Guide
Does seat roughness or seat form control leakage in a metal-to-metal poppet valve?
Leakage in a metal-to-metal poppet valve is the flow that gets past a closed seat because two hard surfaces never touch perfectly all the way round. A University of Strathclyde study of pressure relief valves reports that seat form, not roughness, is the main surface-finish factor behind it, so a drawing that calls only Ra leaves the larger cause uncontrolled.
By Saola Manufacturing engineering team · Published · Updated
On this page
Does seat form or seat roughness control leakage?
In the one metal-to-metal study this guide relies on, form is the main surface-finish factor and roughness is secondary. Roughness still belongs on the drawing as a cap, but it cannot tell you whether the seat closes around its whole circumference.
The study is a 2019 doctoral thesis on pressure relief valve leak tightness. Its abstract states that industry experience points to form, not roughness, as the main surface-finish factor in leakage, and the author combined metrology with that experience to look at poli-lapping a new disc to improve its form, which reduced the leakage rate (University of Strathclyde thesis record, 2019). The numerical model in the thesis treats form, waviness and roughness together, and it reads leakage against the API 527 seat tightness method at a set pressure of 0.5 MPa (thesis repository page).
The three words mean different things on a seat. Form is the departure of the seat from its ideal shape: out-of-round, a bell-mouthed or crowned cone, a tilt to the axis. Waviness is the slower undulation around it. Roughness is the fine texture left by the cutting edge. A seat can have a very low Ra and still be out of round by more than the elastic compliance of the contact can absorb, and then the leak path is the gap that form leaves, not the scratches.
How far does a relief-valve result transfer to a small poppet or coupling valve?
As a direction, not as a number. The thesis concerned relief valve discs and seats, which are larger and loaded differently from a small coupling poppet, and many coupling poppets seal on an elastomer rather than on bare metal.
The thesis author also concluded that the API 527 method falls short of quantifying leakage, which means it indicates leakage rather than measuring it precisely (thesis repository page). So even within its own field, the leak number has limits, and a figure from one valve type should not be copied onto a different one.
What does carry over is the argument. A seat must be round and have a consistent section shape to close all the way round, and the more rigid the two mating surfaces, the less they can deform to close a gap. If your poppet seals on an O-ring or other elastomer seat, the metal surface behind it is a gland or a sealing face, and the groove and finish rules in the O-ring groove guide and the O-ring failure guide apply instead. State which kind of seat you have on the drawing, because the right callouts differ.
What else makes a poppet seat leak besides finish?
Contamination trapped between the seat surfaces and chatter that wears them. NASA names both as the causes of leakage in a metallic poppet-and-seat check valve built for extreme conditions.
A NASA technical brief from 1970 explains that a small opening distance can make the valve chatter under resonant conditions, giving erratic operation and excessive wear on the sealing surfaces, and that the small opening also lets contaminants sit between the seats so the valve cannot close completely (NASA Tech Brief 70-10688, December 1970). The remedy was a double seat that tolerates contamination of a certain particle size and a low-mass finger spring, so that any wear is taken on one sealing interface.
For a machinist that translates into edge and cleanliness requirements near the seat. The Open Compute Project coupling specification sets visual criteria that include no residual shavings or burrs, loose or attached, and no foreign material, rust or other evidence of corrosion, across all inspection zones (OCP UQD v2.0.0, table 10). Debris from machining is a leak mechanism in its own right, so put the burr and cleanliness note on the drawing next to the seat rather than in a general notes block.
Which callouts belong on the seat of a poppet or check valve?
Control the seat form and its relation to the guide with geometric tolerances, then keep Ra as a ceiling. A seat cone is best described by a profile or circularity tolerance plus runout to the guide diameter, with Ra and a no-burr note alongside.
The table lists the callouts, what each controls and why it matters for leakage. Where a row is marked as engineering judgment, it is this guide's translation of the form finding into drawing language, not a published rule. Choose the tolerance values from your own leak requirement and test result; no public source located for this guide gives a number for a small coupling seat.
| Callout | What it controls | Role in leakage | Basis |
|---|---|---|---|
| Circularity of the seat line or cone | Out-of-round of the sealing line | Direct, per the form finding | University of Strathclyde thesis (2019) |
| Profile of a surface on the seat cone | Cone angle and section shape together | Direct, a way of controlling form | Same study; the mapping to profile is engineering judgment |
| Runout of the seat to the guide diameter | Alignment of the closed poppet to its seat | Closure around the whole line | Engineering judgment |
| Ra maximum on the seat | Cutting-edge texture | Secondary cap; the OCP document gives 32 Ra maximum for sealing surfaces | OCP UQD v2.0.0 (2026), tables 1 and 2 notes |
| Edge, burr and debris note | Contaminant and damage sources | Prevents trapped particles and seal damage | OCP table 10; NASA brief (1970) |
| Lay of tool marks | Direction of the finish pattern | Turned seats give circumferential lay; marks crossing the seat can form leak paths | Engineering judgment |
What does an interoperability specification actually call out on sealing surfaces?
