Every reading The Align gives you is relative. It is not telling you the handpiece is at some angle in the room — it is telling you how far you have moved from a position you chose. That difference decides almost everything about how the device behaves, and it is the thing most worth understanding before you use it.
Two ways a device could work
An absolute instrument reports orientation against a fixed external frame: level, plumb, magnetic north. A spirit level is absolute. It needs no setup, because gravity supplies the reference, and it is right the moment you put it down.
A relative instrument reports change from a starting point you define. It needs setup, because you have to give it the starting point, and it knows nothing about the world beyond what you told it.
The Align is relative, deliberately.
Why relative is the right choice here
An absolute device would have to know what the correct angle is. That would mean us deciding, in advance, what angle a preparation should be held at — across every tooth surface, every technique, every operator preference and every clinical situation. We are not in a position to make that decision, and we do not think a device should.
Relative sidesteps that entirely. You decide what is correct. The device measures your adherence to your own decision. It has no opinion about the decision itself.
There is a practical reason too. Absolute orientation from inertial sensing is hard: it requires a magnetometer for heading, which is unreliable near metal, motors and the general electrical environment of a clinic. Gravity gives you tilt reliably; heading does not come free. Working relative to a captured starting point avoids depending on a measurement that a dental surgery is a bad place to take.
What happens when you zero
Holding the handpiece where you intend to work, you tell the device to capture that orientation. It samples while stationary and stores that position as its origin. Everything afterwards is reported as a distance from it.
The stationary part matters. The capture is a measurement like any other, and a measurement taken while the device is moving is a worse measurement. This is why there is a brief settle before the reference is taken rather than an instant snap on button release — the pressure of the press itself moves the instrument.
The consequence people run into
The device will hold you faithfully to a bad reference.
If you zero while the handpiece is in a position you did not intend — resting at an angle, mid-adjustment, not yet settled — everything after that is measured from the wrong origin. The readout will be confident and wrong, in the sense that it correctly reports your distance from a point that was never the point.
There is no way for the device to detect this. It cannot know what you meant. This is the cost of the relative design and we would rather name it than pretend it away: the quality of your reference sets the quality of everything the device tells you afterwards.
Re-zeroing is normal
Because the reference is a decision rather than a fact about the world, it should change whenever your decision changes. Moving to a different surface, a different tooth, a different step in the procedure — each may mean a different intended angle, and the reference should follow.
Holding a stale reference across a genuine change of intent produces alerts that are technically correct and useless. The device reports that you have moved a long way from where you started, which is true, and irrelevant, because you meant to.
Re-zeroing costs one button press. That is deliberate: a workflow where correcting the reference is cheap is more robust than one where the initial capture has to be perfect.
It also answers drift
Inertial orientation estimates accumulate error over time. That is a property of the physics, not a flaw we have failed to fix, and any device that claims otherwise is either correcting against something external or not telling you the whole story.
A relative device with cheap re-zeroing has a good answer: if the estimate has wandered, re-establish the origin and the error goes with it. Rather than promising drift does not happen, the design assumes it does and makes the correction trivial.
What this means in practice
- Take the reference deliberately, in the position you actually intend to hold.
- Let the device settle before capturing. A rushed reference is a bad reference.
- Re-zero whenever your intended angle changes, not only when something feels wrong.
- If the readings stop making sense, re-zero before assuming a fault. It is the cheapest thing to try and it fixes the most common cause.
None of this makes the device more accurate. It makes the number it reports mean what you think it means, which matters more.