I keep seeing Plan 23 specified on pumps running at 160°F with clean water. At that temperature, Plan 21 handles the job with half the components and none of the venting headaches. The industry consensus treats Plan 23 as the universal upgrade, but in my experience, for standard services below 200°F, Plan 21 is the better default — and the simpler system means fewer failure points. Both belong to the broader family of API 682 flush plans.
This page assumes cooling is already on the table — that is, Plan 11’s discharge recirculation is no longer adequate for your seal temperature, and the upstream Plan 11 vs Plan 21 question is settled. If that upstream question is still open, the Plan 11 vs Plan 21 decision walks the temperature thresholds and lubrication margins where adding a cooler at all becomes worth the complexity. Once that is decided, the within-cooling choice — single-pass Plan 21 vs recirculated-loop Plan 23 — is what this page resolves.
How the Circulation Source Changes Everything
The entire difference between these two plans comes down to one question: where does the circulating fluid come from?
Plan 21: Single-Pass Cooling from Pump Discharge
Plan 21 taps fluid from the pump discharge, routes it through an orifice and a heat exchanger, then delivers it to the seal chamber. The orifice controls flow rate. The cooler knocks down temperature. Fluid passes through once and returns to the process. That is the entire system — an orifice, a cooler, and piping.
The simplicity is the point. Flow rate is directly controlled by orifice size and differential pressure. No pumping ring, no throat bushing, no venting. When something goes wrong with Plan 21, you check two things: the orifice and the heat exchanger. API 682 does not list Plan 21 as a preferred plan because of the higher heat load on the exchanger in hot service. That criticism is valid above 200°F — but it assumes high-temperature applications that many standard pumps never see.
Plan 23: Closed-Loop Recirculation with Pumping Ring
Plan 23 isolates the seal chamber from the pump discharge entirely. A pumping ring mounted behind the seal faces circulates a small volume of fluid through a heat exchanger and back into the seal chamber. A close-clearance throat bushing thermally separates the cooled seal environment from the hot pump body.
The critical detail most engineers overlook: Plan 23 circulation is not controlled directly. The flow rate depends entirely on the pumping ring’s performance within the closed loop — influenced by ring design, fluid viscosity, piping layout, and heat exchanger pressure drop. Proper venting is essential. Air trapped in the closed loop kills the thermosyphon effect and starves the seal faces. This is an emergent system, not an engineered one, and that distinction matters when you are troubleshooting at 2 AM.
Efficiency and Cooling Performance Compared
Plan 23 can cut cooling water demand by over 90% compared to Plan 21 in a water-to-water system. That number is real, and I will not pretend otherwise. The reason is straightforward: Plan 21 cools the entire process fluid stream from pump discharge temperature, while Plan 23 only needs to remove seal-generated heat and heat soak from the pump body.
But context determines whether that 90% matters. On a pump running at 350°F with heavy hydrocarbon, the delta-T between process fluid and desired seal chamber temperature is enormous — Plan 23’s closed-loop approach saves significant cooling water and energy. On a pump running at 170°F with clean water, the heat duty on a Plan 21 exchanger is modest. The cooling water savings from Plan 23 might amount to a few gallons per minute that nobody notices on the utility bill.
The efficiency advantage of Plan 23 scales with temperature. Below 200°F, that advantage shrinks to the point where the added complexity of a pumping ring, throat bushing, and venting requirements becomes the dominant cost — not the cooling water.
Plan 21 or Plan 23: A Decision Framework
“Hot applications” without a number is the kind of vague spec that puts the wrong plan on the pump. The thresholds below define hot in measurable terms — temperature, vapor margin, and cooling-water availability — so the decision is auditable instead of intuitive.
Below 140°F: Plan 21 Territory
Plan 23 should not be used if process temperature is below 140°F (60°C). At these temperatures, the heat duty is low enough that Plan 21’s single-pass cooling handles it comfortably. Adding a pumping ring, throat bushing, and venting system adds cost and failure modes with no meaningful cooling benefit. Stick with Plan 21. If Plan 21 is not cooling adequately at these temperatures, the problem is orifice sizing or heat exchanger fouling — not the plan itself.
Above 200°F: Plan 23 Advantage
Above 200°F (93°C), Plan 23’s closed-loop approach starts earning its complexity. The heat duty differential between the two plans becomes significant, high-temperature service demands more aggressive cooling, and Plan 23’s ability to isolate a small fluid volume from the hot process stream delivers real performance gains. For clean hydrocarbon service above 300°F, Plan 23 is not just preferred — it is the only practical option for a single seal arrangement.
The 140-200°F Gray Zone
Three factors drive the decision in this range:
- Fluid cleanliness. Clean fluids with low fouling potential favor Plan 21 — the heat exchanger stays effective longer. Dirty or polymerizing fluids favor Plan 23’s smaller closed-loop volume.
- Cooling water availability. If cooling water is expensive or scarce, Plan 23’s 90% reduction justifies the complexity even at moderate temperatures.
- Viscosity. High-viscosity fluids degrade Plan 23 pumping ring performance. If the process fluid is highly viscous, the pumping ring may not generate adequate circulation, and Plan 21’s pressure-driven flow is more reliable.
For most standard water and light hydrocarbon services in this range, in my experience, Plan 21 remains the right call. The flush plan selection decision should be driven by operating conditions, not by the assumption that Plan 23 is inherently superior.
Where Each Plan Goes Wrong
Plan 21 Failure: Undersized Orifices
I worked on a sour condensate pump application where the seals were lasting seven months against a 36-month design life. Process temperature was running 260-270°F — already above the 246°F design spec — but the root cause was not the temperature alone. The Plan 21 orifices were sized at 0.125 inches, restricting flush flow below 3 gal/min. That is a starvation condition. The seal vendor recommended opening one orifice to full bore to restore adequate flow. The plan was not wrong; the implementation was.
Plan 21 failures almost always trace back to orifice sizing or heat exchanger fouling. Check the orifice first. If flush flow is below spec, do not jump to Plan 23 — fix the orifice.
Plan 23 Failure: Restriction Bush Positioning
A poorly positioned restriction bush at the bottom of the seal chamber creates what I call the swimming pool effect. Instead of isolating a small fluid volume for the closed circuit, the entire seal chamber volume enters the loop. That large volume gets constantly superheated by thermal convection through the pump casing — seal face overheating, or an oversized heat exchanger that costs three times what it should.
The elegant solution uses cartridge seals with integral restriction bushes that minimize the closed-circuit volume. Smaller volume means more frequent passes through the heat exchanger, better cooling, and the possibility of air-cooled systems that eliminate cooling water entirely. But this requires proper engineering during seal selection — not a field decision.
Making the Right Call
Match the plan to the operating conditions, not to the assumption that newer means better. Below 200°F with clean fluid, Plan 21 delivers reliable cooling with an orifice and a heat exchanger — no pumping ring, no venting, no uncontrolled circulation. Above 200°F, Plan 23 earns its complexity with real cooling advantages that Plan 21 cannot match. In the gray zone, let fluid properties and cooling water availability make the decision. The API 682 standard specifies these plans for a reason — both exist because both have applications where they are the correct choice.
