Dosing Pump Problems: The 6 Failure Modes, and the One Cause Behind Them

TL;DR: Most dosing pump problems come from the same architectural fact: a dosing pump has to generate pressure to push chemical into a line that is already pressurised. Everything that wears, fouls, perishes or loses prime exists to serve that one job. The six failure modes engineers call us about are loss of prime, vapour lock, pulsation, fouled check valves, perished peristaltic tubes and calibration drift, and four of the six are silent, meaning the pump keeps running and the panel keeps saying it is fine while nothing useful reaches the water. A dosing valve removes the pressure-generating job entirely: it meters chemical into the line pressure you already have. That deletes the prime, the tube and most of the wear parts. It is not a universal answer, and there is one real constraint, covered below.

Last updated: 17 July 2026

Key takeaways

• Four of the six most common dosing pump problems are silent: the pump runs, the panel is green, and the dose is wrong.
• Pulsation is not a cosmetic issue. A pulsed dose makes the residual swing, so a spot sample can read in spec while the water rarely is.
• Loss of prime and vapour lock are architectural, not brand problems. A pump with a suction lift and a gassing chemical will do this.
• A dosing valve meters into existing line pressure instead of creating pressure, which removes the prime, the tube and most wear parts.
• The honest constraint: a dosing valve needs a pressurised line. Dosing to an open tank is still a pump job.

The root cause behind most dosing pump problems

Before the individual faults, it is worth naming what nearly all dosing pump problems share. A dosing pump has to do two jobs at once: measure a small volume of chemical, and generate enough pressure to push that volume into a line that is already under pressure.

The second job is the hard one. Generating pressure needs a moving diaphragm or a squeezed tube, check valves to stop backflow, and a suction path that must stay full of liquid. Every one of those is a part that wears, fouls, perishes or lets go. The measuring job is comparatively easy. Almost every dosing pump problem in this article is the pressure-generating job failing and taking the measuring job down with it.

Keep that in mind, because it explains why swapping one pump for a newer pump usually buys you time rather than a fix.

1. Loss of prime

The classic. Air gets into the suction side, the pump can no longer draw liquid, and the stroke moves nothing but gas. The motor still turns. The stroke counter still counts. The panel still reports a running pump.

It happens when a drum runs low or empty, when a foot valve leaks back overnight, when a suction line has a high point that collects air, or after any maintenance that opens the suction. On a site where the drum is changed by whoever is on shift, it happens regularly.

The reason it hurts is not the fault itself, it is the detection lag. Unless you have a flow switch or a residual sensor wired to alarm, you find out at the next manual reading, and everything between the last good reading and that moment was undosed.

2. Vapour lock

A close relative of lost prime, but the gas comes from the chemical rather than the air. Sodium hypochlorite is the usual culprit: it off-gasses, especially in warm plant rooms and as it ages, and that gas collects at the highest point in the pump head. Once there is a gas pocket in the head, the stroke compresses gas instead of displacing liquid, and dosing stalls.

Hydrogen peroxide behaves similarly as it decomposes. Both are chemicals people dose continuously and rely on, which is exactly why the failure is costly.

You can fight it with degassing heads, bleed valves and careful pipework. You are managing a symptom of the architecture: any pump with a suction side and a gassing chemical will keep trying to do this. We cover this specific problem on the sodium hypochlorite dosing valve page.

3. Pulsation and residual swing

This one is not a fault. It is how a dosing pump works, and it is still a problem.

A dosing pump delivers chemical in discrete shots. Between shots, nothing is dosed. Downstream, that produces a residual that rises after each shot and falls until the next. The average can sit exactly on target while the actual concentration spends most of its time above or below it.

Now take a spot sample. Depending on when you sample relative to the last shot, you can record a perfect result on a system that is swinging widely, or an alarming result on a system that is broadly fine. Neither reading tells you much, and both go in the logbook with equal authority.

Pulsation dampers exist for this reason. They help. They are another component, on another maintenance schedule, correcting a side effect of the pump being there at all.

4. Fouled check valves

Check valves keep flow going one way. They are small, they have a seat and a ball or a poppet, and they sit directly in the chemical.

Anything that crystallises, precipitates or carries solids will eventually stop one seating properly. Caustic soda crystallises at cold plant-room temperatures. Hard water carries scale. Polymer is viscous and clings. Once a check valve does not fully seat, some of each stroke goes backwards, and the delivered dose is quietly lower than the dial says.

Note the failure signature: not a stop, a drift. There is no alarm for a valve that mostly seats.

5. The perished peristaltic tube

A peristaltic pump squeezes a flexible tube with rollers. It is elegant, the chemical only touches the tube, and there are no check valves to foul. The tube is also a consumable being deliberately destroyed by the mechanism that drives it.

As it hardens it delivers less per revolution, so the dose drifts down. When it finally splits, it releases chemical inside the pump head, which is a maintenance job and, with an aggressive chemical, a safety event. Tube changes are a recurring planned cost and a recurring unplanned one.

6. Calibration drift

A dosing pump is calibrated by setting stroke length and frequency, then verifying against a calibration pot. That calibration is true for the conditions on the day: that back pressure, that chemical strength, that temperature.

