Summary: Two quotations for the same single screw granulator can land on different cost per tonne figures, and the difference usually sits in arithmetic that was never shown. This article builds the model instead of asserting it. Five buckets are defined and priced from published figures: energy at the 0.28 to 0.42 kWh/kg specific energy of the TL-SJ family, labour at your shift pattern, consumables at your screen interval, capital recovery on the published US$12,500 to 46,000 machine band, and the uptime that divides all of them. A worked example runs at 400 kg/h, then the same tonne is costed on three screws to show where scale pays. The closing section answers the commercial question directly: which cost lines make simplicity genuinely cheap, and the three where it quietly stops being cheap.

The formula, written out before any money
Cost per tonne is not a rate a supplier quotes. It is the sum of five lines, each one per tonne or per year, divided by the tonnes that reach the bag. Write it down in this order and every later argument is short.
1. Tonnes per year. Rated kg/h × scheduled hours per year × uptime, divided by 1,000. Uptime is the term people leave out, and it carries screen changes, die cleaning and unplanned stops with it.
2. Energy per tonne. Specific energy in kWh/kg × 1,000 × your tariff. The published TL-SJ figure is 0.28 to 0.42 kWh/kg, so the whole term is decided by two things you control: what you melt and what you pay per kWh.
3. Labour per tonne. Operators per shift × paid hours × loaded hourly cost, divided by tonnes per year. This line is nearly fixed, so it punishes a small machine and rewards a large one.
4. Consumables per tonne. Screen packs and cutter blades per year, divided by tonnes per year. Screen life is set by contamination, not by the machine.
5. Capital recovery per tonne. Machine value divided by years, divided by tonnes per year, plus a spare-parts and maintenance allowance. The published family band is US$12,500 to 46,000 FOB Qingdao, and the equipment side of that price is discussed in what a plastic granulator costs.
What the formula deliberately excludes
Freight, installation, local duties, financing, rent and plant overhead sit outside this model, and they belong there. They are the same for a single screw and a twin screw of the same output, so they do not change the choice. Any supplier who presents a cost per tonne without naming the exclusions is hiding the part of the number that varies between them.
Worked example: 400 kg/h of washed film on a TL-SJ-120
Take the mid-point of the TL-SJ-120 band, 400 kg/h, and fill the model with inputs. Everything in square brackets is yours to replace.
- Tonnes per year — 400 kg/h × 6,240 scheduled hours [20 h/day × 26 days × 12 months] × 90% uptime = 2,246 t/yr.
- Energy — 0.35 kWh/kg [mid of the published 0.28 to 0.42] × 1,000 × US$0.12/kWh [your tariff] = US$42.00/t.
- Labour — 1 operator × 6,240 h × US$6.00/h = US$37,440/yr ÷ 2,246 t = US$16.67/t.
- Consumables — screen pack every 8 running hours at US$28/pack, plus one blade set per 1,000 running hours at US$85 = US$20,133/yr ÷ 2,246 t = US$8.96/t.
- Capital — US$24,000 [firm price from your quotation] over 5 years = US$4,800/yr ÷ 2,246 t = US$2.14/t.
Total: US$69.77 per tonne. Notice what that number is not made of. Capital is 3% of it, consumables 13%, and energy and labour 84% — both dominated by two inputs you set before you buy anything, the tariff and the shift pattern.
Why the model is set on output and not on time
Per-tonne figures reward a machine that runs long, and a small screw running twenty hours a day can produce more tonnes a year than a large screw running one shift. That is why scheduled hours sit inside the first line rather than outside the model.
Three machines, one tonne: where scale actually pays
The same tonne is costed on the smallest, the middle and the largest screw, holding energy and consumables constant per tonne and letting machine value and output scale together. The machine values are inputs inside the published US$12,500 to 46,000 family band, and you replace them with your own quotation.
The result is not what most buyers expect. Capital recovery per tonne is almost flat across the three machines — US$3.26, US$2.14 and US$2.17 — because a bigger screw costs more money and makes more tonnes in roughly the same proportion. The whole scale effect sits in the labour line: US$38.09 a tonne on the 90 mm screw against US$10.67 on the 150 mm. Buying the larger machine is not buying a better melt; it is buying fewer operator hours per tonne.
