Shaft steels

[agpowder]

Hi, I hope someone can help me.
I have designed some specific application machinery and one of the processes is a twin screw extruder for making powder coatings, I designed this machine two years ago, made it and have been using the machine now for over a year. originally it was only meant to be a prototype just to see if it worked, ie did it make powder coatings, anyway it did and I have developed a business around this extruder.
The problem is now the gearbox, the shafts last about 2 months and then fail. I recently redesigned the mechanics basically beefing up the shafts diameters etc. My question is: Previously (whilst living in England) I always used En24T, turned the shaft, installed and ran and I always had good luck with it. Here in the States I have had trouble with steels, Does anyone know a decent equivelent to En24T (is it 4340 - chemically similar but is it equiv in physical properties?) I also used 1144 on advise from our steel stockist, but I think he had another motive. OR - does anyone know where I can get En24T from in USA?

 

[agpowder]

Hi desertfox
You are correct the extruder imparts high torque loads, I was thinking of the loads on the bearings, sorry.
The sort of load would be analogious to an auger conveying a really thick syrup. In fact this is exactly what it is.
I think maybe half the snapped shafts have gone at shoulders and the other half have broken inside the gear. all the breaks can be "fitted" back together.Often the key slot is part of the snap. Often when one output shaft broke upon inspection the other one would have a crack in it. and you could probably predict the breaks by a combination of product throughput and running load.
Sorry I dont understand the last question.by faces do you mean surface. The shaft is 213 mm long. the first 87mm is 20dia then 31mm of 21dia then 44mm of 30dia then 54mm of 25dia then 32mm of 5/8" hexagon. There is no machined tapering on the shaft.

 

[agpowder]

Wow replys come quick here. I wish I could type faster.
TMoose.
You say interference? ie the shaft is half thou bigger than the ID of brg?
Deffinately not, I machine to +0.00mm/-0.01mm which is 0.4thou?? I think. I do not have a cylindrical grinder, I usually machine to +0.01 or +0.02 and stand for a couple of minutes with emery paper whilst the lathe is running fast. I then use a bearing to check the fit and aim for one that "slips on with very light pressure.
I probably should go tighter but I worry about damaging these tiny bearings whilst pushing them on or damaging another bearing that would be under load whilst pushing it.

 

[unclesyd]

desertfox has hit the nail on the head concerning the existence of an axial load on your shaft. We had 6 far larger WP twin screw extruders and initially suffered horrific growing pains with getting the torque to the screws. In the larger extruders the problem with any whipping is controlled by the clearance between the screws and the elements.

On certain polymers there is also a stick/slip or windup problem that will put very large impact loads on the shafts and even transmit them to gear box.

How is the outlet setup on you extruder, like filter, die plate changer, etc?
The reason I ask is that I checked and the only problem anyone could recall was that on some small extruders we had some mechanical failure that were due to a screen changer being too slow.

How are you heating the extruder?

 

[agpowder]

UncleSyd.
I do not have a die at all at the end of the extruder. The melt just pours out of the "figure 8" and falls 3" into a chill roll.
If, on start up I havn't heated enough and the kneaders are still stuck then the drive will dead head and bang, bang bang. This scares the crap out of me and I make sure the plastic is really liquid and normally I will turn the gearbox input shaft pulley (sheave) by hand to make sure before starting, ie: it is not really an issue anymore, I also do not clean with PVC as I felt the impact was too great, I now just clean with Polyester resin.
The clearance between the barrel and screw elements is now about 1.5mm on the radius. WP extruders are notoriously tight on clearances. With Powder coatings this is unnecessary.
Thanks for your reply and any additional help would be appreciated.

 

[desertfox]

Hi agpowder

I think you need to calculate stresses based on the axial and torsional loads the screw actual see's in service, ie a combined stress calculation (Tresca) or (Von Mises)and apply the appropriate stress concentration factor.
If you can fit the shaft ends together that have failed that would suggest a brittle fracture, so it would appear as another poster as suggested that a crack is initiated during operation which grows slowly under repeated cyclic stress untill the reduced area can no longer carry the load and fails rapidly (fatigue).
What I mean by the last part of my previous post is imagine your looking at the longnitudinal length of the shaft, now saw through the shaft dead straight at right angles to the longnitudinal axis, ie the cut faces would be at right angles to longnitudinal axis of the shaft.
Now take the same saw and do the same but this time hold the saw at an angle of say 45 degrees the faces are no longer at right angles to longnitudinal axis.
My guess is your shaft ends looking at the longnitudinal axis have a angle to them, that would be explained by the principle tensile stresses acting along the principle planes due to the combined loading mentioned at the start of this post.
Can you post a sketch of the shaft and its supports along with the axial and torsional loadings that it see's?

regards

desertfox

 

[desertfox]

hi again

Take a look at this link;-

http://www.mech.uwa.edu.au/DANotes/SSS/failure/theories.html

Look under the heading failure modes and thats what I was trying to explain.

regards

desertfox

 

[agpowder]

I will bring my camara to work tomorrow and open the top of the gearbox and photo it. Can you attach photos here?

