A hottest point is one coordinate
A specification line that says the surface reaches a certain temperature is a statement about one thermocouple at one location. A hand covers a palm-sized area, so what the user experiences is the average across that area plus the coldest patch inside it. A unit with a bright hot spot near the element and a cool rim can report an impressive peak and still feel uneven in the hand, while a unit with a lower peak and a flat spread feels warmer overall.
That is the argument for asking for a map instead of a number. A map is not a more precise version of the same claim; it is a different claim, about distribution rather than maximum.
How a map is usually built
The common method is a grid. The outer shell face is divided into cells, for example a few rows by a few columns across the grip area, and each cell gets a reading at a defined moment. Grid density matters: a coarse grid averages away exactly the local variation the map is supposed to reveal, so a four-cell grid on a large face is decoration rather than evidence. The record should state the grid, the cell count and whether readings were simultaneous or sequential.
The instrument is the second variable. An infrared camera gives a full-field image quickly but depends on emissivity, and a moulded shell with a mixed surface finish can produce misleading readings unless the emissivity is set or the surface is coated for the measurement. Contact thermocouples are slower and disturb the surface slightly, but they are easier to defend in a report. Many good records use both: infrared for the pattern, contact probes to confirm a few anchor points.
The third variable is timing. A map captured during warm-up shows the element pattern; a map captured at the plateau shows the settled distribution. Both are useful, and they are not interchangeable. A record that does not say which one it is cannot be compared with another record.
Reading the map
Metrics worth asking for
| Metric | What it describes | What it hides |
|---|---|---|
| Peak value | The single hottest cell | Everything about the rest of the surface |
| Minimum across the grip area | The coldest cell a palm will cover | Whether that cold patch is where the palm rests |
| Spread, maximum minus minimum | The range across the face | Whether the distribution is smooth or has isolated spikes |
| Gradient across the grip area | How fast temperature changes with distance | Localised voids smaller than the grid |
| Ratio of hot to cold area | How much of the face is near the peak | Which cells a thumb actually touches |
A supplier who can supply peak, minimum, spread and a grid drawing has produced a map. A supplier who can only supply peak has produced a number, and the two should not be filed as the same kind of evidence.
Where non-uniformity comes from
The heating element's own pattern is the first source. An etched foil track that doubles back on itself concentrates heat where the track density is highest, and a stamped or printed film can have a similar bias. Designers compensate with a spreader, with track width variation, or by accepting the pattern. Whichever route was taken, the drawing shows it.
The bond line is the second source and the most insidious. A pressure-sensitive adhesive with a void, or one that has crept after cycling, leaves an air gap between the element and the shell. That patch reads cool while the element behind it runs hotter, so a single-point probe on the shell reports a healthy figure and a map shows a dimple. On a dual-sided unit the same defect can exist on one face only, which is why maps should be produced per face.
The third source is the internal architecture. The battery compartment occupies volume behind part of the face, and the cell, the harness and the separator all change the local thermal mass. The area above the pack often reads cooler than the area above bare element, and that is a layout consequence rather than a manufacturing defect. Finally, the control board and the OTP protector sit somewhere; the region above them tends to run cooler because the board blocks the path.
Cold spots are felt more than hot spots
Users describe a warmer as uneven when part of the palm is not warm, not when one spot is slightly warmer. That asymmetry should shape the acceptance criteria: the minimum across the grip area and the size of any cold patch usually matter more than the peak. A face with a moderate spread and no cold island will be described as comfortable; a face with a high peak and one cool corner will be described as patchy, even though the second unit may have the higher average.
This is also why a uniformity criterion belongs in the golden sample signoff rather than only in the design review. A sample that passes on peak but fails on cold-patch size will generate returns for a reason the peak figure never captures.
A peak figure ≠ a uniform surface
Stating the boundary plainly: a high peak does not mean a well-distributed surface, and a flat map does not mean the surface is safe or comfortable. Uniformity is a distribution property; safety comes from the OTP chain, the insulation stack and the failure behaviour, and comfort depends on the absolute level and on how long the unit sustains it. A perfectly uniform surface at an uncomfortable level is still uncomfortable.
There is a second boundary about sampling. A map taken on one golden sample describes that sample. If the element is die-cut, printed or bonded with any process variation, the map should be repeated across a small batch before the distribution claim is written into a purchase order.
Five questions for the supplier
- How many measurement cells were used, and where were they placed relative to the grip area? A grid drawing with cell count is verifiable; "uniform heating" is not.
- Was the map captured during warm-up, at the plateau, or both, and at what ambient temperature? The timing determines which claim the map supports.
- Which instrument produced the map, and how was emissivity handled if infrared was used? Mixed surface finishes are a common source of false readings.
- What are the minimum and the spread across the grip area, not only the peak? The three numbers together describe the distribution.
- Was a map produced for both faces on a dual-sided unit, and on how many samples? One face and one sample is a design illustration, not a distribution record.
Next step
Add a uniformity line to the specification that names the grid, the instrument, the timing and the acceptance band for minimum and spread, then ask for the map at sample approval and again at first production. For the measurement-point question underneath this one, see where a hand warmer's temperature is actually measured; for the general pattern of inflated surface claims, see surface temperature spec inflation. The golden sample signoff form has a place to attach the map. For how a cold patch is graded at inspection, see the defect classification grading table. To have the criterion drafted for your own specification, start at contact.
Written for the Emberly heating-element programme. Specifications are illustrative; confirm every figure against a supplier test record.