Case Study: A Greenhouse Fertigation Injector Retrofit
A 1.2 ha greenhouse replaced a weekly eye-mixed drum with a metering pump driven by a flow pulse, holding 120 ppm from a 15 percent stock.
An eye-mixed stock drum fails on the days when nobody is watching. A 1.2 ha greenhouse dosing at 120 ppm can swing its delivered strength across a 64 ppm band inside one week. The drift shows up as scorched leaf margins on one bed and a pale, hungry crop on the next, not as an alarm. By the time the crop tells the story, the stock is already in the root zone.
A 1.2 ha multi-span greenhouse ran a fertigation program through eight irrigation zones. Each zone placed 1.5 cubic meters of water per day, two zones watered at once, and the header peak reached 10 cubic meters per hour. The daily volume of 12 cubic meters passed in 1.2 hours of pumping. Nutrients arrived from one 60 L drum of stock, mixed by eye each week and fed by gravity. This case covers the fertilizer injector for irrigation system retrofit that replaced that routine, plus the arithmetic behind it. Every figure here is the generic shape of this class, checked against published bands rather than a named plant record.
Case Snapshot
The figures below are the generic shape of this class of fertilizer injector for irrigation system retrofit, not a named plant record.
| Item | Value |
|---|---|
| Header peak | 10 m3/h with two zones running |
| Daily volume | 12 m3 in 1.2 h of pumping |
| Injection target | 120 ppm from a 15 percent stock |
| Stock demand | 8.0 L/h, 9.6 L/day, 67.2 L/week |
| Old practice | One 60 L drum, eye-mixed weekly, gravity fed near 8 L/h |
| New practice | Metering pump stroked by a flow-meter pulse at 80 percent |
The 12 m3 daily volume is the number every other figure hangs from, from the 8.0 L/h stock rate to the 67.2 L weekly draw that the old 60 L drum could not cover.
The Challenge
A 60 L drum is a weekly chore, not a dosing device. A grower mixes it by eye to somewhere between 10 and 18 percent, sets it on a stand, and lets gravity pull stock into the header through a fixed restriction. The header draw near 8 L/h is set by that restriction, not by the water flow. The dose therefore moves with whatever strength the drum happens to hold.
Three failure modes came out of that routine, and each one arrived on a different clock.
| Failure mode | Where it shows | Clock |
|---|---|---|
| Strength swing | Dose moves from 80 to 144 ppm around the 120 target | Within the week |
| Drum shortfall | The last day of each week fed below 120 ppm | End of each week |
| No flow tracking | A fixed draw cannot follow a changing zone count | Every shift |
Fertigation injection systems of this class all fail the same way when the stock rate is fixed and the water rate is not. The strength is a variable, the draw is a constant, and the water flow moves with the valve schedule. Two of those three have to be tied together, or the dose wanders.
The drum shortfall is quieter than the strength swing. The site needed 67.2 L of stock each week and the drum held 60 L. That 7.2 L gap did not show as a fault light. It showed as a crop fed light on the last day of every cycle, week after week.
Why Eyeball Mixing Drifts
The mixed strength is the whole problem. The delivered dose depends on three things at once: water flow, stock rate, and stock strength. An eye-mixed drum pins none of them. The grower pours by volume and judges by colour, and one batch reads near 10 percent while the next reads near 18 percent.
At a near fixed 8 L/h draw, that 10 to 18 percent band moves the delivered dose from 80 to 144 ppm around the 120 target. The published dose relationship carries the swing directly, because the strength sits in the numerator. Halve the strength and the dose halves. The stock rate did not change and the water did not change, yet the crop saw a different feed.
The band is wide enough to matter in the root zone. A bed fed at 80 ppm runs a third below target, while a bed fed at 144 ppm carries a fifth more than the crop needs. Neither extreme announces itself at the time. A grower watching the plants sees a healthy house until the weekly average drifts far enough to show.
A second drift compounds the strength swing. Gravity feed holds near 8 L/h while the header duty changes with the zone schedule. When two zones run, the flow is high and the 8 L/h is close to right. When one zone runs, the same 8 L/h is too much stock for the water present. A fixed draw only holds its target at one flow, and a greenhouse rarely sits at that one flow for long.
