Our Dosing Pumps Last About a Year. We Treat That as a Design Problem.

There is a part of designing a sewage treatment plant that almost never appears in the drawing set, and it is the part I now spend the most thought on. Not the tank sizes. Not the arithmetic. The question of who is going to keep this equipment running in year three, and whether the thing we specified makes that possible or impossible.

I have written before that building the plant is the easy part, and that the drawing is where the money is actually decided. What I want to put down here is narrower and, honestly, less finished: how we choose equipment when the criterion is after-sales support, and the one piece of equipment where we still do not have a good answer.

If I rank the parts of this work by value, calculations come last

Design, equipment specification, commissioning. That is the order in which the work happens, and it is close to the reverse of the order in which people think value sits.

The calculations are the part I worry about least. In fairness we did not have AI when I started. With AI, the arithmetic of a plant is a breeze now — flow, load, chamber volumes, oxygen demand, the carbon-to-nitrate ratio you divide before buying methanol. All of it checkable. I tell clients to put our design through ChatGPT and I mean it. If a number is wrong, I want it found.

Choosing the right specification for the equipment is a whole new ball game altogether. And commissioning a plant is, for us, more important and more critical than building it and installing the right equipment. Knowing what to do next based on the results we are getting on site, the smell we are smelling on site — that is far more valuable than any sheet of calculations.

I can put a price on how much I believe that. On one project we charged about 2.4 million pesos, we paid another consultant 200,000 pesos — roughly eight per cent of the contract — because we were having issues at our commissioning stage. We paid it happily. It is the only real market price for wastewater consultancy in my own records, and in that transaction we were the buyer, not the seller. Every site is different, so read that as one job's number and not a rate card.

The dosing pump is our own unsolved problem

We design based on after-sales support. We try not to use something that will give us problems in the future. Dosing pumps are where that principle is currently losing.

They fail fast. In our experience they last maybe a year. That is not one bad batch on one site; it is a pattern across plants, and it is the single item I get called back about most. A dosing pump is a small thing in the bill of materials and a large thing in the client's life, because when it stops, the chemical stops, and the parameter it was holding starts drifting within days. If that chemical was alkalinity correction, the nitrifiers are the first to notice. If it was coagulant, phosphate turns up in the final sample.

Part of the cause, I suspect, is that these pumps are often left out in the open — sun, rain, salt air — which is wrong to begin with. Cebu is not gentle on a small plastic pump head with a diaphragm in it. So we have started treating the pump's environment as part of its specification rather than as site housekeeping: a shaded, ventilated position with the chemical drum beside it and a man able to stand in front of it. Not a cabinet somebody has to be sent to unlock.

I am not going to claim we have solved it. We are still figuring this out. What we have done is stop pretending it is the operator's fault.

What "specify for after-sales" actually changes

Three things, mostly.

First, we do not specify a unit that only one importer can service. If the plant needs a part and the part takes six weeks and one exclusive agent to arrive, then the plant's compliance depends on that agent's stock level, which is not an engineering position I want a client standing in. The same logic is why we never let compliance depend on a chemical that may not be reliably available locally. A design that only works when the supply chain cooperates is a design with a hole in it.

Second, we size machines rather than multiplying small ones. On a bigger plant, one duty unit and one standby beats six little units — six of anything means six things to service, six sets of spares, and six chances that the one that failed is the one nobody noticed. On small plants we go the other way for a different after-sales reason: we use air pumps rather than air blowers, because blowers are simply too loud, and a noisy machine in a basement next to occupied rooms gets switched off by somebody eventually. Unless the strength is genuinely high, air pumps stay. A plant that is off for three days because the aeration was unbearable is a design failure wearing an operations costume.

Third, we put the chemical lines in the quote honestly. Alkalinity correction and coagulant are shown as estimates and marked "if required", the same way sludge hauling and third-party laboratory testing are always estimates. We would rather the client understand which running costs are conditional than discover them later. That is the same argument as the cheaper quotation that was not the cheaper plant.

The equipment we can service is often the equipment we can also live without

The cleanest way to reduce after-sales exposure is to need less equipment. That is not a slogan; it drops out of the design decisions we make anyway.

A fixed-film stage on FBBR-style media holds biomass in the tank, so the process tolerates the kind of upset that would otherwise be managed with pumps and instruments. On underground tanks we install no weir plate and no scum board, because neither can be levelled or skimmed through a hatch once the cover slab is on — that is confined space entry, and we are not designing a job that requires a man to climb into a tank to do routine maintenance. We use a submerged draw-off cast through the cross wall instead, and a sludge pump in the sump. Fewer moving parts above water, fewer reasons to open a lid.

We do not track MLSS on most of our plants either. We read a cone, because a cone does not go out of calibration and does not need a consumable that a provincial supplier has stopped carrying. An instrument nobody can maintain gives worse information than a simple test somebody actually does.

The operator is part of the specification

Whatever we specify has to survive the person in front of it, and that person changes. Turnover is often the real problem rather than the plant, and a good operator can carry a plant but cannot carry a bad design. So the test we apply to a piece of equipment is not "can it be operated correctly" — it is "what happens when it is operated by somebody who started last month".

Practically: controls that fail to a safe state, a plant that recovers when you switch it back on, no setting that quietly wrecks the biology if it is nudged. We have lost a population of nitrifiers to a pump, and the pump was doing exactly what its wiring told it to. That was our specification's fault, not the operator's.

Risks and assumptions go in front of the client before signing

All of this goes in the quote and the presentation. The equipment we are proposing, what it needs, what it will cost to keep running per cubic metre, and where our assumptions could be wrong. The client has the right to know all of it before signing the contract, including the parts that are not flattering to us — such as a dosing pump we expect to replace sooner than we would like. Those are the questions I would ask a supplier myself, and I would rather answer them in writing than in year two.

If you are trying to work out what a compliant plant will cost you before you talk to anybody, the price estimator and the design generator will get you an indicative shape, and the compliance page sets out what the discharge standards actually ask of you. Our own wastewater practice, the parameters that matter by industry and the plants on our completed projects page cover the rest of the ground. If the difficult parameter is ammonia, nitrate or phosphate, those three pages are where our position on each is written down. And if the incoming water is hard enough to be part of the problem — a kitchen is the usual case — a softener on the inlet is often cheaper than more treatment plant behind it.

Where this stops transferring

These are our conclusions from our own plants, in Cebu and around it, at the sizes we usually work at. A dosing pump that lasts a year for us may last four years in an air-conditioned plant room in a large industrial estate with a maintenance department. A contractor with a different supply network will have different service answers, and may rightly specify things we avoid. Even inside our own work, the right answer changes with the reuse duty, the strength of the wastewater and how far the nearest competent technician is from the site.

So do not take this as a specification list. Take it as the question we ask about every item we put on a drawing: when this fails, who fixes it, with what, and how long is the plant out of compliance while they do. If that question has no comfortable answer, the item is wrong even when the calculation says it is right.

If you want to go through it against your own site, book a free 15-minute consultation.