Summary: A cost per tonne for a plastic pellet mill is five annual terms over one denominator: capital, power, labour, consumables and rework, divided by the tonnes you actually make. The compact machine behaves nothing like a large line, because capital is a small share of the total and labour is a large one. In the worked case below, capital is 11% and power is 56% of US$96 per tonne. This article gives the equation, a full base case for a TL-PM-100 on one shift, and six scenarios you can recompute with your own tariff, wage and running hours. It also shows why a second shift saves only about 6% per tonne, and why 45 minutes of warm-up at every start costs more capacity than any other line on the specification.

The equation, and the number that decides all five terms
Cost per tonne = (capital + power + labour + consumables + rework, all per year) ÷ annual tonnes. Nothing in that expression is controversial; what decides the answer is the denominator. Annual tonnes is the planned rate in kg/h, multiplied by running hours, multiplied by yield. Get the rate wrong and every term per tonne is wrong in the same direction.
The base case used throughout this article is a TL-PM plastic pellet mill TL-PM-100 on one shift: 55 kW main motor, hydraulic screen changer, 200 kg/h planned, which is the bottom of the published 200 to 300 kg/h band. Running hours are 1,600 a year, being 200 days at 8 hours. That gives 320 tonnes. Every figure below is recomputable from those four inputs.
- Capital — US$17,000 installed (machine, freight, electrical connection, commissioning), an assumption inside the published US$6,800-22,000 band, over 5 years is US$3,400 a year, so US$10.6 per tonne at 320 tonnes.
- Power — 0.45 kWh/kg at US$0.12/kWh is US$0.054 per kg, or US$54 per tonne.
- Labour — one operator at US$4.00 an hour fully loaded, divided by 0.2 tonnes an hour, is US$20 per tonne.
- Consumables — knives, screen pack, heater bands and die refurbishment, US$5 per tonne at a wear-kit life of about 120 tonnes.
- Rework — 4% of output recirculated and handled twice, US$6 per tonne.
The total is US$95.6, call it US$96 per tonne. The shares are the useful part: power 56%, labour 21%, capital 11%, rework 6%, consumables 5%. A change to the price of the machine moves a tenth of the answer. A change to how full the machine runs moves all of it.
Capital is the smallest term, so bargain on the right thing
Raise the installed cost by half, from US$17,000 to US$25,500, and the capital term goes from US$10.6 to US$15.9 per tonne. The total goes from US$96 to US$101, a rise of 6%. That is the arithmetic behind an uncomfortable conclusion: on a compact mill, haggling over the machine price is close to irrelevant compared with what the machine does once it is installed.
Now run the same arithmetic the other way. Hold the price and lose 10% of the rate, from 200 to 180 kg/h. Annual tonnes fall to 288, capital per tonne rises to US$11.8, power per tonne rises because part load is less efficient, and labour per tonne rises to US$22.2. The total is US$100. A 10% loss of rate costs about as much per tonne as paying 50% more for the machine.
What that changes in a negotiation
Spend the effort on evidence rather than discount. Ask for the rate on your material, the kWh/kg at that rate, and the downtime record of the model you are buying. Then ask what the wear parts cost and how long they last, because a machine that is cheap to buy and expensive to keep running inverts this whole model. The published price ranges by platform are a starting point, not a comparison.
One caution on the base case: it assumes the machine is loaded at its planned rate for all 1,600 hours. Very few workshops achieve that in year one. If you actually run 1,200 hours, every per-tonne figure above rises by a third, and capital rises fastest.
Power: a two-term model and the part-load penalty
Specific energy on this platform is published as 0.35 to 0.50 kWh/kg. That band is not noise; it is what a two-term model predicts. Write consumption as E = 16 ÷ Q + 0.37, with E in kWh/kg and Q in kg/h. The 0.37 is the marginal energy of melting and pumping a kilogram. The 16 kW is the near-fixed load: heater bands, dewatering blower, water pump and drive losses, none of which shrink when you feed less.
