Datacenter
Fire pump acceptance test field guide for data centers
How the fire pump flow test proves a building has the water its sprinklers and standpipes were designed for, and what the three test points, the certified curve, and the witnessed signoff each have to show.
Direct answer
A fire pump acceptance test proves the pump delivers its rated flow and pressure at three points: churn with no flow, 100 percent of rated flow at rated pressure, and 150 percent of rated flow at not less than 65 percent of rated pressure. NFPA 20 governs the test, but the AHJ and the certified pump curve control acceptance.
Key takeaways
- A fire pump acceptance test checks three points: churn at no flow, 100 percent rated flow at rated pressure, and 150 percent rated flow at 65 percent or more of rated pressure.
- NFPA 20 caps churn (shutoff, no-flow) pressure at not more than 140 percent of rated pressure to avoid over-pressuring downstream piping.
- Net pressure (discharge minus suction) is the value plotted against the certified curve; read discharge, suction, and flow at the same moment at every point.
- A fire pump that fails the 150 percent point is usually starved on suction (cavitating) or not turning rated speed, not worn out.
- NFPA 20 requires the witnessed test be coordinated with the AHJ, and the manufacturer's certified pump curve must be present before the test runs.
The fire pump acceptance test, and what it proves
A fire pump acceptance test proves the pump can deliver the flow and pressure the sprinkler and standpipe systems were designed to receive, measured across its full operating range and compared against the manufacturer's certified curve. It is the field demonstration that the pump on the pad actually makes the numbers the design assumed when someone sized the heads, the risers, and the standpipe outlets.
The pump exists because the water supply usually cannot do the job alone. A municipal main might hold 50 psi static, but a sprinkler system at the top of a tall building, or the far end of a sprawling data center campus, needs far more pressure and flow than the street can push to that point. The pump makes up the difference. The acceptance test is where you find out whether it really does, before anything depends on it.
This is a witnessed test, run once at turnover, not a checkout the installer does alone and reports on. NFPA 20, the standard for the installation of stationary pumps for fire protection, governs the acceptance test. The result becomes the baseline the building is maintained against for its entire life, so the test is run carefully and the numbers are recorded honestly. A pump that was never properly flow-tested is a pump nobody can trust and nobody can trend.
The fire pump system: what you are actually testing
The acceptance test exercises a system, not just a pump, so it helps to name the parts before you run it. The main pump is the centrifugal unit that makes the pressure and flow, driven either by an electric motor or a diesel engine. The driver choice changes the whole test, which is why the two get separate sections below.
Around the pump sit the pieces that make it usable. The controller is the brain: it starts and stops the pump, runs the start sequence, and reports the alarms. The jockey pump, sometimes called the pressure-maintenance pump, is a small pump that holds system pressure between events so the big pump only starts on a real demand. The suction piping brings water in, the discharge piping sends it out, and a check valve and control valves sit in the line. A relief valve protects against over-pressure on diesel and variable-speed installations.
The test gear is its own list. A test header with multiple hose valves lets you dump measured water to atmosphere, or a closed-loop flow meter does the same job back to suction or supply. A calibrated pitot or playpipe reads each hose stream. Gauges read suction and discharge pressure. On an electric set you read volts and amps; on a diesel you read engine speed and the instrument panel. Get the instruments wrong and the whole test is wrong, no matter how well the pump runs.
| Component | What it does in the test |
|---|---|
| Main pump (electric or diesel) | Makes the rated flow and pressure; the thing under test |
| Controller | Starts, stops, sequences the pump and reports alarms |
| Jockey / pressure-maintenance pump | Holds system pressure so the main pump starts only on real demand |
| Suction and discharge piping | Feeds the pump and delivers water; read pressure on both sides |
| Relief valve | Protects against over-pressure on diesel and variable-speed units |
| Test header or flow meter | Lets you flow measured water at each test point |
| Pitot / playpipe and gauges | Measure flow and pressure at each point |
What are the three fire pump test points?
