Water hammer arrestor sizing field guide for plumbers

Water hammer, and what the bang is actually doing

Water hammer is the pressure surge that travels back through the piping when moving water is stopped fast. Water has mass, and a column of it moving down a pipe carries momentum. Slam a valve shut on that moving column and the water cannot stop gently. It piles into the closed valve and the energy goes into a pressure spike that races back up the line at close to the speed of sound in water, hits the next fitting or change of direction, and reflects. The bang you hear is that spike shaking the pipe against its hangers and the structure.

The noise is the symptom people notice. The damage is what they do not. Each surge can spike the line several times the working pressure for a fraction of a second, and the pipe, the joints, the solder, the valve seats, and the fixture supplies all take that hit every cycle. Over months it loosens fittings, cracks solder joints, splits supply tubes, wears out fill valves, and works pipe loose from its straps until it rubs and leaks.

On a real building the trouble is rarely the first bang. It is the thousandth. A solenoid valve on an ice maker or a flush valve in a busy restroom cycles thousands of times, and the failure shows up as a weep at a joint or a fixture supply that lets go behind a wall, long after the install passed inspection.

What causes water hammer?

Water hammer is caused by a quick-closing valve stopping moving water faster than the water can decelerate. Two things have to be present: water moving with some velocity, and a valve that shuts hard and fast. Take either one away and the surge drops off.

The quick-closing valves are the usual suspects, and they are everywhere on a modern job. A washing machine solenoid snaps shut in a fraction of a second. So does a dishwasher fill valve, an ice maker, a boiler feed solenoid, and any electrically actuated valve, because a solenoid has no slow taper to its close. Flushometers, the flush valves on commercial closets and urinals, close fast by design and cycle constantly. A single-lever faucet shut quickly does it too. The common thread is the speed of the close, not the size of the valve.

Velocity is the other half, and it is the half you can design out. The faster the water is moving when the valve shuts, the bigger the surge, so an oversped line bangs harder for the same valve. That is why water hammer and high pipe velocity travel together, and why the fix often starts with the velocity, not the arrestor. The arrestor treats the symptom. The velocity is closer to the cause.

Why velocity matters and the surge math behind it

The size of a water hammer surge is set mostly by how fast the water was moving when the valve closed. The relationship is direct: the pressure spike is proportional to the change in velocity. The Joukowsky relationship puts numbers on it, the surge pressure equal to the fluid density times the pressure-wave speed times the velocity change. Double the velocity at the moment of closure and you roughly double the surge. That is the lever, and it is why a system run too fast bangs even with arrestors in place.

Plumbing design holds water-supply velocity down for exactly this reason, alongside pipe erosion and noise. A common design ceiling is in the range of 5 to 8 ft per second for water-distribution piping, with many designers holding hot copper lower because high velocity scours the copper at the elbows over time. The same velocity discipline shows up across the trade: the recirculation loop is kept slow to protect copper, and the drainage side is sized for its own self-scouring speed, a different number for a different job.

So the first question on a banging system is not which arrestor to buy. It is whether the line is oversped. A run sized too small for its flow carries water too fast, hammers harder, and erodes from the inside, and no arrestor fixes the erosion. Size the pipe for a sane velocity first, then add the arrestor for the quick-close surge that remains.

The air chamber versus the mechanical arrestor

The old fix for water hammer was an air chamber, a capped stub of pipe run up off the supply near the fixture. The trapped air in the stub compressed when the surge hit and cushioned it. It works on the day it is installed, and that is the problem. The air is in direct contact with the water, and air dissolves into water under pressure. Within weeks to months the stub fills with water, the air cushion is gone, and the chamber is waterlogged and dead. Now it is a capped dead leg doing nothing, and the banging is back.

A waterlogged air chamber is the single most common reason an older building still hammers after someone supposedly fixed it. You can recharge one by shutting off the supply and draining the system from the lowest faucet to let air back into the stub, but it waterlogs again, because nothing about the design changed. It is a temporary repair on a permanent problem.

The mechanical water hammer arrestor solves the waterlogging by sealing the air away from the water for good. It is a listed device with a permanently charged cushion of air or inert gas held behind a moving piston or a bellows, so the cushion can compress to absorb the surge but can never dissolve into the line. That sealed, pre-charged design is what makes it maintenance-free and what the code and the standards are written around. If a system has air chambers, the upgrade is to pull them and install listed arrestors, not to keep recharging stubs that will waterlog again by next season.

