EXTRUSION: Venerable Maddock Mixer Still an Extrusion Workhorse

Originally titled ''

Variations to this decades-old mixing section are widely used, but processors should carefully analyze these designs and not assume they will perform better.

Related Topics:

The “Maddock mixer” is arguably the most prevalent mixing section used by screw designers in North America. It was originally developed at the Union Carbide lab in Bound Brook, N.J., by Gene Leroy and then popularized by Bruce Maddock in the early 1970s for improving melt quality in film and cable-jacket processing. It has retained the Maddock name over the years.

Although primarily a dispersive mixing device, the Maddock is now used in all kinds of mixing applications and with many design iterations, the “spiral Maddock” being the most common. In this design, the expectation is that the helical flutes will produce a certain amount of pumping, resulting in a lower pressure drop and some degree of self-cleaning action. Both can be true to varying degrees, but the design effort necessary to optimize these benefits compared with the original axial fluted design is surprisingly complex. And without taking these complexities into consideration in the design of the spiral configuration, any potential benefits may not only be eliminated but the entire performance of the mixer degraded. Simply put, not all spiral-style Maddock mixers are created equal.

Depending on what combination of dimensions is retained from the original design, the resulting output, pressure drop, shear heating, and mixing performance may be either improved or diminished compared with the original straight-flute design. Unfortunately, it seems there are many variations in current use that do not consider the design calculations necessary to make it work optimally, and instead are simply “fit” it into the envelope. As a consequence, many poorly designed spiral-type Maddock mixers are being used by extrusion processors right now.

With the spiral design, some of the dimensions in the cross-section or circumference must be reduced from the original axial designs due to the helix angle. In the accompanying illustration, the helix angle for the spiral design was set at 32.48° (5-in. lead) while the barrier and wiping lands were held constant with the intention of maintaining the dispersive mixing and the channel leakage. This resulted in the groove radius being reduced from 0.485 in. to 0.375 in. in order to fit in the cross-section. 

The choice of helix angle affects every dimension in the mixer. In this case, for example, a 5-in. lead for the spiral resulted in  a 75% increase  in the length of the groove and a 40% reduction in groove volume. This resulted in several times the pressure drop of the axial grooved mixer even with the slight pumping effect from the helix. Analysis of the pumping for a medium-viscosity polymer, even at a modest head pressure, eliminated any pressure-drop advantage of the spiral mixer. No quantitative study has been done to determine if the “self cleaning” effect is an advantage,  but observation of many screws being pulled did not indicate a measurable difference.

The mixer design needs to be matched to the process requirements, just like the rest of the screw design. The overall output of the screw is affected by the barrier gap, width, depth, and length of the grooves. Drag flow over the barrier gap—the primary mixing effect—aids output by overcoming much of the pressure drop over the barrier lands. At the same time, the width and gap of the barrier lands control the degree of dispersive mixing. The number and geometry of the grooves determine a lot of the distributive mixing effects.

The pressure drop and temperature rise are similarly affected. Each change improves one aspect and usually reduces another. Like almost everything in extrusion there is no free lunch.

Analysis of the mixers is very challenging, particularly the estimation of viscosities in various areas of the mixer, as well as the polymer flow in the grooves and over the barrier. The quality and viscosity of the melt being delivered by the screw to the mixer also adds another variable.

That said, some simple logic can be of a great assistance. For example, if dispersive mixing is the main issue, the width of the barrier land in the circumferential direction should not be reduced. If minimum pressure drop is critical, the helix angle and groove volume should be kept as large as possible to shorten the flow path and reduce the average shear.

It appears the spiral Maddock has a certain pseudo-technical appeal compared with the original axial-grooved design, but it does not necessarily have better capability unless very carefully designed so that its positives are magnified and negatives minimized. In many cases, processors are not getting the benefits provided by the original design, let alone any improvement.

ABOUT THE AUTHOR

Jim Frankland is a mechanical engineer who has been involved in all types of extrusion processing for more than 40 years. He is now president of Frankland Plastics Consulting, LLC. Contact  jim.frankland@comcast.net or (724)651-9196.