Fresh Water Aboard a Hunter 290

Every cruiser eventually does the same arithmetic. You look at the water tank, you look at how far you want to go, and you start wondering what it would cost to stop worrying about it. Here's what we found when we actually priced it out for our Hunter 290 — the packaged units, the DIY route, the power budget, and the answer we didn't expect.

There is no $1,000 watermaker

Start with the bad news, because it saves a lot of shopping. Nobody sells a packaged 12 volt marine watermaker for under a thousand dollars. Not on sale, not used, not from a discount outfit. The floor for a real marine unit is roughly three and a half grand.

Prices checked August 2026. Note the second row: $3,480 buys you a unit you operate by hand, at about a gallon an hour, until your arm falls off. That's the price of the technology, not the price of the convenience.

So if $1,000 is the number, you're building it. That's fine — this is an open source boating site, building it is the whole point. But you should know going in what you're signing up for.

How a small watermaker actually works

Reverse osmosis is conceptually simple and mechanically demanding. Seawater sits at roughly 35,000 parts per million of dissolved salt, and its osmotic pressure is around 350 psi. To push water backwards through a membrane against that gradient, you need to beat it convincingly — a small marine system runs somewhere around 800 to 900 psi.

That pressure is the expensive part. Everything else in the system is plumbing.

Note the proportions: most of what you pump goes back over the side as brine. On a small system like this you're recovering only around 15 to 20 percent of the feed water as drinking water. That inefficiency isn't a flaw — it's what keeps the membrane from silting up.

The under-$1,000 DIY build

The standard small-boat design in the cruising community is a Pumptec 107SS plunger pump driving a single 2.5 × 14 inch seawater membrane. It's the smallest sensible seawater RO setup, it runs on 12 volts, and the parts are all available off the shelf. One well-documented build of exactly this puts out about 3 gallons per hour at 900 psi while drawing roughly 14 amps.

The used pump is the whole game. Everything else on this list is close to a fixed cost, and the Pumptec is the single biggest line item. They turn up regularly on eBay and through misting and pressure-washer surplus dealers — watch for a few weeks rather than buying in a hurry, and $1,000 is a real number. Buy it new and you're at about $1,765.

Build it portable, not installed

This part is specific to boats our size, and it's the single best decision in the whole project. Don't plumb it in.

A Hunter 290 has no spare through-hull and no machinery space worth the name. Cutting a new hole in the hull for a watermaker intake, on a 28-foot boat, to feed a system you'll run a couple of hours a week, is a poor trade. So: mount the whole thing in a milk crate or a plastic toolbox — pump, vessel, prefilters, gauge, valve. Intake is a weighted hose over the side. Output is a hose into the deck fill.

It lives in the cockpit locker. It comes out when you want water. No hull penetration, no permanent real estate lost, and — the part people forget — it goes with you to the next boat, or gets sold for most of what you paid. This is exactly the architecture the commercial portable units use. They just charge five thousand dollars for it.

The power math is the real constraint

Here's where the project actually gets decided, and it isn't money.

The high pressure pump pulls 15 amps. The feed pump is a demand pump, so it only runs a fraction of the time and averages well under one — call it 16 amps overall. Two hours of runtime gets you roughly 6 gallons and costs you about 32 amp-hours.

Now look at the other side of the ledger. A stock 290 house bank is likely a couple of Group 27s — call it 100 amp-hours you can actually use without abusing them. So one water session is a third of your usable bank, spent on six gallons.

Which is a nice way of saying the watermaker isn't really a water project. It's a solar project with a plumbing attachment.

And then we asked the obvious question

The Hunter 290 carries 40 gallons. Two people living carefully at sea use about 3 to 4 gallons a day. That's ten to thirteen days of range sitting in the tank already.

We're coastal cruising. We are never more than a few days from a hose.

So the honest conclusion, after all that pricing: we don't need a watermaker. We need maybe four to six gallons a day of top-up, and there are much cheaper ways to get it. Spending $1,000 and 32 amp-hours per session to solve a problem the boat doesn't have is exactly the kind of thing this site exists to talk people out of.