Change the back pressure and the delivered volume per stroke changes. Let the chemical age and the active strength changes. Nudge the stroke knob while cleaning and nobody records it. Each is small. Together, months later, the dose is not what the record says, and the record is what an auditor reads.

Why silent failure is the expensive part

Group the six dosing pump problems by how they announce themselves and the real issue appears. A split tube announces itself. The other five mostly do not. Lost prime, vapour lock, a fouled check valve and calibration drift all present as a pump that is running normally.

So the cost is rarely the part. It is the window: the time between the dose going wrong and someone noticing. On a cooling system, that window is biocide control you thought you had, which matters directly to the Legionella regime described in the HSE Legionnaires’ disease guidance. On a treatment works it is a residual at the far end of the network. On a CIP circuit it is a clean you cannot evidence. In every case the chemical cost is trivial next to the consequence.

This is also why monitoring matters as much as mechanism. If dosing is going to fail silently, something has to be watching that is not the pump’s own opinion of itself.

The alternative: meter the pressure you already have

Go back to the root cause. If most dosing pump problems come from the pump generating pressure, the structural fix is to stop asking it to.

In most plant rooms the line you are dosing into is already pressurised. A liquid dosing valve uses that. It meters chemical into the pressurised line rather than pushing against it, so the pressure-generating machinery is simply absent. No prime to lose, because there is no suction stroke. No tube to perish. No pulsation, because dosing is not delivered in strokes.

The MAC BV214 and BV221 cores AQUAIOT supplies use a balanced poppet, which equalises upstream pressure variation so the volume per pulse stays consistent when supply pressure moves. They are solenoid actuated under PWM or proportional control, so the dose can track a flow signal instead of running at a fixed rate. Body and elastomer options are selected per duty; that selection is an engineering conversation about your chemical, concentration and temperature, not something to read off a web page.

The silent-failure problem still needs answering, and mechanism alone does not answer it. AQUAIOT logs every dose event with a timestamp, the volume and the sensor context around it, so the record is generated as the system runs rather than reconstructed later. That turns a dose from something you assume into something you can export. The full range by application and chemical sits on the chemical dosing valve pillar, and if you want the mechanism explained from first principles, start with what is a liquid dosing valve.

When a pump is still the right answer

A dosing valve is not a universal replacement, and it is worth being straight about where it does not apply.

A dosing valve needs a pressurised line to dose into. That is the constraint. If you are dosing into an open tank, a vented sump or an atmospheric channel with no line pressure to meter, there is nothing for the valve to work against, and a pump is the correct tool. The same applies to high-volume transfer duty, which is a pumping job rather than a dosing job.

Where the line is already pressurised, which covers most cooling circuits, treatment mains, CIP sets, boiler feedwater and closed heating loops, the valve route removes an entire category of failure. That is the question worth asking about your own site, and it is a short conversation. Details of the flow and pressure envelope are on the MAC BV214 and BV221 dosing valve page.

Frequently asked questions

What causes most dosing pump problems?

Most dosing pump problems trace back to the pump having to generate pressure to inject chemical into an already pressurised line. That job requires a diaphragm or squeezed tube, check valves and a primed suction path, and each of those wears, fouls, perishes or admits air. Loss of prime, vapour lock, fouled check valves, perished tubes and calibration drift are all consequences of that architecture rather than faults of a particular brand.

Why does my dosing pump keep losing prime?

Air is entering the suction side, or gas is coming out of the chemical itself. Common causes are a drum running low, a foot valve leaking back overnight, a high point in the suction line collecting air, or maintenance that opened the suction. With hypochlorite or peroxide the gas is generated by the chemical as it off-gasses or decomposes, collecting in the pump head. The pump keeps running throughout, which is why it is usually found at the next reading rather than at the moment it happens.

Does pulsation actually matter if the average dose is right?

Yes. The average being on target is not the same as the water being on target. A pulsed dose makes the downstream residual rise and fall between shots, so at any given moment the concentration is usually above or below the setpoint rather than at it. It also makes spot sampling unreliable, because the result depends on when you sampled relative to the last shot. A non-pulsed dose removes that swing and makes the sample mean something.

Can a dosing valve replace our dosing pump?

In most cases yes, provided the line you dose into is pressurised. The valve meters into that existing pressure, so it needs no prime, no tube and no pulsation damper. Where there is no line pressure to dose into, such as an open tank or vented sump, a pump remains the right choice. The practical test is your line pressure, chemical and duty, which our engineers will go through with you.

How do we catch dosing failures earlier?

Stop relying on the pump to report on itself. Because most dosing failures are silent, the detection has to come from something independent: a residual or flow sensor watching the outcome, thresholds that alarm when the reading drifts, and a dose record produced automatically rather than written up afterwards. AQUAIOT logs each dose event with timestamp, volume and sensor context so a drift is visible while it is still a drift.

Talk to an engineer about your dosing

If you recognise two or three of these dosing pump problems on your own plant, the useful next step is specific: tell us the chemical, the duty and the line pressure, and we will tell you honestly whether a dosing valve fits or whether you are better off keeping the pump. Speak to an expert.