A Dialysis Centre Needs Reverse Osmosis. Most Buildings That Ask Us for It Do Not.

Reverse osmosis is the most asked-for thing in our water and water treatment work and the most often wrongly asked for. Somebody has a water problem, somebody else has told them the answer is RO, and by the time the enquiry reaches us the technology has already been chosen and the only question left is the price — and on the wastewater side we have learned that the technology is rarely what actually decides a job. We have quoted it, installed it, designed around it, and on one of our own jobs we took it out of our own drawing. This note is about how we decide, because the deciding is where the money is.

Two things on the blog already touch this and neither is this post. The one about drinking treated wastewater lists membranes as part of what a potable barrier stack would need, and the one about whether reuse pays says plainly that we decline leachate reuse because doing it properly means membranes and a concentrate stream. This post is reverse osmosis itself — what it removes, what it costs to own, the job where it was not negotiable, and the four problems people bring us that are not RO problems at all.

Start with what it actually removes

Reverse osmosis is the only common process that takes out dissolved salts. That is the whole of its special claim. Everything else we sell removes something you could, in principle, catch another way: a filter catches particles, carbon adsorbs organics and chlorine and most taste and odour, oxidation and filtration take out iron and manganese, a softener exchanges hardness ions for sodium, ultraviolet light and chlorine deal with microbiology, and there are two ammonia routes we will not sell at all. None of those touch the dissolved load. Sodium, chloride, nitrate, sulphate, total dissolved solids — if the thing you need gone is dissolved and your other options cannot be selective about it, you are in RO territory and nowhere else.

So the first question we ask is not what the water tastes like or what the client wants the system to be, and it is never which company is cheapest. It is: what exactly are we removing, and is it dissolved? We get that from a test, in a clear glass if nothing else is available, the same way we open every water treatment quotation. Half the RO enquiries we receive answer that question in a way that points somewhere cheaper, and we say so.

The job where it was not negotiable

We designed and quoted a water system for a hemodialysis centre in Metro Manila, thirty beds, about thirty cubic metres a day. I have written about that centre before from the effluent side, where the question was whether treated water is ever safe to drink and the answer turned on how many independent barriers it takes. Here is the supply side of the same job.

Dialysis water is not a preference. The water contacts a patient's blood across a membrane for hours at a time, so the dissolved load has to come down to a level that no filter and no softener can reach. There is no argument to have. RO is the process, and the only design conversations are about pre-treatment, redundancy and what happens to the reject.

And the arithmetic on that job is the thing I would want a building owner to see. Of roughly thirty cubic metres a day passing through that facility, about half was product water for the dialysis stream and about half was RO reject, which went to drain and was combined with the rest of the discharge before it left the building. Nobody was being wasteful. That is reverse osmosis working exactly as designed.

The reject is not a footnote

This is the part that gets left out of the sales conversation and I have never understood why, because it is the single biggest thing RO does to a building.

A reverse osmosis unit splits the feed into two streams. One is the water you wanted. The other is everything the membrane refused to pass, in a smaller volume of water, which means at a higher concentration than what you fed it. The fraction that comes out as product is the recovery, and on the small and medium commercial units that make up almost all of our enquiries it is well short of everything. We design on the figures the membrane manufacturers publish rather than on anything we have measured ourselves, and the honest statement is that a meaningful share of the feed leaves as reject.

Work it the way a building owner has to. If you need ten cubic metres a day of RO water, you do not buy ten cubic metres a day. You buy that plus the reject, you pay for all of it on the water bill, and you discharge the difference. Three consequences follow, and none of them is on the brochure:

The first is that your incoming supply has to be big enough for the feed, not for the product. If the building is on a metered municipal line, that is a bill that goes up by more than the water you use. If it is on a deep well with a pump already at its duty, the plant may not fit the supply at all.

The second is that the reject becomes part of your discharge. It is water, it carries a concentrated dissolved load, and it leaves the property. On a site with a discharge permit with a declared flow on it, that reject counts towards the number the owner declared, and the number in the permit is a ceiling the owner set himself. We have seen an RO installation added years after a plant was sized, with nobody going back to check what it did to the declared figure or to the parameters being sampled.

The third is that if there is a treatment plant downstream, somebody should ask what the reject does to it. Mostly it is a dilution of the organic load and a lift in dissolved solids, which is usually tolerable. On a small plant it is a hydraulic addition that was never in the capacity the tanks were built for. That is a question with an arithmetic answer, it takes ten minutes, and our price estimator will give you the shape of it before anyone visits.

Where we took reverse osmosis out of our own design

On a Cebu job we were revising our own proposal through several issues. An earlier revision put a membrane stage at the end of the treated water train so the building could reuse water for flushing and washdown. It would have worked. We removed it.

What replaced it was filtration, activated carbon and ultraviolet disinfection we could validate — which is the treatment level that flushing and washdown actually need, and the ladder we set out when we ask a client what the reuse is for. Washdown water is judged by eye. Flushing water goes back inside the building and so it needs disinfection we can stand behind. Neither of them needs the dissolved salts taken out, and we were about to sell a process whose entire special claim was removing dissolved salts.

Look at what came out of the design with it. Concentrate handling, and a destination for it. Antiscalant dosing, which means a dosing pump, and our dosing pumps are our own unsolved problem. Clean-in-place chemicals and the procedure that goes with them. Monitoring feed water quality closely enough to protect the membrane. And a membrane set to be replaced on a cycle of years, which is a capital item arriving in the client's operating budget. That is five recurring obligations deleted by asking one question about what the water was for, and this is the same instinct as talking a client down from the cleanest water we could build: above the governing end goal you are buying quality nobody uses.

