A prefilter lands at twelve dollars a unit. A plumber's callout runs ninety. When those two numbers meet, it's almost always because a filter did something stupid. It wouldn't open. It wouldn't seal. It split at the thread.
If you import water treatment gear, you already know the unit price isn't what hurts. The second container is what hurts, when the installer who talked you up on the first one stops taking your calls and the return rate parks at 4% with no explanation attached.
We make these housings. Over the past few years we've sat in on enough post-mortems with distributors, and read enough of the customer emails they forward, to notice something: the complaints aren't random. Nine of them cover almost everything, and they arrive in roughly the same order of frequency regardless of whose brand is on the box.
So this is those nine, then the part most suppliers skip. Where each one is actually created on a production line, and what we changed in the Legom housing to get out of the pattern. The ten questions near the end are the ones we'd want a buyer to ask us.
What a whole house sediment prefilter is actually for
A mechanical filter on the main line, usually downstream of the meter or the pressure-reducing valve. Sand, rust flakes, chunks of old galvanised pipe, whatever the utility pushed through after a main repair. All of it stops at the screen instead of turning up in the water heater, the dishwasher, the aerators.
That's the whole job. A whole house sediment prefilter doesn't soften, doesn't strip chlorine, doesn't touch dissolved metals. If a supplier starts edging toward the word "purification", ask which NSF standard covers that claim. There isn't one that applies to this product.
Which is also why the category barely moves. There's no performance race to win. Every unit on the shelf does the same thing, so the fight goes to price, and price goes to the two parts nobody photographs: the bowl and the brass head.
Why the shape hasn't changed in fifteen years
Three requirements fight each other. Every design you can buy is a different bad compromise between them.
- The bowl has to come off. Sediment doesn't rinse out of a mesh. It has to be reached, wiped, sometimes scrubbed. Any housing that keeps the bowl permanently attached is paying for convenience with cleaning quality.
- Both faces of the screen have to be cleaned. Dirt lands on the upstream face, fine. But biofilm and silt bind inside the weave and on the downstream side, and one-directional backwash never touches that half.
- The brass has to be lead-free and still machine well. US plumbing rules cap wetted brass at a 0.25% weighted average lead content under NSF/ANSI/CAN 372. Lead-free brass is harder, gummier, and costs meaningfully more at the billet. That's usually the point where "lead-free" quietly becomes "lead-free plated".
Solve one and the other two get worse. A bowl that opens by hand needs looser threads. Looser threads need a fatter O-ring. A fatter O-ring takes a compression set and starts weeping by the third service. You can follow that chain straight to your return rate.
The nine complaints, roughly in the order we hear them
1. "The flush doesn't do anything"
Open the little valve, watch brown water come out, call it maintenance. What's actually happening is the upstream face getting rinsed while the mesh stays half blinded. Flow drops slowly enough that nobody notices, and eight months later someone complains about shower pressure. By then the screen needs to come out, which is a different and much more annoying job.
2. Cleaning it takes twenty minutes and a wrench
Close the main. Depressurise. Find the strap wrench, because the collar hasn't moved since installation day. Crack it, pry the bowl, scrub, reseat the ring, retighten, reopen. Plumbers bill for that. Homeowners skip it entirely, then call you about pressure.
3. Water hammer turns the housing into the weakest part of the line
A solenoid shuts, a washing machine fills, and the surge goes looking for the softest component in the system. That tends to be the bowl thread. Codes deal with this through arrestors made to ASSE 1010, and inspectors start asking about a PRV once static pressure passes roughly 80 psi. Plenty of installs have neither.
It looks like a manufacturing defect when it fails. Hairline crack at the thread root, a seat gone oval, a fitting that only weeps during a surge. You'll warranty it as a defect. It was an installation the product never accounted for.
4. Lead, and the file you can't produce
Brass is the default head material because it machines cleanly and takes chrome. Leaded brass is also the cheapest route to a price point. The WHO keeps a provisional guideline of 0.01 mg/l for lead in drinking water for a boring, uncomfortable reason: most lead in tap water comes from plumbing, not from the source.
When a buyer asks for your lead documentation and all you have is a logo on a datasheet, that conversation ends. NSF listings are searchable in the public database. Check the number before you print the badge.
