Feed Throat Bridging in Profile Extrusion: Why Output Suddenly Drops

What bridging actually is

A bridge is a stable arch of particles that forms across a converging hopper, leaving an empty void beneath it. The screw keeps turning, but it is starved. On a profile line the first sign is not a stoppage; it is a slow fall in head pressure and a profile that gets lighter, smaller or slightly out of tolerance while every setpoint on the panel stays where it was.

A rathole is the related failure. Instead of an arch across the whole hopper, a narrow channel opens above the feed opening and material hangs up on the walls around it. The level indicator still shows a full hopper, the line still runs, and only the centre of the hopper is actually feeding.

Both are caused by the same thing: the bulk solid becomes cohesive enough to support itself. Moisture, fines, warm surfaces, a shallow hopper wall angle and a small outlet all help. Rigid PVC dry blend is a classic case because it is a free-flowing powder at room temperature but contains lubricant and a low softening resin that start to tack at modest temperatures.

The fault is intermittent, which is why it wastes so much time. An operator taps the hopper, the bridge collapses, output jumps back, and the line looks fine for another hour. Nothing in the extruder controls was changed, so the obvious conclusion is that the problem was a one-off.

Downstream the symptoms are easy to misread. A starved screw produces surge, the melt temperature rises because there is less cold feed to absorb heat, and the haul-off is blamed for a wall thickness change that actually began at the hopper. Reading the whole line together is what identifies the feed end as the source.

The feed throat is a thermal boundary

The barrel is heated in zones to melt the polymer; the feed throat is deliberately cooled so that the solids arriving from the hopper stay solid. On a rigid PVC line the throat is usually water jacketed, with the cooling water held somewhere in the 20 to 60 C region depending on the compound and the screw design.

That boundary has to hold under conduction. The first barrel zone can be at 160 to 180 C, and steel conducts heat backwards into the throat casting. If the water channel is scaled, the flow is too low, or the isolation gasket between the throat and the barrel is damaged, the throat slowly warms up during the first hour of the run.

Once the throat wall approaches the softening range of the compound, particles tack to the metal. A thin ring of softened material builds at the outlet, the effective opening gets smaller, and the arch that forms above it becomes stable. Cooling the throat back down usually clears a fresh ring, but a baked-on layer needs a clean-out.

Too cold is also a fault. Very cold throat walls can condense moisture from the air on the compound in a humid plant, and moisture on a PVC dry blend promotes caking. The useful target is a stable temperature that is cool enough to prevent tack and warm enough to stay above the dew point of the surrounding air.

Log the cooling water inlet and outlet temperatures, not just the flow switch. A rising outlet temperature with a constant inlet tells you the heat load is increasing, which usually means heat is leaking back from the barrel or the channel is beginning to block. It is an early signal that shows up before output moves.

Bulk density, regrind and particle shape

A flood-fed single screw extruder takes what falls into it. Output is therefore proportional to the bulk density of the feed as well as to screw speed. A change from a dense virgin dry blend to a fluffy regrind mix reduces the mass delivered per revolution even though the hopper is full and the screw speed is unchanged.

Regrind changes more than density. Ground profile has a wide particle size distribution, with fines that pack into the gaps between larger chips and increase cohesion. It also carries its own lubricant and stabiliser history, so a regrind-rich blend can soften at a slightly lower throat temperature than the virgin compound.

Hot mix discharged from the hot mixer and not properly cooled in the cold mixer is another common source. Blend stored warm in a silo arrives at the hopper already near its tack point, and the residence time in the hopper is then enough for it to cake. Discharge and storage temperature belong on the batch record.

Measure bulk density rather than guessing. A simple weighed litre cup at goods-in and at the hopper takes two minutes, and a trend across a shift will show a regrind ratio change or a mixer problem before it becomes a wall thickness complaint. It also gives the setter a number to use when changing screw speed.

Table 1 - Feed conditions that starve a flood-fed extruder.

