Guide · site & event timelapse
How to set up a site timelapse camera that lasts the build
Timelapse rigs do not fail because of the camera. They fail because someone guessed the power and the storage instead of doing five minutes of arithmetic. A construction timelapse is a small remote installation that has to survive weeks of weather, knocks and flat batteries with nobody standing next to it. Three decisions decide whether it makes it: the shooting interval, the storage, and the power budget. All three are simple maths. This guide works through each with real numbers, then covers mounting, remote check-ins, and the dry run that catches problems while you can still fix them.
Last updated 2 July 2026 · by Alien IT Solutions
The three decisions
Camera choice gets all the attention and deserves almost none of it. Nearly any camera that can shoot on an interval will produce usable frames. What kills long term timelapse projects is everything around the camera: an interval picked at random, a card that fills in week 3, a battery that dies in the first cold snap. Do the arithmetic before you touch hardware. It takes five minutes and it is the difference between six weeks of footage and six weeks of nothing.
Decision one: the interval. Work backwards from the video.
Do not pick an interval and hope. Start from the clip you want and divide.
Say the goal is a 30 second final video, which is plenty to tell a build story. At 25 frames per second, 30 seconds is 750 frames. That is the total number of photos you need. Now spread them over the job: a 6 week build is 42 days, which is 1,008 hours. Divide 1,008 hours by 750 frames and you get one frame roughly every 80 minutes, around the clock.
Most builds do not run around the clock, and a third of your frames being night shots makes a dull video. Restrict shooting to a 10 hour work day, 7 am to 5 pm, and the window shrinks to 420 hours, which puts the interval at about one frame every 34 minutes.
Here is what I would actually do: shoot denser than the minimum and cull in the edit. One frame every 10 minutes across those work days is 60 frames a day, 2,520 across the build, about 100 seconds of footage at 25 fps. From that you can cut a 30 second version, a 60 second version, and drop the rain days entirely. Frames you never shot cannot be recovered. Frames you do not need cost nothing to delete.
Decision two: storage. Frames times file size.
Storage sizing is one multiplication. A stills camera JPEG lands somewhere between 5 and 15 MB depending on resolution and scene detail. Take the 2,520 frame plan above: at 5 to 15 MB each, that is roughly 13 to 38 GB for the entire build. A 128 GB card covers the worst case three times over.
Go aggressive, one frame every 5 minutes running 24/7, and it becomes 288 frames a day, 12,096 over six weeks, roughly 60 to 180 GB. Still one card, but now the top end matters, so do the multiplication before you buy.
When is an on-camera card enough? When someone visits the site weekly anyway and losing the frames between visits would be annoying rather than fatal. When neither is true, and on most long jobs neither is true, push the frames off site as you go. A rig that uploads daily protects the footage from theft, water and card corruption, and, as covered below, it also tells you the rig is alive.
Decision three: power. Budget for the worst week of winter.
This is where most long term rigs actually die. The method is short: work out the rig's average draw, turn it into daily watt hours, then size solar for the darkest week of the job, not the average. It is the same sizing method used for any remote gear: consumption first, then panel, then battery, always against the worst week.
A duty-cycled rig, camera plus a controller that wakes it, shoots and sleeps, can average around 2 W. That is 48 Wh a day. For the panel, divide daily consumption by worst case winter sun, call it 2.5 usable hours a day in the southern half of Australia. That gives about 20 W of panel as the bare minimum, and you double it for dirt, tilt and cloudy runs, so around 40 W. For the battery, three days of autonomy is roughly 150 Wh of usable capacity, more if the site is shaded.
And if the site has a mains feed, use it and stop being clever. A powered site shed or a builder's temporary supply beats any solar setup. Keep a small battery in line so a tripped breaker does not end the project, and spend the saved effort on mounting.
Mounting and framing for weeks
The frames only cut together if the camera never moves and the settings never drift.
Pick a fixed viewpoint
The camera must not move for the whole job. Mount to something permanent, a neighbouring building, a light pole, a dedicated post. Never to scaffold or fencing that will shift or come down mid-build.
Lock everything manual
Auto settings drift as the scene changes. Fix zoom, focus and white balance manually, tape any ring that can move, and the frames will align in the edit instead of breathing.
Secure it properly
A locked, unbranded enclosure mounted out of arm's reach survives. Gear at head height with a visible lens tends to walk off site. Treat theft as a design input, not bad luck.
Seal against weather
An IP rated enclosure, cable glands, a drip loop on every cable and a desiccant sachet inside. Face the lens away from the prevailing weather where the framing allows it.