What the Duty Actually Is
The job is a metering duty, not a fill duty. A metering pump moves a precise volume of liquid in a specified time, and that duty is called metering. The site needed a device that would deliver a set volume of stock for every cubic meter of water, not a device that would pour stock at a set rate.
The dose relationship is published on this site, and it gives the rate in one line. Dose rate in L/h equals water flow in m3/h times ppm, divided by stock percent times 10.
Filling that form with the header peak of 10 m3/h, the 120 ppm target, and the 15 percent stock gives the required stock rate:
(10 x 120) / (15 x 10) = 1,200 / 150 = 8.0 L/h
That 8.0 L/h has to hold only while the header is pumping. The day moves 12 m3 in 1.2 hours, so the stock drawn per day is 8.0 x 1.2 = 9.6 L. Over seven days the weekly draw is 9.6 x 7 = 67.2 L. Two of those three numbers already break the old practice. The drum holds 60 L, and the weekly need is 67.2 L.
An irrigation injection system sized to a rate alone would still drift. The rate has to be tied to the flow, or the same 8.0 L/h overfeeds at low flow and underfeeds at high flow.
The Solution
The retrofit replaced the gravity feed with a metering pump stroked proportionally by a pulse signal from the flow meter. Every pulse stands for a fixed volume of water through the header. The pump takes one stroke per pulse, so the injected volume tracks the watered volume instead of standing still while the flow changes.
| Step | Device | Working setpoint |
|---|---|---|
| Reference flow | Pulsing flow meter on the header | One pulse per fixed volume |
| Injection | Metering pump, diaphragm head | 8.0 L/h at header peak |
| Stroke setting | 0 to 10 L/h head | 80 percent stroke |
| Mixing | Straight run to the bend or mixer | 10 pipe diameters |
| Verification | Metered stock draw per day | 9.6 L |
The 0 to 10 L/h head covers the duty with room to spare. A required rate of 8.0 L/h on a 0 to 10 L/h head is 8.0 / 10 = 80 percent of full stroke, so the pump runs with margin on either side of the target. The fertigation pump selection guide walks the head and signal choices this retrofit used, and the site’s own chemical injection and mixing nozzle guide carries the dose relationship behind the 8.0 L/h figure.
The meter signal is what closes the loop. A metering pump alone runs at whatever rate the stroke sets, and that is fine only if the flow never changes. A greenhouse flow changes every time a zone valve opens or closes. Tying the stroke to the meter makes the injected litres follow the watered litres, so the delivered strength holds at 120 ppm through the whole pump window.
A gravity drum and a metered head are the two ends of the same choice. The drum sets strength by hand and rate by restriction. The metered head sets rate by stroke and strength by the stock recipe. Only the metered head can hold a target while the flow moves.
The Engineering
Three pieces of arithmetic carry the design, and each one is checkable with a flow meter, a stock sight glass, and a scale.
Stock rate at the header. The published dose relationship gives the rate in one line:
(10 x 120) / (15 x 10) = 1,200 / 150 = 8.0 L/h
At the header peak, the pump has to move 8.0 L of stock every hour it runs.
Daily and weekly stock. The day moves 12 m3. At 10 m3/h that is 12 / 10 = 1.2 hours of pumping. The stock for the day is 8.0 x 1.2 = 9.6 L, and the week is 9.6 x 7 = 67.2 L.
Stroke setting. The installed head spans 0 to 10 L/h. A required 8.0 L/h sits at 8.0 / 10 = 0.80, or 80 percent stroke. That leaves 2.0 L/h of headroom at full stroke.
Active mass, route one. A 15 percent stock at a density of 1.0 kg/L carries 150 g of active per litre. The day draws 9.6 L, so the active metered in a day is 9.6 x 150 = 1,440 g = 1.44 kg.
Active mass, route two. The day waters 12 m3 at 120 ppm. In grams that is 12,000 L x 120 mg/L = 1,440,000 mg = 1,440 g = 1.44 kg.
Both routes land on 1.44 kg of active per day, and that agreement is the check that matters. It means the stock strength, the dose target, and the water volume all agree with each other. A retrofit that passes one route and not the other has a fault either in the arithmetic or in the meter factor.
Metering Pump Behaviour and Vapour Lock
The pump class brings its own behaviour, and two faults are worth naming before commissioning.
