
A 304 stainless steel temperature sensor or thermocouple housing has a specification that is unusual on this site: it has to satisfy two measurement systems at once. A resistance thermometer and a thermocouple do the same job and disagree about what they need from the metal around them. A drawing that serves both without deciding which one it is will produce a part that works adequately for one and badly for the other.
Carbon steel, 304 and 316 stainless steel, aluminium and brass, with mill certificates and heat numbers on the first article. So the four alloys are available and the grade is a real decision – and the first decision is not the grade at all.
Which measurement system, because it decides the wall
- A thermocouple measures a voltage, and the housing must not participate in the circuit. So the wires are insulated from the housing, the wires do not touch the tube, and the housing’s alloy must be compatible with the sheath if the arrangement is grounded. And the housing’s conductivity works against accuracy by conducting heat along the tube.
- A resistance thermometer measures a resistance, and the housing must conduct. So the wall is short and the material is conductive, the element is thermally coupled to the process, and the housing’s alloy is a thermal parameter.
- So the same drawn shell, in the same alloy, with the same dimensions, is a good thermocouple protection tube and a poor resistance thermometer probe – and vice versa if the wall is thin and conductive.
Which means the first item on this part’s enquiry is: is this for a thermocouple or for a resistance element? It decides the wall, it decides whether the element sits inside the tube or against its inner surface, and it decides whether the alloy is chosen for the process or for the measurement.
The grounded question again, and it decides the alloy
This is the part of a thermocouple housing that carries the most risk and it is worth being very specific, because the answer changes which alloy the housing can be made from.
- An ungrounded housing is not in the circuit. The element’s junction is enclosed in a ceramic or a MgO fill inside the housing, and the housing is a mechanical and thermal part. Any of the four alloys works, chosen for the process.
- A grounded housing is in the circuit. The junction’s sheath is bonded to the housing’s inner end – welded, brazed or filled with an electrically conductive compound. Now the housing’s alloy and the sheath’s alloy form a junction, and that junction sits inside the temperature gradient.
- And a common sheath is a nickel-chromium alloy. So a grounded junction in a 304 housing has a 304-to-Inconel junction inside the gradient, and it will generate an EMF that changes with temperature and with immersion depth. That error is not a manufacturing defect and no amount of careful making will remove it.
- Which means the housing’s alloy for a grounded thermocouple has to be specified against the sheath, not against the process. That is the opposite of how a housing is normally specified, and it is the single most useful thing to establish on this part.
The welding, wherever there is any
Deep drawing presses plus CNC turning centres for machined faces. So the part is drawn and machined, and on a thermocouple housing there is often a weld as well – either a welded closure or a welded mounting flange. Two requirements come with it.
- 316L rather than 316 wherever there is heat input, because the low carbon resists sensitisation and a sensitised grain boundary in a pressure boundary or a thermowell is a leak site that will develop over years.
- And the weld’s own metallurgical compatibility with the sheath, where the weld is near the junction. A weld that is a dissimilar joint from the sheath reintroduces exactly the junction problem that the grounded design was trying to avoid.
The wall, and the same trade under two names
The wall on this part is a pressure boundary where the process is pressurised, a thermal path always, and a stiffness member where the part is handled or mounted. The thermal argument is the one that is specific to this part, and it goes differently depending on the measurement system.
- For a resistance thermometer, a thin wall in a conductive alloy is what makes the probe fast and accurate. So on this case the material is chosen for conductivity and the wall for thinness, and the pressure calculation is the constraint that stops both being as good as they would like.
- For a thermocouple, a thicker wall in a poorly conductive alloy is what reduces the conduction error, so the material is chosen for low conductivity and the wall for heaviness, and the response time is what suffers.
- Which means the two cases want opposite materials from the same alloy list – and the drawing cannot be written honestly until the case is decided. This is the practical expression of the point at the top of the page and it is worth raising because it is invisible on a dimensional drawing.
A grounded probe adds one finish requirement the open case does not have
Zinc, nickel or chrome plating, anodising, passivation or e-coat, with finish thickness and adhesion checked on the first article. On this part the finish has three requirements and one of them is specific to the grounded case.
- The sealing face, unmasked. It is the process boundary and a plated face will not seal.
- The thread, masked or specified post-finish with an allowance – and on a threaded probe the flakes go into the process.
- The interior, where the fill or the element sits, should be clean and free of any residue or lubricant, because any of it will contaminate the fill or interfere with a bonded sheath.
- And for a grounded housing, the inner end’s surface is the electrical contact – so whether it is masked, plated, or left bare is a functional specification, not a finishing preference. A passivated inner end may not bond reliably to a conductive fill, and a plated one changes the contact resistance. On a grounded design this is one of the few places where the surface finish is part of the measurement circuit.
