10.2 Valve Trim, Leakage Class, and Characteristics

Trim is the part that sees the vena contracta

Trim is the plug, seat, cage, and associated guiding that set capacity, characteristic, leakage, and how the valve fails in flashing or noisy service. The body is a pressure boundary. The trim is the control element. Topic 3.B on the 2027 PE Control Systems outline is where examinees lose items by mixing up shop leakage class, inherent characteristic, and installed gain.

IEC 60534 and ISA-75 are the industrial documents that define these quantities. They are not among the 2027 supplied design standards (those remain ISA-5.1 2024 and IEC 61511-1 2018). You still need the concepts: the handbook will not walk you through FCI 70-2 Class IV versus Class VI, and ISA-5.1 will only show a valve symbol, not a leakage test procedure.

---

Seat leakage classes (industry practice)

FCI 70-2 and IEC 60534-4 define shop leakage tests with air or water at specified ΔP, not “how much it leaks after three years in slurry.” Qualitative classes you must recognize:

Class IV is the default metal-seat control-valve expectation. Class VI is not “better metal”; it is a soft-seat test. Specifying Class VI on 800 °F steam is a materials contradiction, not a tightness upgrade. Class II is what you should expect when you chose a double-port globe for unbalance reasons in 10.1.

Leakage class is not the same as IEC 61511 proof-test leak tightness for a safety block valve. A BPCS globe can be Class IV and still be the wrong final element for a SIF that must be demonstrated tight in accordance with the safety requirement specification.

---

Noise-reduction and anti-cavitation trim

Hydrodynamic noise (liquids) is mostly cavitation: local pressure at the vena contracta drops below vapor pressure, bubbles form, then collapse if downstream pressure recovers above vapor pressure. Flashing is the neighboring case where $P_2$ stays below vapor pressure and the mixture does not collapse back to liquid in the valve. Aerodynamic noise (gases and steam) scales with jet velocity and mass flow; it is a sound-power problem, not a bubble-collapse problem.

Low-noise / anti-cavitation trim (drilled cages, multi-stage stacked disks, tortuous-path labyrinths, downstream diffusers) does three practical things:

1. Splits the drop into stages so no single vena contracta falls as far below vapor pressure.
2. Replaces one large jet with many small jets, shifting frequency and cutting acoustic efficiency.
3. Lowers peak velocity for gas service.

Those trims reduce $C_v$ relative to a simple window cage. If you size on a full-port $C_v$ and then install five-stage trim, the valve will not make the required flow. Noise trim is not a silencer you bolt on after a butterfly is already undersized on recovery; high-recovery rotaries remain poor cavitation machines even with a plate attenuator.

IEC 60534-8-3 / 8-4 prediction methods exist in professional practice. They are not 2027 supplied standards. On the exam, reason qualitatively: high $F_L$ globe plus staged trim for liquid letdown; do not pick a butterfly as the “quiet” high-ΔP steam valve.

---

Inherent characteristic: linear, equal-percentage, quick-opening

Inherent characteristic is the relationship between volumetric flow and fractional travel when $\Delta P$ across the valve is held constant (manufacturer’s bench or a test stand).

• Linear: $f(x) = x$. Equal travel increments produce equal flow increments at constant ΔP. Inherent valve gain $dq/dx$ is constant on the bench.
• Equal-percentage: $f(x) = R^{(x-1)}$ with rangeability $R$ typically about 30:1 to 50:1. Equal travel increments produce equal percentage changes in the current flow. Inherent gain is low when the valve is nearly closed and high when it is nearly open.
• Quick-opening: a disk or similar geometry that delivers most of the flow in the first 25–40% of travel. Inherent gain is huge near the seat and then flat. Use it for on-off, relief bypass, or snap-acting service — not for a PID throttling loop that must modulate across 20–80% travel.

Cage windows are cut to produce these curves. A V-notch ball is usually equal-percentage. A butterfly’s inherent curve is its own family (often between linear and equal-percentage depending on disk angle) and is not a reason to treat every rotary as equal-percentage without looking at the published $f(x)$.

---

Installed characteristic and loop gain

The plant does not hold valve ΔP constant. Piping, orifices, exchangers, and the pump curve take a share of the available drop that increases with $Q^2$ (turbulent). At low flow the valve sees almost all of the available ΔP; at high flow the valve may see only a small remainder.

Flow through the installed valve still follows $Q \propto f(x),C_v\sqrt{\Delta P_v(Q)}$. As $Q$ rises, $\Delta P_v$ falls, so $\sqrt{\Delta P_v}$ shrinks. That shrinkage reduces installed gain toward the open end if $f(x)$ is linear.

Loop gain for a simple flow or pressure loop is the product of controller gain, transmitter gain, process gain, and valve gain $K_v = dq/dx$ installed. If $K_v$ varies 3:1 or 4:1 over the operating range, a controller tuned at one end is sluggish or oscillatory at the other.

---

Worked: why equal-percentage linearizes a varying-ΔP loop

Take a pump discharging to a header. Available ΔP at zero flow is 40 psi (shutoff head minus static). Piping plus exchanger drop is $kQ^2$ and equals 30 psi at the design flow of 100 gpm, leaving 10 psi on the valve at 100 gpm. At 25 gpm, piping drop is $30 \times (0.25)^2 = 1.9$ psi, so the valve sees about 38 psi.

Suppose a linear inherent valve is 40% open at 25 gpm and 80% open at 100 gpm (the actual travels depend on $C_v$; the shape is what matters). Relative to the 25 gpm point, $\sqrt{\Delta P}$ at 100 gpm is $\sqrt{10/38} \approx 0.51$. The linear trim does not increase $f(x)$ fast enough to offset that: installed $dq/dx$ falls as the valve opens. The loop is hot near the seat (high ΔP, still-appreciable $f(x)$ slope) and lazy near wide open.

Equal-percentage trim does the opposite on the bench: $f(x)$ and $df/dx$ rise with travel. That rising inherent gain is intended to cancel the falling $\sqrt{\Delta P_v}$. The installed curve is then much closer to a straight line from 25 gpm to 100 gpm, so $K_v$ is closer to constant and one PID gain works across the turndown.

If instead the valve takes nearly all of the system drop at every flow (oversized pump, short pipe, valve as the only resistance — a level-control drain from a tank with constant head is the textbook case), ΔP is almost constant. Linear inherent trim then stays linear when installed. Equal-percentage would over-compensate and make the loop more aggressive as the valve opens.

---

Characteristic versus process

Exam trap: “equal-percentage is always better” is false. Equal-percentage is better when installed ΔP varies downward with flow. Linear is better when ΔP is essentially constant. Quick-opening is not a throttling characteristic.