Description
Vibrations and wind gusts produce an intermittent and unpredictable defect: the majority of frames in a session are nominal, but some show violently deformed, doubled, or trailed stars spanning a few arcseconds.
Unlike chronic guiding defects (backlash, PEC, flexure), this defect is not systematic: it depends on isolated external events -- a sudden gust, a tripod knocked by vibration, a door slamming on a terrace, a step on a balcony, or even a passing truck nearby.
Guiding corrects slow drifts but stands no chance against an impulsive disturbance whose frequency exceeds the correction loop bandwidth (typically >1 Hz).
The result: a subset of unusable frames that must be identified and discarded before stacking.
Visual signature
On affected frames, stars show short but pronounced trails, often in a zigzag or S-shape, the signature of a damped oscillation following a shock.
The pattern can be directional (all stars deformed along the same axis in the case of a gust) or more chaotic (self-sustaining tube oscillation).
In the worst cases, stars are clearly doubled. On the PHD2/NINA guiding graph, an isolated sharp spike appears on RA and/or Dec, sometimes spanning several seconds of error, followed by a quick return to baseline.
The FWHM measured by SubframeSelector spikes for the affected frames (often 2 to 5 times the median). On a visual blink of the subs, these frames stand out immediately.
Differential diagnosis
Distinguish from a chronic guiding defect, which affects all frames uniformly (stars elongated in the same direction throughout).
Do not confuse with a focus problem (stars blurry but round, not trailed).
Different from field rotation on an alt-az mount (progressive rotation around the center, not a punctual trail).
Do not mix up with thermal focuser drift or mechanical flexure (slow deformation over several consecutive frames, not sudden).
If all frames are deformed along the same axis, it is likely sustained dominant wind -- a site exposure problem, not isolated vibrations.
Also verify that it is not a satellite or aircraft crossing the field: the trail would then be a straight, thin line, not a deformation of the stars themselves.
Probable causes
- Wind gusts exceeding 15-20 km/h on an unprotected tube (Newtonians and SCTs particularly exposed)
- Vibrations transmitted through the ground on a balcony, wooden floor, or suspended structure
- Tripod placed on an unstable surface (gravel, decking, loose soil)
- Accidental impact on the tripod or mount during the session (passing animal, observer movement)
- Cables under tension or pulled by the wind, transmitting force to the mount
- Dew heater strap or flexible cable poorly secured, acting as a sail in the wind
- Leg cap or column cap not tightened securely, play in the telescopic legs
- Nearby vibration source (boiler, air conditioner, subway, busy road)
Course of action
- Systematically sort subs with SubframeSelector (PixInsight) or Blink before stacking
- Define a FWHM or eccentricity threshold beyond which frames are rejected
- In PHD2, monitor the RMS error graph during the session and log abnormal spikes
- Install a windbreak around the mount for sessions in windy conditions (>10 km/h)
- Add weight to the mount with extra counterweights or a sandbag on the tripod
- Shorten individual exposures in windy conditions (60s rather than 300s) to limit damage
- Use a dome or partial shelter if wind exposure is a recurring issue at the site
- Check cable routing and add cable clamps to the mount to absorb strain
- Cancel or postpone the session if wind gusts exceed 25-30 km/h
The Doc's advice
A gust cannot be corrected; it can only be culled. There is no point lamenting 5 frames out of 80: run SubframeSelector, set thresholds on FWHM and eccentricity, and discard without hesitation. The real trap is keeping a borderline sub out of reluctance: it degrades the stack SNR more than it contributes. In astrophotography, 60 clean frames beat 80 frames with 10 questionable ones every time.
Think you can see this defect in your image?
Run a diagnosisFrequently asked questions
How do I objectively determine which frames to reject?
SubframeSelector (PixInsight) or its equivalent in Siril measures several metrics per frame: FWHM (star size), eccentricity (roundness), SNR weight, noise. Sort the subs by FWHM and eccentricity and examine the distribution: frames affected by wind stand out clearly from the median. A pragmatic threshold: reject anything exceeding 1.5 times the median FWHM or eccentricity > 0.65. Never rely on purely visual inspection for large datasets.
Can guiding compensate for vibrations?
No, not rapid vibrations. Guiding typically corrects at 0.5-2 Hz, which makes it effective against slow drifts (PE, thermal flexure) but blind to impulsive disturbances at several Hz. Worse, after a sharp error spike, the guiding may over-correct and add its own instability. The only real remedy against vibrations is mechanical: dampen, add weight, shelter the setup.
Are Newtonians more susceptible to wind than refractors?
Yes, noticeably. The Newtonian tube acts like a full sail perpendicular to the wind, whereas a small-diameter refractor presents a much smaller wind cross-section. SCT and RC closed tubes are also exposed, but their more concentrated mass makes them more stable. On a 150-200mm Newtonian on a balcony, a single 15 km/h gust is enough to ruin several frames; a fabric windbreak stretched on the windward side makes a dramatic difference.
Should I redo a guiding calibration after a strong wind gust?
No, unless the wind physically displaced the mount or altered the polar alignment. Wind gusts do not change the guiding parameters themselves: they simply introduce impulsive disturbances that the guiding absorbs passively. However, monitor the PHD2 history over the long term to detect a calibration drift that has nothing to do with wind but could resemble a continuous degradation.