The short answer: a ring main has no fixed flow direction, so treat the flow path as something to be measured
On a ring main, water is not obliged to travel past the polyphosphate dosing unit. A loop is two parallel paths between the same two points, and the proportion of flow taking each path changes with demand, valve position, pump status and even the time of day. If the dosing vessel sits on one leg of that loop, it may see full flow, half flow, reverse flow or almost no flow — and the same outlet may be fed by treated water in the morning and untreated water in the afternoon.
That is why polyphosphate treatment on ring mains so often disappoints: the media is working correctly, but only part of the loop is actually in contact with it. Identifying the real flow path is the first engineering task, before specifying media, dose or vessel size. This article sets out what drives flow in a loop, how to measure the true path, and what the result means for slow-dissolving polyphosphate media such as the crystals commonly marketed under names like Siliphos.
What polyphosphate treatment does — and what it does not do
Ring main failures are frequently misdiagnosed because several water-treatment functions are spoken about as if they were interchangeable. They are not, and the distinction affects how you interpret flow-path data.
| Function | What it actually does | What it does not do |
|---|---|---|
| Scale inhibition (threshold treatment) | Interferes with crystal growth so hardness salts stay dispersed rather than depositing on hot surfaces | Does not remove hardness from the water; total hardness is unchanged |
| Softening (ion exchange) | Removes calcium and magnesium ions, replacing them with sodium or potassium | Not achieved by polyphosphate media |
| Scale removal | Dissolving or breaking down existing deposits | Polyphosphate media are preventive, not a descaling agent |
| Filtration | Physical retention of suspended solids | Polyphosphate crystals are a dissolving medium, not a filter bed in the conventional sense |
| Corrosion inhibition | Some phosphate-based formulations form a thin protective film on metal surfaces | Performance is formulation- and water-chemistry-specific and must be verified for the exact product |
| Disinfection | Inactivation of microorganisms | No polyphosphate product disinfects water |
Slow-dissolving polyphosphate media work by releasing a small quantity of phosphate into the water as it passes through the bed. The release rate depends on water passing over the crystal surface, contact time, temperature, water chemistry and the surface area of the bed. If water is not flowing, the dose delivered to the loop is effectively zero — although crystals in a wetted vessel may continue to dissolve slowly, which can produce localised over-dosing once flow resumes.
How a ring main actually behaves
Pressure-driven loops
Where the loop is supplied directly from a pressurised main, the driving force is the incoming supply pressure. Water leaves the supply point and splits between the two legs of the loop in inverse proportion to the hydraulic resistance of each leg. When a draw-off on the “near” leg opens, the pressure at that point falls and most flow arrives by the shortest, least resistant route. A draw-off on the far leg reverses the balance. Flow in the leg containing the dosing unit can therefore reverse several times a day without any valve being touched.
Pumped recirculating loops
Where a circulation pump is fitted — common in hot water services and in process cooling circuits — the pump imposes a direction and a relatively stable flow rate. This is the easier case to design for, because full-flow contact can be achieved by placing the dosing point on the return leg ahead of the pump suction. Even here, though, branch take-offs and balancing valves can starve sections of the loop, and variable-speed pumps change the picture at part load.
Mixed loops, dead legs and low-turnover sections
Many real installations are hybrids: a pumped circulating loop with a mains-fed top-up, isolation valves that are nominally open, and branches that were added after commissioning. These are where the flow path departs most sharply from the drawing. Dead legs — capped or rarely used branches — may never receive treated water at all, and they are also where stagnation and biofilm risk concentrate. A flow-path survey should record them explicitly rather than treating them as a detail.
A practical method for identifying the actual flow path
- Redraw the loop as built. Walk the system and record supply points, the dosing vessel, pumps, non-return valves, isolation and balancing valves, draw-off points and dead legs. Mark anything on the drawing that does not match what is installed.
- Identify the driving force under each operating mode. Mains pressure, pump, gravity or a combination. Note which modes occur at night, at peak demand, and during maintenance. Flow path is a set of conditions, not a single answer.
- Measure rather than infer. Choose a method from the table below that suits the pipe material, access and whether the loop is hot or cold.
- Map residence time, not just direction. Inject a tracer at the dosing point and record the time until it appears at each sampling location. Residence time tells you whether the water reaching an outlet has passed through the media, and how long the contact was.