What we're doing instead — the $400 water plan

The foot pump is the headline. It plumbs into the existing galley line in an afternoon, keeps the pressure system intact for when you want it, and two people on a foot pump land around 2 gallons a day. That alone takes 40 gallons from about ten days to about twenty.

Add the jugs and you're at three weeks of independence. Add a rain catcher and you're at "we stopped counting." For coastal work, that's more range than we'll ever use before the next fuel dock — for a third of the price of the DIY watermaker and none of the amp-hours.

Rain catchment, in parts

This is the best gallon-per-dollar on the boat and it isn't close. No amps, no membrane to poison, no pump to service — just a surface you already own and a hose.

How much you actually get

Practical Sailor measured a real installation at 0.44 gallons per square foot per inch of rain — about 70% of what lands on the surface, with the rest lost to wetting, wind and splash. That number makes the whole thing easy to plan.

Our bimini is somewhere around 30 square feet on its own. Add the solar panels and the new arch and we're closer to 50. One inch of rain on that is 22 gallons — better than half the tank, in an afternoon, for nothing. The same squall would take a 3 gal/hr watermaker seven hours and about 120 amp-hours.

The parts

Two tiers. The first is what we'd do this weekend with a trip to the hardware store; the second is what it looks like once you're catching enough that it's worth doing properly.

The $60 line in the kit table above is the middle of this range — hose, funnel and a corner drain, without the filter and the treatment. Add those when the tank is being filled by rain often enough to care what's in it.

Five rules that decide whether it works

1. Throw the first flush away. Everything that has landed on the bimini since the last rain — salt, dust, bird mess, and on new canvas the water-repellent finish itself — comes off in the first minute. Dump it. Rinsing the canvas with seawater first is fine; a towel wipe afterward takes care of the residual salt.
2. No ¼″ fittings anywhere. They clog with exactly the debris you're trying to keep out, and they throttle the flow when it's raining hardest. Keep everything ½″ or bigger.
3. No low spots in the hose. Water sitting in a sag goes stale and slows the flow. Route it so it drains completely, and stow it dry.
4. Taste before it goes in the tank. Same discipline as the watermaker: catch into a jug, taste it, then commit it. A gallon in a jug is a nuisance; a bad 40-gallon tank is a weekend of scrubbing.
5. Rig it before you need it. Squalls do not wait while you find the funnel. Ours lives in a labelled bag with the hose, in the cockpit locker, ready to deploy in about two minutes.

Part two — we're building it anyway

Everything above still stands. The $400 plan is the right answer for coastal cruising, and we're doing that too. But the parts list was sitting there, the used pump turned up, and the boat's mission has a way of changing. Most of the components are aboard. This section is the build log — the questions that came up once the boxes were open, and the answers we found.

Why two prefilter housings

The short answer: one 5 micron cartridge in coastal water blinds off fast, and a blinded prefilter is the most common way this exact build destroys itself.

Run the stages in series — 20 micron first, 5 micron second — and each one does a different job:

• The coarse stage does the dirty work. Silt, plankton, the general soup of a coastal anchorage — that all loads onto the 20 µm cartridge, which costs a couple of dollars and gets changed often. The 5 µm behind it sees relatively clean water and can last a whole season.
• It protects the pump. The 107SS has ceramic plungers and stainless valves running at 900 psi. Grit is what kills them. Two stages mean a single cartridge channeling or tearing doesn't put sand straight into the pump.
• It's the starvation guard. This is the one people miss. As a cartridge loads up, the pressure drop across it climbs, and the pump inlet starts to go into vacuum. A positive-displacement pump that can't get fed cavitates — the discharge flow falls below its rated 0.25 GPM, and cross-flow at the membrane falls with it. Two stages mean that pressure drop builds gradually and predictably instead of all at once.

Could you run one housing? Technically yes. But then you'd want to change it religiously and you'd have no warning before it starved the pump — and the second housing costs about $30. Keep both.

Pick the media, not just the micron rating

Two cartridges at 20 and 5 microns is the right staging. What matters almost as much is which kind of cartridge goes in each housing, and the answer is different for the two stages.

