How we test Locksure

1. Designed around real locks

Locksure fits beside the thumb turn or key on a suitable euro-cylinder lock. A small magnet turns with the lock, and three magnetic sensors sit around it at measured angles.

Because the sensors are spread out, the monitor can tell which way the lock turned, not just that something moved.

3 sensors across a 107° arc

2. The turning rig

A motorised rig turns the thumb turn mechanism past the sensors of a Locksure monitor, again and again, while we log every result. A lock can be turned twice from the same side, or from either side of the door, so we always test clockwise and anticlockwise.

Our scripted suite covers 100 turn patterns, from slow, careful turns at 10 rpm to rushed turns at 340 rpm, plus part turns and over-turns. A single overnight run is around 750 turns.

10 to 340 rpm, both directions

3. Detection under stress

A lock turn can sweep the magnet past all three sensors in about a fifth of a second, so the monitor has to wake and read fast.

When a test turn is missed, we treat it as something to explain, not as noise. One example: we found that reading the sensors at the very instant the monitor went to sleep could set its wake-up the wrong way round, so we added a pause of about a tenth of a millisecond to let the reading settle first.

about 0.2 s for a fast turn

4. Battery and power

Battery life depends on how little the monitor uses while it waits, so we measure it with a precision power analyser instead of estimating. Asleep, the Original monitor draws around 10 microamps (millionths of an amp) or less.

We even unplug the programming lead before measuring, because the lead on its own adds about as much current as the sleeping monitor uses.

The result: a replaceable CR2032 coin cell, typically 18 months depending on use.

around 10 µA or less asleep

5. Radio and range

When the monitor is close to its hub it talks in a fast, low-power mode. If its messages stop being answered, it raises its own power and switches to a long-range mode without waiting to be told, then steps back down once messages are getting through again.

We test this with the hub moved to eight spots around a real house, and on the bench with simulated near and far positions. In one three-hour bench run that included three planned hub outages, all 281 of 281 events were delivered.

281 of 281 events delivered

6. Long soak runs

Some problems only appear after hours, so we run long soak tests. One monitor can act as several virtual doors, sending a lock or unlock event every few seconds through the same code a real turn uses.

In one run our display hub worked for 13.9 hours straight and handled more than 2,500 events with no restarts.

13.9 hours, 2,500+ events, 0 restarts

7. Security

Messages between an Original monitor and its hub are encrypted, and the hub ignores an old message if someone records it and plays it back later. Matter sensors use the security built into the Matter standard.

We test these rules during long runs, including hub restarts, so a genuine door keeps reporting. Our statement under the UK PSTI product security rules is on our PSTI page.

8. Safe updates

Before a firmware update is installed, the downloaded file is checked against its expected size and two independent checksums. If any check fails, the update is refused.

New firmware goes to our own test devices first. We only release it after it has passed there and our download server has returned the identical file several times in a row.

size + 2 checksums

9. Real homes

A rig turns very precisely and the same way every time. Real doors rattle, spring back and get half turned, and some problems only show up there.

So monitors in homes can report diagnostic details about how each turn was read, and those reports shape the next update. Every fix then goes back on the rig, so a change made for real homes does not break the basics.