The three test points are churn with no flow, 100 percent of rated flow at rated pressure, and 150 percent of rated flow at not less than 65 percent of rated pressure. Together they define the working shape of the pump curve, from shutoff down to the overload end where the pump has to keep making useful pressure while moving half again its rated water.
Churn, also called shutoff or no-flow, is the pump running with the discharge valve closed and nothing flowing out. It reads the highest pressure the pump makes. NFPA 20 caps it: churn pressure should not exceed 140 percent of the rated pressure at rated flow, because a pump that churns too high can over-pressure the system piping and components downstream. The rated point, 100 percent flow at rated pressure, is the design condition, the flow and pressure the system was sized around.
The 150 percent point is the stress test and the one pumps fail. At 150 percent of rated flow, the pump must still hold at least 65 percent of its rated pressure. This is the overload end of the curve, where a marginal or worn pump runs out of capacity and the pressure collapses. A pump that sails through churn and the rated point can still die at 150 percent, and that is exactly the point that proves the pump has reserve when a real fire pulls more water than the design day assumed.
| Test point | Flow | Pressure requirement |
|---|---|---|
| Churn (no-flow / shutoff) | 0 gpm | Not more than 140 percent of rated pressure |
| Rated (100 percent) | 100 percent of rated flow | At least rated pressure |
| Overload (150 percent) | 150 percent of rated flow | Not less than 65 percent of rated pressure |
The pump curve: field test against the certified curve
A pump curve plots discharge pressure against flow, and the whole acceptance test is really an exercise in laying your field-measured curve over the manufacturer's certified shop-test curve and seeing whether they agree. The manufacturer runs every pump on a test stand before it ships and certifies a curve for that specific pump by serial number. That certified curve is the reference. Your field points either land on it or they do not.
The shape carries the meaning. A healthy centrifugal fire pump curve starts high at churn, stays relatively flat through the rated point, and then falls off toward 150 percent, still holding 65 percent of rated pressure at the overload end. A curve that droops early, or that cannot reach the 150 percent pressure floor, is telling you the pump is undersized, worn, starved on the suction, or not turning its rated speed. The three required points are the minimum needed to draw and check that shape, though many witnesses take extra points to fill it in.
Field points rarely land exactly on the certified curve, and they do not have to. They have to be close, and they have to meet the code floors at each point. A common rule of thumb in the field treats results that fall below about 95 percent of the certified flow and pressure as a flag that something is degraded and worth investigating before acceptance. Confirm the acceptance band against NFPA 20 and the project specification, because the test exists to catch the pump that ships a little tired, not to demand a perfect overlay.
How is flow measured during the test?
Flow is measured one of two ways: a test header that dumps hose streams to atmosphere while you read each stream with a pitot, or a closed flow meter loop that pipes the water back to suction or the supply and reads flow directly. Both methods are accepted, and both have to be calibrated for the reading to mean anything.
The test header is the traditional method. It is a manifold of hose valves, each fitted with a smooth-bore nozzle or playpipe. You open valves to reach each target flow, then read the velocity pressure of each stream with a pitot gauge held in the center of the stream just outside the nozzle. The pitot pressure converts to flow through the standard discharge equation, roughly Q equals 29.84 times the nozzle coefficient times the diameter squared times the square root of the pitot pressure. Add the streams to get total flow at that point, then read the discharge and suction gauges at the same moment.
The flow meter loop is cleaner because it does not throw water everywhere, but a meter drifts, and a meter that reads high lets a weak pump pass. The trustworthy practice is to confirm the meter against a pitot at least once: flow a known stream, pitot it, and check that the meter agrees. If the meter and the pitot disagree, the pitot wins, because it measures the physics directly. A meter nobody has verified against a pitot is a number, not a measurement.
Net pressure, and why you read both gauges
Net pressure is discharge pressure minus suction pressure, and it is the number that actually belongs on the curve, because it is the pressure the pump itself added. Read only the discharge gauge and you fold the supply pressure into the result, which flatters the pump on a strong supply and punishes it on a weak one. The pump did not make the suction pressure. The supply did.