The model checks itself against the published band. At 200 kg/h it gives 0.45 kWh/kg; at 300 kg/h, 0.42; at 140 kg/h, 0.48. All three land inside 0.35 to 0.50, which is a good sign that the band comes from real machines at different loads rather than from a rounding exercise.
Why part load is expensive on a small machine
The fixed 16 kW is the reason. Run at 140 kg/h instead of 200 and the power term rises from US$54 to US$58 a tonne while making 30% less product, so capital and labour per tonne rise with it. Total cost climbs to US$113, up 18%. The instinct with a compact mill is to run it gently and protect the knives. The model says the opposite: run at the rate you have confirmed on your material, and stop when the feed runs out rather than idling through to the end of the shift.
Electricity price is the other lever, and it belongs to your site rather than the supplier. At US$0.20/kWh the power term goes to US$90 and the total to US$132, a 38% rise with nothing else changed. If your tariff is time-of-day metered, shift the run into the cheap window; a 200 kg/h machine moves 1.6 tonnes in an eight-hour shift, which is enough mass to care about.
Labour, shift pattern and where a second machine stops paying
Labour per tonne is wage per hour divided by tonnes per hour, which makes it the term most sensitive to rate. At 200 kg/h and US$4.00 an hour it is US$20 per tonne. Halve the rate to 100 kg/h and the same person costs US$40 per tonne for the same wage. This is why two workshops with identical machines and identical wages can report costs 30% apart, and neither is wrong.
Now add a second shift on the same machine. Capital spreads over twice the hours and falls from US$10.6 to US$5.3 per tonne. Labour does not fall, because a second shift needs a second operator at the same rate per hour; power, consumables and rework are unchanged per tonne. The total is US$90, a saving of US$6 per tonne, or about 6%.
The break-even that actually matters
Compare that with buying a second identical machine and running both on day shift. Two machines cost twice the capital, two operators, and the same power and consumables per tonne: US$95.6. So the second shift is cheaper by US$3,400 a year and buys you nothing but hours. The second machine is more expensive per tonne and buys redundancy, a maintenance window, and the ability to run two materials without a changeover.
That is the real break-even, and it is not a formula. Run the second shift while demand is steady and supervision can cover it. Buy the second machine when a breakdown stops your deliveries, or when the two materials you run cannot share a barrel without a purge that costs half a shift. In the first case you are buying capacity; in the second you are buying uptime, and uptime is worth more than 6%.
Start-stop: the cost that hides in the schedule
A compact mill is not left hot overnight. From cold, the barrel and the 200 mm die plate need 40 to 60 minutes to reach the 180 to 240 °C working band, and shutdown adds 15 to 20 minutes of purge and clean. Two starts a day therefore consume about two hours in which the machine is drawing power and producing nothing.
Put that against 1,600 paid hours and it is 400 hours, or 25% of the year. At 200 kg/h those hours represent 80 tonnes of capacity that never appears on an invoice and never appears in the cost model either, because the model divides by the tonnes you made rather than the tonnes you could have made. This is the single largest avoidable loss on a small machine, and it is a scheduling decision rather than a capital one.
The arithmetic of one start a day
Consolidate to a single start and you recover roughly 40 tonnes of capacity, about 12% of annual output, with no capital at all. It also removes one cold start from the die, and a cold die with old polymer in the capillaries is the most common cause of blocked holes. Fewer starts is cheaper, faster and gentler on the wear parts at the same time, which is rare enough to act on.
The counter-argument is inventory. One start a day means you hold a day of feed and a day of product, and in a small workshop that floor space may not exist. Where it does not, price the two-start schedule honestly: it costs 80 tonnes of capacity a year, and that number belongs in the comparison against any larger machine that would run continuously. plastic recycling equipment sets out how the mill sits in a small recovery operation.