• Coarse stage: pleated polyester, 20 µm, washable. A pleated cartridge has several times the surface area of a spun cartridge in the same can, so it holds far more dirt at a lower pressure drop — and you can pull it out, rinse it in the sink and put it back, many times over. On a small boat where locker space for spares is real estate, a washable sacrificial stage is worth a lot.
• Fine stage: melt-blown depth, 5 µm. Here you want depth filtration, where particles are captured throughout the thickness of the media. A pleated cartridge at 5 µm loads on its surface, and a surface-loaded cartridge can release a slug of accumulated dirt downstream when the pressure drop climbs — straight at the pump and the membrane. Depth media doesn't do that. This one is a consumable; don't try to wash it.

Don't go finer than 5 µm. It's tempting to fit a 1 µm as insurance, but the membrane doesn't care about the micron number so much as the fouling index of what reaches it, and a 1 µm cartridge blinds several times faster for very little gain.

One nice thing about this build: at 0.25 GPM you're running standard 10″ cartridges rated for 1–2 GPM at something like 15% of their rated flow. Pressure drop will be trivial and cartridge life long. The filters are the one part of this system with generous headroom — no reason to upgrade to Big Blue housings.

The feed pump: how we got this wrong

The parts list above used to say "12V diaphragm feed pump, $60." That was not a specification, it was a shopping hint, and it cost us a pump.

What turned up was an AUBIG DC40A-1245 — 12 V, 1.2 A, 620 L/h, and on the face of it a bargain. It is also a brushless magnetic-drive centrifugal submersible pump: an aquarium and fountain pump. It cannot do this job, for three separate reasons.

• It doesn't self-prime. Submersible means it has to sit in the water. A watermaker in a portable crate has to lift from over the side, and this pump can't lift at all.
• Its maximum head is 4.5 m — about 6.4 psi. Pumptec's own manual asks for +20 psi at the inlet for optimum performance. Even flooded, this pump is a long way short.
• Centrifugal pumps collapse under back-pressure. That 620 L/h is the figure at zero head. Put a strainer and two cartridges in front of it and you have a trickle.

The result would have been the plunger pump sucking through two prefilters instead of being fed — which is the exact failure the manual warns about in capitals: DO NOT STARVE THE PUMP OR RUN DRY.

What the 107SS actually asks for

Straight from the Pumptec service manual, and worth reading before you buy anything upstream of it:

• +20 psi inlet pressure for optimum performance; most will run on flooded suction with good inlet plumbing.
• Inlet supply must exceed the flow the pump delivers — so more than 0.25 GPM, with margin for a loading prefilter.
• Inlet line at least one size larger than the pump inlet fitting, and it must be flexible hose, not rigid pipe. Your ¼″ inlet on ½″ hose is right.
• Avoid tees, elbows and valves in the inlet line to reduce the risk of cavitation.
• Above 130 °F feed you need a pressurised inlet regardless.

That fourth point sounds like it forbids the gauge tee and the accumulator tee. It doesn't, as long as they're full-bore and nothing flows through the branch — a dead-ended branch costs you essentially no pressure. But it does mean: put the isolation valves upstream of the filters, use full-bore ball valves rather than gate valves, and keep the last run into the pump short, straight and flexible.

Demand or bypass? Demand — and this one is counterintuitive

Whatever pump you fit will deliver far more than 0.25 GPM, so something has to absorb the difference. There are two ways, and they are not equivalent.

A demand pump has a pressure switch and only runs when pressure drops. Feeding a 0.25 GPM draw from a 1.8 GPM pump, it runs roughly 14% of the time — so its average current is a fraction of its running current. A bypass pump runs continuously and diverts the surplus back to its own inlet, which means it draws full current for the entire session while doing the same useful work.

Bypass is the obvious-looking answer because it never cycles. But if you care about amp-hours per gallon, demand wins by a wide margin — and the only real drawback of demand, thousands of rapid on-off cycles, is exactly what an accumulator fixes.

So: a demand pump, plus a 1–2 quart accumulator teed in between the last prefilter and the HP pump inlet. The accumulator stretches each cycle from seconds to minutes, gives the plunger pump a steady pulsation-free inlet, and costs about $45. You get low average current and long gentle cycles at the same time.