In the OCP quick disconnect document, a roughness cap, linear tolerances, a plug-end chamfer and visual defect limits. The socket and plug figures show diameters, angles and linear tolerances, and no circularity or profile frame appears in them.
The sealing and interoperability surfaces are called out as machined to 32 Ra maximum, with the note giving no unit; confirm it with the document owner, because 32 microinches is the usual American reading, which converts to about 0.8 micrometers (derived: 32 x 0.0254) (OCP UQD v2.0.0, notes to tables 1 to 4). Its visual criteria allow surface defects up to 0.3 mm² in some zones and 1.0 mm² in others, and a version appendix warns that a plug end with minimal or no chamfer may damage socket seals.
That is a sensible interchange specification, because it controls what two suppliers' parts must agree on. It leaves the internal valve seat to the coupling supplier, so a buyer writing a drawing for a custom poppet and seat has to add the form and alignment controls that an interchange document does not carry.
How should a supplier verify the seat?
Measure form and alignment on the machined seat before assembly, then confirm closure with a functional leak test on the assembled valve. The measurement tells you the metal is right; the leak test tells you the valve works.
Circularity is checked with a roundness trace or a coordinate measurement of the seat line, seat angle and profile with an optical or contact method, and runout to the guide diameter with the part located on the guide. The circularity guide and runout guide explain how to write and read those callouts. Leak testing methods and their limits are covered in the leak testing guide, which concerns housings but applies the same method logic.
For acceptance values, borrow the structure, not the numbers. The OCP document holds a hydrostatic test for a minimum of 3 minutes with no visible leakage and runs helium leak tests in accordance with ASTM E493 (OCP UQD v2.0.0, sections 9 and 9.2.4). Those criteria belong to that coupling family and its test conditions; they show how an acceptance note is built, not what limit your valve needs.
When is a lapping step worth asking for?
When a turned seat does not meet the leak requirement in testing, ask for lapping or another form-improving step as a separate, quoted operation. In the thesis, poli-lapping a new disc to improve its form reduced leakage.
Lapping is not turning or milling, so say in the RFQ whether the seat is to be delivered turned or lapped and quote the step on its own line. Saola turns the seat cone and guide diameter by CNC turning, ideally in one setup so both share one axis, and the first-article leak test shows whether that is enough; CNC milling covers cross holes and flats on the same body. If a secondary seat finishing step is needed, raise it in the RFQ as its own line so the quote can state whether it is offered, who performs it and where.
Ask for form data before and after the step on the first article, so the cost of lapping is justified by a measured change and not assumed. If the valve passes without it, leave it off the drawing.
Frequently asked
Is a finer Ra always better on a poppet seat?
No. The study points to form as the larger factor, so a very fine Ra on an out-of-round seat still leaks. A cap is sensible, and very coarse marks that cross the seat are worth avoiding, but do not spend your tolerance budget on Ra before form is controlled.
Which GD&T symbol fits a conical seat?
Profile of a surface describes a cone, and circularity describes the seat line. Runout to the guide diameter ties the seat to the axis the poppet moves on. Pick the combination that matches how the part will be measured.
Does this apply to O-ring seals?
Not directly. Elastomer seals close small gaps by deforming, and their gland and sealing face rules are different; see the O-ring guides linked above.
What leak rate should I specify for my valve?
Take it from your system requirement and write the test method with it. Published criteria, such as the hydrostatic and helium tests in the OCP coupling specification, belong to those products and their test conditions.
Can a CNC lathe hold the seat form needed?
A seat cone and guide diameter turned in one setup share one axis, which helps alignment. Whether the form meets your leak requirement is proven by measuring the first article and leak testing it.
Sources
- Investigation and model development of pressure relief valve leak tightness, doctoral thesis record, University of Strathclyde, University of Strathclyde, 2019
- Thesis repository page, Strathclyde, University of Strathclyde, 2019
- Low Leak Rate Poppet-and-Seat Check Valve, NASA Tech Brief 70-10688, NASA, December 1970
- Open Compute Project, Universal Quick Disconnect (UQD) Specification, version 2.0.0, Open Compute Project, document dated September 22, 2026