What the model leaves out, and the three numbers that replace it
Five things are excluded. Freight and duty, which vary by destination and can reach 15% of machine value. Installation labour beyond the electrical connection. Financing cost, which on a machine at this price is usually small but not zero. Spare-parts stock, which for a compact mill is one wear kit rather than a warehouse. And the cost of off-spec pellets rejected by your customer, which does not appear as a cost at all until the order is lost.
The three numbers to ask for
Confirmed rate on your material in kg/h. kWh/kg at that rate, not at the rated maximum. And the wear-kit interval in tonnes. Those three replace every assumption in this article with a measurement, and all three come out of the factory test run, where output, melt pressure, melt temperature and kWh are logged every minute and the report ships with the machine. how every line is factory tested describes what that run consists of.
Then rebuild the equation with your own tariff and wage. The structure does not change; only the shares do. In a market where power costs US$0.06/kWh, capital and labour dominate and a larger machine is hard to justify. At US$0.25/kWh, the rate you run at matters more than the price you paid, and the part-load penalty becomes the first thing to fix.
| Scenario | What changes | Cost per tonne | Movement |
|---|---|---|---|
| Base case | 200 kg/h, 1,600 h/yr, US$0.12/kWh, one operator at US$4.00/h | US$96 | — |
| Machine price +50% | Installed capital US$17,000 to US$25,500 | US$101 | +US$5 (+6%) |
| Electricity at US$0.20/kWh | Power term US$54 to US$90 per tonne | US$132 | +US$36 (+38%) |
| Rate falls to 140 kg/h | Power at 0.48 kWh/kg, labour and capital over 224 t | US$113 | +US$17 (+18%) |
| Second shift, same machine | Capital spread over 3,200 h, second operator | US$90 | −US$6 (−6%) |
| Abrasive filler, wear-kit life halved | Consumables US$5 to US$10 per tonne | US$101 | +US$5 (+5%) |
Frequently asked questions
Q1: What does it cost per tonne to run a plastic pellet mill?
In the worked case here, US$96 per tonne for a TL-PM-100 at 200 kg/h on one shift, with power at US$0.12/kWh and one operator. Power is 56% of that, labour 21% and capital only 11%. Your own figure depends mainly on three inputs: the confirmed rate on your material, your tariff, and your running hours.
Q2: Is electricity or labour the bigger cost on a small pellet mill?
Electricity, by roughly 2.6 to 1 in the base case: US$54 against US$20 per tonne. Power dominates because specific energy is 0.35-0.50 kWh/kg and the heater bands and blower draw a near-fixed 16 kW whatever the feed rate. Labour only overtakes power at low rates or high wages, typically below 100 kg/h.
Q3: Does a cheaper machine lower the cost per tonne?
Barely. Cutting the installed price in half saves about US$5 per tonne, or 5%, because capital is only 11% of the total. Losing 10% of output rate costs about the same. On a compact mill, the rate confirmed on your material and the hours you actually run matter far more than the purchase price.
Q4: Is a second shift cheaper than buying a second machine?
A second shift on the same machine gives about US$90 per tonne against US$96 for a second machine on day shift, so roughly 6% cheaper. The second machine costs more per tonne but buys redundancy, a maintenance window and parallel materials. Choose the shift for capacity and the machine for uptime.
Q5: Which three numbers should I ask a supplier for?
Confirmed rate on your own material in kg/h, kWh/kg at that rate rather than at the rated maximum, and the wear-kit interval in tonnes. All three come from the 4-hour factory test run, logged every minute by the PLC. With those you can rebuild this cost model in a spreadsheet in about ten minutes.
Engineer’s note
Buyers open with the price of the machine and only later ask what it will do on their material, which is the wrong order for this calculation. Two thirds of the cost per tonne is power and labour, and both are set by the rate. That is why we run four hours on the customer’s own feed before the figures go on the quotation: a realistic rate changes the model more than any discount can.
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 target output and your electricity tariff, and we will return the confirmed rate for a TL-PM configuration so you can run this model on your own numbers. Request the quotation and the test report follows the run.