Tolerance Stack-Up in Sensor Magnet Assemblies: How to Protect Your Sensing Margin
A practical guide to mechanical and magnetic tolerance stack-up in sensor magnet assemblies, covering air gap variation, field strength margin, and how to prevent false triggers or missed events in production.
Decision brief
Who this is for
Mechanical engineers, sensor integration engineers, and quality teams designing assemblies that include a permanent magnet and a magnetic sensor.
What you can decide
- How to calculate worst-case air gap variation from component tolerances.
- Whether the magnetic field margin survives the full tolerance stack.
- When to tighten magnet tolerance vs. housing tolerance vs. sensor placement.
Evidence included
- Air gap stack-up calculation example with real dimensions.
- Field margin analysis at nominal vs. worst-case gap.
- Tolerance contributor ranking table.
- Mitigation strategy decision matrix.
Practical boundaries
- This guide uses simplified 1D stack-up models. Complex 3D assemblies with angular misalignment or thermal expansion require finite element analysis.
- Field estimates are based on dipole approximation and should be validated with gaussmeter measurement on the actual assembly.
The most common sensor magnet failure I see is not a bad magnet — it is a good magnet in an assembly where the air gap varies more than the engineer expected. The magnet meets spec, the sensor meets spec, but when production tolerance, adhesive variation, and housing dimension spread combine, some units work and some do not.
This guide explains how to calculate and control tolerance stack-up in sensor magnet assemblies. It is written for engineers designing Hall effect sensor assemblies, reed switch modules, magnetic encoder systems, and any application where a permanent magnet must reliably trigger a sensor across a mechanical gap.
What Is Tolerance Stack-Up?
Every component between the magnet and the sensor has a dimensional tolerance. The air gap — the distance from the magnet surface to the sensor sensitive area — is not a single number. It is a range defined by the combined tolerances of every part in the path.
A typical sensor magnet assembly has these contributors:
| # | Component | Nominal | Tolerance | Contribution to gap variation |
|---|---|---|---|---|
| 1 | Magnet thickness | 3.00 mm | ±0.05 mm | Thinner magnet = larger gap |
| 2 | Adhesive bond line | 0.10 mm | ±0.05 mm | Thicker adhesive = magnet sits higher, may help or hurt |
| 3 | Housing wall thickness | 1.50 mm | ±0.10 mm | Thicker wall = larger gap |
| 4 | PCB thickness | 1.60 mm | ±0.10 mm | Thicker PCB = sensor sits further from housing face |
| 5 | Sensor package stand-off | 0.50 mm | ±0.05 mm | Varies by IC package and solder height |
| 6 | Sensor die position in package | — | ±0.10 mm | Defined by sensor manufacturer; not controllable |
Nominal air gap: 1.50 + 0.10 + 0.50 = 2.10 mm (housing wall + adhesive offset + sensor stand-off)
This is the distance from the magnet face to the sensor die, assuming the magnet fills its pocket completely.
Worst-Case Analysis
In worst-case analysis, assume every tolerance goes in the direction that maximizes the air gap:
| Contributor | Worst-case addition to gap |
|---|---|
| Magnet 0.05 mm thinner | +0.05 mm |
| Adhesive 0.05 mm thicker (pushes magnet away from sensor) | +0.05 mm |
| Housing wall 0.10 mm thicker | +0.10 mm |
| PCB 0.10 mm thicker | +0.10 mm |
| Sensor stand-off 0.05 mm larger | +0.05 mm |
| Sensor die offset 0.10 mm deeper | +0.10 mm |
| Total worst-case gap increase | +0.45 mm |
Worst-case maximum gap: 2.10 + 0.45 = 2.55 mm
That is a 21% increase over nominal. For a magnet producing 30 mT at 2.10 mm, the field at 2.55 mm drops to roughly 22–24 mT (depending on magnet geometry). If the sensor operates at 15 mT, the margin just went from 2:1 to about 1.5:1 — before temperature derating.
RSS (Statistical) Analysis
Root Sum Square analysis assumes tolerances follow a normal distribution and are independent. The statistical gap increase is:
RSS gap increase = √(0.05² + 0.05² + 0.10² + 0.10² + 0.05² + 0.10²) = √(0.0025 + 0.0025 + 0.01 + 0.01 + 0.0025 + 0.01) = √0.0375 = 0.194 mm
RSS maximum gap (3σ): 2.10 + 0.194 = 2.29 mm
RSS gives a tighter bound than worst-case, but it assumes all contributors are statistically independent and normally distributed. In practice, some contributors (like adhesive thickness) may not be normally distributed, so RSS can be optimistic.
| Method | Max gap | Gap increase | Use when |
|---|---|---|---|
| Worst-case | 2.55 mm | +0.45 mm (21%) | Safety-critical, low volume, no SPC |
| RSS (3σ) | 2.29 mm | +0.19 mm (9%) | High volume, SPC in place, rejects acceptable |
Field Margin Analysis
The real question is not "how much does the gap vary?" but "does the sensor still trigger at worst-case gap?" Here is how to check:
Step 1: Measure or calculate field at worst-case gap
For a small cylindrical NdFeB magnet (3 mm diameter × 3 mm thick, N42 grade), approximate surface field is about 350 mT. At distance d from the surface, the on-axis field for a short cylinder drops roughly as:
B(d) ≈ Br/2 × [d / √(R² + d²) − (d+L) / √(R² + (d+L)²)]
Where R = radius, L = length, Br = remanence.
