We Took the Pump Out of the First Chamber. Now the Plant Has No Equalisation Tank at All.

We drew a plant for an existing office building with an equalisation tank in it, the way we draw nearly every plant. By the time the design basis was finished, that plant had no equalisation tank at all. Nobody deleted it to save money and nobody argued it out of the scope. It went because of a chain of three decisions, each of which was right on its own, and the first one was not about equalisation at all. It was about a screen that did not exist.

I have written separately about how we size an equalisation tank and why we size it off the shape of the day rather than a percentage, and that post says plainly that equalisation only works if what leaves the tank is genuinely constant, which means a transfer pump on level control and not gravity. This post is the case where we did the opposite, deliberately, and what we had to put in its place. The front-end chambers themselves — screening, grit and the fat — I have already covered as three separations we pour in concrete rather than buy as machines. And the plant in question is the same one whose clarifier return runs on gravity off the cone; that post mentions in passing that the chamber ahead of the clarifier discharges on level control. This is what that sentence is actually hiding.

The building's sewer was already in the ground, and there was nothing on it

This was a retrofit, which is a different job from a new build in a way that people underestimate. On a new building we draw the sewer and the plant together, so the screen channel sits where we put it and everything downstream of it is protected by definition. Here the building's drainage was laid years ago, the line arrives where it arrives, and the first chamber of our plant is the first thing in the whole system that anybody has ever thought about.

Which means the line arriving into that first chamber carries everything the building puts into it. Rags, wipes, sanitary items, plastic, the general solid traffic of occupied toilets. That is not a surprise and it is not a complaint about the client. It is simply what an unscreened sewer off an existing building contains, and the survey said so. The question the drawing then has to answer is what the first piece of equipment in the plant is going to be, and whether it survives contact with that.

Without a screen, that pump is a sitting duck

The conventional arrangement puts a submersible transfer pump in the first chamber. It is the obvious thing to draw and I have drawn it many times. On this plant the sequence of the chambers meant the screen chamber sat after the first chamber rather than before it, because of where the incoming invert landed and how the chambers had to be stacked in the space available.

So the pump would have been working in unscreened sewage, in the dirtiest position in the plant, on a building with no spare pump on site and nobody on staff who would notice it struggling. A pump in that position does not fail cleanly. It wraps, it loses duty, it runs hot and draws more current for less flow, and then it stops on a Saturday. I have watched one pump failure take the biology out of a plant, and I have watched plants where the problem was never the plant but who was left looking after it. Putting an impeller in front of an unscreened line is volunteering for both.

So we took it out. Not moved it, not upgraded it — removed it from the drawing. The first chamber now discharges onward by gravity, through a pipe, into the chamber that carries the bar rack. The solids reach the rack before they reach anything with a moving part in it.

A chamber that discharges by gravity is not equalising anything

This is the part I want to be honest about, because it is a consequence I did not choose and would not have written into the brief.

An equalisation tank is not a volume. It is a volume with a constant outflow. The whole point is that the peak arrives, the level rises, and what leaves the far side of the tank is the same steady rate it was leaving at two hours earlier. Take the pump out and you have a chamber that passes forward exactly what arrives, a few minutes later. The level barely moves, because an unrestricted gravity outlet does not let it. The chamber is still useful — it is still a quiet volume where grit begins to drop and where a surge is softened slightly — but it is not performing equalisation, and I am not going to call it an equalisation tank on a drawing because it holds water.

If I had insisted on keeping equalisation at that position, I would have been insisting on the pump, and the pump was the thing most likely to put the plant out of service. Between a design feature and a plant that stays running, the feature goes. That is the same test as drawing for the end goal rather than for more: every item in a plant is there because somebody asked for a result, and a chamber that creates a failure point to deliver an averaging benefit on a small building is not earning its place.

What runs the plant instead: a float, a differential, and a burst

Equalisation moved downstream. The chamber that does carry a pump — after the rack, where the water has already had its solids taken out of it — runs on level control, and it feeds the rest of the train in bursts rather than continuously.

The arithmetic of a burst is simple and it is worth writing out, because people assume level control is a detail in a panel rather than a design decision. The float differential on that chamber is about a hundred millimetres, which on its plan area is roughly a cubic metre of working volume. So every burst moves about a cubic metre, and it takes a quarter of an hour or so to move it at the transfer rate we sized. The number of starts a day is then just the daily flow divided by that cubic metre, which on a building of this size is tens of starts a day, not hundreds — and that is the number you check against the pump's permitted starts per hour before you accept the differential. A narrow differential gives you gentler bursts and more starts; a wide one gives fewer starts and a bigger slug of water arriving each time. You are trading one wear mechanism against another, and the decision belongs on the drawing, not to whoever wires the panel.

The float itself sits in a stilling tube. A bare float in a chamber that is being filled by an incoming line chatters, and a chattering float is a pump that starts and stops ten times in a minute. The tube costs nothing and it is the difference between a level signal and a guess.

Two settings that stop this arrangement going wrong quietly

Burst operation has a failure mode that is specific to it, and it is not mechanical. If the building goes quiet — a weekend, a holiday, a floor that empties out — the level never reaches the start float, so the pump never runs, and the water sits. Raw sewage held with no air goes septic, and then the slug that eventually arrives downstream is not just late, it is different water, with sulphide in it and a smell that reaches the building before it reaches the lab. I read plants by smell before the results come back, and this is one of the smells that has a cause you can design out.

