Guide
How do spring rate, seat diameter and spring-seat depth set a check valve's cracking pressure on the drawing?
Cracking pressure is the minimum inlet-to-outlet pressure differential at which a spring-loaded check valve first passes flow, and on a small machined valve it comes from the spring preload acting over the seat. The drawing controls it indirectly through the spring-seat depth, the seat diameter and the spring you buy, and the practical way to hold a window is a functional test written into the requirement, not a longer tolerance chain.
By Saola Manufacturing engineering team · Published · Updated
On this page
What is cracking pressure on a spring-loaded check valve?
It is the lowest differential pressure between inlet and outlet at which flow starts, and the valve is not fully open at that point. In a spring-loaded valve it is the upstream pressure needed to overcome the spring tension.
A valve maker's technical article defines cracking pressure as the inlet pressure at which the first sign of flow is present, expressed in psi, bar or kPa, and says it depends on the check valve design (DFT Valves, page last modified April 2026). A general reference describes it the same way for spring-loaded in-line valves and lists unseating head and opening pressure as other names (Wikipedia, Check valve).
How the first flow is detected is part of the definition, and the definition above leaves the method open. "First sign of flow" can mean a bubble stream on a gas test, a drip on a liquid test or a flow-meter threshold, and the closing point, often called reseal, is a separate reading on the way down. A requirement that says "cracking pressure" without a detection method and a fluid leaves room for two labs to measure two different numbers.
Which machined dimensions set the cracking pressure?
To a first approximation, cracking pressure equals spring preload divided by the area the inlet pressure acts on. The dimensions that matter are therefore the spring-seat depth, which sets how far the spring is compressed with the poppet closed, and the effective seat sealing diameter, while the spring rate and free length come from the purchased spring.
This is a derived relation, not a published tolerance. Preload is the spring rate k times the compression with the valve closed, and that compression is the spring's free length minus the closed length. Written out, P(crack) is approximately k x (L free - L closed) / A seat, where A seat is the area enclosed by the seat's sealing line. It ignores friction, seal drag and flow forces, so it ranks the contributors; it does not predict the number.
The closed length is a stack: the spring-seat depth in the body, the poppet length, and the position of the seat shoulder the poppet closes on. Two consequences follow from the same relation. First, for a given preload a softer spring compressed over a longer travel is less sensitive to a depth error, because the force error is the rate times the depth error. Second, a larger seat area lowers the pressure produced by a given force and also lowers its sensitivity to force errors. The effective sealing diameter is set where the poppet touches the seat, so seat angle and seat form, covered in the circularity guide, move it slightly.
No public source located for this guide gives a tolerance on spring-seat depth for a target cracking window, and the valve-maker articles and NASA brief give none, so none is invented here. Work the sensitivity from your own spring rate and your own seat area instead.
Why verify a cracking window with a functional test instead of tolerancing every dimension?
Because spring rate, free length, friction and seat geometry all stack, and none of the sources for this guide gives a dimensional tolerance that guarantees a cracking window. Set the window as a range and confirm it by testing assembled valves, with reseal as a second reading if closing matters.
The valve-maker article says plainly that cracking pressure depends on the check valve design, so a window taken from another maker's valve is not evidence for yours. Each contributor in the relation above has its own scatter: the spring rate and free length from the spring maker, the closed length from several machined features, and friction from the guide fit and the surface condition. Adding those up on paper tends to give either a window too wide to be useful or a tolerance chain too tight to machine economically.
A functional test cuts through that. It measures the result of the whole stack on the real spring, in the real fluid and orientation, and it catches the effects the first-order relation ignores. The table lists what a test-based requirement has to state, and where each item comes from. Conditioning cycles before the reading are worth stating because a valve that has been sitting closed may not read the same on its first actuation as on its tenth; the right number of cycles is yours to set from your own test data.
| Item | What to write | Basis |
|---|---|---|
| Window | Minimum and maximum inlet-to-outlet differential | Definition as first sign of flow, DFT Valves (2026) |
| Detection criterion | What counts as first flow: bubble stream, drip or flow threshold | Engineering judgment; the definition leaves the method open |
| Test fluid and temperature | Gas or liquid, and the temperature of the test | Engineering judgment |
| Orientation | Mounting direction, for example vertical flow-up | Weight of moving parts can overcome the spring flow-down, DFT Valves (2026) |
| Conditioning | Number of actuations before the reading | Engineering judgment |
| Reseal | Closing reading and allowed seat leakage, if closing matters | Engineering judgment |
| Spring | Part number, or rate, free length, wire specification and end style; mass if chatter is a concern | Stainless spring wire per ASTM A313/A313M; low-mass spring against chatter, NASA (1970) |
What should the drawing and the RFQ say about cracking pressure?
Write the requirement as a functional note with its test conditions, show the machined features as controlled dimensions, and call the spring out as a purchased part. Say whether the valve is fixed, adjustable or selected at assembly.
