Factory monitor roofs are the hardest common skylight job: 20-35 degree pitch, heavy dust from processes, and glass that bakes in summer. A robot handles them if you plan lanes, dust removal and heat timing first.
Walk any industrial estate and you see monitor roofs, those raised glazed strips that give daylight to a production floor. They are efficient for light and miserable to clean, which is why so many are dingy for years at a time.
What makes a factory monitor roof different?
- Pitch of 20-35 degrees on both glazed faces.
- Process dust that bonds to glass with oils and fumes.
- High glass temperature under direct sun.
- Limited roof access, often a single hatch.
- Overhead hazards: extract ducts, cable trays, aerials.
That last one matters. A robot lane is not just glass; it is an obstacle course. Map the ducts and trays before the first run, because a machine that meets a cable tray at speed is a repair bill.
How do you remove bonded industrial dust?
You cannot brush oil-bound dust off dry. Do a pre-soak with a mild degreaser at the dilution the maker states, let it dwell, then run the robot with a fresh-water rinse. Skipping the pre-soak is the number one reason factory skylights come out streaky after a robot pass.
| Soil type | Method | Notes |
|---|---|---|
| Dry cement dust | Brush + rinse | Do not wet first, it sets |
| Oil-bound fume film | Degreaser dwell + rinse | Two passes typical |
| Metalworking haze | Mild acid-free cleaner | Test on a small pane |
| Algae near gutters | Biocide dwell + rinse | Repeat after two weeks |
Note the cement dust row. Wetting it first turns it to a thin grout that then cures on the glass. Sequence matters as much as the chemical choice.
What does a realistic schedule look like?
A 6,000 m2 monitor roof cleaned twice a year is about 12,000 m2 of annual work. With a machine running 250-400 m2/h in practice on pitched, dusty glass, that is roughly 30-50 machine hours a year plus setup. Split it into spring and autumn to catch pollen and pre-winter grime. Our throughput guide explains why real rates sit below demo numbers.
Who should not put a robot on a factory roof?
If the roof has loose fixings, sagging frames, or glazing already cracked, fix the structure first. A robot adds downforce and vibration; on a tired roof that can turn a small crack into a panel replacement. Also skip robots where access is one narrow hatch and no fall protection exists, because you still need a human to place and retrieve the machine. Lingfeng S1 and Lingyun Y3 are common picks for these roofs, but only after the structural check.
Key Takeaways
- Monitor roofs run 20-35 degrees and carry oil-bound dust that needs a pre-soak.
- Never wet dry cement dust first; it sets and bakes on.
- Map ducts and cable trays before the first robot lane.
- A 6,000 m2 roof twice a year is roughly 30-50 machine hours.
- Fix cracked glazing and weak frames before adding robot downforce.
How do you split a monitor roof into robot lanes?
Work each glazed face as its own set of lanes, running along the slope rather than across it. That keeps the machine climbing a single pane band instead of crossing mullions. On a long roof, split into 10-15 metre sections so a paused machine never blocks the whole run, and keep the section boundary at a mullion where the pads can reseat cleanly.
Mark the lane start points with a small floor sticker at the hatch. Crews rotate and forget the plan; a sticker keeps the second shift doing what the first one did.
What does a factory roof clean actually cost to run?
| Item | Per 1,000 m2 | Note |
|---|---|---|
| Machine hours | 3-4 h | At 250-350 m2/h |
| Water | 60-120 L | Depends on soil load |
| Chemistry | Low | Pre-soak drives most of it |
| Labour | 1-2 people | Operator plus spotter |
The chemistry line is small; the pre-soak dwell time is what makes it larger. Budget the labour honestly, because a supervised run is not the same as a solo one.
When is the right season for a factory roof?
Spring, before pollen seasons peak, and mid-autumn, before winter grime settles. Summer sound good on paper but the glass is too hot, and winter brings short days and ice risk on the access route. Two well-timed passes beat four rushed ones on this kind of roof.
How do you keep a factory roof clean longer?
Attack the source where you can. Extract ducts that discharge near the roof deposit the same film back within weeks. If a local exhaust can be extended away from the glazing, or a filter upgraded, cleaning intervals stretch. The robot is the cure; ventilation design is the prevention.
Where the source cannot change, accept that this roof will always be a repeat job and price it as an annual program rather than a one-off event.
What is the biggest cost nobody budgets?
Access. Getting a machine and water onto a factory roof through one hatch takes longer than the first cleaning lane. If there is no lift or stair, crews hand-haul everything, and that labour can exceed the cleaning time on a small roof. Price access before you price the robot pass, or the job will run at a loss.
What does success look like a year later?
Daylight through the monitor roof improves measurably, the schedule runs without emergency callouts, and the pre-soak habit is standard rather than special. If cleaning still takes two unplanned passes a year later, the soil source or the chemistry is wrong, not the machine. Revisit the ventilation before you buy a second robot.