Flake has a bulk density of 350 to 450 kg/m3, about a third of pellets, so a gravity hopper leaves the screw only half full. The TL-FK uses a vertical force feeder whose auger speed is slaved to screw torque, holding the flights 85 to 95% full instead of half empty.
Torque is a better control signal than hopper level because it reports what the screw is actually receiving. A full screw builds stable pressure, reaches the published 0.26 to 0.36 kWh/kg and gives the vacuum vents a predictable amount of material to work on.
Why the feeder matters
- Low bulk density starves a gravity-fed screw and makes output swing.
- Torque rises as the flights fill, so it is the right feedback signal.
- A full screw holds melt pressure and specific energy in the published range.
- A sawtooth torque trace means the auger control needs tuning.

Flake is the easiest feedstock to run and the hardest to run well
Flake flows, which sounds like an advantage until you look at the screw. Washed HDPE and PP flake has a bulk density of 350 to 450 kg/m3, roughly a third of the density of pellets. A gravity hopper delivers that light, irregular material into a Screw designed for a much denser feed, and the flights arrive only half full.
A half-full screw cannot build pressure consistently. Melt pressure fluctuates, output swings, and the same nominal speed produces different quantities from hour to hour. The material is easy to melt, but the machine cannot make what it is not being given.
The TL-FK solves this with a vertical force feeder whose auger speed is slaved to screw torque. The feeder pushes flake into the screw at the rate the screw can actually accept, and the flights stay 85 to 95% full instead of half empty.
How torque feedback works

Screw torque is a direct measure of how much material is in the flights. When the screw is starved, torque falls. When it is full, torque rises to a stable working value. The force feeder reads that signal and adjusts its auger speed to hold the screw at the target load.
That is a better control variable than hopper level. A hopper can look full while the material bridges above the throat, and a level sensor cannot tell the difference between flake that is flowing and flake that is standing still. Torque tells you what the screw is actually getting.
The result shows up in output stability and in specific energy. A full screw does the same work on each revolution, so the machine reaches the published 0.26 to 0.36 kWh/kg. A starved screw wastes revolutions moving air and produces less for the same energy input.
What the feeder changes downstream
A consistently full screw gives a stable melt pressure at the die, which means the cutter receives a steady strand. Pellet weight variation falls, and a melt pump becomes an optional refinement for plus or minus 0.4% pellet weight rather than a necessity.
The vacuum vents also work better. With a steady feed, the 3 m pre-heating section sees a predictable amount of material and a predictable amount of surface water. A surging feed lets wet flake bypass the venting window and carry moisture into the compression zone.
Filtration benefits too. A double-piston changer sized by dirt load assumes a known throughput of contamination per hour. When feed surges, the screen sees a dirty load in bursts and blinds earlier than the schedule predicts.
Sizing the feeder with the screw
The TL-FK range uses 150, 165 and 180 mm screws with 90, 160 and 250 kW motors, producing 300 to 450, 500 to 700 and 800 to 1,000 kg/h respectively. Each screw size needs a matching feeder, because an auger sized for the smallest machine cannot keep the largest one full.
The feeder is specified as part of the line, not as an accessory. A buyer who sources a granulator without the matching force feed will reproduce the starvation problem regardless of which brand of screw they buy.
Flake size also matters. The published feedstock window is 8 to 30 mm washed flake. Material outside that range can bridge in the feeder throat or compact in the screw, and it should be tested before the machine is specified.
Checking that the feed is right
Watch the torque trace on the control screen during a steady run. A flat trace with small variations means the feeder is holding the screw full. A sawtooth trace means the feeder is chasing the load and the auger control needs tuning.
Then check output against the rated figure over a four-hour window. A line that reaches nameplate speed but cannot hold output is being starved somewhere. The cause is usually the feeder, the flake size or a bridge in the hopper.
Our factory test procedure includes a material trial on the customer’s own flake, and the torque trace from that trial should travel with the machine. It is the baseline against which later feeding problems can be judged.
Gravity feed versus force feed on flake
Flake behaves nothing like pellet or regrind. The table compares what happens with and without torque-following force feed.
| Property | Gravity hopper | Torque-following feeder |
|---|---|---|
| Screw fill | Around 50% | 85-95% |
| Melt pressure | Fluctuating | Stable |
| Output | Swings by hour | Holds rated range |
| Specific energy | Above reference range | 0.26-0.36 kWh/kg |
| Sensitivity to flake size | High | Controlled within 8-30 mm |
Common questions, with short answers
Q1: Why does flake starve a screw?
Washed HDPE and PP flake has a bulk density of only 350 to 450 kg/m3. A gravity hopper cannot keep a screw designed for dense pellets full, so the flights arrive half empty and melt pressure fluctuates.
Q2: How does torque feedback control the feeder?
Screw torque rises as the flights fill and falls when the screw is starved. The force feeder reads that signal and adjusts its auger speed to hold the screw at the target load, keeping fill at 85 to 95 percent.
Q3: Does force feeding affect energy use?
Yes. A full screw does useful work on every revolution, so the line reaches its published 0.26 to 0.36 kWh/kg. A starved screw spends energy moving air and produces less for the same input.
Q4: What flake size does the TL-FK accept?
The published window is 8 to 30 mm washed HDPE, PP, PS or ABS flake. Material outside that range can bridge in the feeder throat or compact in the screw, so it should be tested before the machine is specified.
Related pages on this site: flakes pelletizing machine and size reduction equipment.
If you are working through the same machine family, these are the other Running 4% Moisture Flake Without a Separate Dryer, 80 or 100 Mesh? Size the Screen by Dirt Load, Where the Energy Goes in Flakes Pelletizing and 24/7 Flake Pelletizing: The Screen and Blade Rhythm.
About the author
Zhou Gong (周工) builds recycling lines at TENGLONG Plastic Machinery, a factory in Linyi, Shandong, with 18,000 m² of floor area, 30-plus staff and 820-plus lines delivered to 46 export markets. A hopper can look full while the screw is starving. Torque never lies about that. The numbers in this article come from test runs and from dated published sources, not from a sales sheet.
If you want a recommendation rather than a catalogue, send the material form, the contamination level and the target output. We return a model and a price range within one working day, and the test run on 10–20 kg of your own material is free before the balance is paid.