---
title: "Holding focus through the night"
canonical: https://stackingstarlight.com/troubleshooting/focus-drift
description: "If your early subs are crisp but later ones soften over hours — stars slowly bloating as the air and equipment cool — focus is drifting with temperature. Here's how to confirm it and keep focus locked."
---

# Holding focus through the night

> If your early subs are crisp but later ones soften over hours — stars slowly bloating as the air and equipment cool — focus is drifting with temperature. Here's how to confirm it and keep focus locked.

## 1. Confirm it's thermal, not something else

- **Gradual, whole-frame softening that tracks falling temperature** → thermal focus drift. Keep
  reading.
- **Sudden focus loss** → drawtube slip, mirror flop, or a bump — mechanical, not thermal.
- **One-sided star distortion (not uniform softening)** → alignment/tilt, see
  [diagnosing distorted star shapes](/troubleshooting/collimation).

## 2. Diagnose

Plot **HFR/FWHM (star size) against time and temperature** across the session. A steady rise in star
size that correlates with dropping temperature is the signature: most tubes shrink as they cool,
shifting the focal point. Note how many degrees of cooling it takes to visibly soften (it varies by
tube material and focal ratio — faster scopes are less forgiving).

## 3. Keep focus locked

- **Refocus on a cadence** (e.g. every N minutes or every few °C) — the simplest fix.
- **Temperature compensation:** if you have a motorized focuser, set a compensation coefficient so it
  nudges focus as temperature falls.
- **Autofocus routine** between targets/filters (each filter may need its own offset).
- **Reduce mechanical slip:** lock the drawtube, address mirror flop, balance the train so the
  focuser isn't fighting gravity.

## 4. Verify

After enabling a refocus cadence or compensation, the HFR-vs-time plot should stay flat instead of
climbing. Confirm across a temperature swing, not a short clip.

## When it isn't focus

If stars are distorted directionally rather than uniformly soft, that's optical, not focus — see
[distorted star shapes](/troubleshooting/collimation).

---

How focusing works: [focusing in the imaging train](/equipment/focuser).
Considering a motorized solution to automate this? [Choosing a focusing approach](/buy/focuser).

---

Source: [https://stackingstarlight.com/troubleshooting/focus-drift](https://stackingstarlight.com/troubleshooting/focus-drift) · Part of [Stacking Starlight](https://stackingstarlight.com) by Michael Kalika.

LLM resources: [LLM index (llms.txt)](https://stackingstarlight.com/llms.txt) · [Complete LLM text (llms-full.txt)](https://stackingstarlight.com/llms-full.txt) · [Markdown homepage (index.md)](https://stackingstarlight.com/index.md)