What to actually buy

A Shurflo 8000-543-236, 12 VDC, about $123, ticks every box:

• 1.8 GPM, 50 psi maximum, 6.4 A maximum draw — averaging well under 1 A at this duty cycle
• Self-priming to 8 vertical feet, which covers a crate anywhere on a boat this size
• Can run dry without damage — which, on a portable rig where the intake hose can lift clear in a chop, matters
• Santoprene diaphragm and Viton valves in a plastic head: the right materials for seawater
• ⅜″ FNPT ports, the same thread as your filter housings

You come out ahead on power as well as capability — roughly half an amp average against the aquarium pump's 1.2 A continuous, while actually delivering the pressure the system needs.

Two small valves worth fitting

Isolation valves either side of the filter bank, so changing a cartridge doesn't drain the low-pressure side into the bottom of the crate. And put the housings vertically, caps up, mounted below the vessel, with enough clearance underneath to drop the sumps off — that way air purges upward on start-up and the whole low-pressure side drains for storage.

What each part has to be

The feed pump got through because the parts list named a category and a price instead of a specification. So here is the same list again, written the way it should have been the first time — with the criteria that rule a part out. Print this and shop from it rather than from the price table.

The trap, in actual numbers

Here's the cleanest way to see what the design is asking of the membrane.

A FilmTec SW30-2514 — the standard 2.5 × 14 inch seawater element — is rated at 150 gallons per day, and the test conditions on the data sheet are 800 psi, 32,000 ppm salt, 77 °F, and 2% recovery. Two percent. That implies something like five gallons a minute of feed water sweeping past the element to carry away the salt it rejects.

Your pump delivers a quarter of a gallon a minute. So to get 3 gallons an hour of product out of 15 gallons an hour of feed, you're running the element at 20% recovery — roughly ten times the recovery it's characterized at, with a twentieth of the cross-flow.

That's the trap. It isn't a plumbing mistake you can fix with a fitting; it's the bargain you struck when you chose a pump that draws 15 amps instead of 50. What you can do is decide how hard to push it — and that decision is a straight trade against amp-hours.

There is no free option here. Being gentle on the membrane costs amp-hours; saving amp-hours costs membrane life. At $150–200 an element, running at 20% and replacing it every couple of seasons is a defensible choice — just make it deliberately rather than by accident.

The ratio rule

You never need to know what the pump is actually delivering. Recovery is product divided by the total, and both of those come out of hoses you can point into a jug — so catch both streams for the same length of time and compare volumes. The ratio does the work, which means an extra second on the stopwatch affects both equally and cancels out.

One part product to seven parts brine is a good place to live. That's the whole method — no spec sheet, no trust in the gauge, and no stopwatch you have to read to the second.

How to actually hit it

1. Measure, don't guess. Brine hose into a bucket, product hose into a measuring cup, both started at the same moment. A minute is plenty. Then use the ratio table above — and give it five to ten minutes to settle after every adjustment before measuring again. The membrane doesn't respond instantly and you will chase your tail otherwise. Two or three iterations gets you there.
2. Pressure is the only control. Winding the regulator up raises pressure, which raises product flow, which raises recovery. Wind it down and everything falls. There is no separate "recovery knob" — resist the urge to chase gallons per hour by adding pressure.
3. Re-measure when the water temperature changes. This is the one that catches people. Membrane flux rises roughly 3% per °C, so a setting that gave a clean 12.5% in 60 °F spring water will be pushing 18–20% in an 80 °F August anchorage — same pressure, same everything, and you would never know without measuring. Rule of thumb: back off about 50 psi for every 10 °F rise in feed temperature, then verify with the jug. Check at the start of every trip, not once at commissioning, and always log feed temperature alongside pressure — a reading without it can't be compared to anything.
4. Treat 800–850 psi as a ceiling, not a home. This is the correction that matters most. Pressure is a consequence of the setting, not the target — at 10–15% recovery you will probably settle somewhere around 650–780 psi, and at 800 you are more likely near 18% than 15%. Set the regulator by the jug ratio and let the gauge read whatever it reads. The 1000 psi shared by the vessel and the 12 V pump is the hard limit above that.
5. Sweep during the run, not just at the end. Winding the regulator all the way down for two minutes every half hour gives you the full 0.23 GPM across the element at 0% recovery, which flushes loosely-held deposits before they consolidate. Given how little cross-flow you have the rest of the time, it is the cheapest membrane insurance in the build — it costs a few minutes of production and nothing else.
6. Finish every run with the same sweep, for five minutes. Pressure drops, product goes to nearly nothing, and the element face gets the best cross-flow this pump will ever give it.
7. Then freshwater flush — and see the warning below about what "fresh water" means.
8. Keep a log. Pressure, product rate, feed temperature, product TDS, hours. When it takes more pressure to make the same water at the same temperature, fouling has started and you still have time to do something about it.