Functional note. Give the cracking window as a minimum and a maximum differential, the test fluid, the temperature, the detection criterion, the number of conditioning cycles and the mounting orientation. For example: cracking pressure [min] to [max], first steady bubble stream, [fluid], [temperature], after [n] conditioning cycles, vertical flow-up. The brackets are yours to fill; a window pulled from a catalog for a different design is not a basis for yours.
Spring. List the part number or the rate, free length, wire material and end style, and state that the spring is a purchased item whose tolerance is held by its maker. For a stainless spring, name the wire specification; ASTM A313/A313M covers stainless steel round wire intended especially for springs. If the machining shop buys the spring, require approval of the source.
Spring-seat depth. Dimension it from the same datum the poppet seat is measured from, so depth and seat position are not two separate chains. Control the squareness of the spring seat floor to the bore axis, because a tilted seat puts a side load on the spring and adds friction that shows up as scatter in the cracking pressure.
Seat. Call out the seat diameter, angle and form, plus a burr-free edge note. How to write the roughness and form callouts is in the Ra versus Rz guide and the runout guide.
Adjustable or selected. If a screw sets the preload, give its adjustment range and how it is locked. If springs or shims are selected at assembly, say so, since the machined depth then only needs to stay inside the window that selection can absorb.
What goes wrong when cracking pressure is too high or too low?
Too high raises pressure drop and can leave the valve only partly open at normal flow, where it may chatter and wear. Too low slows closing and, in some orientations, lets the poppet weight overcome the spring.
The valve-maker article says a cracking pressure that is too high causes a higher pressure drop and a valve that may not open fully, and that a valve not fully open during normal flow may pop or chatter, leading to premature wear. Too low reduces closing speed, and a valve in a vertical flow-down position may fail to close fully because the weight of the moving parts can overcome the spring (DFT Valves, What is cracking pressure). Orientation therefore belongs in the functional note.
A NASA technical brief on a metallic poppet-and-seat check valve describes the same mechanism from the seat side. A very small opening distance can drive resonant chatter, which causes erratic operation and excessive wear on the sealing surfaces, and the small opening also traps contaminants between the seats; the design answer was a double seat and a low-mass finger spring whose natural frequency sits above the range that excites chatter (NASA Tech Brief 70-10688, December 1970). The spring is not a passive number: its mass and natural frequency are design choices that the machined seat depth cannot fix.
What can a build-to-print supplier hold, and what stays with the valve designer?
The supplier can hold the machined dimensions, seat surfaces, edge condition and cleanliness on the drawing and report them. The cracking pressure itself is a system property that the design owner validates on assembled valves with the intended spring.
Saola machines bodies, poppets and spring seats by CNC turning and CNC milling at its lead plant in Vietnam. In practice that means reporting the spring-seat depth, the seat dimensions and the burr condition per lot, so that if a valve window shifts you can correlate it with the machined data instead of guessing. A burr on the edge of the spring seat bore is a small feature with a large effect, because it can hang up the spring or the poppet.
A first article that includes your spring is the sensible checkpoint: the supplier measures the metal parts, you or a lab run the cracking test on assembled valves, and the two data sets are compared before the order is released. If the test needs a gauge or fixture at the supplier, the functional gauge guide explains how to specify it, and the first article guide covers the report. Supplier-side testing of cracking pressure is an item to agree in writing rather than assume.
Frequently asked
Is cracking pressure the same as opening pressure or set pressure?
Opening pressure is listed as another name for cracking pressure in general check valve references. Relief valves use set pressure for their own opening point, so confirm which term your drawing means and which test defines it.
How tight can a cracking pressure window be?
No public tolerance data was found for spring-seat depth against a cracking window, and cracking pressure depends on the valve design. Set the window from tests on your own assembled valves with your own spring rather than assuming a narrow one.
If I double the spring rate, does cracking pressure double?
At the same compression, preload is the spring rate times the compression, so the force doubles, and the cracking pressure rises in the same proportion by the first-order relation above. Real valves also change with friction and spring stack, so check on assembled valves.
Why state conditioning cycles before the reading?
A valve that has sat closed may not read the same on its first actuation as after several. Stating the number of actuations before the reading makes results from two benches comparable; set the number from your own test data.
Do I need a functional gauge for cracking pressure?
Only if the supplier will test it. Many buyers test assembled valves themselves; if the supplier is to test, the fixture, fluid and detection method must be written into the requirement.
Sources
- What Is Cracking Pressure?, DFT Valves technical article, DFT Valves, page last modified April 2026
- Check valve, Wikipedia, Wikipedia, accessed October 2026
- Low Leak Rate Poppet-and-Seat Check Valve, NASA Tech Brief 70-10688, NASA, December 1970
- ASTM A313/A313M-18, Standard Specification for Stainless Steel Spring Wire, ASTM International, 2018, reapproved 2025