So at every test point you read three things at the same moment: discharge pressure, suction pressure, and flow. Subtract suction from discharge to get net, and plot net against flow. This is what makes the field curve comparable to the certified curve, which was generated on a test stand where the manufacturer also accounted for suction conditions.
Watch the suction gauge for its own sake too, not just the subtraction. If suction pressure sags hard as flow climbs toward 150 percent, the supply or the suction piping is the limit, not the pump. A pump can be perfectly healthy and still fail the test because the water cannot get to it fast enough at high flow.
Electric fire pump specifics
An electric fire pump is driven by a motor through the controller, and the test adds an electrical layer to the hydraulic one. At each flow point you record voltage and current on all phases alongside the flow and pressure. Current climbs with flow, peaking near the 150 percent point, and that reading confirms the motor is not overloaded and the conductors and overcurrent protection were sized for the real load the pump pulls at the overload end.
The power-source test is the part people forget and the part that matters most in a data center. An electric pump is only as reliable as its power, so where a transfer switch and an alternate source feed the controller, the acceptance test proves the transfer. You drop the normal source, confirm the automatic transfer switch moves the pump to the alternate or generator power, and confirm the pump starts and runs on that source. A fire pump that cannot start when the utility goes down, which is exactly when fires and outages travel together, is not a fire pump at all.
Watch the controller's starting method too. Across-the-line, reduced-voltage, and soft-start controllers each behave differently on the inrush, and the locked-rotor and starting current are real loads the upstream gear has to take. The controller, the transfer switch, and the motor are tested as one chain, because the failure usually lives in the handoff between them, not in any single box.
Diesel fire pump specifics
A diesel fire pump carries its own power source, the engine, so the test shifts from electrical readings to engine readings: speed, oil pressure, coolant temperature, and the instrument panel through the run. The pump still has to make all three hydraulic points, but the engine now has to prove it will start and keep running when there is no utility to lean on, which is the whole reason a diesel gets specified on a critical building.
Starting is the headline. On the automatic start, the engine should reach rated speed quickly, commonly cited at within about 20 seconds of the start signal. The acceptance test runs a series of starts, both automatic and manual, to prove the engine starts reliably and that the manual emergency start works when the automatic path fails. Confirm the exact start count and sequence against the NFPA 20 acceptance procedure and the controller listing, because the number of required starts is edition-dependent.
Two systems on a diesel deserve their own attention. The engine has two independent battery banks so a dead or failed battery does not leave the pump unable to crank, and the test confirms the engine starts on each bank. The fuel and cooling have to hold up for the full run, commonly a minimum 30-minute run during acceptance, and the day tank is sized for it, often around 1 gallon per rated horsepower plus an allowance for sump and expansion. Confirm fuel volume, exhaust back-pressure, and the cooling loop against the engine listing and NFPA 20, since these are the items that look fine for ten minutes and fail at thirty.
The controller test: automatic start, manual start, and the run timer
The controller is tested as hard as the pump, because a pump that makes its curve but will not start on its own is useless. The headline function is the automatic start on a pressure drop. You bleed system pressure down through the sensing line until it crosses the start setpoint and confirm the controller calls the pump, on its own, without anyone touching a button.
The manual start gets proven separately. You confirm the pump starts from the controller's manual control, and on a diesel you confirm the mechanical emergency start handle works independent of the electrics, because the emergency start is the fallback when everything automatic has failed. The controller also has to report its alarms, and the test drives them: loss of power or phase, pump running, controller trouble, and on a diesel the engine-specific alarms. Each should annunciate locally and reach the monitoring station.
The minimum run timer is the quiet, important one. Most fire pump controllers run for a minimum period once started, so the pump does not short-cycle and so it actually gets to work on a real event. The acceptance test confirms the pump, once started, runs out its minimum timer and does not stop the instant the start condition clears. A controller that stops the pump too soon defeats the point of automatic operation.
The jockey pump and pressure maintenance
The jockey pump exists to keep the main fire pump from starting on nothing. A sprinkler system leaks a little, all systems do, and pressure also drifts with temperature. Without a jockey pump, every small pressure drop would call the big pump, which would start, run its minimum timer, and stop, over and over. The jockey is a small pump that makes up that small leakage and holds system pressure in a tight band so the main pump only starts on a real demand.
