How long before the stars streak
The 500 rule and the NPF rule side by side, because they disagree by about a factor of two and only one of them knows what sensor you are using.
Untracked exposure length
Stacking session
Nothing leaves this device
Where these rules come from
The 500 rule: t = 500 / (focal length × crop factor). A darkroom-era approximation that predates anyone examining files at 100%. It knows nothing about your sensor's pixel density or your aperture.
The 200 rule is the same formula with a stricter constant, offered by people who found 500 too generous on modern sensors.
The NPF rule: t = (35N + 30p) / f, where N is the aperture, p the pixel pitch in micrometres and f the focal length in millimetres. Published by Frédéric Michaud of the Société Astronomique du Havre and popularised by PhotoPills. It accounts for pixel density and aperture, which is why it usually returns about half the 500 rule's answer and why frames shot to it hold up when you zoom in.
Declination divides by cos(δ). Stars near the celestial pole trace shorter arcs, so exposures there can be longer; at the celestial equator the correction does nothing. Within a couple of degrees of the pole the cosine collapses, so the figure is capped rather than reporting minutes of untracked exposure as achievable.
Pixel pitch comes from sensor width divided by pixel count. Without a pixel count NPF is left blank rather than guessed.
Why the two rules disagree, and which to believe
They are answering slightly different questions. The 500 rule asks when trailing becomes visible in a print of moderate size. NPF asks when a star stops landing on essentially one pixel. On a 12-megapixel camera viewed as a web image, 500 is fine. On a 60-megapixel camera examined at full size, it is optimistic by a factor of two and the stars are short dashes.
The practical answer: start at NPF, take a frame, and zoom all the way in on a star near the edge. If it is a point, you have room to lengthen. If it is a line, shorten. Two minutes of testing beats any rule, and both of these exist only to get you close enough that the first test frame is usable.
Stacking changes the question
If you are stacking, the length of one frame stops being about noise and becomes only about trailing. Twenty frames of ten seconds gather the same light as one frame of two hundred, but with round stars and far less noise after averaging. That is why the session block above reports total integration time — it is the number that determines how clean the result is.
The trade is that more frames means more storage, more time in the field, and more processing. Sixty frames at 25 MB is a gigabyte and a half for one image.
What none of this covers
Light pollution, which usually limits how long you can expose before the sky itself washes out — often well before trailing does. The moon. Atmospheric steadiness. And tracking: on a star tracker, the constraint moves from sky rotation to how accurately the mount is aligned, and these rules stop applying entirely.
Related
Timelapse planner for star trail and holy-grail sessions. ND and long exposure for daylight work.
Questions people actually ask
Which rule should I actually use?
Start with NPF and check a test frame. It knows your pixel pitch and aperture; the 500 rule knows neither. But no formula can see your tripod, the wind, or how large you will print — take one frame, zoom to 100% on a corner star, and adjust from there.
Why does the 500 rule give twice the exposure?
Because it was calibrated for film and modest enlargements, long before anyone inspected pixels. It asks when trailing becomes visible in a print; NPF asks when a star stops landing on one pixel. On a 45-megapixel body those are very different questions.
What is pixel pitch and why does it matter?
The physical size of one photosite, in micrometres — sensor width divided by pixel count. A star's image moves a fixed distance across the sensor per second; whether that distance covers one pixel or four decides whether it looks like a point or a dash. Two cameras with the same sensor size but different resolutions have genuinely different limits.
Does declination really let me expose longer?
Yes. Stars near the celestial pole trace small circles while those on the celestial equator sweep the full 15 degrees an hour. Pointing at Polaris genuinely allows several times the exposure. Pointing at Orion, which sits near the equator, does not — leave declination at zero for the worst case.
Is it better to shoot one long frame or stack many short ones?
Stack, almost always. Twenty ten-second frames gather the same light as one two-hundred-second frame but keep the stars round and average away most of the noise. The only costs are storage and processing time.
Do these rules apply on a star tracker?
No. Once the mount is following the sky, sky rotation is no longer the limit — polar alignment accuracy and periodic error are, and both are properties of your particular mount. Trackers routinely allow minutes where these rules allow seconds.