When writing a purchase order or contract for a new extruder, new screw, or re-cut screw, be sure to include a copy of the screw drawing on the PO.  The dimensions are crucial to any analysis of the machine performance, and it is very expensive to shut down the machine (lost production), pull the screw, and make detailed measurements of the dimensions. 

The friction on the barrel must be greater than that on the screw for good solids conveying. 

Friction depends greatly on surface temperature.

The temperature on the barrel and screw surface will likely be variable.

Therefore, a "net" friction factor is defined.  A "net" friction factor is the average friction factor that exists over the screw length in accordance with the variation in surface temperatures.  Making the barrel friction greater than the screw friction means keeping the screw surface cooler further along the axial length than the barrel surface.

Screw breakage is most often attributed to lack of barrel straightness.  However, also consider the possibility of thermal stresses in the screw.  These can be significant if the machine undergoes frequent and rapid speed/product changes, or it is being operated at maximum power.

Freezeout of polymer on the barrel walls can occur when the screw speed is made too high.  This is most often done for rate increases, but the effect also requires additional barrel cooling to maintain the same product melt temperature.  The additional cooling can reach a point that polymer will freeze onto the barrel walls.  Inordinate power increases are also often observed which are caused by the diminished flight clearance created by the frozen layer of polymer.  A redesigned screw is a solution to this.

A key element for flow stability for an extruder screw for continuous product is that the (flow rate produced by solids conveying section) equal (the melting rate of the melt section) equal (the pumping rate of the metering section.)  Each of these flow rates must be at individual pressures that sum to the delivery pressure of the extruder.  Any significant mismatch will result in erratic output rate.

Flow rate is easily calculated for a continuous process by multiplying the density, cross-sectional area, and line speed.  For sheet products, the cross-sectional area is simply the width x average thickness of the sheet.  Of course, some unit conversion will likely be necessary to obtain the final product of flow rate in typical units.  

Copper gauze is better than stainless steel gauze or steel wool for cleaning screws as it will not scratch the surface nearly as easily.  This applies to chrome plated screws and polished stainless screws.

Die bolt length is important to the range in rotation for torquing the die bolts.  That is, the desired torque will be within a range of bolt rotation (degrees).  If the die bolt is short, then the angular range of the torque wrench rotation is small.  A longer bolt gives a greater range, which makes the proper torque more reliably obtained.  More uniform die clamping is, therefore, obtained with die bolts of longer length.

Product melt temperature should be nominally greater than the barrel zone temperature of the last barrel heat zone.  This means that the melt is being slightly cooled at the end of the extruder, and there will be minimum thermal gradients in the melt.  Temperature uniformity is thereby optimized.  Above all, avoid a final barrel heat zone temperature that is much higher or lower than the desired product temperature for the best thermal uniformity.  This may require a new or re-cut screw.

First, check frequency of the instability.  If it is higher than the screw rotational frequency, then the thrust bearing is suspect.  If it is at or near the screw speed frequency, then the screw straightness is possibly the problem.  If it is lower than the screw speed frequency, then solids conveying or melting instability are likely causes.  For solids conveying stability, be sure to check the hopper and hopper level stability.  Solids conveying in the extruder and melting stability problems can be investigated by experimenting with the barrel zone temperatures one at a time in those areas.

There is an interesting ambiguity to solids conveying models in that the choice of the pair of friction factors (screw and barrel) is not unique.  This means that there are many pairs of friction factors for the barrel and the screw which will predict the same flow and pressure.  This has been observed for the Tadmor and Klein model and for other models.   However, more elaborate models that predict temperature will show a change in the thermal effects even though the calculated flow and pressure are the same for two different pairs of friction factors.

In order to design for highest flow rate from a given extruder, the maximum power of the drive much match the power required by the screw at the maximum speed of the screw.  The maximum power of the screw should be about 90% of the maximum drive power, and the maximum speed of the screw should be 90% of the maximum drive speed.  This will provide leeway for control of the process.

The product temperature must be achieved at the end of the screw at the desired flow rate.  This will most typically occur only at one point of speed operation with optimized barrel zone temperature settings.  Operating the extruder at several different speeds will require different barrel temperatures.  This means that a different temperature profile will result in the product for different speeds.  The net result is a different flow distribution in the die and cast product at different speeds even though the "average" product temperature has been maintained. 

Proper channel depth for desired product temperature.  The channel depth in the pumping or metering section of the screw will determine the product temperature at a given screw speed (and flow rate.)  If the desired temperature is too low, the channel depth should be made smaller.  However, this will lower the flow rate.  The lower flow rate can be offset by lengthening the lead length of the screw.  Vice versa, if the temperature is too high then the channel depth can be made larger and the lead length shortened.   

If cast sheet has a periodic flow non-uniformity, check to seek if it is at the same frequency of the screw speed. (line speed/pitch of non-uniformity = screw rotational speed.)  If so, examine or change the thrust bearing.  A slightly damaged thrust bearing will result in a significant pulse in flow (the screw moves back and forth) that is proportional to screw speed.  

Transients in the flow from the extruder will cause thickness nonuniformity in the sheet.  This can be minimized in the design of the extrusion line by putting the filter close to the die.  This maximizes the volume of plastic upstream of the filter, and this volume will help absorb (smooth) fluctuations in flow so they do not reach the exit of the die. The large pressure drop of the filter serves as a barrier to rapid flow change.  

Of course, the extruder must be fully warmed before turning over.  This is best determined by manually turning the screw, even for large extruders.  Use a suitable lever or large wrench to turn the screw by hand to be sure that it is not frozen to the barrel at some point.  Severe damage can occur if the motor is started and the screw is frozen to the barrel.   

Gauging OD of New Screws.  Make a brass ring, about 1 to 1.5 L/D, to match the ID of your barrel.  Slide it over existing screws that properly fit to get the feel of how loose the brass ring should be.  Then, when you receive a new screw you can easily judge the size of the OD of the screw by the same procedure.  The brass ring should have the same "feel" in terms of looseness for the new screw as for the old screws.   

Thermal Steady State after Speed Change:A sudden change in screw speed will require time for the screw temperature to achieve a new steady state value of temperature.  The amount of time is proportional to the screw diameter, D, and it can be approximated by "hours = 50 x D^2" , where D is in meters.

Solids conveying of amorphous polymers can be difficult is that the material  has no distinct melting point.  However, good results have been obtained by boring the screw  for a distance up to the compression  or transition section.  This lowers the axial heat transmission rate in the screw from the melted section of the screw to the solids conveying section of the screw.  A cooler screw surface results which aids solids conveying.  Care must be taken to not make the bore in the screw so large as to weaken it, especially for small screws.