Optical sorter payback math shifts as retrofit modules undercut new-line spend
Processors are increasingly buying a single AI-assisted ejection unit rather than rebuilding a sort line, cutting typical payback from around 40 months to under 20 on the same recovery gain.
The economics of optical sorting have changed less because the machines got better than because they got smaller. A growing share of installations are single-module retrofits dropped into an existing sort line, and the payback arithmetic on those looks nothing like the arithmetic on a full line rebuild.
RecyclerWorld reviewed installed-cost and recovery-gain figures shared by seven mid-sized processors. The pattern is consistent: retrofit projects delivered 60 to 75 percent of the recovery improvement of a full line replacement at roughly a quarter of the capital cost and a fraction of the downtime.
What the numbers look like
A full sort-line rebuild for a mid-sized plant typically runs into seven figures once structure, conveyors, air systems, controls and installation are counted, and it takes the line out for weeks. On a recovery gain that generates, say, $340,000 a year of incremental product value and avoided residue disposal, that is a payback in the high thirties of months before financing.
A single ejection module — one camera bank, one valve block, one control cabinet, installed over a long weekend into an existing belt — carries a fraction of that cost, needs no structural work if the belt geometry cooperates, and captures the largest single increment of the recovery gain because it is placed at the worst loss point in the line.
“The mistake is thinking you need to fix the whole line,” said Ana Kirchhoff, plant engineer at Quarrick Materials, which retrofitted an ejection module onto its container line last year. “We measured where the metal and the good polymer were leaving, put one machine there, and stopped talking about the rest of it. The rest of it was never the problem.”
Where the gains actually come from
Two things drive most of the improvement.
Placement. The recovery benefit is dominated by the single worst loss point, which is usually a residue line or a fines transfer that nobody had instrumented. Plants that measured before they bought consistently reported better outcomes than plants that bought a machine and then looked for somewhere to put it.
Classification quality on dirty material. Modern classification handles overlapping, wet and partially obscured items materially better than the previous generation, which is what makes residue-line placement viable at all. Older sensor packages effectively required a clean, singulated presentation to work.
The cautions
Three failure modes came up repeatedly.
Compressed air is the hidden cost. An ejection module can add meaningfully to plant air demand, and several operators found they needed compressor capacity they had not budgeted. One plant discovered its existing system could not sustain the duty cycle and spent nearly as much on air as on the sorter.
Belt speed and burden depth have to be right. A module fed an uneven, deep burden will underperform its specification regardless of what the classification can do, and fixing belt loading is a mechanical problem, not a software one.
Maintenance capability matters. Valve banks, optics cleaning and calibration are ongoing obligations. Plants without a maintenance culture see performance decay within a year and then blame the machine.
“It is a piece of process equipment, not an appliance. If nobody owns it, it will quietly stop earning.”
For plants that get the placement right and can keep the thing clean, the retrofit route has moved optical sorting from a strategic capital decision to an operational one.