- Repeat at a second operating condition. A single measurement taken at midday with all valves open is not a flow-path survey. Test at low demand and high demand, and with the pump on and off if both are possible.
| Method | What it reveals | Practical limits |
|---|---|---|
| Clamp-on ultrasonic flow meter | Flow velocity and direction in the section being measured | Needs a suitable straight run and pipe access; gives a spot reading that may reverse later in the day |
| Differential temperature tracing (hot loops) | Direction of travel and relative flow between legs | Only works where there is a sustained heat source or draw-off; not applicable to cold loops |
| Tracer injection with timed sampling | Actual path, residence time and short-circuiting | Requires sampling points; an existing phosphate residual complicates conductivity-based tracers |
| Isolation and draw-off testing | Which sections are fed from which side of the loop | Disrupts supply and needs coordination with building users |
| Differential pressure across valves and sections | Relative balance between legs | Indirect; needs a baseline and stable conditions to be meaningful |
What the flow path tells you about dosing performance
Once the path is known, three design questions usually answer themselves.
- Is the dosing point on a full-flow route? If a significant share of the loop bypasses the vessel, the phosphate residual at the far end will be lower than at the near end, and scale protection will be uneven. Moving the dosing point to a position where all flow must pass is usually more effective than increasing the media quantity.
- What is the real contact time? Short contact times limit how much phosphate can be released in a single pass. This is a function of the loop’s residence time and the media’s dissolution behaviour — both must be confirmed against the media supplier’s data for the specific product rather than assumed.
- Are there stagnant or low-turnover sections? These will receive little or no treatment and may also cause ongoing slow dissolution from wetted media, producing locally high phosphate concentrations once flow resumes.
A note on reversion and hot loops
Polyphosphates can revert toward orthophosphate under certain combinations of temperature, time and water chemistry. Where reversion occurs, the scale-inhibiting behaviour changes and deposition risk at hot surfaces may increase. The extent to which this matters depends on the specific medium, the loop temperature profile and the residence time you have just measured. Reversion behaviour should be requested from the media supplier as product-specific data, not inferred from another product’s literature.
Common mistakes when dosing a ring main
- Installing the dosing vessel on a branch or on one leg of a loop and assuming the whole loop is treated.
- Sizing the media charge from total loop volume instead of from the flow that actually passes the bed.
- Commissioning on a single flow measurement taken under one set of conditions.
- Treating a phosphate residual reading at one outlet as representative of the whole system — the residual at the far end is the number that matters.
- Ignoring dead legs, which appear in no flow-path measurement and in no dosing calculation, but still affect water quality.
- Assuming a slow-dissolving polyphosphate medium softens water or removes existing scale. It does neither.
When the flow path cannot be resolved
Some loops are genuinely too complex, or too poorly documented, to characterise economically. In those cases the practical alternatives are: relocate the dosing point to a position of guaranteed full flow, such as the incoming supply or a pump suction; install a dedicated side-stream treatment loop with its own small pump so the media sees a controlled, predictable flow; or move to a flow-paced dosing arrangement where the dose responds to measured throughput. Each has cost and maintenance implications, and none removes the need to understand the loop’s hydraulics at least approximately.
Frequently asked questions
Does the position of the dosing unit matter on a ring main?
Yes, more than on a branched system. On a loop, water can reach the same outlet by two routes, so a dosing unit that is not on a full-flow route will treat only part of the circulating water. Placing it where all flow passes — typically on the return leg ahead of a circulating pump, or on the incoming supply — gives the most predictable contact.
Can I use a phosphate residual test to confirm the flow path?
Only partially. A residual reading confirms that treated water has reached a sampling point at some time, but it does not tell you the route, the residence time or whether the reading is stable. Use residual monitoring alongside a tracer or flow measurement, not instead of one.
Is Siliphos the same as every polyphosphate crystal?
No. “Siliphos” is a trade name associated with a family of slow-dissolving polyphosphate-silicate glass media, and products sold under similar names are not necessarily interchangeable in composition, dissolution rate or recommended application. Specifications should be taken from the datasheet of the exact product being supplied, not transferred from another source.
Is food-grade polyphosphate media suitable for drinking water?
Not automatically. Food-grade status describes a material’s suitability for contact with food in defined conditions; it is not the same as a verified drinking-water approval for a specific product in a specific market. Drinking-water suitability must be confirmed for the exact product, the application conditions and the destination market before use.
How often should a flow-path survey be repeated?
After any change to the loop: new branches, valve alterations, pump replacement, changes in occupancy pattern or a switch between pumped and mains-driven operation. Otherwise, a review at the same interval as the media service is usually reasonable.
Send the loop details with your enquiry
If you are specifying slow-dissolving polyphosphate media for a ring main, the flow path determines the specification — not the other way round. When you contact ENVO CHEMICAL, include the loop layout as built, the driving force (mains pressure, pump or both), the pipe material and diameter, the location you are considering for the dosing vessel, the source-water analysis including hardness, pH and temperature, the expected flow rate and residence time, the treatment objective (scale inhibition, corrosion film, or both), the destination market and the required quantity. With that information we can discuss which media grade and vessel arrangement is worth evaluating for your system, and which points still need to be confirmed by measurement on site.