How a mechanical arrestor works

A mechanical water hammer arrestor is a small sealed cylinder that tees into the supply and gives the surge somewhere to go. Inside is a charge of air or inert gas held under pressure behind a moving piston with an o-ring seal, or behind a sealed bellows. When the quick-closing valve slams shut and the pressure spike races back up the line, the piston or bellows moves against the gas charge, the gas compresses, and the surge energy is absorbed in that compression instead of slamming the pipe. When the spike passes, the charge pushes the piston back and the device is ready for the next cycle.

The sealed charge is the whole point. Because the gas never touches the water, it cannot dissolve away, so the device holds its cushion for its rated life with no draining, no recharging, and no maintenance. A piston unit relies on its o-ring to keep the seal, and a quality unit uses a pre-lubricated o-ring rated for the cycle count it will see. A bellows unit has no sliding seal to wear.

Size, not just presence, decides whether it works. The gas volume in the device has to be big enough to swallow the surge from the load it protects, which is why the standards size arrestors by fixture units rather than letting one small unit cover everything. An undersized arrestor takes the edge off and still leaves a knock. A correctly sized one kills it.

How do you size a water hammer arrestor?

You size a water hammer arrestor by the fixture-unit method in PDI WH-201, the Plumbing and Drainage Institute standard that established the rating and sizing of these devices. Total the supply fixture units on the section of pipe the arrestor protects, then pick the arrestor whose fixture-unit range covers that total. The standard sorts arrestors into six sizes, A through F, with size A the smallest and size F the largest, each covering a band of fixture units.

The fixture-unit ranges below are the WH-201 sizing bands. Confirm them against the current edition of the standard and the manufacturer's literature, because manufacturers publish their model numbers against these PDI sizes and the edition has been revised over the years. The supply fixture-unit values for the fixtures themselves come from the plumbing code's water-supply tables, so use the code's fixture-unit values, not a guess, when you build the total.

The method is the same whether you protect one fixture or a whole branch. For a single quick-closing valve, total that fixture's units and read the size, which usually lands at the small end. For a branch with several quick-closing valves, total all of them on the protected section and read the larger size that covers the sum. The trap most people fall into is sizing the arrestor for one washer when it sits at the end of a branch feeding several, and undersizing it as a result.

Field example: sizing for a flush-valve restroom branch

Take a commercial restroom branch feeding three flushometer water closets and one flush urinal, all quick-closing valves on the same supply branch. Pull the supply fixture-unit value for each from the adopted code's water-supply table and total them. For public flushometer fixtures those values run high, and a total in the mid-thirties of fixture units is realistic for this group.

With the total at roughly 35 fixture units, the PDI WH-201 band that covers 33 to 60 is size C, so a single size C arrestor at the end of that branch protects the whole group. One correctly sized C beats four undersized A units scattered along the branch, both on cost and on performance, because the surge from the branch adds up at the end of the run where the C sits.

Now change the job. Drop it to a single residential washing machine on its own supply, total a handful of fixture units, and you are in the size A band, the small unit that fits a washer box. The method did not change. Only the fixture-unit total did, and the total is what drives the size every time.

Where do water hammer arrestors go?

A water hammer arrestor goes as close as it can to the quick-closing valve that causes the surge, on the supply that feeds it. The surge starts at the valve that slams shut, so the arrestor has to sit near that valve to catch the spike before it travels into the rest of the piping. Put it across the building from the valve and the spike has already hammered every fitting between the two before it ever reaches the device.

On a single fixture, that means the arrestor sits right at the fixture supply, ahead of the stop, on the branch feeding the quick-closing valve. The washer box arrestor and the under-sink unit are exactly this: the device an arm's length from the solenoid or the stop it protects.

On a branch serving several quick-closing valves, the arrestor goes at the end of the branch, downstream of the last fixture takeoff, where the surge from the whole branch concentrates. The far end of a dead-ending branch is where the spike has nowhere left to go, so it is where the energy piles up and where the single correctly sized arrestor does the most good. A long branch with quick-closing valves spread along it can need a unit at the end and, on a really long or hard-hitting run, additional units, but start with the end of the branch.