Hopper geometry either helps or hurts. A steep cone with a large outlet and a smooth internal finish encourages mass flow, where the whole contents move together. A shallow cone with a small outlet encourages funnel flow, where only the centre moves and the walls hold material until a bridge forms.

Flood feeding and starve feeding are not interchangeable

Most rigid PVC profile lines flood feed: the hopper is kept full and the screw takes what it can drag in. Output is set by screw speed, feed bulk density and the pressure the die imposes. A separate gravimetric feeder that meters a controlled mass flow is a starve-fed arrangement and behaves quite differently.

With a starve feeder, a bridge upstream shows up immediately as a loss of feed rate at the weigh belt or the loss-in-weight hopper, and the controller alarms. With flood feeding there is no such signal, which is why a bridge can run for an hour before anyone notices a dimensional drift.

Adding a level sensor to a flood hopper does not solve this. The sensor sees material at the top while the void is at the bottom. A low-level alarm near the outlet, or a simple load cell on the hopper, is more informative because it detects the moment material stops moving rather than the moment the hopper empties.

Agitators, vibrators and air pads are legitimate remedies but blunt ones. Continuous vibration can compact a cohesive powder and make the bridge stronger, and an air pad that injects humid plant air into a hygroscopic blend creates the caking it was meant to prevent. Use them intermittently and with dry, clean air.

Where bridging is chronic, a mechanical change is more reliable than an operator with a mallet. A steeper hopper cone, a larger outlet, a polished liner or a short vertical spool piece above the throat all reduce the chance of a stable arch. These are inexpensive compared with the scrap from an unexplained surge.

Reading a feeding fault from the line

Start with the pressure trace at the die. A feeding interruption produces a fall in head pressure with a delay of roughly the residence time between the hopper and the die, followed by a recovery when the bridge collapses. A screw wear or temperature problem produces a much slower drift, and a screen pack blockage produces a rising pressure, not a falling one.

Motor load helps separate the cases. A starved screw draws less torque because it is moving less material, while a cold feed or a blocked screen raises load. Record torque and head pressure on the same time axis; the pair is far more diagnostic than either number alone.

Melt temperature often rises during a bridge, because there is less cold solid arriving to absorb the heat input. If melt temperature and head pressure move in opposite directions, look upstream before touching the barrel zones.

Check the profile itself. A short period of starvation shows up as a light section, a narrower width or a slight bow, and because the calibrator is downstream the part records the event as a length of off-spec material. Mark the metre position when the operator taps the hopper and compare it with the trace.

Do the cheap physical checks before any disassembly. Feel the throat casting, read the cooling water temperatures, look into the hopper for a rathole, and weigh a litre of feed. Those four checks take five minutes and identify most feeding faults.

Table 2 - A diagnostic matrix for an output drop on a profile line.

Write the finding into the shift log even if the fix was a tap on the hopper. A repeated entry at the same time after start-up is the classic signature of a throat that warms up slowly, and it is much easier to fix from a log than from memory.

A feeding routine that prevents restarts

Start with the blend. Record the batch number, the cold mixer discharge temperature, the regrind ratio and a weighed bulk density. Keep the hopper covered so the compound is not sitting in humid air, and do not top up a hot blend straight from the mixer.

At start-up, confirm cooling water flow and log inlet and outlet temperatures before the barrel zones reach setpoint. Heat soak the screw and barrel as the line manual requires, then start at low screw speed and watch head pressure stabilise before bringing the haul-off up to production speed.

During the run, watch the trend rather than the instantaneous value. A head pressure that oscillates with a period of several minutes is usually feeding, not control. A steady climb in throat water outlet temperature is a warning even while the profile is still in tolerance.

Clean the feed path on a planned stop. Remove the ring of softened material at the throat outlet, brush the hopper walls, and check the cooling channel for scale. A light clean at a planned stop is far cheaper than an emergency strip-down in the middle of a run.

Finally, cap the regrind ratio during any trial. Regrind changes bulk density, fines and the softening point together, so a feeding problem that appears after a regrind campaign should be investigated with a fixed ratio. Once the line is stable again, the ratio can be raised deliberately and the result recorded.