Loading the vessel

The element goes in one way round. Get it wrong and you will spend a weekend wondering why you're making no water.

Which way the element faces

Spiral-wound elements carry a brine seal — a chevron-shaped ring near one end that seals against the inside of the vessel. Its job is to stop feed water sneaking around the outside of the element instead of going through it, and it only works in one direction. DuPont's loading manual is explicit: the seal opens in the upstream direction, and the element is always loaded into the feed end of the vessel.

Put it in backwards and the seal collapses instead of expanding. Feed water bypasses the element, product flow goes to almost nothing, and you can damage the seal getting it back out. So: identify which port on your vessel is feed before you load anything, and load from that end.

Silicone, not glycerin

Lubricate the brine seal and the end-cap O-rings with a thin film of silicone grease — DuPont specifically recommends Dow Corning 111, which carries NSF approval for potable water. Glycerin is often to hand on a boat and it works for about five minutes before it washes out. Use it sparingly: too much silicone is its own problem.

Shim it so it can't move

This one is easy to skip and it matters on a single-element vessel. If the element can slide axially inside the tube, it will hammer against the end cap every time you start and stop the system — DuPont calls this out by name in the loading guidelines. Push the element home, then shim the gap between the face of the element and the end-cap hub so there's no free travel. Don't overdo it in the other direction either: force the element too far in and the end plate won't seat properly.

Never dead-head the permeate line

The data sheet says to avoid static permeate-side back-pressure at all times. The membrane is a thin film glued to a backing; push on it from the wrong side and it delaminates. In practice this means the product line must always have somewhere to go while the pump is running.

So the diverter on the product side should be a genuine three-way valve — dump or tank, always one or the other — not a two-way ball valve you could accidentally close. Keep the product line short and unrestricted, and if you fit an inline TDS probe make sure it isn't a choke point.

Mounting

Horizontal is standard and fine. If you can give it a degree or two of tilt, put the feed end low so air migrates out with the brine on start-up and the tube drains for storage. Leave 15–20 inches clear at one end so you can actually withdraw the element — easy to forget when you're laying out a crate, and miserable to discover later.

The plumbing, junction by junction

Two rules make the fitting list almost write itself. Everything upstream of the high-pressure pump is ½″ hose and barbs — cheap, flexible, forgiving, and easy to route around a crate. Everything downstream is ¼″ NPT in 316 stainless, because that's what the pump, the vessel and the regulator all speak.

Also budget for: about 15 ft of ½″ ID reinforced hose, 6 ft of ⅜″ hose, a dozen or so 316 hose clamps, PTFE tape and a high-pressure thread sealant. Call the whole fitting list $120–180 depending on how much of it you buy in stainless. The three additions from the configuration work above — accumulator (~$45), carbon cartridge and housing for the flush leg (~$55), two isolation valves (~$25) — add roughly $125 on top.

Why two tees, and why a hose off the pump

A single ¼″ cross does the same job in one fitting, and it's the obvious-looking choice. Two instrument tees in series are usually cheaper, easier to source, and they space the gauge and the relief valve apart instead of clustering four connections at one point — which makes the gauge readable and the relief valve reachable when everything is packed into a crate. It costs you about an inch of length.

More important: don't hang that assembly off the pump on a rigid nipple. A plunger pump pulses several hundred times a minute, and a rigid cluster of tees, gauge and relief valve bolted to the discharge port is both a fatigue concentration and a couple of pounds of overhung weight on a casting — in a box that gets carried, dropped and bounced. Use a short high-pressure hose from the pump to the first tee and bracket the tees to the crate. The hose isolates the vibration, takes the load off the pump, and does half the gauge-damping the snubber is there for.

