Views: 0 Author: J-VALVES Publish Time: 2026-11-11 Origin: Site
A gate valve is close to symmetrical: the wedge sits between two seat faces and either flange can be the upstream one. A globe valve is not. Its seat is machined on an internal web, and the disc arrives at that seat from one face only, guided by the stem. That asymmetry is the reason a globe can regulate, and it is also the reason the body is directional and carries an arrow. The arrow is not a preference. It states which side of the disc the pressure is meant to act on.
In the usual arrangement the medium enters below the seat and leaves above the disc, so pressure acts on the underside of the disc and pushes it away from the seat, in the opening direction. Catalogues call this flow to open. The valve opens easily and evenly, which suits a handwheel. The cost is paid at the other end of the stroke: to shut, the stem must push the disc down against that same pressure, so the last turns are the hardest ones, and an actuator has to be sized on closing thrust rather than on running torque.
Set the valve the other way round and the medium enters above the disc. Pressure now presses the disc onto its seat, so shutoff improves as line pressure rises and a smaller operator can close it. That is why the arrangement is used on high pressure steam where a tight shut is the whole duty. The cost moves to the opening stroke: the first turn carries the full load, the disc can break away suddenly once it cracks open, and throttling at very small openings becomes unsteady.
The arrangement also decides what the stuffing box sees while the valve is shut. With flow under the disc, the seat lies between upstream pressure and the bonnet, so the packing and the bonnet joint sit on the depressurized side. A gland weep is then a nuisance rather than a pressure boundary failure, and in steam the packing runs nearer the downstream temperature. With flow over the disc, everything above the seat is at full line pressure while the valve is closed: the same weep is a hazard and the packing runs at line temperature.
Reversing a globe valve does not stop it working, which is why the mistake survives commissioning and appears a year later. Four things change at once: the gland is permanently pressurized, the disc is no longer steadied by the pressure at small openings, the jet through the throttling gap strikes the other face of the seat, and the effort curve of the handwheel is inverted, so the operator's feel for the valve is wrong. In high differential service the seat pays, and the damage looks like erosion or wire drawing rather than an installation error.
State which side you intend to use as upstream, and why. A globe ordered for throttling is normally built flow to open; one ordered for a tight shut at high pressure may be built flow to close, and the two are not interchangeable even when the drawing looks the same. If the line can be fed from either end, or a pump can reverse the flow, that joint does not want a globe valve at all; a gate valve, which seals in both directions, is the honest answer.
| Valve | Is direction fixed? | What decides it | What reversing does |
|---|---|---|---|
| Globe | Yes | Seat web and the face the disc approaches | Packing permanently pressurized, unsteady throttling |
| Gate | No | Wedge guide and seat faces | Little, unless the body is marked for one side |
| Check | Yes | Hinge position and disc travel | Disc hangs off the seat, valve never closes |
| Butterfly | No | Disc and seat retention | Still seals, but the load path through the shaft changes |
| Strainer | Yes | Screen pocket and blow-down branch | Debris washes back onto the screen |
J-VALVES builds globe valves from 1/2 to 48 inch in ASME Class 150 to 2500, in WCB, CF8 and CF8M bodies with 13Cr or stainless trim, bolted bonnet or bellows sealed, shell tested at 1.5 times rating and seat tested at 1.1 times, and issued with an EN 10204 3.1 certificate. Every body carries a flow arrow and the order states the direction the valve is to be used in, so the two cannot drift apart. The globe valve range covers straight, angle and Y patterns; the bolted bonnet globe valve is the pattern most often built flow to open for throttling duty. What that valve is allowed to hold when a line is pressure tested is set out in our note on valves as test boundaries.
Can a globe valve be installed backwards? It will work for a while, and that is the problem. The valve still shuts, the handle still turns, but the packing sits under full line pressure whenever the valve is closed and the disc is no longer held steady at small openings. The first sign is usually a gland weep, then seat damage on a high differential line.
Why does the gland weep only when the valve is closed? Because that is when the stuffing box is pressurized. With flow over the disc this is expected behaviour, since the whole bonnet is at line pressure at shutoff. With flow under the disc it means the seat is passing, the valve has been reversed, or the packing was tightened cold and has since relaxed.
Does the arrow matter on a low pressure water line? Less than on steam, but it still matters. The differential is small, so the seat survives being on the wrong side, yet the gland is still pressurized at shutoff and the packing still wears on one side. Fit it the way the arrow points and the question never comes back.