For our example:
- At 2.10 mm (nominal): ~28 mT
- At 2.55 mm (worst-case): ~21 mT
- At 2.29 mm (RSS): ~24 mT
Step 2: Compare against sensor threshold with temperature derating
If the Hall sensor operates at 10 mT (BOP) and the magnet loses 12% of its field at 85°C:
- Field at worst-case gap, hot: 21 × 0.88 = 18.5 mT
- Margin over BOP: 18.5 / 10 = 1.85x ✓ (acceptable for most industrial applications)
If the Hall sensor BOP is 15 mT:
- Field at worst-case gap, hot: 21 × 0.88 = 18.5 mT
- Margin over BOP: 18.5 / 15 = 1.23x ⚠️ (too thin — risk of missed triggers)
Which Tolerances to Tighten First
Not all tolerances contribute equally. Rank them by impact:
| Contributor | Tolerance | Impact on gap | Cost to tighten | Tighten first? |
|---|---|---|---|---|
| Housing wall thickness | ±0.10 mm | High | Medium (mold quality) | Yes — highest impact, moderate cost |
| PCB thickness | ±0.10 mm | High | Low (specify tighter PCB) | Yes — cheap fix |
| Sensor die position | ±0.10 mm | High | Cannot control | No — design around it |
| Magnet thickness | ±0.05 mm | Medium | Medium (grinding cost) | Maybe — depends on volume |
| Adhesive bond line | ±0.05 mm | Medium | High (dispensing control) | No — hard to control |
| Sensor stand-off | ±0.05 mm | Low | Cannot control | No — use sensor datasheet value |
Key insight: The two largest contributors (housing wall and PCB) are also the cheapest to tighten. Specifying ±0.05 mm housing tolerance and ±0.05 mm PCB thickness cuts the worst-case stack from ±0.45 mm to ±0.30 mm — a 33% reduction without touching the magnet.
Practical Mitigation Strategies
| Strategy | Effect | When to use |
|---|---|---|
| Reduce component count in gap path | Eliminates tolerance contributors entirely | Redesign stage — most effective |
| Use a datum surface | Controls magnet-to-sensor distance directly | When housing design allows a shared reference |
| Increase magnet volume | Higher field compensates for larger gap | When space allows and magnet cost is acceptable |
| Use higher grade (N42 → N52) | ~15% more field | When temperature allows (N52 max 80°C) |
| Tighten housing tolerance | ±0.10 → ±0.05 mm | When mold quality supports it |
| Use press-fit instead of adhesive | Eliminates bond-line variation | When magnet geometry and housing material allow |
| Add magnetic circuit (steel flux concentrator) | Increases field at sensor through flux focusing | When assembly complexity and cost are acceptable |
Assembly Design Rules of Thumb
-
Target 2x field margin minimum at worst-case gap and maximum operating temperature. For automotive and safety applications, target 3x.
-
Measure the actual assembly, not just the magnet. A magnet that passes incoming inspection can still fail in the assembly if the gap is larger than expected.
-
Control adhesive thickness if bonding is used. A 0.1 mm adhesive variation on a 2 mm gap is a 5% field change — significant for tight-margin designs.
-
Ask the sensor manufacturer for die position tolerance. This is often the largest contributor that engineers forget because it is inside the IC package.
-
Test at temperature extremes. NdFeB loses ~0.11%/°C. At 125°C above room temperature, you lose ~14% of the field — on top of the gap variation.
-
Document the stack-up in the magnet drawing. When you send us an RFQ, include the nominal gap, worst-case gap, and minimum field requirement. This allows us to recommend the right grade, geometry, and coating without guessing.
What to Include in Your RFQ
When you contact our engineering team for a sensor magnet quotation, the following stack-up information helps us recommend the right solution:
- Nominal air gap (magnet surface to sensor die)
- Stack-up tolerance (or the component dimensions so we can calculate it)
- Sensor model and operate/release thresholds
- Operating temperature range
- Required field margin (or we will recommend based on application class)
- Whether you need us to supply the magnet as a bare component or as an assembly
Assemblies with controlled datum surfaces — where we press-fit or bond the magnet into a carrier with the sensor mounting reference built in — typically cut the effective gap tolerance by 40–60% compared to loose magnet + separate housing builds.
More Posts
Sensor Magnet Failure Modes and Incoming Inspection Plan
A practical guide to sensor magnet failure modes, inspection gates, magnetic field checks, coating risks, packaging controls, and supplier documentation for OEM buyers.
Section 301 Tariff on Sensor Magnets: Why Magnetic Assemblies Reduce Your Landed Cost
How 2026 US tariff changes affect sensor magnet sourcing, why magnetic assemblies face lower duty rates than bare magnets, and what procurement teams should evaluate before the next PO.
Sensor Magnet Material Selection: NdFeB, SmCo, AlNiCo, and Ferrite Compared
A side-by-side engineering comparison of the four permanent magnet material families used in sensor applications, covering magnetic properties, temperature limits, corrosion behavior, and cost trade-offs.