The flagged list is the ends and their relationships, accuracy included
+/-0.01 mm on sealing faces and +/-0.05 mm on drawn profiles, with dimensional reports accompanying the first article. On this part the flagged dimensions are the ends and the relationships, and one of them is the accuracy.
- Sealing face: flatness and perpendicularity, measured as two values, against a named datum.
- Bore concentricity to the sealing face, where the part is located by its outer surface or by a spigot, because an eccentric tube reads differently at different immersion depths.
- Length, as a functional dimension with a tolerance, because it is the immersion depth and therefore the error budget.
- And the requirements that are not dimensions: grounded or ungrounded, the sheath’s material, the minimum wall, and the tube’s bore relative to the sheath. Those four determine whether the part measures and none of them is a tolerance.
A part serving two systems must say which one it is verified against
First articles and batch samples on the CMM, with dimensional reports accompanying the first article. On a part that serves two measurement systems the inspection plan has to say which system it is being verified against.
- Sealing face and concentricity, as above.
- A functional test where the part is supplied with an element – an EMF check at a reference temperature for a thermocouple, or a resistance check against the curve for a resistance element. On a drawn and machined housing, that is the only test that verifies the element’s position and the housing’s thermal coupling together.
- Material verification with the heat number, which on a part where the alloy may be chosen against a sheath’s compatibility is not a paperwork requirement but a correctness one.
- And where the part is grounded, a continuity or contact check at the inner end, because a grounded design with a bad contact produces a reading that is wrong in a way nobody will diagnose as a contact problem.
Thermocouple or resistance element, and grounded or not?
Is this for a thermocouple or for a resistance element, and is the thermocouple arrangement grounded or ungrounded? Three answers, and they decide the wall’s thickness, the alloy, the interior’s treatment, the surface finish on the inner end, and the tolerance scheme. It is worth asking because the two systems want opposite things from the same housing and a drawing that does not decide will deliver a compromise that is acceptable for neither. And if it is a thermocouple, the second question – grounded or not – determines whether the alloy is chosen against the process or against the sheath, and the sheath is the part of the assembly nobody discusses. On a grounded design with a nickel-chromium sheath, that single unasked question is the difference between an accurate instrument and one that reads a degree low and drifts with immersion depth.
You will find this part in our Sensor Housings range. Process notes, common materials and inspection points for every product in the family are collected across the sensor housings pages.
Specifications
| Item | Stainless Steel Thick-Wall Open Cylinder Sleeve |
|---|---|
| Brand | AUNOV |
| Designing | Send drawings, or send a sample to be measured – a manufacturability review comes back with the quotation; revisions are re-quoted once rather than silently absorbed. |
| Tooling | Dies are made and maintained in house, which keeps tooling lead time and cost under control. |
| Samples | Pre-production samples come with a full inspection report, so the approved first article becomes the gauge reference, and production starts only after your written sign-off. |
| Materials | Carbon steel, 304 / 316 stainless steel, aluminium and brass are all held in stock; mill certificates and heat numbers accompany the first article. |
| Surface finish | Zinc, nickel or chrome plating, anodising, passivation or e-coat; finish thickness and adhesion checked on the first article. |
| Processing | Drawn shells receive machined seats and threads where the print asks; filed batch records let a repeat order start from approved data. |
| Equipment | Deep drawing presses plus CNC turning centres for machined faces; first articles and batch samples are checked on the CMM. |
| Tolerance | +/-0.01 mm on sealing faces, +/-0.05 mm on drawn profiles; dimensional reports accompany the first article. |
| Delivery terms | FOB Ningbo or Shanghai, payment by T/T, L/C at sight or D/P; call-off schedules are supported. |
Related Products
Frequently Asked Questions
What is unusual about a housing that serves both a resistance element and a thermocouple?
This housing has to satisfy two measurement systems at once, and a resistance element and a thermocouple want different things from the same metal, so a drawing that serves both without deciding will suit one badly. Carbon steel, 304 and 316 stainless steel, aluminium and brass are held, with mill certificates and heat numbers on the first article.
How does the grounded question decide the alloy?
On a grounded design with a nickel-chromium sheath, that single unasked question is the difference between an accurate instrument and one that reads a degree low and drifts with immersion depth. Three answers decide the wall’s thickness, the alloy, the interior’s treatment, the surface finish on the inner end and the tolerance scheme.
What do the three answers decide about the wall and finish?
Deep drawing presses plus CNC turning centres produce the machined faces, and on a thermocouple housing there is often a weld as well, either a welded closure or a welded mounting flange. Zinc, nickel or chrome plating, anodising, passivation or e-coat is applied with thickness and adhesion checked on the first article.
Against which system must a part serving two be verified?
It has to say which one it is verified against, at +/-0.01 mm on sealing faces and +/-0.05 mm on drawn profiles with dimensional reports accompanying the first article. A part serving two systems without that statement is only half specified.
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).




