can a worn extruder screw cause surging?

By griffex
Published: November 9th, 2009

I've heard this for years but can't see how a linear variation with a period of a minute or two can be caused by a slow radial change over many weeks or months or even years.  If anyone can come up with a reasonable and technically sound explanation, I would like to see it.


Perhaps the question should

Perhaps the question should be restated. Changes in feed screw diameters or even root diameter dimensions are not likely to cause melt direction changes in output.

What I can envision are changes in surface finish caused by wear, aggressive cleaning or mechanical damage. I believe that maintaining the frictional difference that allows the material to slip on the feed screw while sticking to the barrel is critical to maintaining consistent feed.

So it all depends on the definition of worn screw.

All interesting, and asks the

All interesting, and asks the question When is the wear? as well as my usual question, Where's the wear?
Going back to how I started, is there any reason why wear on an extruder screw (or barrel, or both), however or wherever caused, can cause surging, defined as a cyclical but not smooth variation in mass flow with period around 1 to 2 minutes?

Continued thoughts - on

Continued thoughts - on extruder surging

Oops, engaged fingers faster than brain. The intent is not to be aggregating total screw and barrel wear as being a one side or the other phenomenon and regret it appearing to be so one sided - to continue;

In singles, we did see wear in a brief spike at the beginning of each screw and barrel rebuild for a short duration while all of the items "wore in" so to speak (remember, we were looking at situations where multiple screws and materials went into each machine).

We would observe a long period of wear which would be fairly predictable when on lines committed to more or less one material, and then accelerate nearer the end as the wear began to accumulate enough to cause significantly greater changes to process 'output' and we applied more and more resultant torsion to the screw and barrel interfacial region with other process variables. Our feeling was that since the screws were captive at the hub and near the gate, the torsion resulted in a greater arc in the "bend" to the screws and opened up the clearances at an accelerating rate in the 'middle' of the processing area.

Other factors coming into play were the over all rate as a percentage of "top end" capability since the observations were made on 2 1/2 and 3 1/2" machines. I guess the anecdotal way to communicate this concept is the same as saying that when painting, using an 18" roller versus a 9" doesn't double the output at the same intended quality.

The major difficulty area for our data may have been the constant change from one material to another, the various screw designs with and without various mixing elements in differing locations down the screw. Rather than having to overcome "wear" in one part of the screw and barrel with ONE screw and barrel, things were changing up and down the interfacial area probably from just after the feed section until well into the metering sections from a number of the combinations.

Of course once we began to be confronted with leakage back over the flights, and the corresponding reduced outputs etc, the natural tendency was to try to put a toe or two out over the edge without leaping into the abyss - increase screw speed. Most find that dealing with the resultant increase shear, and "residence" time for a portion of the material along with the general increase in melt temperature required rethinking how to pump the material down a screw and barrel (with increasing wear) in differing manners versus suffering the resultant degradation, adhesive versus cohesive, flow related and sometimes aesthetic issues related to the hotter melt.

This tug of war over the abyss was generally continued until we would be delivering high rate, continual output in terms of rate but dealing with stratified material plug flow after the screw tip with a lower level of melt homogeneity through the breaker plate/screens area. Instead of relying on the breaker plate and screens to help move ribbon to plug flow, we found ourselves relying on these to play an increasing role in melt mix distribution as well.

Sorry, all this once again is beyond the scope of this discussion. The short course was that it was (well past?) time for a screw and or barrel rebuild or replacement; and remember first that as processors, we need to look holistically at all the the interelated variables and manage them as just that - a whole.