5. It only fits one way
Most prefilters take water from a single direction. On a retrofit where the main drops out of the wall instead of rising from the floor, the installer either stacks elbows - every one a future leak and a pressure loss - or hangs the unit sideways with the bowl facing somewhere useless. He remembers the brand that cost him an extra hour.
6. The O-ring gives up on the third opening
Nobody engineers the seal. It gets sized from whatever groove is left after the thread is drawn. Then over-compression is used to cover loose tolerances, the rubber takes a set, and the unit weeps after a few services. A bowl that seals perfectly on the test bench and drips under a kitchen sink is a tolerance-stack problem. Tolerance-stack problems are designed in, not shipped by accident.
7. Cracks at the pressure ring
The two thinnest sections of any bowl are the thread root and the shoulder against the head. That's also where a badly placed injection gate leaves a weld line, which becomes the weakest line in the part. Add municipal chlorine and polycarbonate starts stress-cracking. A bowl that's gone hazy has already lost most of its ductility.
8. Numbers on the box that nobody measures
"40 micron" is cheap to print. So is "3 tons per hour". Two filters can both claim 40 micron while one has a third less open area, and almost nobody publishes open area, screen diameter, or a pressure-drop curve. If the only number you have is the headline flow rate, you don't have data. You have a design brief for the carton.
9. Cross-threaded on the second service
Fine-pitch plastic threads are unforgiving. One person cocks the bowl, another overtightens it to stop a drip, and the thread is done. A thread designed for a factory torque wrench won't survive ten years of annoyed homeowners.
Where all nine come from on the line
This is the part trading companies can't discuss, because they aren't in the building. Every failure above traces back to five decisions, and four of them are locked in before the first unit is moulded.
- The tool. Whether the gate sits on the pressure ring or off to one side gets decided at the tooling quote, long before anybody sees a sample. Moving it later means slides, a longer cycle and a new tool. It is also the first thing that disappears when a project is priced to hit a target FOB.
- The material, and the regrind. "PC" on a datasheet means nothing until you know the grade, the UV package, and how much regrind the moulder is allowed to feed back in. High regrind ratios run fine on the bench and fail in cold weather. The regrind ratio is never on the datasheet. Ask for it anyway.
- The tolerance stack. Bowl thread, head thread, groove depth, rubber hardness. Hold each one loosely to save cost and the seal absorbs the total. That's complaint six, every time, in every brand that does it.
- Brass process control. Lead-free brass eats tool inserts faster. A shop that changes inserts on a fixed count ships clean threads on Monday and torn ones on Friday, and both pass inspection. Consistent lead control is a process discipline, not a purchase order.
- The test protocol. Batch sampling tells you the batch was probably fine. A 100% pressure hold on every unit tells you about that unit. A rig that opens and closes the assembly a few thousand times tells you about the design. Ours was built in-house rather than bought, which is not elegant, and it found two problems no certificate would have caught.
What we changed, and why none of it is cosmetic
The Legom housing carries patents across the mechanism, the valve and the water path. Patents on their own sell nothing. What matters is which complaint each one deletes.
Mesh blinds and flow decays
Two screens alternate in service; the idle one gets cleaned properly. Flow stays stable instead of drifting down over a season.
Service needs a wrench and the main valve
Gear linkage releases the bowl by hand, plus a local shutoff that isolates the housing. Service drops to minutes and the homeowner can do it.
"Tool-free" that isn't
Assembly specified for hand operation. Removes the "call a plumber to clean a filter" objection at retail.
Surge cracks bowls and takes appliances with it
Internal structure built to dissipate water hammer before it reaches the downstream line. Fewer cracked bowls, fewer warranty claims that were never defects.
Lead risk in the water path
Lead-free brass on wetted surfaces. A file you can hand over when a buyer asks.
One inlet direction limits the layout
Universal inlet that switches direction at installation. Fewer elbows, faster retrofit, less installer pushback.
Bowl deforms or cracks under cycling
Alloy specified for the pressure-bearing parts. Margin against surge and pressure cycling.
Nothing on the shelf to look at
Two-tone space-element design language. Differentiation that survives sitting next to identical housings.
Three of those deserve a proper explanation, because they're the ones a competitor can't lift from a product photo.
Two screens instead of one
Run one screen in service and keep the other isolated, then rotate. The one being cleaned isn't fighting line pressure, so the flush actually removes what's in the weave. It changes cleaning from a hopeful rinse into a defined operation, and it kills the failure mode where a "clean" filter still chokes flow.
