Section 28 · Install & Run
Monitoring, records & handover
A ram has no meter, no controller and no alarm. "How is it doing?" is a question you have to answer by measuring — and an hour of setting up, plus a five-minute habit, turns a silent machine into one that tells you when it needs attention.
On this page
Measuring what it delivers
Commissioning gives you a number — the delivery measured at the tank must match the design estimate. Everything after that is a comparison against it, and both of the methods below need nothing more than a container and a clock.
Method 1 — time a container
The simplest measurement in water engineering. Close the outlet, let the delivery fill a bucket or drum of known volume, and time it.
A 10 L bucket that fills in 15 s is 10 ÷ 15 = 0.67 L/s. Measure twice — the first reading usually includes the air you purged. For very small deliveries, time three bucketfuls and divide by the total.
Method 2 — watch the tank
Where the delivery is too large to bucket, stop the draw-off, mark the level, and time how fast the tank rises. Multiply the fall or rise in level by the tank's plan area.
A 3 m × 3 m tank whose level falls 50 mm in 10 minutes is delivering 9 × 0.05 = 0.45 m³ in 600 s — 0.75 L/s. Watch rise with the outlet shut for a delivery figure, and fall overnight for a demand figure; mixing the two is the classic mistake.
Run your site's numbers through the sizing calculator and expect to see roughly 70–80% of the theoretical figure in the field — friction, tuning, entrained air and wear all take a cut (derating). Delivering less than about two-thirds of the estimate is not "a slow ram"; it is a fault worth chasing (troubleshooting).
The cheapest permanent instrument is a marked dipstick and a clock: a notch cut into a stick at the normal tank level, and a fixed time each week to read it. Where mains power or a small solar panel is available, a cheap flow or level sensor with a text alert does the same job remotely — useful for a scheme whose caretaker lives a valley away, but it measures the same two quantities.
The log and what it tells you
A ram fails slowly far more often than it fails suddenly. Worn valve rubbers, a silting intake and a chamber losing its air all show up first as a gentle decline in delivery — invisible day to day, obvious on a page of numbers. Five minutes a week is enough.
| What to write | How |
|---|---|
| Date and time | Fixed day and hour, so readings are comparable |
| Delivery (L/s or m³/day) | One of the two methods above — the same method every week |
| Beat rate (beats/min) | Count for 15 s and multiply by four |
| Source and tank level | Dipstick marks, or just "high / normal / low" |
| Weather and stream condition | Whether it is the dry season, or there has been a storm |
| Anything done | Screen cleaned, valve re-tuned, chamber recharged, spares used |
Read as trends, those columns are a diagnostic. The maintenance page carries the full symptom → cause → fix matrix; at the level of a log, four patterns cover most of it.
| The trend | Usually means | First thing to do |
|---|---|---|
| Steady decline over weeks or months | Worn valve facings, fouled screen, or a chamber slowly losing its air | Clean the screen first — it is free — then check the rubber and recharge the chamber |
| Sudden stop after a storm | Debris on the intake, or a scoured pit | Clear the screen; check the intake has not been undermined |
| Faster beat than it used to be | Less flow reaching the valve, or the weight has shifted | Check the source level, then the weight position |
| Delivery good but the line bangs | Waterlogged chamber | Recharge the air (the procedure), then look at the snifter |
The as-built record & handover
This is the cheapest intervention in the whole project and the one most often skipped. The ram will outlive the person who installed it, and its next caretaker cannot re-tune a pump whose settings nobody wrote down.
Write this down before you leave
Pair that sheet with the weekly log and you have something sturdier than either: the record says what the machine is, and the log says whether it is still doing it. Together they are the difference between a scheme that can be maintained by someone who has never seen it built, and one that depends on a single person's memory.
Both the site survey checklist and the record above print cleanly at 100% — the print stylesheet drops the chrome and keeps the page. A laminated sheet on the pump-house wall gets filled in; a file on someone's laptop does not.
With a number on the log and a record on the wall, the rest is routine: the maintenance schedule, and the seasonal work in cold climates.