On twins - yes we saw the most amount of wear at the end of the process since the forces exerted in the conveying mechanism and screw tempering as you mentioned gave us a good many more options over the life of the two elements -

Just my two cents
Jim Wilson

Jim/eagertask Thank you

Thank you for your "two cents. It may be worth thousands of dollars to people who read it. I've been saying for years that "speed kills," that is, it kills precision and requires more thickness to avoid field failure. And if you can't sell more, why run faster? Prudent management may even buy from competitors to resell, if they think the need for more is temporary, rather than rush to put in new equipment to meet the (temporary) need.
As for changing every X lb, that reminds me of the "proverb" that says "Beat your child every day. If you don't know what he did wrong, he does."
Your item (a) in interesting. I have felt that rate of wear should be greater (in mils/month as well as percentage) at the beginning when screws are new and clearances are smaller. This is for singles rather than twins, where design and materials of construction are more variable, and we have to deal with conical vs parallel, flood vs starve feed, and different wasy to control temperature inside the screws. You found wear fastest at the end of the screw life. Was this with singles or twins?

I tend to agree with you that

I tend to agree with you that the slow evolution of “wear” is not likely to create a situation where we have good melt pumping on Tuesday, only to have it fall apart on Wednesday. Screw melt pumping conditions DO change over time with wear, and particularly when aggressive screw and screw tip cooling are employed to “throw material around” in profile tools. To be fair, we have to acknowledge here that there are a number of critical process areas that contribute perhaps a much to "surging" based on material delivery (aspect ratio of material components, mix of materials, conveying method/stability, existence of fines, drying temperature, air flow, residence time in the dryer, feed throat temperature control, zone 1 and or 2 temperature control, etc etc etc) to the feed section of a screw that are ignored or not very well understood in too many shops, although this is way beyond the scope of this particular discussion.

On the other hand, in custom plants where we measured screw and barrel wear faithfully by quarter on 10 or more extrusion lines that ran a variety of materials from week to week (aka not just one material from beginning to end of each screw/barrel rebuild life) we observed the following –

a) the most amount of wear from quarter to quarter occurred at the END of the screw and or barrel life measurement period, not at the beginning. (by production hours)
b) the R&D staff tended to prefer to “run in new items on new screws and barrels” and in short order, the “process” was not as stable in future runs and required “tweaking”
c) many US firms have gotten caught in the trap of increasing rates to “increase contributions per hour” (false economy of course if no new sales fill the new empty machine time) at the expense of significantly closing the ‘window of process ability’ with any particular brand of elements
d) we have observed that batch to batch variation in raw materials from manufacturers quite frankly (despite their “certifications” on lab materials) DOES exist, sometimes in great measure
e) business conditions (and perhaps lack of cash flow?) in custom extrusion in poorly run shops affects the run to run use of varying raw material virgins and regrinds – rather than running a standard mix of regrinds back into product from the beginning of the run until the end, some houses accumulate and “rev the meter” from nearly 0 to nearly 100%
f) tweaking in general to re-achieve proper melt thixotropy with varying amounts of regrind can result in some rather aggressive approaches to achieving a stable melt pumping mechanism at the correct melt temperature and pressure to have the material “be one” with the intended tooling flow paths –

In the case(s) above, we find that the requirements for process adjustment in a process where we KNOW that all else is in good working order and under “control” end up being “justified” due to “changing outputs”. The bottom line is that although a good deal of data can be accumulated in “screw and barrel wear” verus "surging" it is meaningless, unless the contribution to wear can some how be tied back to the requirements for any number of the destabilizing run approach problems above in the plant as well.

Your question is an excellent one, because many shops quite frankly do not do a very good job of monitoring, mixing, drying etc the input of raw materials and then “change the process and blame the extruder” aka worn barrel and screw.

Quite frankly, I would guess that the pipe folks who will routinely rebuild screws and change out barrels on an “X” pounds schedule on lines running the same product 24/7 for months on end would have a totally different answer than the “custom gang”.

Just my two cents
Jim Wilson

© 2011 UBM Canon | please visit these other sites

UBM Canon | Design News | Test & Measurement World | Packaging Digest | EDN | Qmed | Appliance Magazine | Powder Bulk Solids | Canon Trade Shows