That's why the build takes three hoses rather than two: 12–18″ from pump to gauge tee, about 18″ from switch tee to the vessel, and about 24″ from the vessel's brine port to the regulator. Don't go shorter than about 12″ on any of them — bend radius at the fittings is what kills hydraulic hose.

Why a regulator instead of a needle valve

Earlier versions of this build used a needle valve on the brine line as the pressure control, with a relief valve on a branch as the safety. That works, and it's the cheaper way in. We changed it, and the reason is worth understanding because it's about which failure you're designing out.

A needle valve is a throttle. It does exactly what you tell it, including the wrong thing — wind it a quarter turn too far while chasing gallons per hour and a positive-displacement pump has nowhere to send its flow. Everything on the high-pressure side is then relying on a relief valve, on a branch, working correctly on the one day it's ever asked to.

A back-pressure regulator is a different animal. It holds a set pressure by opening further as pressure rises, so it cannot be wound shut. The most likely failure in this whole system — a human over-closing the throttle — simply stops existing. That is worth $199 on its own, and it buys repeatable pressure settings as a bonus: you dial the spring to a number rather than feeling for it.

Two things to confirm with the supplier before ordering the regulator, because the listings are thin: that it is genuinely a pressure-sustaining regulator rather than an unloader, and that it will regulate at 0.23 GPM. Ours is advertised to 5 GPM, and a device running at 5% of its rated flow can regulate coarsely or hunt.

One thing the regulator does not fix: temperature drift. It holds pressure, not recovery. Warm water still gives more product flow at the same pressure, so you will still be re-trimming it seasonally and re-doing the jug measurement. See the recovery section.

Wiring it

This is the part of the build we hadn't written down at all, and it's the part with a fire risk in it.

Add it up: the high-pressure pump draws 15 amps and the feed pump can peak at 6.4, so the crate wants a circuit good for about 21 amps, running for two or three hours at a stretch. That is not a cigarette-lighter socket. Those are typically rated 10–15 A, the plug contact is a spring against a tube, and at 20 A the contact resistance turns into heat in a fitting nobody is looking at. Don't.

What the circuit needs

• A dedicated circuit from the house bank, not a spur off something else. A 25 or 30 A breaker or fuse, sized to protect the wire, with the fuse within about 7 inches of the battery terminal as ABYC asks.
• 10 AWG tinned marine cable for runs up to roughly 10 feet each way, which keeps voltage drop near 3%. Go to 8 AWG beyond that. Voltage drop isn't just tidiness here — the pump's output falls with supply voltage, so a sloppy run costs you water.
• An Anderson SB50 connector (or similar) where the crate plugs in. Keyed, rated well past the load, and it lets the whole rig come out of the boat in one piece — which was the point of building it portable.
• Separate switches for the feed pump and the high-pressure pump, so you can prime and flood the filters before the plunger pump ever turns.
• Ventilation. Don't seal the crate. A 180 W motor running three hours in a closed plastic box gets hot, and hot motors draw more current.

Wiring the high-pressure cut-out

The cut-out switch on the second tee is the only thing standing between a blocked hose and a burst fitting, so wire it as a safety device rather than a convenience:

• Normally closed, in series with the high-pressure pump's supply. Fail-safe: a broken wire or a pulled spade stops the pump instead of silently removing the protection.
• Manual reset. An auto-reset switch faced with a blockage will restart the pump the moment pressure decays and keep doing it. You want it to stop and stay stopped until you have found out why.
• Set about 900 psi, with the regulator working at 800. That leaves the conventional 10% margin above working and still keeps 100 psi under the 1000 psi ceiling shared by the vessel and the 12 V pump.
• Switch the pump, not the whole crate. Leave the feed pump running when the cut-out trips, so the high-pressure pump is never restarted against a dry or unprimed inlet.
• Stainless wetted parts. A plated-steel hydraulic switch body will corrode on a seawater branch, even one that gets freshwater-flushed. If the only affordable switch you can find has a steel body, mount it on a 316 nipple, keep it high in the crate, and treat it as a wear item.

Your needle valve becomes a spare. Don't be tempted to plumb it in series with the regulator "just in case" — that puts a device back in the line that can be wound shut, which is the exact failure this architecture removes. Keep it in the spares bag; if the regulator ever fails you can fall back to it as a manual control, carefully.

