Media reclassification: the cheapest way to cut abrasive spend

Why a working mix drifts before anyone notices

Abrasive does not wear out in one direction. Every impact fractures a few grains, and each fracture adds another fragment to the population. Two things then happen at once: the mean particle size falls, and the size distribution spreads out. Neither is visible from the blast window. What the operator notices first is that the machine sounds different, the dust plume gets thicker, and the same cycle time no longer produces the same grey finish.

That is why abrasive spend is usually managed by accident. Fresh media is added in whatever quantity the last operator felt like adding, spent media is carried out to the skip on a rough visual judgement, and the loss shows up as a creeping cost line rather than as something anyone can point at. The media is not being wasted through carelessness so much as through the absence of a number.

The alternative is to treat the mix as a controlled inventory with a measurable state. Media reclassification, meaning putting the working mix through a sizing step that removes dust and broken grains and returns only the particles still sharp enough to cut, is the cheapest intervention available, because it recovers material you are already buying.

What a sieve analysis actually tells you

The starting point is a plain sieve analysis on a well blended sample. Stack the screens in descending order, fill a graduated cylinder to a fixed volume, shake for a couple of minutes, and read the volume retained on each screen. Multiplying each retained volume by two gives the percentage of the mix in that size band, which is enough to catch a system that is drifting.

The standards bodies agree on the screen series. ANSI B74.12, FEPA 43G and JIS R6001 all describe grit sizing on the same mesh basis, so a cut size quoted in mesh travels between suppliers without conversion. Steel shot is graded separately under SAE J444, and that document is worth reading if you buy metallic media, because it specifies the shot size designation and its tolerance rather than a generic grit number.

The thresholds that matter in a shop are plain:

• Oversize particles above roughly 30 mesh hit hard, consume paint, and leave a surface profile above what the coating specification can hide. Keep them out of a general cleaning mix.
• The 50 to 80 mesh band is what does the work when cleaning new steel and when stripping rust and old paint. Losing that fraction is what flattens the anchor pattern.
• Particles below 140 mesh are dust in practice. Once that exceeds a few percent in a recycled mix, the material is being carried into the dust collector instead of the workpiece.
• A rising fines rate over time is the earliest reliable indicator that your separator is no longer capturing the small particles.

If a single number has to survive contact with a shift supervisor, make it the fines percentage. One figure, a known sample volume, recorded weekly, tells you more about abrasive economics than a month of purchasing data.

Building classification into the recovery loop

Removal alone is half the job. Fines that are not captured stay in the circuit, so classification has to sit where the material is still moving predictably.

• Vibrating multi-deck screening handles the coarse end. It suits hard, regular media such as steel shot and aluminium oxide, and it is the practical choice when the contaminant is metallic swarf or broken castings.
• Gravity air wash separates by terminal velocity. Light dust and light debris are carried to the filter while sound grains drop into a hopper. The key advantage is low handling velocity, which means less secondary fracture caused by the recovery equipment itself.
• Magnetic separation goes before screening. It removes ferrous contamination from steel media and stops swarf from travelling back to the blast chamber, where it would become an embedded inclusion.
• Cartridge filtration on the dust side is what keeps the fines fraction from becoming a real loss. Sealed hoppers working under negative pressure also keep humidity and recirculated dust out of the working mix.

Storing grades separately matters as much as separating them. Media that share a hopper share a fate, and a batch kept dry and segregated still cuts evenly many cycles longer than one scooped into a damp corner.

Failure modes worth recognising

Dust is a performance problem before it is a safety problem. A dusty chamber means the fines are carrying the load, which means longer cycles to reach the same finish. The operator compensates by raising the pressure, which fractures more grains, which raises the fines rate again. That loop is self feeding, and it is the reason media spend climbs while nobody changed anything.

Compensating for dull media with more pressure is the same mechanism from the other side. Media fracture rate rises sharply with pressure, and a grade chosen for a low pressure application will not survive a high pressure schedule. Match the pressure to the grade the classifier is actually holding.

Moisture is a quiet killer. Damp media clump, and a clump fractures along its own edges rather than at the point of impact, while clumped fines defeat a well tuned air wash. Organic media such as walnut shell and corn cob are especially absorbent. Dry air costs less than recovered media.

Mixing media types makes separation unreliable, because different densities and different fracture behaviour do not sort cleanly on one stack of screens. When you change grade or type, purge the pit, the recovery line and the separator fully before charging.

There is a counter intuitive conclusion worth stating plainly. Fresh media is usually the expensive answer and tired media is usually the cheap one. A working mix that is properly classified holds its cutting edge far longer than an unclassified one, so the money saved is simply the media you never had to buy. Track the fines percentage, grade the mix on a fixed cycle, top up rather than replace, and the media line on the job sheet stops being the line nobody can explain.