A robot vacuum’s advertised threshold height only makes sense with the geometry used for that claim. A combined two-step height is not permission to cross a single vertical lip of the same height. Measure the rises and horizontal spacing separately, then compare them with the exact model’s footnote.
We checked three Roborock global product pages on October 11, 2026. Their conditions illustrate the problem clearly. These are manufacturer test claims, not results from our own doorway trials, and they do not establish local availability or guaranteed performance.
Three claims that need their conditions
| Exact model | Advertised combined height | Geometry in its manufacturer footnote |
|---|---|---|
| Saros 20 | Up to 8.8 cm, double layer | First rise at most 4.5 cm; second at most 4.3 cm; gap deeper than 12 cm and no deeper than 14 cm |
| Saros 20 Flow | Up to 8.8 cm, double layer | First rise at most 4.5 cm; first-layer width at least 20 cm; second rise below 4.3 cm |
| Qrevo Edge 3 Pro | Up to 8.8 cm, double layer | First rise at most 4.5 cm; second at most 4.3 cm; gap deeper than 12 cm and no deeper than 14 cm |
Sources: Saros 20 footnote 1, Saros 20 Flow footnote 2 and Qrevo Edge 3 Pro footnote 2.
The Flow page specifies first-layer width where the other two specify a gap. Keep that difference in your record. A shared headline figure does not make the test layouts interchangeable.
Why adding the rises changes the answer
The headline sum is 4.5 + 4.3 = 8.8 cm. That describes two rises under the documented conditions. It does not describe a robot driving into an 8.8 cm vertical wall.
Consider an illustrative 5 cm single lip. Its overall height is lower than 8.8 cm, but that fact alone does not place it inside any of these double-layer test conditions. A doorway with two 4 cm rises also needs the relevant horizontal dimensions checked. Matching the sum while ignoring the spacing leaves the compatibility question unanswered.
These are comparisons of claim scope, not predictions of success or failure. Your surface, approach and actual use can differ from the manufacturer’s test.
Make one record for each transition
Record the first vertical rise, any second rise, the horizontal tread or gap, approach width and surface profile. Add the exact robot variant and the linked footnote. A wide, sloped transition and a narrow upright lip are different shapes even when a tape measure gives the same overall change in floor level.
Then classify the result:
- Documented geometry appears to match: keep the exact conditions and the manufacturer’s performance caveat together.
- A dimension is outside the condition: ask the manufacturer about that transition rather than treating the larger headline number as approval.
- The condition or measurement is unknown: leave the fit unresolved before buying for whole-home access.
A threshold can separate reachable rooms from the dock. Plan around the actual routes the robot can use; a large cleaning-area claim does not establish that it can enter each room.
Keep the remedy within the instructions
Do not force the robot over a lip, cover cliff sensors or improvise a ramp from this comparison. Any accessory or altered transition needs its own manufacturer compatibility and household-access assessment. Separately managed cleaning zones can be a more honest plan than promising uninterrupted coverage across an unverified doorway.
Method and limits
The original asset is the normalized footnote table and measurement record. No robot was handled or tested. The examples explain how to read specifications; they do not certify a doorway or rank these models. Recheck the exact regional model and current documentation before purchasing.