That has a direct consequence for how the machine is run. A 90 mm screw on one operator is a two-operator-per-machine business unless the plant runs long hours, because the operator is there whether the screw is full or not. If your annual tonnage is small or your shifts are short, the labour line dominates your quote and the smaller screw may be right for a different reason than the price list suggests.
Where the published cost figures sit
The equipment side is published rather than estimated: the family band of US$12,500 to 46,000, the specific energy of 0.28 to 0.42 kWh/kg, and the 600 kg/h comparison at US$28,000 against US$68,000 for a twin screw line. Those economics are set out on single screw versus twin screw.
Where simple is cheap, and the three lines where it stops being cheap
The commercial argument for a single screw granulator is real, and it is worth stating precisely so that it can be tested rather than repeated.
- Capital per kg/h is the strongest line — the published 600 kg/h comparison puts a single screw line at US$28,000 against US$68,000 for a twin screw, and the twin-screw platform starts at US$38,000 on its own.
- Wear parts and spares are half the count — one screw, one barrel, one screen changer against a set of intermeshing elements, and a machine with fewer parts to hold and to reorder.
- Service and training are simpler — one screw and one barrel with bearings at the 3 to 5 year horizon, and 18 months of warranty from commissioning with 3-year spare-parts availability.
- Energy is not a strong line — 0.28 to 0.42 kWh/kg is already close to a twin screw, so do not sell the single screw on energy.
Three lines break the argument, and each breaks it for a specific material. First, mixing. A single screw melts; it does not disperse. Filler above 15%, glass fibre, flame retardant, or colour that must disperse below 20 µm need the twin screw, and the premium you earn on the better dispersion is the only thing that pays for it. Second, colour change. The published changeover figures are 2 to 3 hours on a single screw against 40 to 60 minutes on a twin screw, so a plant running many short colours loses more in purge material than it saves on capital. Third, throughput and dirt. The family tops out at 700 kg/h, so above that, or on dirty post-consumer feed, the double-stage platform that filters twice is the correct machine. That is where the tree in how to choose a plastic granulator ends up.
The one thing simplicity never buys back
None of these lines is settled by the machine, and that is the point worth stating to a sales team. The capital line is real but it is also the smallest line in most plants. Labour, energy and screen interval decide the total.
The two inputs that move the answer more than the machine does
Sensitivity analysis on this model is short, because only two inputs move it materially.
Uptime. At 80% instead of 90%, output falls from 2,246 t to 1,997 t, labour per tonne rises from US$16.67 to US$18.75 and capital recovery rises from US$2.14 to US$2.40. Consumables per tonne do not move, because the packs are replaced per running hour and the tonnes fall at the same rate. Total cost per tonne goes to US$72.11. A plant that loses ten points of uptime to screen changes and stoppages is paying more per tonne than a plant that buys a better screen pack.
Energy. At the bottom of the published band, 0.28 kWh/kg, the energy line is US$33.60 and the total is US$61.37. At the top, 0.42 kWh/kg, the energy line is US$50.40 and the total is US$78.17. A range of more than US$16 a tonne, which is more than the entire capital line.
Everything else is second order: the screen interval moves consumables in proportion, the shift pattern moves labour, and the machine price moves capital. Send the tariff, the loaded labour rate, the scheduled hours, the screen interval and the annual tonnage, and the model comes back with your numbers in it. The published figures are on the TL-SJ single screw granulator page, and the test protocol on how every line is factory tested.
A model that survives a change of tariff, shift pattern and machine price is one you can defend to a customer. If the total moves sharply when you change one of those inputs, the machine decision was never the binding constraint.
| Cost line | TL-SJ-90 at 175 kg/h | TL-SJ-120 at 400 kg/h | TL-SJ-150 at 625 kg/h | What moves it |
|---|---|---|---|---|
| Tonnes per year at 6,240 h and 90% uptime | 983 t | 2,246 t | 3,510 t | Scheduled hours and uptime |
| Energy at 0.35 kWh/kg and US$0.12/kWh | US$42.00 | US$42.00 | US$42.00 | Polymer, moisture and the screen pack |
| Labour, one operator at US$6.00/h | US$38.09 | US$16.67 | US$10.67 | Shift pattern, staffing and output rate |
| Consumables, packs and blades | US$8.96 | US$8.96 | US$8.96 | Contamination level and screen interval |
| Capital recovery over 5 years | US$3.26 | US$2.14 | US$2.17 | Firm price from your quotation |
| Total cost per tonne | US$92.31 | US$69.77 | US$63.80 | Everything above, taken together |
Frequently asked questions
Q1: How do I calculate the cost per tonne of pellets from a single screw granulator?
