Troubleshooting

Why Your Mass Finishing Numbers Are Wrong: 8 Costly Mistakes

The eight mistakes that most often throw off tumbling and deburring numbers, each with its symptom, root cause, and a numeric fix.

Symptom: parts come out nicked, lodged, or under-deburred despite a full cycle. The most common root cause is a media-to-part ratio below 3:1 by volume, which lets parts contact each other instead of media. For gears, threaded fittings, and anything with mating surfaces, run 4:1 to 6:1, and push to 8:1 for delicate stampings under 2 mm thick. Check it before you trust anything downstream, because the Media-To-Part Ratio calculator uses volume, not weight. Loading 40 lb of steel parts against 40 lb of plastic cones is not 1:1 by volume, it is closer to 1:3, and that mismatch is where impingement marks start.

Symptom: your Bowl Load Capacity says you can fit more, but throughput drops when you top it off. The root cause is confusing static fill with working fill. A vibratory bowl runs best at 55 to 70 percent of gross volume so the mass can roll. Fill past 85 percent and the toroidal action stalls, cycle time climbs 20 to 40 percent, and the bottom layer barely moves. If you rated the bowl at its brochure liters, subtract media displacement and the 15 to 30 percent freeboard first. A 300 L bowl is a 180 to 210 L working chamber, not 300.

Symptom: Tumbler Cycle Time estimates that never match the floor, always short by 15 to 30 minutes. Usually the input skipped the ramp and the water or compound flow rate. Cut times published on media datasheets assume steady-state amplitude of 3 to 5 mm and a set flow, often 2 to 4 L/min for through-flow. Run it dry or at half flow and cut rate can fall by half, so a 45 minute nominal cycle becomes 75. Log actual amplitude with a vibration card and confirm flow at the valve, then feed real numbers in rather than the spec sheet ideal.

Symptom: Compound Usage looks fine on paper but drums foam, parts tarnish, or cost creeps. The classic error is dosing by volume of parts instead of by water volume and flow. Liquid compounds are dosed at 2 to 5 percent concentration of the water in a recirculating system, or 20 to 60 mL per liter of make-up in flow-through. Mixing a batch-tank ratio into a flow-through machine overdoses by 5x or more. Track make-up water, not part count, and the Compound Usage number stabilizes. A common unit slip is reading mL/L as oz/gal, which is off by a factor of 7.8.

Symptom: Part Separation Yield below 95 percent, with good parts riding out in the media reclaim. Root cause is a screen deck sized to nominal part dimension, ignoring orientation and burr flash. A 12 mm part can pass a 10 mm slot on edge, and cones sized to part diameter let flat parts nest. Size separation gaps to 60 to 70 percent of the smallest part dimension and add a second deck. Also weigh reclaim: if 3 percent of part mass leaves with the media, your true yield is not the 99 percent the counter shows, and Part Separation Yield should be run on mass, not piece count.

Symptom: Media Wear Rate looks acceptable but finish quality drifts batch to batch. The missed variable is media shape breakdown, not just mass loss. Ceramic media loses 0.5 to 2 percent mass per hour, plastic 1 to 4 percent, but the corners round long before the mass number alarms. Rounded media cuts 30 to 50 percent slower and polishes instead of deburring. Measure edge radius or track cut time per batch, not only weigh-in versus weigh-out. Topping off worn media with fresh cones also shifts the size mix and changes lodging behavior, so re-check the Media Wear Rate and separation together.

Symptom: Unload Labor Time on the quote is half of what the floor actually books. The root cause is timing only the dump and ignoring dry, separate, inspect, and handling. A 200 lb batch might dump in 4 minutes but take 18 to 25 minutes end to end once drying and inspection are counted. If Finish Throughput is derived from cycle time alone, it overstates capacity by 25 to 50 percent because the machine sits idle during load and unload. Bracket the whole man-and-machine loop, then divide good parts by total elapsed time, not just spindle-on time.

Symptom: Deburr Cost Per Batch quoted low, then margin evaporates on reruns. The failure is treating rework as zero. If 4 percent of parts need a second pass, your effective cost is not the first-pass number, it is roughly 1.04 to 1.10x once you add re-handling and re-inspection. Feed a realistic scrap and rework fraction in, and use Rework Reduction Savings to price the fix rather than absorb it. A 1 percent drop in rework on a 50,000 part run at 12 cents per part is 60 dollars of pure recovery per run, which compounds fast across a year of production.

Published 2026-07-01.