Four problems that arrive labelled as RO problems

These four account for most of the enquiries where we end up quoting something else.

Hardness. Scale on fittings, soap that will not lather, white deposits in the kettle. That is calcium and magnesium and it is a softener's job, sized twice — once on capacity and once on service flow rate. Hardness also does things people do not connect to hardness, like turning kitchen fat into soap in a grease trap. RO will remove hardness. It is an expensive way to do it and it brings all five obligations above with it. In a commercial kitchen the softener is almost always the answer.

Iron and manganese. Orange staining, brown water on first draw, black specks. That is an oxidation and filtration job, and we have our own media for it. Fed to an RO membrane untreated, iron is a fouling agent rather than a thing the membrane cheerfully removes.

Chlorine taste and odour. Carbon, and nothing more. Worth knowing in the other direction too: chlorine damages the common thin-film membranes, so any RO fed from a chlorinated municipal supply needs dechlorination ahead of it whether anyone minds the taste or not.

Bacteria. Ultraviolet or chlorination, selected and sized for the duty. RO does hold back microbiology, but using a membrane as a disinfection barrier means trusting its integrity, and integrity is a thing you have to monitor rather than assume, the same way we treat anything that has to keep working after handover.

There is a pattern in all four. People reach for RO because it removes everything, and removing everything feels like the safe choice. The cost of removing everything is paid monthly, by somebody who may not be the person who signed — and a price difference between two quotations is a scope difference first. Our comparison of systems exists for exactly this conversation.

What it costs to own, in items rather than pesos

I am not going to put a figure on an RO plant here, because the figure depends on the feed water and I would be inventing it. What I can do is list what an owner takes on, which is more useful anyway and is how we argue cost of ownership on the wastewater side too.

Pre-treatment, always. A membrane is the most delicate thing in the plant and everything upstream exists to protect it, which is why pre-treatment ahead of RO is a subject in its own right rather than an accessory. Dechlorination where the feed is chlorinated. Antiscalant dosing, with a pump that is a wear item. Periodic chemical cleaning, with the chemicals, the storage and somebody who knows the procedure. Watching feed quality, because a membrane fouls slowly and the first evidence is usually a pressure reading nobody was recording. A high pressure pump, which is where the electricity goes — RO has a real energy cost per cubic metre, unlike a softener or a filter. A membrane replacement cycle measured in a small number of years. And remineralisation if the water is for drinking, because RO water is aggressive to pipes and tastes of nothing.

None of that is a reason not to buy RO. It is the reason to buy it on purpose, for a dissolved-solids problem you have actually identified, rather than as a general upgrade. And it is the reason we price from rates we have paid rather than from an estimate nobody has tested.

Where it genuinely earns its keep

I would be misread if this came across as an argument against the technology, so here are the cases where we put it forward without hesitation.

Dialysis and other clinical water, as above. Water refilling stations, where the product being sold is the water itself and the dissolved load decides whether there is a product at all — that whole business is built on RO, which is why we keep a dedicated page for refilling station operators. Boiler feed and process water with a specification written in dissolved solids. Brackish or saline sources, where no other process will get you there. And genuine potable treatment where the raw water carries a dissolved contaminant such as nitrate that filtration cannot touch, and where drinking water standards leave no other route.

What those cases share is that the requirement is stated in dissolved terms, by somebody with authority to state it. When that is true, the five obligations are simply the cost of the only process that works, and we design them in properly rather than hoping the client will not notice them. Where the plant that follows is a sewage treatment plant, the same discipline applies to how we scope that.

What we ask before quoting one

Five things, none of which is a warning and all of which are ordinary survey items on the consultation:

A water analysis, or at least a sample we can look at. What the water is for, in the client's own words, because that is what sets the target. How much product water per day is genuinely needed, as against the number that came from a rule of thumb. Where the reject is going to go, physically, and whether the site has a discharge permit it affects, which for an industrial site means knowing which parameters your industry is actually held to. And what the incoming supply can actually deliver, because the feed is larger than the product and more RO plants are limited by supply than by anything else.

If those five have answers, the design takes very little time and the quotation is honest. If they do not, any number we gave would be a guess dressed up as an engineering document, and we would rather you checked our reasoning than took our word for it.

Where this stops being transferable

Every site differs, and the parts of this that are arithmetic travel while the parts that are judgement do not.

Recovery is a property of the specific membrane, the feed water and the configuration, so the reject fraction on your plant is not the reject fraction on ours — get it from the proposal you are holding, and if it is not stated, that is itself worth asking about. On a very large installation the economics change: staging, energy recovery and concentrate management become their own engineering subject and the rules of thumb here are too small to apply. Where the municipal supply is already low in dissolved solids, RO is sometimes the cheapest route to a clinical standard simply because the pre-treatment burden is light. And the reverse of my own argument holds honestly: if somebody has already identified a dissolved contaminant and been told to solve it with a filter and some carbon, they are being sold the wrong thing in the other direction, and that is the more expensive mistake because it does not work at all.

We work on both the wastewater and the water treatment sides out of Cordova, Cebu, which is why these two questions keep arriving in the same conversation — our Cebu work and our completed projects are there to look through. If you have a water analysis and a problem you can describe, bring both and we will tell you which process it is, including when that answer costs us the bigger sale. Book a free 15-minute consultation.