Mount the pump on rubber isolators if you value your ears. A plunger pump bolted rigidly to a plastic crate turns the crate into a soundbox.

First start-up, and the log

1. Regulator wound all the way down. Feed pump on, HP pump off. Let seawater flood the prefilters and push the air out — you'll hear it settle.
2. Note the inlet gauge reading with clean cartridges. Write it on the crate. That's your reference forever.
3. Start the HP pump with the regulator still wound down. Run several minutes at low pressure. Product will be salty and that's fine — it goes to the dump port.
4. Wind the regulator up slowly, watching the high-pressure gauge, until you reach about 700 psi — a starting point, not the target. Slowly is the operative word; a fraction of a turn does more than you expect.
5. Let it stabilise for ten minutes, then do the jug measurement and trim to one part product to seven parts brine. Wherever the gauge ends up is where it ends up.
6. Check product TDS. Under about 500 ppm and it's drinkable; a new element should do much better than that.
7. Discard the first hour of permeate on a new element — that's DuPont's instruction, not a rule of thumb. The element ships wet with preservative and it takes that long to wash out. Only after that does the diverter go to the tank.
8. Log pressure, product rate, feed temperature and TDS. That first entry is what every later one gets compared against.

Start a log with the first run and keep it in the crate. Everything that tells you the membrane is going off — rising pressure for the same output, falling product rate, creeping TDS — is only visible as a change against an earlier reading. One page of this is worth more than any amount of guessing later.

Three columns do most of the work. Inlet psi falling means the prefilters are loading. HP psi climbing for the same product rate at the same temperature means the membrane is fouling. TDS climbing means the membrane is failing rather than fouling. Feed temperature is there because it moves everything else — a reading without it can't be compared to anything.

The install video goes up on YouTube once it's all in the crate and making water.

Sources

• Sailboatdata — Hunter 290 specifications (40 gal water, 20 gal fuel, Yanmar 18 hp)
• SeaWater Pro — 12V single membrane watermaker ($4,295, 17 GPH, 600 W)
• Defender Marine — Katadyn Survivor 35 ($3,480, 1.2 gal/hr manual)
• Abroad Reach Travel — DIY Watermaker build log (~3 gal/hr at 900 psi, ~14 A, parts list)
• Misting Direct — Pumptec 107SS 12V pump and motor ($629, 0.25 GPM, 15 A at 12 VDC)
• Pumptec — 107SS service and operating manual (+20 psi optimum inlet pressure; flexible inlet line one size larger; avoid tees, elbows and valves in the inlet; do not starve the pump or run dry)
• Pumptec — 107SS series (every variant is 0.25 GPM; the 12 V model is 1000 psi)
• Shurflo 8000-543-236 12 VDC diaphragm pump (1.8 GPM, 50 psi, 6.4 A max, self-priming to 8 ft, runs dry without damage, Santoprene diaphragm and Viton valves, ~$123)
• Pentair Shurflo — 8000 series overview (demand versus bypass behaviour, flow and material options)
• Hydrocomponents — 2.5 × 14 in 1000 psi pressure vessel ($396)
• Hydrocomponents — DIY watermaker guidance (low feed flow and membrane scaling)
• DuPont — FilmTec SW30 seawater element data sheet (SW30-2514: 150 GPD rated at 800 psi, 32,000 ppm, 77 °F and 2% recovery; operating limits — <0.1 ppm free chlorine, SDI 5, 15 psi max element pressure drop, avoid permeate back-pressure, discard the first hour of permeate)
• DuPont — Loading of Pressure Vessels, preparation and element loading (brine seal opens upstream, load from the feed end, silicone lubricant, shimming to prevent hammering)
• Practical Sailor — Install a Water Saver: A Galley Foot Pump
• Defender Marine — Whale Gusher MK3 galley foot pump
• All At Sea — Harvesting Rainwater on Your Boat
• Practical Sailor — The Rain Catcher's Guide (0.44 gal per sq ft per inch of rain; white potable hose, no ¼″ fittings, no low spots; Pentek FloPlus-10 and chlorine dosing)

All prices checked August 2026 and will drift. Treat them as a shape, not a quote.