Optical Sorter Buying Guide | AISORT
Buyer's Guide
How to Evaluate and Select an Optical Sorter for Recycling
Selecting an optical sorter is one of the highest-stakes decisions in a recycling facility investment. The right machine can pay for itself within 12-18 months through improved purity premiums, higher throughput, and reduced labor. The wrong machine — or the right machine in the wrong position — becomes an expensive bottleneck.
This guide walks through the key technical, operational, and commercial factors to evaluate when comparing optical sorting systems, with practical criteria you can use regardless of manufacturer.
Step 1: Define the Sorting Objective
Before comparing equipment specifications, clarify exactly what the sorter needs to accomplish:
- Positive sorting (recover target material): Extract valuable material from a mixed waste stream. Example: recovering PET bottles from mixed rigid plastic. Key metric: recovery rate (% of target material correctly ejected into the accept fraction).
- Negative sorting (remove contaminants): Eject specific contaminants from a primarily clean stream. Example: removing PVC flakes from rPET flake. Key metric: contaminant removal efficiency (ppm of contaminants remaining in accept).
- Quality upgrade (both): Simultaneously recover target material while ejecting contaminants. Requires balancing recovery vs. purity — two metrics that trade off against each other.
The sorting objective determines everything downstream: sensor selection, valve pitch, chute width, and whether you need single-pass or multi-pass configuration.
Step 2: Understand Your Feedstock
The most common reason sorters underperform is that the buyer did not fully characterize their feedstock before specifying equipment. Gather data on:
| Feedstock Parameter | Why It Matters | How to Measure |
|---|---|---|
| Particle size distribution (min, max, D50, D90) | Determines chute width, valve spacing, and sensor resolution requirements. Particles smaller than valve pitch are ejected imprecisely. | Sieve analysis (dry or wet depending on material) |
| Material composition (% of each target and contaminant type) | The baseline composition determines how many sorting stages are needed to reach purity targets. | Hand-sort audit of 50-100kg representative sample |
| Moisture content | Wet material clumps, sticks to chute surfaces, and can cause false sensor readings (especially for NIR). | Moisture balance; target <5% for dry sorting, >95% for wet/wash sorting |
| Bulk density | Affects throughput calculation — a sorter rated for 5 t/h of PET flake handles a very different volume of film or foam. | Weigh a known-volume container of feedstock |
| Presence of fines (<2mm fraction) | Fines coat sensors, clog valves, and create dust that interferes with optical detection. | Sieve analysis; consider pre-screening if fines >5% |
| Surface condition (clean, coated, wet, oxidized) | NIR and RGB sensors rely on surface reflectance. Coatings, labels, dirt, and oxidation can change the spectral signature enough to cause misclassification. | Visual inspection + trial sort on a lab-scale unit |
Step 3: Match Sensor Technology to Your Material
| Sensor Type | Best Applications | Not Suitable For | Approximate Cost |
|---|---|---|---|
| RGB Camera (Visible) | Color-based sorting of rigid plastics, glass cullet, e-waste, construction debris | Material of same color but different composition (e.g., clear PET vs clear PVC) | $ — baseline |
| NIR (Near-Infrared) | Polymer identification (PET/HDPE/PP/PVC/PS), paper/cardboard sorting, textile fiber identification | Black or very dark materials (absorb NIR); metals; wet materials with surface water film | $$ |
| Hyperspectral / SWIR | Dark plastic discrimination, food-grade purification, distinguishing similar polymers (e.g., HDPE vs LDPE) | Metal-only streams; applications where RGB+NIR is sufficient | $$$ |
| X-Ray Transmission (XRT) | Heavy metal separation, removing aluminum from heavy fraction, mineral/ore sorting | Light materials (plastics, paper); organic materials | $$$ |
| Eddy Current / Induction | Metal detection in flake and granulate streams; copper vs aluminum separation | Non-metallic materials; very fine particles (<2mm) | $ — usually combined with optical |
| 3D / Laser Triangulation | Shape-based sorting (e.g., wire vs. granulate, 3D vs 2D objects); thickness measurement | Fine powders; materials that don't need shape discrimination | $$ |
| AI / Deep Learning Camera | Complex objects with variable appearance; brand-specific packaging identification; material-in-context recognition | Simple color-only sorting tasks; applications where training data is unavailable | $$ — software premium over camera hardware |
Step 4: Evaluate Throughput vs. Purity Trade-offs
For any given sorter, higher throughput reduces purity because each particle spends less time in the detection zone and the ejection system has less time to respond. The relationship is approximately:
- 80% of rated capacity: Optimal purity and recovery; ejection system has comfortable response time.
- 100% of rated capacity: Design-point performance; meets spec but with reduced margin for feedstock variability.
- 120% of rated capacity: Purity drops 1-3 percentage points; recovery drops 2-5 percentage points. Acceptable only for non-critical applications.
- 150%+ of rated capacity: Significant performance degradation; material bridges across the chute, sensors miss particles, valves cannot cycle fast enough.
Practical rule: Size your sorter for 70-80% of rated capacity at your design throughput. This provides headroom for feedstock variability and peak-period surges without compromising sort quality.
Step 5: Calculate Total Cost of Ownership
Purchase price is typically 40-50% of 5-year TCO. Factor in:
| Cost Category | % of 5-Year TCO | What to Ask the Vendor |
|---|---|---|
| Capital equipment | 40-50% | What's included? Chute, sensors, ejection system, control cabinet, mounting frame, installation supervision? |
| Installation and commissioning | 5-10% | How many days on-site? Is training included? Is the first-week performance guarantee in writing? |
| Spare parts and consumables | 15-25% | What is the expected life of valves, belts, lamps, and sensor windows? What's the cost and lead time for each? |
| Energy consumption | 3-5% | What's the connected load and typical running load? Is compressed air required (and at what pressure/quality)? |
| Preventive maintenance labor | 10-15% | How many hours per week of operator attention? What skills are required? Is remote monitoring included? |
| Downtime cost (risk) | 5-10% | What is the demonstrated MTBF? What's the guaranteed response time for on-site support? |
Step 6: Validate with Real Material
Never buy a sorter without running your actual feedstock through it. A lab-scale or pilot-scale test with 200-500kg of representative material will reveal:
- Whether the sensor can actually distinguish your specific materials (lab NIR spectra from clean reference samples are not the same as real-world material with coatings, aging, and contamination)
- How throughput affects purity on YOUR material (manufacturer spec sheets use clean, ideal feedstock)
- How much false ejection your material generates (false ejects go to the wrong fraction and either cost you revenue or require re-processing)
Ask the vendor to provide the test data, including per-frame images of eject decisions so you can independently verify what was correctly sorted vs. misclassified.
Red Flags When Evaluating Sorters
- Specifications without context: "99% purity" is meaningless without specifying the feedstock composition and throughput at which it was achieved.
- Refusal to test your material: Any reputable manufacturer will test your material. If they won't, walk away.
- Proprietary consumables with single-source supply: Valves, lamps, and sensor windows that can only be sourced from the manufacturer create a long-term cost trap.
- Closed control system: If you can't adjust sorting parameters yourself without calling the vendor, you'll lose days of uptime over the machine's life.
- No local service presence: For production-critical sorters, same-day or next-day on-site support matters. Ask for the locations and response times of the nearest service engineers.