So there are two settings in the panel that are part of the design and not part of the commissioning. The first is a minimum-run: if the pump has been idle for about four hours, it runs for ten minutes regardless of level. That moves the water and the air with it. The second is a low-low permissive, which simply refuses to let the pump run below a level where it would be pulling air, because a pump running dry on a plant with one pump in it is the whole plant.

Both of those have to be written into the commissioning record as values, not as intentions. This is the same lesson as the ball valve on that plant's transfer line: a setting that exists only in the head of whoever started the plant is a setting that will be restored to factory default in good faith by the next person, and nobody will connect the symptom to it. Commissioning finishes on a result, not on a date, and the record of what was set is part of the result.

The screen is now the biggest single risk on the job

Once the pump is downstream of the rack, the rack is carrying the whole plant. Everything that is going to protect that pump has to be caught by a bar rack that somebody rakes, and if it is not raked, the solids pass through and reach the one pump the plant has.

So the screen stopped being a modest front-end chamber and became the item I think about most on that drawing. We specified a fine bar rack at about six millimetres of clear spacing, in both tiers of the quotation. It is not a cheap-tier trade and it is not available as a saving, and I say so in the quotation rather than leaving it to be discovered. Wet wipes are the specific reason. A twenty-millimetre rack lets wipes through all day; they do not break down, they travel to the far end of the plant, and they find the pump. A price difference between two quotations is almost always a scope difference first, and a bar spacing is exactly the kind of line that makes two plants look identical on paper and behave nothing alike.

The other half of that commitment is that screenings now have a destination and a bin at the lift point, which is the part of a plant nobody draws. A rack that is raked daily only works if there is somewhere to put what comes off it. Specifying the rack and leaving the operator with no bin is how a rack stops being raked.

Why I am not worried about the level swing

The obvious objection to running a plant on bursts in a building is what happens if the pump stops and the level keeps rising. On an office building the answer cannot be that sewage backs up into the ground floor.

It does not, because every plant we draw carries an emergency overflow off the last chamber, routed to the point of discharge. We do not back up into the building. That overflow is not a treatment path and it is not something anybody wants to use — it is untreated water leaving at the far end, which is a compliance event and has to be treated as one. But it means the failure mode of a stopped pump is a recorded discharge rather than a flooded plant room, and it means the level swing in a burst-operated chamber has somewhere to go that is not upwards into the building. That is the difference between a design that is tight and a design that is brittle, and it is worth checking on any plant you are being asked to buy: ask where the water goes if the pump does not start, and ask for it on the drawing.

What the burst train costs downstream, and what it buys

Bursts are not free. A chamber fed in slugs sees a higher instantaneous rate than its average, which is why the upflow rate through that plant's clarifier had to be worked out during a burst rather than on the daily average, and why getting the basis of that calculation wrong mattered. Carryover over a clarifier weir shows up as a solids failure that looks like bulking and is actually hydraulic. Anything dosed on flow has to be dosed against the burst, not the average. And on a small plant every disturbance is proportionally larger, so a slug that would be invisible in a large tank is a real event in a 20-cubic-metre one.

What it buys is that the plant has one pump in it, in a clean position, behind a screen, with one thing to re-specify in three years instead of three. On a building where nobody will hold a spare and the person who looks after the plant also looks after the generator and the lifts, that is not a small thing. A good operator can carry a plant, but he cannot carry a design that needs him to be lucky.

Where this stops being transferable

I would not generalise any of this, and there are several places it falls over.

If the incoming invert lets you put the screen chamber first — which is the ordinary case on a new building, and the reason we normally do not have this problem — then keep the pump in the balance tank and keep real equalisation. Nothing here is an argument against equalisation; it is an argument about what to do when the sequence of the chambers is decided by levels you did not set. If the building has a genuinely sharp peak and a process load behind it, losing equalisation is expensive and the answer is probably to find the levels for a proper balance tank somewhere else, even at the cost of a second pump and a standby. If the flow is large, bursts stop being a tolerable way to feed a train at all. And if the plant is a conversion of something that is already there with an oversized existing sump in it, you may already have equalisation without drawing any, which is worth checking before you draw anything new.

The one part I would carry anywhere is the order of the reasoning. The screen decided the pump, the pump decided the hydraulics, and the hydraulics decided which chamber was doing equalisation. Nobody asked for a plant without a balance tank. It fell out of a question about bar spacing, three decisions upstream. If you are reading a plant somebody has drawn for you and a feature you expected is missing, the useful question is not why it was left out — it is which earlier decision made it impossible, and whether that decision was the right one.

If you want the background on how we size and sequence a plant before any of this comes up, the wastewater treatment services page is the place to start, the design generator will give you chamber volumes for your own flow, and the price estimator will put a number against it. The limits those volumes are designed to meet are set out on the DENR discharge standards page, and which parameters apply to your industry is often the first thing worth settling. If a plant is already built and running, maintenance and operation and our operations support are where that sits, and if you are starting from nothing, getting a sewage treatment plant sets out the sequence. The plants themselves are on completed projects, and if you are in Cebu, wastewater treatment in Cebu is worth ten minutes. Where the design flow number itself should come from, I have written separately — the water bill has usually been measuring it for years, and what you declared on the permit is what you are held to.

If you have a plant on a drawing and you cannot tell where the equalisation actually happens, or what runs if the one pump stops, send it over and we will read it with you. Book a free 15-minute consultation.