Add five lines and divide by the tonnes that reach the bag. Tonnes per year come from rated kg/h × scheduled hours × uptime; energy is specific energy × 1,000 × your tariff; labour is operators × hours × rate divided by tonnes; consumables and capital follow the same pattern. On the published TL-SJ specific energy of 0.28 to 0.42 kWh/kg the energy line is usually the largest single term.
Q2: What is the biggest cost item when granulating plastic on a single screw?
Energy and labour together, not the machine. In a worked TL-SJ-120 case at 400 kg/h, energy at 0.35 kWh/kg and US$0.12/kWh comes to US$42.00 a tonne and one operator at US$6.00 an hour comes to US$16.67. Capital recovery is under 3% of the total. That is the arithmetic behind buying the larger screw: you are buying fewer operator hours per tonne, not a better melt.
Q3: Why does the cost per tonne fall as the screw diameter goes up?
Because labour is spread over more tonnes, not because the machine is more efficient. Costing the same tonne on three published model bands gives US$92.31 on the 90 mm screw, US$69.77 on the 120 mm and US$63.80 on the 150 mm, and almost all of that difference is labour: US$38.09 a tonne against US$10.67. Capital recovery per tonne barely moves at all.
Q4: How much do screen packs and cutter blades add to the cost per tonne?
On a worked 400 kg/h case with a pack replaced every 8 running hours and one blade set per 1,000 running hours, consumables came to US$8.96 a tonne — about 13% of the total, and more than the capital line. The interval is set by contamination, not the machine, so blocking a 40 to 80 mesh pack every two hours roughly doubles this line.
Q5: When does a single screw stop being the cheaper machine to run?
On three conditions. When the recipe needs dispersion — filler above 15%, glass fibre, flame retardant, or colour below 20 µm — a single screw cannot do it at any price. When colour changes are frequent, the published 2 to 3 hour changeover against 40 to 60 minutes on a twin screw eats the capital saving. And above 700 kg/h or with dirty post-consumer feed, a double-stage line is the correct machine.
Engineer’s note
I would rather quote a cost per tonne from your own material than from my own spreadsheet, because the two inputs that decide it — moisture and screen interval — are only visible on a running line. The published 0.28 to 0.42 kWh/kg is the machine family; the number on your quotation comes from the 4-hour factory test run on your material, where output, melt pressure, melt temperature and kWh are logged every minute. A model built on that log is a measurement.
Engineering note: the numbers on this page are our own published reference ranges for the machine family, not a specification guarantee for any individual unit. Final output, specific energy and pellet quality are set by the incoming material, its moisture and contamination level, and the screen pack fitted. We write the confirmed figures onto the quotation after a factory test run on your own material.
Who wrote this
Zhou Gong (周工) is a senior plastic machinery engineer at TENGLONG Plastic Machinery in Linyi, Shandong. The factory was registered as Linyi Hedong District Tenglong Plastic Machinery Factory on 7 June 2016 and grew out of a screw and barrel workshop, so his work is rooted in the parts that decide whether a granulator holds output: screw geometry, nitriding case depth, barrel alignment and melt filtration. He prepares the technical specification for customer quotations, reviews the four-hour factory test data before shipment, and runs commissioning support on site. Screws are nitrided in-house to a 0.5–0.7 mm case at HV 900–1000 and straightened to 0.03 mm runout over the full length; barrel, screw, gearbox and frame are aligned to 0.05 mm before the test run.
Figures quoted here are engineering reference ranges published on our own product pages, not guaranteed values for any specific machine. Real output, specific energy and pellet quality depend on the incoming material, its moisture and contamination level, and the screen pack fitted. Final values are written after a factory test run on the customer’s own material, and the PLC-logged test report ships with the machine. For a quotation that reflects your material, send the feed form, contamination level and target output and a configuration comes back within one working day.
Send the polymer, the tariff, the shift pattern, the scheduled hours and the expected screen interval, and we will return the five-line model with your numbers in it and the confirmed figures from a test run on your material. Request the cost per tonne model and the test protocol — a reply comes back within one working day.