove it. When everything is aligned, the magnified reflections collapse into one, and you have a positive confirmation that the telescope is truly collimated — not just close.

The two errors you are correcting

Collimation removes two independent alignment errors. This guide refers to them by these names throughout:

Keeping these two straight is the whole game. Every reflection you are about to read moves in response to one or both of them, and the autocollimator's value is that it lets you tell them apart.

How the autocollimator works

Light from the primary's center spot travels to the secondary, which sends it up the focuser to the autocollimator's flat mirror. That mirror reflects it straight back down to the secondary and on to the primary, where it reflects again — and so on. Each round trip produces another, fainter image of the center spot. Because each pass adds the alignment error again, the images separate by a large multiple of the actual error.

Reading the view

First look

At first glance the view resembles the Cheshire: the bright reflection of the center spot should sit roughly centered in the autocollimator mirror, with the autocollimator's small pupil near the center spot's perforation. What's new is everything behind that bright spot — the fainter reflections that carry the magnified error.

What you start with, and where you're going

The reflections, named

Up to four images of the center spot can be visible. Each is fainter than the last:

• P — the bright, frontmost reflection: the actual center spot. This is what you would see without the autocollimator.
• The upright flanking reflection — moves 2× the focuser error.
• The bright inverted flanking reflection — moves 4× the focuser error.
• The faint second inverted reflection — moves 2× the focuser error; this is the one used to set the focuser axis in the CDP method.

The triangle center spot is what makes these legible: it lets you tell an upright reflection from an inverted one, and it is what forms the tell-tale "hexagram" in the CDP method. A round dot hides all of that.

The one diagnostic that matters most

Before adjusting anything, read which error you have:

The single-pupil rule

Before you start

Three things must be true before the autocollimator will mean anything:

1. The primary mirror is center-marked with a triangle. The procedure reads the triangle's orientation, so a round dot will not work as well.
2. The telescope is already collimated with the laser and Cheshire (the previous layer). The autocollimator shows what those tools left behind; it is not a starting point.
3. The primary cell holds its adjustment — knobs turn easily, settings stay put, and lock. If it drifts, fix that first; no optical procedure will stick on a cell that won't hold.

Insert the autocollimator and seat it squarely and flush in the focuser. If it sits cocked in the drawtube, every reading will be wrong.

Method 1 — Iterative

The straightforward method: nudge the reflections into a stack, re-true the primary with the Cheshire between passes, and repeat. Recommended if you are new to the tool.

1. Confirm the telescope is laser- and Cheshire-collimated.
2. Insert the autocollimator. You will see a close jumble of reflections.
3. Adjust the focuser or secondary mirror to bring the reflections into a stack.
4. The secondary adjustment will have disturbed the primary, so re-align the primary with the Cheshire.
5. Repeat steps 3–4. Each pass tightens the stack.
6. Continue until the reflections stack into one and the Cheshire still reads correct.

Method 2 — Carefully Decollimated Primary (CDP)

The precise method. By deliberately decollimating the primary first, you separate the two errors so they can be set one at a time without chasing each other.

1. Start laser- and Cheshire-collimated. Insert the autocollimator.
2. Decollimate the primary on purpose: turn the top primary screw only (a small amount) until the reflections separate — an upright and an inverted reflection flanking the bright center spot, plus a fainter inverted reflection floating behind it. Using only the top screw keeps the motion to a simple tilt.

3. Set the focuser axis: adjusting only the focuser or secondary, move the faint inverted reflection until it sits directly behind the bright center spot. With a triangle spot, the two form a small six-pointed star (hexagram). Their separation is twice the focuser error, so the hexagram is your "focuser is now zero" signal.

4. Set the primary axis: with the focuser now correct, the separation of the two flanking reflections equals eight times the primary error — a very sensitive reading. Adjust the top primary screw to bring those flanks in behind the center spot. As they close up, all the reflections disappear into a single triangle.

5. Verify with the Cheshire and laser (see the single-pupil rule above).

Finish and verify

The telescope is collimated when all three agree at the same time:

• the autocollimator reflections stack into a single triangle,
• the Cheshire shows the center spot in its ring, and
• the laser return is centered.

Then lock the primary mirror and look again — locking can shift the alignment slightly. If it moved, correct it and re-lock.

Troubleshooting

• Only the center spot in a bright field: the focuser is far off, or the autocollimator is not seated flush. Re-seat it; if that doesn't help, redo the laser/Cheshire step.
• Reflections too jumbled to read: re-true the primary with the Cheshire to spread them, or use the CDP method to separate them deliberately.
• You can't see the faint inverted reflection: use the iterative method instead of CDP.
• The stack looks perfect but stars still look off: re-check the Cheshire — this is exactly the case the single-pupil cross-check is meant to catch.
• Collimation won't hold after locking: the problem is mechanical (springs/knobs), not optical — address the primary cell.
• A reflection drifts as you move your eye: that's parallax in your viewing, not a real error — keep your eye centered on the pupil.

How good is good enough

Collimation tolerance tightens sharply with focal ratio — roughly with its cube. A long, slow scope (f/8 and up) has a generous margin and may not need the autocollimator at all; a fast scope (f/5 and faster) has a tiny margin where the autocollimator's magnified view earns its place. Don't chase a perfect stack past the point your optics can use it on a slow scope, and don't trust a "close enough" laser-and-Cheshire result on a fast one without confirming it here.

Appendix — the reflections in full

For readers who want the underlying geometry. With A = primary error and B = focuser error, the reflections sit at these offsets from the autocollimator pupil (the same numbers Nils Olof Carlin derived by ray tracing):

From these, the practical rules in this guide follow directly. With the focuser corrected (B = 0): P sits at 2A (off the pupil), reflection 3 lands on P to make the hexagram, and reflections 1 and 2 sit symmetrically at ±4A about P — a separation of 8A, the eight-times primary-error reading. With the primary corrected (A = 0): P sits on the pupil and the reflections fan out at 2B, 2B, and 4B. The reflections are produced by 5-, 9-, and 13-bounce paths through the optics, which is why each successive one is fainter.

The figures in this guide are generated from these offsets, so they are geometrically faithful rather than artists' impressions.