A 3d tooth scanner is judged on three numbers and one workflow. The numbers are accuracy, depth of field and scan speed. The workflow is what happens to the file after the scan stops. Buyers who compare only the first number end up with a device that scans beautifully in a demo and slows the chair down in week three, because the software, the tip set and the export path did not match the cases the clinic actually sees.
This page sets out how to read those three numbers honestly, and how to run a trial scan that predicts your own results rather than the demonstrator's.
What a 3d tooth scanner is actually measuring
An intraoral scanner projects a pattern onto a surface and reconstructs geometry from how that pattern deforms. Everything the specification sheet reports follows from that: the size of the field it can capture in one shot, how deep it can still resolve a pattern, and how fast it can repeat the capture. Accuracy is the end result of those three plus the registration software that stitches frames into one model.
That is why two scanners with the same stated accuracy can behave differently on a full arch. The one with a wider field and deeper working range needs fewer frames, and fewer frames means fewer registration decisions and less accumulated error. A 3d tooth scanner spec sheet that lists only accuracy is giving you the conclusion without the mechanism.
The three numbers that matter and how to read them
Accuracy is usually quoted as a single figure, and it is usually the best-case figure. Aident publishes ≤10μm for the AI-30 across a full arch, and ≤7μm for a single tooth on the AI-50. Read those as different claims: one is a whole-arch number, the other is a per-unit number, and comparing them directly is a category error that vendors quietly encourage.
Depth of field is the number buyers skip and later regret. Both AI-30 and AI-50 are specified at 0–25mm, which is deep enough to keep a molar preparation in focus without repositioning the wand mid-prep. Scan speed is quoted in frames per second: ≥30 fps on the AI-30, 38+ fps on the AI-50, with a full arch typically under 60 seconds on the AI-30.
| Specification | AI-30 | AI-50 | What it changes at the chair |
|---|---|---|---|
| Accuracy | ≤10μm full arch | ≤7μm single tooth | Margin capture on multi-unit and implant cases |
| Weight | 198 g | 145 g | Hand fatigue across a full day of scanning |
| Scan speed | ≥30 fps | 38+ fps | Frames per second, and how much the patient moves |
| Depth of field | 0–25 mm | 0–25 mm | Repositioning during deep preparations |
| Field per capture | Full arch workflow | 12×10 mm and 16×15 mm | Number of frames needed for a quadrant |
| Output | STL / PLY / OBJ | STL / PLY / OBJ | Whether your CAD software opens it without a converter |
Where scan speed stops helping
Faster capture helps when the patient is a child, when the patient has a strong gag reflex, or when the case is a full arch and every extra second invites movement. It stops helping when the bottleneck is the operator's path around the arch. A demonstrator who scans daily makes 38 fps look like the reason the scan was fast; in a real chair, the same device in a less practised hand takes twice as long.
So test speed with your own staff, on your own patient mix, and count rescans rather than seconds. A scanner that finishes in 40 seconds and needs one rescan per four cases is slower than one that finishes in 60 seconds and needs none.
A trial scan protocol any clinic can run
Ask for a unit for a working day, not an hour. Scan five cases you would normally send out, and keep the raw files. Then check four things:
- Margin visibility on the hardest preparation you saw that day, not the easiest.
- Whether the scan stitching held across the palate without a manual re-registration.
- Whether the exported STL opened in your design software without repair work.
- How long washing the wand and switching tips added per patient, which is real chair time.
Three intraoral scanning tips make that day more informative. Scan in the order you normally work rather than the order the demonstrator suggests, so the comparison is fair. Keep the same lighting you work in, because some systems behave differently under operatory LEDs. And log one number per case: did the lab accept the file first time.
Run a scanner accuracy check before you sign
Brochure figures describe a device in a controlled setup. A scanner accuracy check describes the device in your hands. The version worth running takes ten minutes: prepare a typodont tooth with a clear finish line, scan it three times in a row without moving the model, then overlay all three meshes in your design software. If the three surfaces separate anywhere along the margin, that separation is your repeatability, and no specification figure will be better than it.
Repeat the check with the wand at the angle you actually hold it, and again with the patient's head turned, because the angle of incidence changes what the projector can resolve. A scanner that holds its margin from the 3 o'clock position but smears it from 9 o'clock is telling you something about the tip design and the lighting, not about its sensor.
Write the result down. Six months later, when a batch of crowns starts coming back, that note is the difference between diagnosing a scanner drift and replacing a device that was never at fault.
Why 3d tooth scanner software decides chair time
The capture is the visible part. The software is where the case is won or lost: automatic patch scanning fills holes the operator did not notice, edge detection flags a margin that was never captured, and the export path decides whether the lab receives a usable file or a support ticket. Aident ships AIDENT Scan for Clinics with lifetime software updates, AI patch scanning, edge detection and 5 GB of cloud storage per account, which matters less for storage than for the fact that the file exists off the workstation.
Ask two questions about software in any 3d tooth scanner evaluation. Is the update free after year one, and does the export include the colour texture or only geometry. The second question separates systems that support shade communication from systems that only support model production.
Powder-free scanning and what it changes
Both AI-30 and AI-50 scan without powder and capture true colour. In practice this removes a step from every appointment and removes the taste complaint that follows it. It does not remove the need for scan powder entirely: reflective restorations, zirconia and polished metal still benefit from a thin matting layer, and labs scanning stone models still use it daily. Treat powder-free as a chairside convenience, not a reason to skip surface preparation on reflective work.
Frequently asked questions
Is a 3d tooth scanner the same as an intraoral scanner?
Yes. The terms describe the same device class, along with digital impression scanner and mouth scanner. What differs between products is the capture technology, the field size, the software and the export formats, not the category.
How accurate should a scanner accuracy check be?
Run the check on a prepared tooth with a visible finish line, scan it three times, and overlay the three models in your design software. Variation between your own repeated scans tells you more about real-world accuracy than any single laboratory figure on a brochure.
Do I need the higher-accuracy model for crown and bridge?
For single units, a full-arch accuracy of ≤10μm is generally sufficient because the cement space absorbs small deviation. The higher single-tooth accuracy of the AI-50 pays off on multi-unit implant work, full arches and cases where the margin is deep or subgingival.
What should digital impression capture produce for the lab?
A watertight mesh in STL, PLY or OBJ, with the margin visible and the opposing arch registered. If the lab has to repair the mesh before design, the time saved at the chair is spent twice at the bench, and the case is later than it would have been with a physical impression.
If you want to run the trial day described above on your own cases, talk to our team and tell us the case mix you scan most. You can compare the two current models on the intraoral scanner page, read the dental tooth scanner specification page, work through the wireless scanner buying guide, or see how the software compares in our interface comparison.
