Source & site assessment

Gauging what the stream actually carries

"About a bucket a second" is not a measurement, and it is the number the whole design rests on. A ram's delivered output is directly proportional to drive flow, so a 50% error in flow is a 50% error in the water you plan to drink. Gauging properly costs an afternoon.

Fall is easy — flow is not

The fall H is the vertical drop from the water surface at the intake to the ram's waste valve, and a level, a long clear hose, or a builder's level and staff will give it to within a few centimetres. Flow Q is the hard one: it changes day to day, it hides in weed and boulders, and it is almost always estimated by eye.

Bucket and stopwatch

The honest method, and the only one that needs nothing but a container and a phone. Divert the whole flow into a bucket, time how long it takes to fill a known volume, and repeat three or four times so a single splash cannot skew the answer. If the stream is too big to divert, it is too big for this method — move on.

Float and velocity-area

Choose a reach that is straight, uniform and free of weed or boulders, and mark two lines a few metres apart. Measure the channel's cross-section (width and depth at several points) to get an area A in m², then time a floating stick or an orange over the marked distance to get a surface velocity.

Surface water moves faster than the water below it, so a single float over-reads. Multiply the surface velocity by roughly 0.8–0.9 for a rough channel and nearer 0.9 for a smooth one: Q ≈ A · v_surface · k. This is why the method is quoted at ±20–30%: it is a good sanity check and a poor design number.

Weirs — the accurate option

A sharp-crested weir turns flow into a depth, and depth is easy to measure accurately. For a 90° V-notch (Thomson) weir, with the head H measured upstream of the notch and clear of the draw-down:

For larger flows a suppressed rectangular weir (no end contractions) follows the Francis form Q = 1.84 · L · H^3/2, again in m³/s with L the crest length in metres.

Either way, a weir only reads as accurately as it is built to the geometry the equation assumes. These are the requirements worth knowing before you spend a day on the structure:

• Measure the head upstream, at least 4h back from the crest and in still water. Reading the depth at the crest itself under-reads badly.
• Keep the notch sharp — 0.8–2 mm of plate in the V, with a chamfered downstream edge so the nappe cannot cling to it.
• Guarantee free fall. The water surface downstream must be at least 6 cm below the bottom of the V. A drowned or backing-up weir is not measuring anything.
• Ignore heads below about 6 cm. At small heads the nappe clings and measurement error dominates, so a ram designed on a 30 mm head is designed on noise.
• Keep it fully contracted — the approach channel wide relative to the head (roughly h/B ≤ 0.2) — or use the partially contracted coefficients instead.

Salt dilution

Where the channel is too rocky to dam, inject a known mass of salt (or a known volume of strong brine) into the flow and measure how much it is diluted by the time it reaches a downstream conductivity probe. It copes with turbulent, boulder-strewn streams where nothing else works, needs no structure, and is the standard method for upland catchments. It is worth hiring or borrowing a meter for a day if your stream fits that description.

Designing for the dry season

This is the single most common way a ram pump project fails, and it fails invisibly for a whole season before anyone notices. The ram is sized on a spring measurement, it works beautifully from October to April, and in August it clacks slowly, delivers a trickle, and eventually stops. Nothing is broken; the stream simply does not carry what it did in April.

The design case is the lowest flow, not the average

A ram has no throttle and no switch: it takes whatever arrives and stops when the drive flow drops below the point where it can slam the waste valve. Size for the low flow you expect to see in the worst month, not for the flow you happened to measure. In hydrology that number is often written Q95 — the flow exceeded 95% of the time — and it is the correct basis for a water supply that must not fail.

Finding it in the field, in order of preference:

• Measure again, and again. Four readings spread across a year beat any single estimate. If the project is not urgent, start now and gauge monthly.
• Read the evidence on the channel. Dry-season watermarks, the wetted width, the lower limit of algae and moss, and the position of the perennial pools all record where the flow sat at its lowest.
• Ask what happens in the worst year. Local knowledge is often excellent and almost never volunteered unless asked directly: "has this stream ever stopped?"
• Check for a local gauge. National hydrology agencies publish long flow records for many catchments; a nearby gauged catchment with similar geology can bracket your low flow.

Leave water in the channel

A ram is unusual in that its waste stream returns to the stream bed, so the net abstraction is only the delivered flow — which is typically a small fraction of what flows past. That makes a ram an easy case to justify environmentally, but it is not a licence to take the lot:

• Flows return, but not where they left. The waste water joins the channel below the intake, so the reach between intake and outfall loses the whole drive flow. Fish, invertebrates and downstream users care about that reach.
• Leave a residual flow — often called a compensation or environmental flow. A common starting point is to plan never to take more than a modest fraction of the low flow, and to keep the intake screen and a shallow channel submerged.
• Check the rules. In many jurisdictions any abstraction, however small, needs a permit or a landowner's agreement. Confirm before you dig, not after.

Floods, silt and the intake

You size the ram for the dry season, but you must build the intake for the wet one. Floods arrive with the energy to move boulders, silt up a forebay and rip an intake off its mounting. The design consequence is a split personality:

• The flow-carrying design (drive pipe diameter, ram size) is set by the low flow.
• The intake structure is set by the flood — robust, anchored, screened, with a settling forebay that can be cleaned out, and a bypass that lets flood water past rather than through.

Silt is the quiet killer: an intake that takes its water from just above the bed draws grit into the drive pipe, where it erodes valve seats and wears the delivery check valve. Draw from just below the surface, not from the bottom (see intake & forebay).

Next: turn those two numbers into a machine — see Design & Sizing, or check the arithmetic with the sizing calculator. Once water is delivered, the next question is where it goes: Storage & distribution.