
A thermistor probe is usually specified by what it is made of. It should be specified by how long it takes to tell the truth. The material sets the temperature range and the stability; the packaging sets the response, and the packaging is where most of the disagreement between a sensor datasheet and a real installation comes from.
Every measurement has a lag, and this lag is not an error you can calibrate away — it is a phase shift, and it changes what the reading means. A probe that lags by four seconds will faithfully report the temperature it measured four seconds ago, and any control loop built on it will be tuned to a lag it was never told about.
Thermal time constant is a packaging property, not a component property
The thermal time constant is defined as the time for the probe to reach 63% of the way to a step change in temperature. It is set almost entirely by the ratio of two quantities: the thermal mass of everything inside the sensor that has to heat up, and the conductance from that mass to the outside world.
Stated that way, the design levers are immediately clear.
- Increase conductance. This means a smaller sheath, a thinner wall, a shorter length, or a thermally coupled tip. A sensor with a heavily plated and thick-walled sheath can be three or four times slower than the same sensor in a thin-wall version.
- Decrease thermal mass. This means less material to heat up: a smaller bead, less potting compound, a smaller lead bundle. A long lead wire inside the sheath is thermal mass that the heat has to travel along before it reaches the sensing element.
- Do not confuse the two. A longer probe in a thinner sheath is usually faster than a short probe in a thick one, despite being longer — because mass scales with volume while conductance scales with surface area.
The practical consequence for a drawing is that the sheath wall thickness and the internal potting are performance-determining dimensions, not just manufacturing dimensions. Changing either one changes the sensor. If a package is released with a “typical” time constant on the datasheet, the number is measured on a specific construction and it does not survive an unqualified material substitution.
Where the sensing element actually sits decides what it reads
A thermistor is a small resistive element. The question the package has to answer is how that element gets thermally connected to the outside world.
Three arrangements are used, and they measure different things.
- Element against the closed tip. The element sits at the end of a closed tube. This is the fastest and the most accurate arrangement, because the heat path from outside to element is a short, well-defined metal path. This is the construction to use wherever the tip can contact the medium.
- Element concentric in the bore. The element sits in the middle of a tube, surrounded by air or filler. This is slower by a factor that depends on the bore diameter, because heat has to travel radially inward as well as axially. It is used when the sensor must be insulated from the medium, or when the tip is not the sensing surface.
- Element in a surface mount. The element is bonded to the outside of a flat face. Extremely fast, because there is no enclosure at all, and entirely dependent on the bond and on mechanical mounting pressure.
A drawing that says “probe with thermistor” has not specified any of these. The bore diameter, the element diameter and the fill material together determine the answer, and they determine whether two nominally identical probes from two suppliers will read the same.
The stainless steel sheath is a decision, and thin-wall stainless steel is harder than it looks
Stainless Steel is chosen for corrosion resistance and mechanical robustness, and it has consequences at every thickness.
At 0.5mm wall and below, the practical limit is not the metal — it is the forming and joining. Thin-wall stainless steel tube cannot be cut and re-formed freely; it work-hardens and cracks. It cannot be closed by simple stamping of a cup unless the cup is very small. And the wall is a pressure boundary in a way that a thicker wall is not, which means the closure method has to be metallurgically sound rather than merely geometric.
The grades behave differently at thin wall in a way that matters here. The lower-carbon variants are preferred wherever welding or brazing is involved, because they are less prone to sensitisation and their weld ductility is better. Where the probe will be brazed closed at the tip, this is not a preference — it is the difference between a joint that survives cycling and one that cracks at the braze filler.
Also worth stating: stainless steel conducts heat far worse than copper. A stainless steel sheath will be measurably slower than a copper one of identical geometry. If a specific response time is a requirement, the material is a design variable, not a compliance item.
Lead resistance turns the sensor into a different sensor
Thermistor resistance is high — 10kΩ, 50kΩ and upwards — and that is a feature: it makes the lead resistance a small fraction of the total. But it is a fraction, not zero, and it varies.