The settings are what make it work, and they stack in a specific order. The jockey stop pressure is set at or above the main pump churn pressure plus the minimum static supply pressure, so the jockey alone can satisfy the system. The jockey start is set below its stop, commonly by at least 10 psi. The fire pump start is set below the jockey start, commonly by about 5 psi, so the jockey gets first crack at a small drop and the fire pump only steps in when the drop is real and the jockey cannot keep up. With multiple pumps the steps stagger downward.
Sizing matters as much as setting. The jockey is meant to make up allowable leakage in about 10 minutes, or 1 gpm at minimum, and no larger. Oversize it and it masks a real demand by keeping pressure up when the fire pump should be starting, which is a dangerous kind of false calm. The classic field fault is short-cycling, the jockey snapping on and off every few minutes, which means the start and stop are set too close together or the system has a real leak the jockey is chasing. Set the band wide enough to stop the cycling, then find the leak.
Suction and the water supply
The suction side decides whether the pump can do its job, and it is where good pumps get blamed for bad supplies. The pump can only move the water that reaches it, so the test watches suction pressure across the whole flow range, especially as flow climbs to 150 percent and the demand on the supply is highest. Suction pressure that drops toward zero or goes negative at high flow is the warning.
Negative suction means cavitation, and cavitation is the supply telling you it cannot feed the pump. When suction pressure falls below the vapor pressure of the water, vapor bubbles form at the impeller and collapse violently, which sounds like gravel through the pump, eats the impeller, and crashes the discharge. A pump cavitating at 150 percent will never make its overload point, and the cause is on the suction side, not in the pump. NFPA 20 generally does not allow the suction pressure to fall below a defined floor during the test, so confirm the suction requirement against the adopted edition.
The supply itself gets proven, not assumed. A supply flow test, often a hydrant or supply-side flow test taken around the same time, establishes what the source can actually deliver, because the design assumed a supply curve and the field has to confirm it. Where a backflow preventer sits on the supply, it adds pressure loss the pump has to overcome, and that loss belongs in the analysis. A backflow device installed after the pump was sized can quietly eat the margin the design counted on.
The relief valve and over-pressure protection
A main relief valve protects the system from pressures higher than the piping and components are rated for, and it shows up mainly on diesel and variable-speed installations because those are the cases where the pump can over-pressure the system under an abnormal condition. The classic case is a diesel engine running fast: a diesel can overspeed, and at higher speed the pump makes more pressure than the system was built to take. The relief valve dumps the excess to protect the pipe, the fittings, and the heads.
Whether one is even required comes from the numbers. A common screening method takes the churn pressure, multiplies it by about 1.21 to account for engine overspeed, adds the maximum expected suction pressure, and compares the total to the pressure rating of the discharge components. If the total exceeds the rating, a main relief valve is required. Confirm the method and the trigger against NFPA 20 and the project documents, because this is exactly the kind of clause that shifts between editions.
Do not confuse the main relief valve with the circulation relief valve. The circulation relief is a small valve that bleeds a little water to keep the pump cool when it runs against a closed discharge, and it operates in normal service. The main relief is a large safety device that should not be passing water in normal operation. A main relief valve dumping during the test, when nothing is abnormal, is a finding, not a feature. It usually means the valve is set wrong or the pump is over-pressuring the system, and either way the curve you measured downstream of an open relief is not the real pump curve.
Who has to witness a fire pump acceptance test?
A fire pump acceptance test is a witnessed test, and the parties that need to be present are the authority having jurisdiction, the pump manufacturer's representative, the installing contractor, and on a commissioned project the commissioning agent and the engineer of record. NFPA 20 requires the acceptance test for every new, rebuilt, or relocated pump, and it requires coordinating the date, time, and location with the AHJ so the AHJ can witness or review it.
The manufacturer's representative is there for a reason beyond watching. Good practice, and the standard, expects the manufacturer's rep to have inspected and pre-tested the installation, made any governor or controller adjustments, and provided the certified shop-test curve for that specific pump before the witnessed test proceeds. You cannot judge a field curve against a certified curve that is not in the room. If the certified curve never showed up, the test is not ready to run.