Two placement rules carry weight. The arrestor mounts so its piston or bellows can move freely, which for most piston units means within the manufacturer's allowed orientation, and it must sit where the device and its connection stay accessible per the listing. And it goes on the pressurized supply side, never on a drain. An arrestor plumbed in the wrong spot is a part that passed the invoice and fails the job.

Single-fixture units versus branch sizing

There are two ways an arrestor gets applied, and mixing them up is how systems end up undersized. The single-fixture unit protects one quick-closing valve and is sized for that fixture's units alone, which usually puts it at the small end of the PDI range. The washer-box arrestor, the mini under-lavatory unit, and the dishwasher arrestor are all single-fixture devices.

The branch or multiple-fixture approach protects a run feeding several quick-closing valves with one larger arrestor at the end of the branch, sized for the sum of all the fixture units on that branch. This is the commercial pattern: a bank of flushometers, a row of solenoid-fed equipment, a laundry with several machines on one supply. One size C or D at the end of the branch does the work of a fistful of small units and does it better, because the surge adds at the end.

The decision is not preference. It is the layout. Scattered single fixtures on separate supplies each get their own small unit at the fixture. A group of quick-closing valves on a shared branch gets one larger unit sized to the group. Read the piping, total the fixture units on the protected section, and let that total pick the size.

Appliance arrestors: washers, dishwashers, and ice makers

Appliances with solenoid fill valves are the most common source of residential water hammer, and they get their own purpose-built arrestors. The washing machine is the loudest offender, because its solenoid snaps shut hard and the machine cycles its fill many times per load. The standard answer is a washer-box arrestor, either a pair of mini units that thread onto the hot and cold supply nipples behind the machine or a recessed outlet box with arrestors built in.

The dishwasher and the ice maker are the same physics at smaller flow. A dishwasher fill solenoid bangs a kitchen wall, and the fix is a small arrestor on the supply at the dishwasher stop or under the sink where that branch ties in. The ice maker on a refrigerator is notorious for a single sharp knock every time it fills, and a mini arrestor at the ice-maker stop kills it.

Put the appliance arrestor at the appliance, not at the manifold across the room. Each solenoid is its own surge source, so each gets its cushion close by. And size even the small ones by the method: a single appliance lands in the size A band, but a laundry with several washers on one supply is a branch, and that branch totals up past size A.

Does a water hammer arrestor have to be accessible?

The plumbing code requires a water hammer arrestor where quick-closing valves are used, and it has long required these devices to be accessible. The historical reason was the air chamber: it needed access because it waterlogged and had to be drained and re-aerated, so the code wanted to be able to get at it. The arrestor requirement and the accessibility language sit in the water-distribution provisions of the adopted code, commonly in the International Plumbing Code around Section 604.9 and the corresponding Uniform Plumbing Code provisions. Confirm the section against the edition the jurisdiction adopted, because the numbers shift between cycles.

The mechanical arrestor changed the accessibility picture, and this is where field practice and the old rule have to be read together. A listed, permanently sealed, maintenance-free mechanical arrestor has nothing to service, so many codes and manufacturer listings now permit it to be concealed without an access panel, where the old air chamber could not be. That allowance is specific to listed maintenance-free devices, and it is exactly why the listing matters.

Either way, the device has to be listed. The recognized standards are ASSE 1010, the performance requirement for water hammer arresters, and PDI WH-201, which covers the certification and sizing. An arrestor with neither listing is not a code-recognized device, regardless of what the box calls it. Specify a listed unit, and confirm whether the adopted code lets it be concealed before you bury it.

High-velocity systems and the real fix

When a whole building hammers, not just one fixture, the cause is usually velocity, and arrestors alone will not cure it. A system run too fast surges hard at every quick-closing valve and erodes its copper from the inside at the same time. Scatter arrestors across it and you treat the symptom at each valve while the underlying overspeed keeps working the pipe and the elbows.

The first move on a high-velocity system is to find out how fast the water is actually moving. High street pressure pushes more flow through a given pipe, which raises velocity, so a pressure-reducing valve set to a sane working pressure often drops the velocity and the hammer together. Undersized piping is the other cause, and that one is a bigger fix, but on new work it is avoidable by sizing the supply for velocity in the design range rather than the smallest pipe that meets demand.