A metering pump moves a precise volume of liquid in a specified time. Piston and plunger types are positive displacement units that hold a practically constant averaged flow across a wide range of discharge pressure. That is why the 8.0 L/h rate holds even as the header pressure moves with the zone count. The stroke length or the piston speed sets the metering rate, which is why the setting lands at 80 percent rather than at the target.
| Fault | What happens | Prevention |
|---|---|---|
| Vapour lock | Gas bubbles in the head compress instead of moving liquid | Degas or prime the suction |
| Backflow through the head | Outlet pressure below inlet pressure opens both check valves | Rate a positive pressure differential check valve downstream |
| Suction starvation | Lost volume per stroke on a starved suction | Flooded suction and adequate line size |
Every pump head runs an inlet check valve and an outlet check valve, often ball type. Those valves stop liquid moving the wrong way, and they are also where two faults begin.
Vapour lock is the quiet one. Gas bubbles in the head cause the pump to compress the bubble instead of moving liquid. The pump can therefore stop metering while still stroking. An operator hears a running pump and reads no fault. Degassing or priming prevents it, and that is why a flooded suction matters more here than on a transfer pump.
The backflow path is the other trap. If outlet pressure sits below inlet pressure, both check valves open and liquid flows through the head uncontrolled. A correctly rated positive pressure differential check valve downstream prevents this. On a fertigation header the pump discharge can see a pressure drop when a valve closes, so the differential valve is not optional.
Diaphragm and peristaltic heads are the other common families. This retrofit uses a diaphragm head because the diaphragm keeps the stock out of the moving parts, which suits a fertilizer solution that carries dissolved salts. A peristaltic head suits lower pressures and easy tube changes, but it wears the tube against the roller and needs more attention on a duty that runs every day.
Where the Eductor Still Fits
A metering pump injects, but it does not mix. The stock still has to be made in the tank ahead of injection, and a motive driven eductor is a common way to do that without a second pump. An eductor entrains stock in a 1:1 to 4:1 band, litres of motive water per litre of stock, so it will draw and blend a drum from the header flow itself.
For this greenhouse the drum mixing stayed manual, but a larger site that pours a batch tank would find an eductor pays for itself by removing the stirring paddle. The eductor pump effect comes from the motive water jet inside the venturi, not from a rotating part, so there is no seal face to wear. Where a batch is made in one tank and then metered, the eductor selection guide covers the motive pressure the venturi needs.
The split of duties is simple. The eductor makes the stock, and the metered head delivers it. A site that tries to make the stock with the injector pump ends up with a tank that is never fully blended.
The Results
No named plant record appears below. These are working outcomes for this class of fertilizer injector for irrigation system retrofit, built from the closed arithmetic above and kept inside the published bands.
| Metric | Before (eye-mixed drum) | After (metered head) | Basis |
|---|---|---|---|
| Delivered strength | 80 to 144 ppm | 120 ppm held | Published dose relationship |
| Weekly stock for the site | 67.2 L needed | 67.2 L metered | 9.6 x 7 |
| Drum cover | 60 L holds 6.25 days | 9.6 L metered daily | 60 / 9.6 |
| Weekly shortfall | 7.2 L short | 0 L short | 67.2 minus 60 |
| Stroke setting | Near fixed 8 L/h draw | 80 percent stroke | 8.0 / 10 |
| Active per day | Drifts with the mixed strength | 1.44 kg | Metered and by ppm |
| Flow tracking | None | One pulse per fixed volume | Meter signal |
The weekly shortfall is the number that hid in plain sight. The site needed 67.2 L of stock each week and the drum held 60 L, so every week ran 7.2 L short. At 9.6 L a day the drum covers 60 / 9.6 = 6.25 days, and the last fraction of each week drew on an emptying drum. A crop on the last day of a seven day cycle was fed below target on a schedule, not by accident.
The strength band and the shortfall together explain the whole complaint. One week the crop took 80 ppm on a weak batch, and the next week it took 144 ppm on a strong one. Neither number came from a decision. Both came from an eye-mix and a fixed draw.
Reading the Numbers Against Published Ranges
None of the sizing sits on invention. Each figure rests on a published band or a published relationship.
