A Clean Paint Filter Proves Less Than You Think

Three loads left the site on a Tuesday with a paint filter result of zero millilitres on every one. On Thursday all three were back, and the tipping hall supervisor sent a photograph: a pool of dark liquid spreading from the base of the pile, with the material sitting in it like a sponge on a wet floor. The test had not lied. It had answered a narrower question than the one the site thought it was asking.

Understanding what that test does and does not cover is worth more than any amount of arguing at the gate.

What a Drainage Test Measures

A paint filter or filter cone test measures gravity drainage from a sample over a fixed period. Water that will fall out of the material under its own weight, through that particular mesh, in that particular time, at that particular temperature. That is all.

It does not measure what happens when 25 tonnes of the same material sits on top of it in a trailer for two days. It does not measure what happens when the sun raises the load to 45 °C. It does not measure what happens when the material is compressed by a compactor blade, or when it is dropped two metres from a tipper. And it does not measure the unhydrated polymer in the middle of a lump, because gravity will not make that polymer release anything.

The test still earns its place — it is fast, cheap, repeatable and it catches gross under-dosing — but it is a screening test, not a guarantee.

The Three Failures It Misses

Syneresis under load. A gel network that is only just holding its water will contract when pressed and push water out. A sample in a cone is not pressed. A sample under 25 tonnes is. This is the most common cause of a load that passed and then wept.

Thermal separation. Absorption falls as temperature rises, and a gel near its capacity at 18 °C is over its capacity at 45 °C. Water that was bound in the morning is free by the afternoon. Sites in hot climates see this every summer, and it is usually diagnosed as a bad batch when it is a bad sample temperature.

Free liquid from unhydrated powder. A batch with 10 percent unhydrated polymer has 10 percent less absorbent than the recipe assumed, so the remaining water is under-bound. The cone may still pass because the dry lumps act as a blotter for the first hour. Two days later they are hydrated and there is nothing left to blot.

Tests That Cover the Gap

You do not need a laboratory. You need a heavier version of the same idea.

• Load test: 2 kilograms of gel in a 150 millimetre cylinder, 5 kilograms of weight on top for 24 hours, collect and measure what comes out. Cheap, brutal and much closer to a trailer.
• Warm hold: seal 2 litres in a jar, keep it at 40 °C for 48 hours, invert once, and look for a water layer. If there is one, the batch is marginal even if the cone passed.
• Core sample: cut a gel lump in half and look at the centre. Dry powder, glossy unhydrated patches or a colour difference means the batch is not homogeneous.
• Time-lapse photograph of a retained sample at 0, 24 and 72 hours. Three pictures on the batch file settle most arguments.

None of these replace the test your regulator accepts. They sit alongside it and stop you relying on it for a question it was never designed to answer.

The Sampling Error That Makes the Whole Thing Moot

A cone test on a bad sample is worthless, and most samples are bad in the same way: they come from the top of the mass. The top is where the paddle throws the driest material and where evaporation does its work. A sample taken there can pass a test that a sample from 300 millimetres down would fail.

Take the sample from the discharge stream across the whole discharge, or from the middle of the batch with a core. If you can only get one sample, take the one from the middle, because that is the material that ends up in the middle of the trailer where it stays wettest longest.

What to Say at the Gate

When a receiving facility rejects a load, the productive conversation is about which test they ran and on what sample, not about whether your test was right. Ask for the method, ask where the sample came from, and ask what the holding time was. In a surprising number of disputes the two sites are running different tests and both are correct.

Then run their test on your material before you send the next load. It costs an afternoon and it aligns the two operations. The acceptance testing workflow sets out a sample plan that covers gravity drainage, load release and thermal hold together.

One practical note: keep the retained sample from every load for as long as the acceptance arrangement requires, and store it somewhere that does not freeze or bake.