The installer runs the test under the witnesses, opening valves and recording readings, while the witnesses confirm the procedure and the results. On a data center the commissioning agent ties this into the larger integrated test, since the pump feeds the same fire scheme the detection, alarm, and suppression all serve. The witness signatures on the report are what make it a record the AHJ accepts and the owner can defend later. An unwitnessed test is a contractor's word, not an acceptance.
How often is a fire pump tested after it is installed?
After acceptance, the owner inherits a recurring test cadence under NFPA 25, the standard for the inspection, testing, and maintenance of water-based fire protection systems, and it runs on two main clocks: a frequent no-flow churn run and an annual full flow test. The acceptance test is the one-time proof at turnover. NFPA 25 is the ongoing proof for the life of the building.
The no-flow churn run starts the pump and runs it against a closed discharge for a set period to confirm it starts and does not overheat. The frequency splits by driver. A diesel pump is commonly run weekly for at least 30 minutes. An electric pump is commonly run monthly for at least 10 minutes, though specific configurations such as vertical turbine pumps, high-rise systems, and certain controllers are run weekly. These intervals have shifted between editions, so confirm them against the adopted edition of NFPA 25.
The annual flow test is the one that matters most over the years, because it repeats the acceptance test. The pump is run across its curve at no-flow, rated, and 150 percent of rated flow, and the points are compared back to the acceptance-test curve to see whether the pump still makes its numbers. This is a trending exercise, and the acceptance test is the baseline it trends against. A pump quietly losing capacity shows up as an annual curve that drifts below the original, and only the acceptance baseline lets anyone see the drift. Skip the acceptance flow test and every future annual has nothing to compare to.
The data center angle: the pump in a mission-critical building
In a data center the fire pump feeds the sprinkler and standpipe systems that protect the structure and the people, and it sits inside a building where the whole design philosophy is that nothing single fails the load. That changes how the pump gets reviewed. The pump is part of the life-safety chain, not an isolated mechanical box, and it gets commissioned with the same witnessed, scripted rigor as the power and cooling plant.
Redundancy and power are the data center concerns. A large or tall campus may run multiple pumps, and the electric pump's power source is the soft spot, since the same event that threatens the building can threaten the utility. Proving the automatic transfer to standby or generator power during acceptance is not a formality here. It is the difference between a pump that works during an outage and one that is dead exactly when the building needs it. This is the same reason the electrical plant runs its own integrated systems test on the power side.
The fire pump test also belongs inside the building's integrated fire test, because the pump is one link in a scheme that includes detection, the alarm cause-and-effect matrix, and the suppression it feeds. The broader fire and life-safety picture, including how the pump fits the suppression strategy, lives in the data center fire and life-safety overview. The point here is narrow: the pump has to make its curve and prove its controller on its own first, before it can be trusted as a link in the larger chain.
Why can't my fire pump make its 150 percent point?
A pump that fails the 150 percent point is almost always running out of water or running out of speed, not running out of impeller. The 150 percent point is the overload end of the curve, where the demand on the supply is highest and any weakness on the suction side shows up first. Before you condemn the pump, watch the suction gauge at full flow.
Starved suction is the usual culprit. If suction pressure sags toward zero or goes negative as you push to 150 percent, the supply or the suction piping cannot feed the pump, the pump cavitates, and the discharge collapses. That is a supply problem wearing a pump problem's clothes. The other common cause is speed: a diesel not turning rated rpm, or a worn or mis-trimmed impeller, will undershoot the pressure floor even with a healthy supply. Check the tachometer and the certified speed before you blame the wet end.