 

[desertfox]

Hi agpowder

Yes you can, but make them pdf's otherwise i can't open them.But we still need loadings

regards

desertfox

 

[agpowder]

Ok, I need to come clean here, I have never been to college and I basically left school at 14. I have no idea how to calculate mpg in my car let alone what you are suggesting. I have however "designed" a complete line of powder grinding and classifier machines, cyclone collectors, dust filters and I can muddle my way through electronic controls, I have sold the grinders to over fifty powder coating companies worldwide (they range in price from $40,000 up to $250,000 a piece. In a powder manufacturing line the grinders are only half the storey, the extruder is the real workhorse. They start at about $200,000 and I decided I could make one to my own design that was much lower cost. I made the prototype we are talking about and it works good (the process side) I couldn't find customers so I decided to start making powder coatings myself with my own machines. I found good customers and I have a really nice business doing it.
I know what I want a machine to do and I can design a mechanism to do it but I dont know how to calculate anything to do with it. I suppose my original question is still valid, If you had to take a shot in the dark what steel would you expect to be good for a low speed, high torque application. For example, 1018 would be like a rubber band. Spring steel would be to brittle and impossible to machine? There must be a good go to steel for general purpose low speed (approx500rpm) and high torque?

 

[desertfox]

Hi agpowder

Well lets look at the photo's anyway, if you can post a pic of a failed shaft it may help.
How did you size the motor's etc if you cannot do calculations I can only assume someone else did them for you in which case can he help you now.
I can't take a shot in the dark I haven't even seen the machine or what it looks like.
Generally ordinary mild steel is a common widely used steel for lots of different applications, but certain components like yours may require a slightly better steel.
I am afraid the only sure fired way is to go back to the design and get someone to do calculations, I suppose you can get the maximum rated torque from the motors which is a starting point but then we would need still the axial thrust on the screw.

regards

desertfox

 

[unclesyd]

desertfox
I was just about to post concerning the open ended aspect of this extruder. Without some restriction on the outlet like a breaker plate or die the only significant axial forces he will see would come from the inlet section of the screw where the polymer is trying to melt.

Several things mentioned by the OP is the "bang bang" aspect his startup procedure. This sounds like windup or a stick slip situation that can put tremendous loads on the screw which would be passed to the prime mover. On our big WP's we always start with the low melting component of the polymer compound. The PM is interlocked with the barrel temperature sensors to prevent premature powering up

agpowder,
I am wondering how you get effective kneading with no back pressure in the kneading area?

Anecdotal:
You mention that The WP's are know for their very tight internal clearance which is correct. I help set up all the tools necessary to measure the barrel and and elements on a 120 WP that was being used for experimental work. The had run the machine about 3 months on a new product and decided to dimensionally check the elements and barrels while some factory representatives were on site. As they pulled the screws my partner I looked at each other as something was way out whack. As they demounted the screw elements we started checking the dimensions. We were setup to measure to 0.001" on the OD. What we measured was in 0.250" loss on the elements and when we measured the bore it's wear added another 0.125" of loss. The bad part of this we had inspected another machine a month before and had tossed elements that showed 0.020" wear. These elements aren't cheap. I sure wish I could have been in the meeting between WP and company.

 

[agpowder]

No I did not have anyone to calculate stuff. I picked 7.5kW for the motor because when I first decided to make the prototype I knew I would be starting it in my garage and that was the biggest motor I could run on the electric supply going to the garage.
In the past I had my father to check I wasn't doing anything stupid but I am in the states now and he is in England and unfortunately I tend not to be very organised on drawings either and he cannot travel anymore. He is or was a chartered engineer (retired). I always new the 10Hp was too big for the original gearbox but the machine was just made originally to see if it would make powder coat with a view to developing a line of machines to sell. I think now I can make more money by using the machines rather than selling them. Perhaps I should hire someone to design the gearbox proper. You know what as I have come up against a brick wall maybe I should just use the 4340 steel as discussed earlier and make a load of spare gearboxes just in case.

 

[Artisi]

Hi agpowder. Make sure you get close-up pictures of the failure - preferably use a camera with macro function so as to get good clear close pics, people want to look at the failed surfaces - not necessarily "the gearbox".

