High pressure trips the switch, the solenoid drops, the actuator vents, the process valve shuts. Pressure falls back, the switch resets, the solenoid re-pressurizes, and the process valve opens again with nobody in the room. That auto-reopen is the fault: the trip path has no memory. Nothing between the pressure switch and the actuator stores the fact that a trip happened, so the circuit follows the switch up and down all night.
Stop Shopping for a Single Latching Pushbutton
The part being looked for does not exist as one catalogue item, and the reason matters. A latch needs two inputs and a memory: the process trip sets it, the operator clears it. A pushbutton valve only has one input, the operator's thumb. Detented and push-pull versions latch on their own actuation, not on a pressure signal, so they can never be tripped by the process. Miniature pneumatic logic ranges such as Kuhnke sell the pieces of this function; the assembly is yours to build, and it is two or three fittings' worth of work.
| Part | What it does | Latches on | Verdict |
|---|---|---|---|
3/2 pushbutton, momentary, NC |
Passes air while held | Nothing | Use as the reset input |
3/2 pushbutton, detented or push-pull |
Holds its own position | Operator's push | Wrong function — no process input |
5/2 or 3/2 double air pilot spool |
Holds last commanded position with both pilots vented | Either pilot signal | This is the memory element |
3/2 normally open, air piloted |
Inverts a pressure signal | — | Converts loss of pressure into a trip pilot |
| Non-return valve + manual bleed on the pilot line | Traps pilot pressure until bled | Trip pilot | Temporary latch, leaks down |
| Non-return valve in the actuator supply | Traps air in the actuator | Nothing | Lock-up — holds the process valve open on trip |
Before ordering anything, prove the actuator direction: bleed the actuator and watch which way the process valve goes. Air-to-open with spring return is what the described circuit implies, and the whole design below assumes venting the actuator closes the valve.
Map the Trip Path Before You Cut Any Tube
The trip presents as a falling signal — pressure disappearing from the actuator line. Memory valves set on a rising pilot. That mismatch is the piece everyone misses, and it decides what you buy.
- Put a gauge on the solenoid outlet, at the actuator port.
- Record instrument air supply pressure at the branch, running.
- Record solenoid outlet pressure with the process healthy, and again with the pressure switch tripped.
- Time the closing stroke with a stopwatch or the closed limit switch. Write it down; this is the number the modification must not spoil.
- Read the minimum pilot pressure off the datasheet of whatever double-pilot valve you plan to fit, and compare it to the supply pressure you just measured.
If the solenoid outlet holds pressure in the tripped state, the solenoid is energize-to-trip rather than energize-to-hold, and the inverter in the next stage is not needed — you pilot the memory directly. Do not guess this from the wiring drawing; read the gauge.
Plumb the Memory Valve Between the Solenoid and the Actuator
Fit the double-pilot valve in the actuator branch, downstream of the existing solenoid. Cutting the solenoid's supply upstream instead leaves the actuator connected to a dead line with no exhaust path when the solenoid is energized, and the valve then closes only as fast as the circuit leaks.
- Port
1from the solenoid outlet, port2to the actuator, port3to a full-bore silencer. - Match or exceed the existing tube and port size. Every added elbow and metre of tube is volume the spring has to push out.
- If the closing time grew against your baseline, fit a quick-exhaust valve directly on the actuator port and re-time it.
Check it before the pilots are connected: isolate the process, pulse pilot 12 from a hand line, then vent it. The spool must stay where you put it. Pulse 14, vent it, same result. A spool that drifts back with both pilots vented is not bistable and is the wrong part.
Make the Trip Set the Latch and Beat the Reset Button
Generate the trip pilot with a normally open, air-piloted 3/2 used as an inverter. Pilot it from the solenoid outlet line, feed its supply from clean instrument air upstream of the solenoid, and land its output on the memory valve's trip pilot. When the pressure switch trips and the solenoid outlet collapses, the inverter opens and sets the memory. The memory vents the actuator and stays vented when the solenoid comes back.
Trip dominance costs nothing if you take it from the same line: feed the reset pushbutton's supply from the solenoid outlet. The button then has air only while the process is healthy, so a trip strips its supply and pressing it does nothing. The pushbutton must be momentary and spring-returned. A detented button left pushed holds one pilot while the inverter holds the other, and a double-pilot spool with both pilots pressurized stalls between positions — you get an actuator that is neither open nor closed.