Then there is the valve. A quick-closing valve that can be swapped for a slow-closing one removes the cause entirely, because the surge comes from the speed of the close. Where a slow-close valve is an option, it beats any arrestor, since it stops the surge from forming instead of absorbing it after the fact. The order of operations is set the pressure, size the pipe for velocity, slow the close where you can, and arrest the surge that is left.

Surge in tall buildings and long runs

The longer the run of moving water and the higher the pressure, the bigger the surge a quick-closing valve can make, which is why tall buildings and long horizontal mains get specific attention. A long column of water carries more momentum, so stopping it abruptly releases more energy into the line. High-rise risers and long site mains both fit this case.

Pressure zoning in a tall building is partly about this. The same pressure-reducing valves and zone breaks that keep the lower floors from seeing crushing static pressure also keep velocities and surge potential in check zone by zone. A flushometer on a lower floor of an unzoned high-rise sees both high pressure and high velocity, the worst combination for hammer, and it bangs accordingly.

On long runs and large mains the surge can be a system-level event, not just a fixture knock, and the design may call for engineered surge control sized by analysis rather than a fixture-unit arrestor at a valve. That is engineering territory, governed by the project design, not a field rule of thumb. The field point holds, though: long, fast, high-pressure runs hammer harder, so the velocity and pressure discipline matters more on them, not less.

Is water hammer the same as thermal expansion?

Water hammer and thermal expansion are two different problems that both raise pressure in the piping, and they get confused because both can make a system bang or weep a relief valve. They are not the same, and the same device does not fix both.

Water hammer is a fast, transient surge from stopping moving water with a quick-closing valve. It happens in a fraction of a second, it comes and goes with the valve, and the cure is to slow the velocity or absorb the surge with a water hammer arrestor. Thermal expansion is a slow pressure rise that builds as water heats and expands on a closed system with nowhere to go. It happens over a heating cycle, it stays until the water cools, and the cure is an expansion tank or another approved expansion control, not an arrestor. The thermal-expansion side is its own subject, covered in the water heater recirculation and sizing guide.

Telling them apart in the field is straightforward. A bang on the instant a valve closes is water hammer. A relief valve that weeps after the water heater has been running, with no valve slamming, is thermal expansion. Put an arrestor on a thermal-expansion problem and it does nothing, because there is no surge to absorb, only a slow rise the arrestor's small cushion cannot hold.

Diagnosing and locating water hammer in the field

You diagnose water hammer by tying the bang to the valve, because the surge fires the instant a quick-closing valve shuts. Run the suspect fixtures one at a time and listen. Start the washer fill and let it shut off, flush the closet, cycle the dishwasher, and note which close makes the noise and where in the building you hear it. The valve that makes the bang on its close is the source.

Locating where to put the arrestor is the next step, and it follows the surge, not the noise. The bang often rings out at a fitting or a loose section of pipe some distance from the valve, because that is where the spike shakes the pipe against the structure, but the device goes near the valve that caused it, not where it rattles. Trace the supply back from the noisy fixture to the quick-closing valve and place the arrestor on that supply.

On a system that already has arrestors and still hammers, check the obvious failures first. An air chamber that has waterlogged is doing nothing and needs replacing with a listed arrestor. An arrestor that is undersized for a branch it was never meant to cover needs upsizing to the fixture-unit total. And a system that bangs everywhere points back to velocity and pressure, not to a missing device. Add or replace the arrestor for a single-valve knock. Chase the velocity for a building-wide one.

Commissioning the fix and confirming it held

Commissioning a water hammer fix is a listen test, and it is quick. With the arrestor installed and the system at working pressure, cycle the quick-closing valve it protects several times and confirm the bang is gone, not just quieter. A knock that is softer but still there usually means the arrestor is undersized for the load on that section, so check the fixture-unit total against the size you installed.

Confirm the working pressure while you are there, because the surge scales with pressure and velocity, and a system above a sane working pressure will keep hammering through a correctly sized arrestor. Read the static pressure at a hose bibb, and if it is high, a pressure-reducing valve set into range often does as much for the hammer as the arrestor did.

Write down what you installed and what it fixed. The size, the location, the valve it protects, and the confirmed result are the record that the next person reads when the noise comes back, so they know whether they are looking at a failed device, a new load, or a velocity problem that was never addressed.

