Stacking Starlight

Stabilizing erratic autoguiding

If autoguiding over-corrects — the star bounces back and forth across the target, the error graph saws up and down, total error stays high — the system is fighting itself. Here's how to read the pattern and calm it down.

1. Read the pattern

  • Fast zig-zag / overshoot both directions → aggressiveness too high or you're chasing the

seeing. Tuning problem. Keep reading.

  • Slow one-way drift on top of the noise → polar alignment or flexure (mechanical), not

aggressiveness.

  • Sudden spikes → wind, cable snag, or a balance issue.
  • Dead spot right after reversals → that's mechanical play, see

reducing mechanical play in the mount.

2. Diagnose

Run the guiding assistant/calibration and read the guide log: separate RA and DEC, look at

whether corrections consistently overshoot (over-aggressive) or lag (under-aggressive), and compare

the residual error to the measured seeing — you cannot guide out atmospheric blur.

3. Calm it down

  • Lower aggressiveness and raise minimum-move so the guider stops reacting to every seeing

wobble.

  • Match guide exposure to the seeing (often 2–4 s) — too short chases turbulence.
  • Re-calibrate near the celestial equator/meridian and confirm the calibration is clean.
  • Tighten polar alignment to reduce the standing DEC drift the guider has to correct.
  • Rule out differential flexure (guide scope vs imaging scope moving relative to each other) — if

present, an off-axis guider helps; see choosing a guiding setup.

4. Verify

After tuning, the graph should flatten and RMS drop toward (not below) the seeing limit. Confirm over

a full sub, not a 30-second sample.

When it isn't tuning

If it's a clean dead-zone only at reversals, fix the mount play first

(here); a guider can't tune away mechanical backlash.

---

How autoguiding works: why long exposures need guiding.

Hardware causing flexure? Choosing a guiding setup.