Prove it with the trip still active: raise the pressure switch with a test pump, hold it there, press and hold the reset button. The actuator must not move. Tell operations the other consequence while you are there — after a total air failure, the inverter asserts trip as soon as air returns, so the valve comes back closed and latched.
Overnight Option: Check Valve and Manual Bleed
Get it running, then fix it properly. With a single-pilot, spring-return 3/2 in the actuator branch, run the trip pilot in through a non-return valve and put a small manual bleed valve on the pilot line as the reset. The check traps the pilot pressure, the trapped pressure is the memory, and the bleed clears it. Every part is on the shelf.
It is a time-limited latch, not a latch. A seat leak past the non-return, a pilot spool seal, or one imperfect fitting drops the trapped pressure below the valve's minimum pilot pressure and the spring returns the spool — the process valve reopens silently and nothing in the circuit says why. A small added volume on the pilot line stretches the hold time; it does not fix the mechanism.
Do not confuse this with a non-return valve in the actuator supply. That is a lock-up: it traps air in the actuator and holds the process valve open through the trip, the exact opposite of what is wanted.
Test before you leave it: latch the circuit, put a gauge on the trapped pilot line, and read it at the end of the shift. Anywhere near the minimum pilot pressure and the bistable valve gets ordered in the morning.
Run the Full Trip-and-Reset Test
Isolate the process and stroke on air only. Then work through the sequence without shortcuts.
- Raise the pressure switch with a test pump. The solenoid drops, the memory shifts, the actuator vents.
- Time the close stroke and compare with the pre-modification baseline. Slower means restriction or trapped volume in the new run.
- Remove the simulated pressure. The process valve must stay closed. This is the whole point of the job.
- Press reset with the trip still simulated. Nothing moves.
- Clear the trip, press and release reset. The valve opens and stays open with the button released.
- Close the air supply isolation, then reopen it. Record what the valve does; expect closed and latched.
- Soap every new joint, silencer and pilot fitting at working pressure.
- Latch it and leave it for a shift. No self-reset.
Mark up the P&ID and the loop sheet, tag the reset button at the point of use, and add the latch to the periodic trip proof test so it gets exercised rather than discovered. Keeping the reset pneumatic means no new electrical device, no new cable and no new certified enclosure in the area — the existing solenoid and pressure switch keep their certification — but the added components still belong in the site's hazardous-area records, with exhausts routed clear of enclosures and away from personnel.
Stop here if this trip is part of a certified safety function or a documented overpressure protection path. Adding components in series with a trip changes the stroke time, the failure modes and the proof-test case, and none of that is yours to sign off from the shop floor. Take the marked-up circuit to the valve and actuator OEM's technical support and to whoever holds the site's functional safety and Ex dossier before it goes back into service.
FAQ
Can I buy a 3/2 pushbutton valve with the latch built in?
No. A detented or push-pull pushbutton latches on the operator's own action and has no process input, so a high-pressure event cannot set it. Build the function from a double air pilot memory valve plus a momentary normally closed pushbutton on the reset pilot.
Does the latch hold through an air supply failure?
A double air pilot spool holds its last position mechanically with both pilots vented, so it stays where the trip put it. A check-valve-and-bleed latch does not — the trapped pilot decays and the spring returns the spool once pressure falls below the valve's minimum pilot pressure.
Can an operator reset the valve while the pressure switch is still tripped?
Not if you feed the reset pushbutton's supply from the solenoid outlet line, because a trip removes air from the button. Also avoid a detented reset button: a held reset pilot plus an active trip pilot stalls a double-pilot spool between positions.
Does adding a memory valve slow the trip stroke?
Yes — the extra tube, fittings and valve add volume and restriction to the exhaust path. Re-time the closing stroke against the pre-modification baseline and fit a quick-exhaust valve at the actuator port if it has lengthened.
Can a non-return valve on its own act as the latch?
Only in the pilot line, where it traps the trip pilot until a manual bleed clears it. Placed in the actuator supply it becomes a lock-up that holds the process valve open during the trip, which is the failure you are trying to prevent.