The rest of the common findings each have a tell. A controller that will not auto-start is usually a sensing-line or setpoint problem, not a pump problem, and you find it by bleeding pressure and watching whether the start ever fires. A main relief valve dumping during the test masks the true curve and points to a wrong setting or genuine over-pressure. A jockey pump short-cycling means the pressure band is set too tight or there is a real system leak. The discipline is the same every time: read the symptom, then chase the cause to the right component instead of swapping the pump.
| Finding | Likely cause |
|---|---|
| Pump cannot make the 150 percent point | Starved suction or pump not at rated speed |
| Gravel sound, crashing discharge at high flow | Cavitation from negative suction pressure |
| Controller will not start automatically | Sensing-line or pressure-setpoint problem |
| Main relief valve dumping during the test | Wrong relief setting or genuine over-pressure |
| Jockey pump short-cycling | Pressure band too tight, or a real system leak |
What to document
The acceptance test record is the baseline the building trends against for its whole life, so it has to capture each test point with enough detail that a future annual test can be laid over it. At each point record the flow, the suction and discharge pressures, the net pressure, and the pump speed, plus the electrical readings on an electric set or the engine data on a diesel, and the pass or fail against the certified curve.
Beyond the points themselves, capture the pump nameplate data and serial number, the certified curve reference, the controller settings, the jockey settings, the suction supply condition, the start tests performed, and the witness signatures with the date. If the relief valve operated, record it and why. The record is only as good as the worst number someone left off it, because the question years later is always whether the pump has slipped from where it started, and you can only answer that against a complete baseline.
| Test point | Flow (gpm) | Suction (psi) | Discharge (psi) | Net (psi) | Speed (rpm) | Electric: volts/amps or diesel data | Pass/fail vs curve |
|---|---|---|---|---|---|---|---|
| Churn (no-flow) | 0 | record | record | record | record | record | record |
| Rated (100 percent) | record | record | record | record | record | record | record |
| Overload (150 percent) | record | record | record | record | record | record | record |
Common mistakes
- Running the test with no certified shop-test curve in the room, so there is nothing to compare the field points against.
- Reading only the discharge gauge and skipping suction, so the recorded curve includes the supply pressure the pump did not make.
- Accepting a flow-meter reading nobody verified against a pitot, letting a weak pump pass on a meter that reads high.
- Testing the pump but never proving the controller auto-starts on a pressure drop, or never proving the transfer to alternate power.
- Letting starved suction get blamed on the pump when negative suction pressure at 150 percent is the supply cavitating the pump.
- Setting the jockey pump start and stop too close together, so it short-cycles, or oversizing it so it masks a real demand.
- Running a diesel for ten minutes and calling it done, missing the fuel and cooling problems that only appear over the full 30-minute run.
- Running the test without the AHJ and the manufacturer's representative, so the result is a contractor's word, not an acceptance.
Field checklist
Want this checklist to run itself on every job — with photo proof and a signed record crews can hand the customer? That's FieldOS.
Standards and references
NFPA 20, the Standard for the Installation of Stationary Pumps for Fire Protection, governs the pump installation and the field acceptance test, including the three test points, the churn limit, the start tests, the relief valve, and the requirement to coordinate the witnessed test with the AHJ. It is the controlling document for everything that happens at turnover.
NFPA 25, the Standard for the Inspection, Testing, and Maintenance of Water-Based Fire Protection Systems, takes over after turnover and sets the recurring cadence: the frequent no-flow churn run and the annual full flow test against the acceptance curve. The electrical side leans on NFPA 70, the National Electrical Code, for the pump feeders and controller, and on NFPA 110 for the emergency and standby power that backs an electric pump. Where a diesel drives the pump, the engine listing adds its own requirements on top of NFPA 20.
Every number here is edition-dependent. NFPA revises on a multi-year cycle, the jurisdiction adopts a specific edition that is often a few cycles behind, and amends it locally, while the property insurer can impose criteria stricter than code by contract. Confirm the test points, the churn ceiling, the start counts, the suction floor, and the relief-valve trigger against the adopted edition and the project documents before you cite a clause, and let the AHJ and the manufacturer govern where documents conflict.
Units, terms, and acronyms
Fire pump work carries a vocabulary that travels across the certified curve, the controller submittal, and the acceptance report, and the same idea can read differently on each. The terms below are the ones that show up on every test.