The dose relationship behind the 8.0 L/h rate is the site’s own published form. The same page shows a 50 m3/h line at 2 ppm on a 10 percent stock needing 1.0 L/h, and running the same form here gives (50 x 2) / (10 x 10) = 100 / 100 = 1.0 L/h. The fertilizer injector for irrigation system duty in this greenhouse is the same calculation at a smaller header and a stronger dose.
The mixing rule is published as well. Keep 10 pipe diameters of straight run from the injection point to the bend or the mixer. That rule is why the retrofit kept the injection quill upstream of a long straight header. Where a static mixer is absent, the published example shows 10 m of 100 mm pipe at 50 m3/h giving 5.6 s of residence time, the entire mixing window. A fertigation header at this scale carries a shorter run, so the straight-run rule does more of the work than a mixer would.
The injection point matters because the dose is only as even as the blend. A worn injection orifice that is 0.2 mm oversize roughly doubled a 0.8 ppm reservoir dose to 1.5 ppm on a published case, which shows how far a small bore fault can move a small ppm target. The same wear logic applies to the stock line, so the stock orifice is checked against its size at each service.
On the sanitation side, this greenhouse ran no chlorination, but the historic dose is worth naming as a boundary. Continuous chlorination of a US water supply at Boonton Reservoir in 1908 ran at doses of 0.2 to 0.35 ppm. A fertilizer injector for irrigation system head often shares the same meter signal, the same straight-run rule, and the same check-valve logic as a chlorine injection point. A site that already meters hypochlorite can reuse the pulse signal for the stock pump.
The materials differ, though. Sodium hypochlorite solution density runs 1.093 g/mL at 5 percent concentration and 1.21 g/mL at 14 percent at 20 C, and the solutions are alkaline at pH 11 or higher. Solutions above 500 ppm available chlorine are corrosive to some metals and many thermoplastics including acetal resin, and hypochlorite attacks stainless steel and aluminium. Compatible materials include PE, HDPE, PP, PVC, PTFE, PVDF, Viton and titanium, with glass also safe. Household bleach is a 3 to 8 percent solution at manufacture, and strength falls in storage. Food contact surfaces may be sanitised at 200 ppm available chlorine or less without a rinse.
Two standards frame the pump and the injection practice. ISO 9906 covers pump performance testing, and ASAE EP734 and ANSI/ASABE F444 frame injection-system practice. Both are named here as public references, not as a site specification.
Why This Matters
The transferable asset is the closed arithmetic, not the pump model. Any fertilizer injector for irrigation system study runs on the same four numbers: the header dose rate, the daily stock draw, the weekly stock demand, and the active mass check. A retrofit that skips one of the four will surprise somebody at the crop check.
The eye-mixed drum failed two of the four on its own. Its strength was a variable, not a setpoint, and its 60 L volume fell short of the 67.2 L weekly need by 7.2 L. Neither fault needed a new injector to be seen, only the arithmetic. That is the point worth carrying forward: measure the four numbers before buying any hardware.
The pulse link is the second transferable lesson. A metering pump on a fixed stroke is a fixed-rate device, and a fixed rate cannot follow a flow that changes with every zone valve. Tie the stroke to the flow meter and the injected volume tracks the watered volume, which is the whole definition of proportional dosing.
A smaller retrofit can reuse the same logic. A row crop boom sprayer retrofit runs a proportional dose against a moving boom, and the agricultural spraying application pages cover the nozzle side of the same duty. Where the stock is made in one tank, the agricultural product range covers the injection and spray hardware.
The verification line matters most. A metered stock draw of 9.6 L a day is a checkable number, and a scale under the stock tank reads it directly. A dose that is set, not mixed, is a dose that a grower can audit.
The Final Call
BoreJet supplies nozzles and does the sizing arithmetic behind them: header dose rate, daily stock draw, weekly demand, and the active mass check for this class of retrofit.
Send the zone count, the daily water volume, and the stock strength through the contact page. The response comes back with the stock rate, the head and stroke setting, and a mixing rule that holds your target dose.
Your Duty May Differ
Send the Vessel Drawing and the Pump Curve.
We size the head class, the count and the operating point for your tank, not a catalogue guess. A coverage test on the actual vessel beats any estimate.