Remote check-ins: find out on day 2, not week 6
The worst timelapse failure is the silent one. The rig stops on day 2, nobody looks at it, and the discovery happens in week 6 when someone asks for the video. The fix is to make the rig report in: push each day's frames to a server or storage bucket, and treat the upload itself as a heartbeat. Photos arrived this morning, the rig is alive. Photos missing, someone drives out today.
The data load is small. Sixty frames a day at 10 MB is 600 MB, well within a modest mobile data link. If the site already has a network running cameras, access control or site offices, the timelapse rides on it for next to nothing. Sizing that connection properly is its own exercise, and we have written the maths up in how much internet does an event need.
Event timelapse vs construction timelapse
Same camera, completely different job. Rerun the arithmetic every time.
Events: short and dense
A 10 hour festival bump in, show and pack down compressed into 30 seconds is 750 frames over 36,000 seconds, one frame every 48 seconds. At that timescale power and storage stop being decisions. The work is in framing and protection: a locked off wide shot nobody bumps, started before bump in, because the site transforming is the best footage.
Construction: long and sparse
Weeks or months, frames tens of minutes apart, and the rig unattended between visits. Here power, storage and check-ins dominate exactly as above, and the camera itself is the least of it.
Never reuse settings
An event interval of 48 seconds is roughly 40 times too dense for a six week build, and a build interval of 34 minutes gives a 10 hour event under one second of video. Different job, different numbers, same five minutes of arithmetic.
The dry run: prove it before site day
Never let site day be the first time the full rig runs. A week out, assemble everything on the bench and run it exactly as it will run on site:
- Run the full stack for 24 hours. Camera, controller, storage, upload and power, wired as they will be wired on site, not a partial test of the camera alone.
- Verify frames land in both places. On the card and at the remote end, with correct timestamps. A rig that shoots but does not deliver has already failed.
- Check exposure across a full day. Morning, midday and dusk. Adjust now, before the framing is locked and the enclosure is sealed.
- Measure the battery after the run. Compare against the power budget. If it drained faster than the arithmetic says, find out why now, not on site.
- Lock, label and photograph. Every setting fixed, every connector labelled, one photo of the assembled rig so it can be rebuilt identically on site.
- Pack spares for the cheap stuff. Cards, cable ties, desiccant, fuses and glands cost nothing to carry and a return trip to fetch.
Do the dry run and site day is 30 minutes of mounting plus one confirmation photo. Skip it and site day is where you discover the problems.
Who runs it
Event Networks is the event-connectivity service of Alien IT Solutions, an Australian IT, networks and connectivity company with more than 18 years of building networks that have to hold under load. Long term rigs like this report in over exactly the kind of links we build. For sizing the connection itself, see how much internet does an event need, and for the wider build, our event network guide.
Questions people ask
What interval should a construction site timelapse use?
Work backwards from the final video. A 30 second clip at 25 frames per second is 750 frames. Spread over a six week build running around the clock, that is one frame roughly every 80 minutes. Restrict shooting to a 10 hour work day and the interval tightens to about one frame every 34 minutes. In practice, shoot denser than the minimum, every 10 to 15 minutes during work hours, and cut the excess in the edit.
How much storage does a long term timelapse need?
Frames multiplied by file size. Shooting every 10 minutes across 10 hour work days for six weeks produces about 2,520 frames. At 5 to 15 MB per photo, that is roughly 13 to 38 GB, so a 128 GB card covers the worst case three times over. Storage is cheap. The real decision is whether frames also upload off site, so a stolen or drowned rig does not take the whole record with it.
Can a timelapse camera run on solar power?
Yes, and the sizing is arithmetic, not guesswork. Take the rig's average draw, multiply by 24 for daily watt hours, then size the panel for the worst week of winter sun and the battery for a few days of no sun at all. A duty-cycled rig averaging 2 W needs about 48 Wh a day, which a 40 W panel and around 150 Wh of usable battery will carry through winter. If the site has a mains feed, use it, and keep solar for the sites that do not.
How do I know the timelapse rig is still running?
Make it tell you. A rig that uploads its frames daily is also sending a heartbeat: if this morning's photos never arrive, you know the same day and drive out to fix it. A rig writing silently to an SD card can die on day 2 and nobody finds out until week 6. The upload itself is small, tens of frames a day is well under 1 GB, so a modest mobile data link on site covers it.
What is different about an event timelapse?
The timescale, which changes every setting. A 10 hour event compressed to 30 seconds needs a frame every 48 seconds or so, not every 80 minutes. Power and storage stop being problems, a battery pack or a mains feed lasts the day and the frames fit on any card. What matters instead is a locked frame that nobody bumps, and starting the interval before bump in, because the site transforming is the best footage.
Filming a build or an event? Get the link right first.
Tell us the site, how long the rig runs and what else needs to connect. We will size the uplink and the check-ins so it reports in every day. No pressure.
Get a quote