Build log

Assembly as it happened, including what did not go to plan. No scale was used: volumes were measured with 200 ml jars.

Column body

PVC pipe, 3 in nominal, cut to 750 mm. The pipe is opaque, so each layer is photographed from above as it goes in rather than through the wall.

With the cap on, the depth from the cap floor to the rim is 770 mm. Layer surfaces are checked with a tape measure from the rim. The column is taped to a mop handle leaning against a wall.

White PVC pipe with a tape measure along it reading 75 cm — Cut to 750 mm.

Outlet

A bore through a 3 in PVC end cap carries a length of clear ½ in hose into a ½ in ball valve. The cap slides onto the pipe over PTFE tape rather than being cemented, so the column can be taken apart.

The first seal was hot-melt glue, which does not hold on wet PVC. It was stripped out and replaced with two-part epoxy.

PVC end cap with clear hose passing through its centre into a ball valve — End cap, hose and ball valve, with the first hot-melt seal.

Leak test

The capped outlet held water for two hours, then the cap on the full pipe held too. Not a drop at either joint, and the valve flows freely.

Layer separators

Five discs cut from plastic insect mesh: one under the support layer and one between each pair of layers above it.

Five roughly circular discs of dark insect mesh — Five mesh discs.

Charcoal

Lump charcoal, crushed by hand inside nested bags. The fraction held on the insect mesh was kept and the powder discarded.

Stage	Volume
Crushed, dry	≈ 500 ml
After topping up	≈ 550–600 ml

Rinsed and soaked overnight so it sinks instead of floating up through the layer above. Some pieces would not crush below gravel size, so it is loaded as a graded layer: coarse at the bottom, fine on top.

Tub of wet, washed charcoal granules — Washed and soaked.

Scoria

Sorted with the insect mesh and by hand into two fractions.

Fraction	Volume	Destination
Coarser	≈ 300 ml	Layer 1, support
Finer	≈ 500 ml	Layer 4, coarse top layer

Fine sand

Hand crushing stopped at gravel size, so neither scoria fraction is fine enough for layer 3. A fine carbonate sand already on hand was tried first and set aside: it kept releasing fines into the rinse water, and there was not enough of it.

Layer 3 uses aquarium-grade fine silica sand instead. Quartz does not break down into fines the way carbonate does, and it leaves water chemistry unchanged. A 3 kg bag covers the ≈ 1.3 kg the layer needs and leaves enough to reload it.

Rinsed in a tub until the water ran clear.

Tub of rinsed fine silica sand under clear water — Rinsed, before loading.

Loaded column

Loaded bottom up per RWL-001, each layer placed into standing water and checked with a tape measure from the rim.

Four views down the pipe as each layer went in: dark support scoria, black charcoal, pale silica sand and red scoria on top — Left to right: support scoria, charcoal, silica sand, coarse scoria.

Loaded white PVC column taped to a mop handle, with the ball valve hanging free below — Loaded and taped to the mop handle.

Still to do

Build the photometric rig and shoot the calibration series