A linkage instead of a thread
Torque is what makes these housings hard to live with. The linkage converts a modest hand turn into an even axial release, so the user isn't fighting thread friction on the way out or over-tightening on the way back. One mechanism deals with complaints two, six and nine.
A local shutoff instead of the main valve
Asking a homeowner to find and close the main valve is the single best reason prefilters never get cleaned. Isolating the housing locally means the bowl comes off while the rest of the house stays live. Small thing. Watch someone do it once and you won't go back.
Ten questions for any supplier, ours included
Print these. Ask the next three quotes the same way, and the answers will sort themselves.
- Who moulds the bowl, and where is the gate? Ask for the tooling layout. A gate on the pressure ring is a decision you're about to inherit.
- What grade, and has it been chlorine-exposed? "PC" is not a specification.
- Is the pressure test 100% or sampled? Get pressure, hold time and pass criteria in writing.
- How many cycles does the release mechanism survive? With the report, not the claim.
- How do you test water hammer? No surge rig means the protection is a drawing.
- Is the brass lead-free by alloy or by plating? Plating wears through. Ask for wetted-surface lead content.
- Which NSF listings exist today, and what are the numbers? Verify each one in the public database.
- What compound and durometer is the O-ring, and what compression set does it hold? This predicts leak returns better than anything else on the sheet.
- What's the measured micron rating and open area, on what bench? Ask for the curve, not the headline.
- Can I see the service procedure written for a homeowner? If it needs a wrench, a valve or a professional, the product won't get maintained, and an unmaintained prefilter becomes a complaint.
The math almost nobody runs before the first order
Say the landed gap between a commodity housing and a serviceable one is six dollars. Two thousand units. A callback costs ninety where you sell. The twelve thousand you saved is gone after 134 callbacks, which is under 7% of the shipment. Most commodity prefilters clear that bar on the first container, before word of mouth starts working the other way.
Put your own numbers in. The conclusion doesn't move much, and it stops being a debate about unit price.
FAQ
What micron rating should a whole house sediment prefilter be?
Most residential mains protection sits between 40 and 100 microns. Finer captures more and clogs faster; 40 is the usual compromise on municipal supply, while heavy sand from a well often needs a coarser first stage.
Can you clean one without shutting off the main valve?
Only if the housing has its own isolation valve. Most don't, so the main has to come off first. A local shutoff lets the bowl open while the rest of the house keeps running.
Do I need a wrench to open a prefilter bowl?
On a conventional collar, usually yes. Hand-serviceable designs use a linkage or cam. If a supplier says tool-free and the sample needs a strap wrench, the claim is marketing.
What does NSF/ANSI 372 actually certify?
The lead content of wetted surfaces, at a 0.25% weighted average. It's a materials certification, not a performance rating. Sediment reduction claims live under NSF/ANSI 42, which is a separate listing.
Why does my prefilter lose pressure over time?
Mesh blinding, nearly always. Silt and biofilm bind inside the weave and a one-direction backwash can't reach them. It's progressive, and shower heads are usually where it gets noticed.
Can water hammer damage a prefilter?
Yes. Surges from fast-closing valves run well above static pressure, and the bowl thread is the thinnest part of the assembly. Arrestors made to ASSE 1010 are the code answer; a housing that dissipates surge internally just gives you margin on top.
Is a transparent bowl better?
Better for merchandising, worse for material stability. Polycarbonate is more sensitive to chlorine stress cracking and UV yellowing than opaque engineering grades. If you sell on visibility, specify the grade and ask for an exposure test.
What should a first order look like for private label?
One SKU at full depth, a second at trial depth, and both from the same production run as the samples you approved. Then ask for that run's pressure-test records instead of a generic certificate.
Before you place the order
Nine complaints, five decisions on the line, three mechanisms that address them. None of it is exotic. The category is old and the fixes are mostly mechanical patience: put the gate where the stress isn't, spec the material you actually need, hold the tolerances tight enough that the seal isn't doing structural work, and build a test rig that fails things.
If you're sourcing a whole house sediment prefilter for distribution, private label or an installer programme, send us the market, the certification list and the volume. You'll get the test reports, the tooling logic and a quote you can compare line by line.