- Churn (shutoff / no-flow)
- The pump running against a closed discharge with no flow, reading its highest pressure; capped at 140 percent of rated pressure
- Rated point (100 percent)
- The design condition, the rated flow at rated pressure the system was sized around
- 150 percent point
- The overload test point, where the pump must still hold at least 65 percent of rated pressure at 150 percent of rated flow
- Net pressure
- Discharge pressure minus suction pressure, the pressure the pump itself added and the value plotted on the curve
- Pump curve
- Pressure plotted against flow; the field curve is compared to the manufacturer's certified shop-test curve
- Jockey / pressure-maintenance pump
- A small pump that holds system pressure between events so the main fire pump starts only on a real demand
- Cavitation
- Vapor bubbles forming and collapsing at the impeller when suction pressure falls too low, damaging the pump and crashing the discharge
- Main vs circulation relief valve
- The main relief is a large safety valve for over-pressure; the circulation relief bleeds a small flow to cool the pump at no-flow
- NFPA 20 / NFPA 25
- NFPA 20 governs the installation and the acceptance test; NFPA 25 governs the recurring inspection, testing, and maintenance
- AHJ
- Authority having jurisdiction, the official who adopts the editions, witnesses or reviews the test, and settles conflicts
FAQ
What are the three fire pump test points?
The three points are churn with no flow, 100 percent of rated flow at rated pressure, and 150 percent of rated flow at not less than 65 percent of rated pressure. Churn is also capped, at not more than 140 percent of rated pressure. Together the three points draw and check the pump curve.
What is the 150 percent point on a fire pump test?
The 150 percent point is the overload test, where the pump must still deliver at least 65 percent of its rated pressure while flowing 150 percent of its rated flow. It is the stress point that proves the pump has reserve when a real fire pulls more water than the design assumed. Marginal pumps fail here.
What is churn pressure on a fire pump?
Churn pressure, also called shutoff or no-flow pressure, is the pressure the pump makes running against a closed discharge with nothing flowing out. It is the highest pressure on the curve. NFPA 20 caps it at not more than 140 percent of the rated pressure, because too high a churn can over-pressure the system downstream.
How often is a fire pump tested after it is installed?
Under NFPA 25, a diesel pump is commonly run weekly for at least 30 minutes at no-flow churn, and an electric pump monthly for at least 10 minutes, with some configurations weekly. Both get a full annual flow test across the curve. Confirm the intervals against the adopted edition, since they have shifted.
Who has to be present for a fire pump acceptance test?
NFPA 20 requires the witnessed acceptance test to be coordinated with the authority having jurisdiction, who witnesses or reviews it. The pump manufacturer's representative and the installing contractor are present, and on a commissioned project the commissioning agent and engineer attend. The manufacturer's certified pump curve must be available before the test runs.
What is the difference between an electric and a diesel fire pump test?
An electric pump test records volts and amps at each point and proves the automatic transfer to standby or generator power. A diesel test records engine speed and panel data, proves automatic and manual starts on two battery banks, and runs the engine its full time, commonly 30 minutes, to confirm fuel and cooling hold up.
Why can't my fire pump make its 150 percent flow point?
Usually the pump is starved on suction or not turning rated speed, not worn out. Watch the suction gauge at 150 percent flow: if it sags toward zero or goes negative, the supply cannot feed the pump and it cavitates. If suction is fine, check the speed, since a worn impeller undershoots the pressure floor.
What is a jockey pump and how is it set?
A jockey pump is a small pressure-maintenance pump that makes up minor leakage so the main fire pump starts only on a real demand. Its stop is set at or above churn plus minimum static supply, its start below that, and the fire pump start below the jockey start. Set the band wide enough to stop short-cycling.
What is net pressure on a fire pump test?
Net pressure is discharge pressure minus suction pressure, the pressure the pump itself added. It is the value plotted on the curve, because reading discharge alone folds in the supply pressure the pump did not make. At every test point you read discharge, suction, and flow together, then subtract to get net before comparing to the certified curve.
People also ask
Codes cited in this guide
This guide is written and reviewed against the published standards below. Always confirm the current adopted edition with the authority having jurisdiction.