 

[desertfox]

Hi unclesyd

I know very little about these machines so anything you can add will be very welcome.
In the meantime i'll try find a bit more about them.


regards

desertfox

 

[agpowder]

Thanks Unclesyd
I agree with your first sentance, the axial loading is the reverse thrust from the augers/screw elements.
The bang bang is on new or cleaned screws when we switch the machines off we purge with epoxy resin that melts low and is easy to start, a purged machine does not bang bang.
Effective kneading: I have never seen any "dispersion" problems at all, in fact I originally set the 32 neader blocks (the elliptical beaters) up as 10 30degrees, 10 60degs and 12 90degs. Now I have 31 30degs and 1 0deg. I manage to get 30% more capacity by doing this and I have not seen any degredation in dispersion quality.
I once stripped an APV machine and the config was half screws, half beaters and half of the screws had completely worn away down to the core of the screw flight. This machine was still working. This was one of the reasons I figured an idiot like me would have a chance of making one myself and getting it to work. I also knew the non back pressure would work because I once bought a Thermo Prism extruder and that did not have anything on the end. I also think Powder Coatings is a very easy application for a twinscrew extruder.

 

[unclesyd]

desertfox,
These machines are basically very high shear rate compounders in nearly all plastic manufacturing facilities. There is a lot of science and well as a little black magic. The amount of power required to run the larger machines is quite high. One the WP 120, early model, had 1000 hp motor.

Here is a short article from Coperion who gobbled up Werner Pfiederer. The old WP site did have a lot of technical information, of which I hope Coperion has keep up. If you ever see a booklet called "Technical Manual by WP grab it.
I tried to check the Coperion site but it is extremely slow tonight.

http://www.coperion.com/up_doc/Veroeffentlichungen/2007/CWP/CWP_200703_Masterbatch production.pdf

 

[EngineerTex]

I was looking on the McMaster-Carr website. You had mentioned that your shafts are their 1144 round stock.

According to the catalog page, their 1144 is treated to a condition of 100ksi (100,000 psi). The 4340 that they offer is in the annealed condition and only offers 68,500 psi. I would suggest finding another supplier where you can get 4340 heat treated to a higher strength. 4340 heat treated (quenched and tempered) to over 125,000 psi has excellent toughness and impact properties and I would suggest finding a distributor who can offer that.

Although McMaster offers 8620 at 100,000 psi minimum yield, I would shy against it. 8620 is pretty similar to 4340, but with less carbon and a lot less nickel. The nickel is what's important here, as adding nickel as an alloying agent is where you get your impact strength properties and 4340 has about 3 times as much nickel as 8620 does.

FYI -- in the AISI numbering system, the last two digits designate carbon content in hundredths of a percent. 4340 has 0.40% carbon and 4330 has 0.30% carbon. Carbon is the overriding factor in heat treatability. An alloy with 0.20% carbon or less (8620, for example) has almost no response to quench and tempering operations. If you see an "L" in the designation like 12L14, it contains lead, which increases ease of machining significantly, but a huge problem for you if any of your extrusions end up in toys, food processing or certain consumer goods. Just an FYI.

-T

Engineering is not the science behind building things. It is the science behind not building things.

 

[EngineerTex]

I was just looking through the McMaster site and it appears that they have ASTM A564 stainless steel hardened to a number of different conditions.

Since it's not going to be about the calculations, you might just experiment with different hardnesses. Their H1150 condition starts at a minimum yield strength of 145,000 psi, which is a 45% strength improvement over your current AISI 1144. If that's too hard, i.e., it breaks even sooner after installation, then go softer. On the other hand, if it fatigues again, but after a longer period of time, then go with the H900 condition rod.

Another question -- are you sure that you have your bearings lined up dead-nuts true? Even a slight amount of misalignment in your frame will allow a fully-reversed bending on the shaft with every revolution, having a devastating effect on any type of shaft that you install in it.

Oh, and by the way -- about not having gone to college, remember that an amateur built the Ark while experts built the Titanic.

-T



Engineering is not the science behind building things. It is the science behind not building things.

 

[CoryPad]

1144 is terrible in fatigue and impact. Go with heat treated 4340 to improve both of these.

Regards,

Cory

Please see FAQ731-376 for tips on how to make the best use of Eng-Tips Fora.

 

[agpowder]

Thanks Engineertex and corypad, this advise is exactly what I was hoping for.
H1150, I can't find that listed maybe from the yeild strength quoted they now call it15-5PH. It is listed as hard to machine and from the earlier posts I can give myself a much better chance by improving the make quality and smooth the finish better.
I will try to source the 4340 in 125,000PSI yield strength.
I remember speaking to the chief engineer of thremo prism when they come to commission the extruder and he slipped that they used En24T for the gearbox input shaft.
Corypad, sorry if I broke the rules.

Thank you all for taking the time to baby me through this discussion.

Oh by the way I think the titanic was engineered OK it was the driver that hit the iceberg and why did the driver of the arc include mosquitoes?