What to document

An arrestor that nobody recorded is one the next plumber has to rediscover behind the wall. The record is what answers the callback when a system that was quiet starts banging again, and it is what tells you whether the device was sized right in the first place.

Capture the quick-closing valve or branch the arrestor protects, the total fixture units that drove the size, the PDI size you installed, the location and orientation, the listing, and the confirmed result after cycling. If you reduced pressure or upsized pipe to bring velocity down, record that too, because it is part of why the system is quiet and the next person needs to know it was done.

Common mistakes

• Installing a capped air chamber instead of a listed mechanical arrestor, so it waterlogs and the hammer returns.
• Recharging a waterlogged air chamber again and again instead of replacing it with a listed device.
• Sizing an arrestor for one fixture when it sits at the end of a branch feeding several quick-closing valves.
• Placing the arrestor far from the quick-closing valve, so the surge hammers the piping before it reaches the device.
• Ignoring high velocity and high pressure, then scattering arrestors that never cure a building-wide knock.
• Burying a unit where the listing or the code requires access, or mounting a piston unit so it cannot move freely.
• Putting an arrestor on a thermal-expansion problem, which it cannot fix, instead of an expansion tank.
• Specifying a device with no ASSE 1010 or PDI WH-201 listing, so it is not a code-recognized arrestor.

Field checklist

Standards and references

Two product standards govern the device. PDI WH-201, from the Plumbing and Drainage Institute, established the certification and the fixture-unit sizing method, including the A-through-F size designations and their fixture-unit ranges. ASSE 1010 is the performance requirement for water hammer arresters, covering the permanently sealed air or gas cushion and the maintenance-free service the modern arrestor provides. A device meeting these listings is what the plumbing code recognizes; confirm the current edition of each, since both have been revised.

The installation requirement lives in the plumbing code. The International Plumbing Code and the Uniform Plumbing Code both require water hammer protection where quick-closing valves are used and call for the device to be accessible, commonly in the water-distribution provisions around IPC Section 604.9 and the corresponding UPC sections. The exact section and the accessibility allowances for listed maintenance-free devices shift between editions, so verify against the adopted edition and any local amendments before citing a number on a submittal.

The velocity discipline that keeps surge in check ties to the water-supply sizing in the same code chapters and to the design references the engineer works from. The surge physics itself follows the Joukowsky relationship between velocity change and pressure rise. Cite the standard that controls the point, and let the project design and the equipment listing govern over any rule of thumb here.

Units, terms, and conversions

Water hammer work crosses a few units and a few names for the same parts, so the same idea reads differently across a spec, a cut sheet, and a code book.

The arrestor itself goes by water hammer arrester or arrestor, shock arrestor, or hydraulic shock absorber, and PDI sizes it A through F. Load is counted in supply fixture units, sometimes written WSFU, from the water-supply tables, not the drainage fixture units used to size drains. Velocity is in feet per second, with supply piping commonly held in the 5 to 8 ft per second range. Pressure and surge are in psi, and the static working pressure is commonly held at or below 80 psi by a pressure-reducing valve. The pressure-wave speed in water that carries the surge runs on the order of several thousand feet per second.

Water hammer

The pressure surge from stopping moving water abruptly with a quick-closing valve

Water hammer arrestor

A sealed device with a permanent air or gas charge that absorbs the surge, listed to ASSE 1010 / PDI WH-201

Air chamber

The old capped pipe stub fix that waterlogs as its air dissolves into the water and then fails

Quick-closing valve

A valve that shuts fast, such as a solenoid, washer or dishwasher fill, or flushometer, that creates the surge

PDI WH-201

The Plumbing and Drainage Institute standard for arrestor certification and fixture-unit sizing, sizes A to F

ASSE 1010

The performance standard for water hammer arresters with a permanently sealed cushion

Supply fixture unit (WSFU)

A load value for a fixture's water demand, used to total the load and pick the arrestor size

Joukowsky relationship

Surge pressure equals fluid density times wave speed times velocity change, so surge scales with velocity

Related tools

Calculators and readiness checks for this work

Compare your options

FAQ

People also ask

• What slope does a drain need?
• How do you size a hot water recirculation pump?
• How do you size a water heater?
• What is a water supply fixture unit?
• How do you size a grease interceptor?