A thermistor probe with a two-wire connection has its lead resistance in series with the element, so the reported resistance is always higher than the true value, and the offset changes with temperature and with lead length. For a 10kΩ element with 1Ω of lead, the error is a tenth of a percent and irrelevant. For a 50kΩ element at the end of a long cable run, the same 1Ω becomes negligible too — but a marginal connection, a partially corroded joint or a lead resistance that changes with temperature can shift the reading by an amount that matters at the bottom of the operating range.
The robust arrangements add leads rather than remove ambiguity:
- Three-wire. A third lead measures the resistance of the connection to the sensor and lets the measurement subtract it. Standard where the lead run is long or the accuracy requirement is high.
- Four-wire / Kelvin. Separate current and voltage paths eliminate the lead resistance entirely. The correct arrangement for a precision measurement, and the one to specify when the sensor is being used for a calibration or a reference rather than for control.
Which one is required follows from the lead length, the element resistance and the accuracy needed, and none of those three are visible on a mechanical drawing. They belong in the specification.
Predicting the response instead of measuring it
If a fast response is required, it is worth working out what the achievable time constant is before committing to a package. Two approaches are used.
The first treats the sensor as a lumped mass with a thermal resistance to ambient. The time constant is then the product of the two, and halving the mass or doubling the conductance halves the lag. This is adequate for a rough feasibility check and it catches the common error of specifying a long sheath in a thick material.
The second treats the sheath as a distributed thermal system, which matters once the probe length is comparable to the diffusion length inside the material. At that point the tip is not at ambient temperature even though the base is, and a lumped calculation understates the lag badly. For a stainless steel probe, the diffusion length grows with the square root of time, so a long probe of the same construction has a much slower tip response than its lumped time constant suggests.
Neither calculation replaces a measurement. But both of them catch the case that matters — a package that is dimensionally perfect and functionally too slow for the control loop it is being put into.
Questions worth settling before the package is released
- Is the thermal time constant a requirement, and if so at which temperature and to which step definition? A datasheet “typical” is not a specification.
- Where exactly is the sensing element relative to the closed tip, and what fills the space between them?
- Does the sheath wall thickness appear on the drawing as a toleranced dimension, given that it sets the response?
- Which low-carbon grade is specified if the sheath is welded or brazed closed?
- Is the connection two-wire, three-wire or four-wire, and is the requirement derived from the lead length rather than inherited from a previous design?
- Is lead length stated as part of the part, with its resistance tolerance, rather than left to the installer?
The last two are cheap to specify and expensive to get wrong. A probe that is correct at the bench and wrong in the field is almost always a lead problem, and the fix is impossible after the harness is built.
Deep Drawn Tubes & Sleeves is the parent range for parts like this one; the overview page gathers the tooling logic, finishing options and tolerance notes that apply across the family.
Specifications
| Item | Stainless Steel Open Tubes, Mixed Lengths Eight-Piece |
|---|---|
| Brand | AUNOV |
| Designing | Quotation follows the print, not a catalogue – unusual walls, threads or tolerances are feasibility-checked first; revisions are re-quoted once rather than silently absorbed. |
| Tooling | Die design and toolmaking stay in house, so a die can be re-cut against the approved drawing, and the charge is refundable against agreed volumes. |
| Samples | Prototype and pre-production samples are provided against the drawing, so the approved first article becomes the gauge reference, and the signed sample is kept for future reorders. |
| Materials | Stainless Steel 304 and 316L, pickled and annealed, 0.5-3.0 mm nominal wall – one grade chosen for corrosion duty and formability of this probe. |
| Surface finish | Surface treatment matched to the duty: plating, coating or polishing; masking and laser marking available where needed. |
| Processing | Tube drawing, straightening and turning in one flow; every stage is logged against your part number. |
| Equipment | Tube drawing benches with in-line annealing furnaces; first articles and batch samples are checked on the CMM. |
| Tolerance | +/-0.03 mm on machined features, +/-0.1 mm on formed ones; critical dimensions carry CMM data with the shipment. |
| Delivery terms | FOB Ningbo or Shanghai, T/T or L/C, with D/P accepted on repeat business; repeat orders reuse the approved first article. |
Related Products
Request a Quote for This Part
Send a drawing or sketch — get material advice, tooling cost and a per-piece price within 24 hours (working days).




