PDA

View Full Version : Plastic stress relief



beanbag
06-19-2012, 04:11 AM
I machined a bunch of polypropylene blocks and now the pieces have a bow in them. It turns out that polypro comes in "as extruded" and "annealed" form, and I must have ended up with the former. I also read that the recipe for stress relief is 50F per hour, and a few hours at 185F. My question is if I can just half ass it and dunk the piece, clamped flat, in a pot of hot water at 185f and then let it slowly air cool. What bad thing will happen if I bring it up or down in temperature rapidly?

macona
06-19-2012, 05:30 AM
I think the soak is the most important part. Not doing that and not cooling properly will end up having a piece that has not been annealed.

Bob Fisher
06-19-2012, 02:57 PM
Years ago, I was working in Ireland for one of the big three. They had a injection molding facility and were getting parts that would not meet requirements. After some investigation, we discovered that annealing the parts brought them back in spec. I still suspect that if the molding parameters were adjusted the annealing would not have been needed. The plant had a tendency to increase the cycle rate to meet production. The point of all this is that it does work. Bob.

beanbag
06-20-2012, 08:09 AM
The plastic supplier said that the sheet was supposed to be the "stress relieved" kind. So now I don't understand how cutting a .25" depth pocket in a .75" block with .02 finish pass can possibly cause enough stress to bow the piece. I've machined nearly the same piece with plastic from a different supplier, and I don't remember this problem.

bob_s
06-20-2012, 10:28 AM
The plastic supplier said that the sheet was supposed to be the "stress relieved" kind. So now I don't understand how cutting a .25" depth pocket in a .75" block with .02 finish pass can possibly cause enough stress to bow the piece. I've machined nearly the same piece with plastic from a different supplier, and I don't remember this problem.


Thermal stresses developed during machining!

beanbag
06-20-2012, 04:35 PM
I had an air blast going and the part immediately after the cut (that is, right behind the end mill) was not even warm.

Grind Hard
06-20-2012, 07:16 PM
Wouldn't be the first time the supplier was wrong! Or maybe they are "right" and are shipping by what they were sent. "It says relieved plastic stuff, so it MUST be."

My blade supplier sent me a bunch of "certified" blanks made out of an exotic stainless.... Certs said so. Two weeks and two independent metallurgical analyses later, we have determined that they in fact sent me shiny CRS.

I could have told them that :roll:

But hey the CERTS SAID THIS WAS STAINLESS so that's what they sent me.

*grumble*

darryl
06-20-2012, 11:04 PM
I think it's pretty easy to build a stress into a part by machining it. The bulk of the material may not heat enough to warp, but it will hold the surface in place while the surface undergoes some plastic deformation from the machining process. In effect, the unexpanded underlying material squeezes the surface when it tries to expand with the heat. When the surface cools, it tries to squeeze the underlying material. Something like that anyway.

Something I've wondered about many times is what happens when a material undergoes an extremely rapid heating and cooling- think of it as a very thin material rotating first through a hot flame, then into a bucket of cold water. Does the molecular change still occur on the same short time schedule, or is that something that takes time once the material is up to temperature?

Reading about glassy metals suggests that the temperatures can be brought down so fast that the material doesn't have time to change state- you end up with the same material but different properties. In a machining process you could have the same thing more or less- the molecules in the skin don't have the same structure as the underlying material. I don't know how this applies to plastics, but it probably does.

beanbag
07-15-2012, 12:24 AM
My conclusion at this time is that the first batch of plastic I got a few months back was better stress relieved than this current batch. The reason I say this is that the parts are nearly identical, and I cut them with nearly the same parameters.

But I have a problem. I took a bowed piece, clamped it flat, and put it thru the anneal cycle of slowly going to 185f, sitting there for a while, and then slowly cooling. The problem is that the piece still comes out slightly bowed, and after about a day, returns almost to its initial position. So in other words, the internal stresses are still there somehow.

Next I will try clamping the piece bent in the opposite direction, but I'm not sure how this is going to result in a consistent final piece.

Also, I am wondering why the anneal temp is suggested at 185f, while the "working temperature" of the material is only 180F, and the melting point is all the way up there at 327f.

Darryl's post reminded me of the trick where you take dents out of sheet metal by heating up the area, and then quickly quenching it. So I suppose another thing I could try is to heat up the convex side and then quickly quench it.

Evan
07-15-2012, 12:32 PM
Anneal temps vs maximum temp depend on crystalline phase changes in the material. Polypropylene is a semi crystalline plastic with several phase changes. Crystalline plastics have a temperature knee somewhere below the melting temperature. Above that knee they turn amorphous prior to melting (phase change). When they cool various sections of the part recrystallize depending on the thermal conduction of the part. Parts that crystallize before other area will cause deformations as the sections undergo volumetric changes at different times. This is very hard to deal with in a part with thick and thin sections.

You can try very slow cooling in a water bath with good agitation. You can also try extremely fast cooling with ice water but it probably won't work. Also make sure that the entire part is heated to slightly above the anneal temperature. Raise the temp by maybe 5 degrees (190f) and hold it there for an hour to ensure total phase change.

To answer your direct question about working temp: The working temp must be kept below the phase change temperature or the material will distort.

Make sure your thermometer is accurate. It matters.

beanbag
07-15-2012, 02:20 PM
Thanks for the explanation. I had already tried the procedure you suggested, which was to put the piece, clamped on both sides with Al plates, in a pot of water, bring it to a near boil (maybe 195-ish), sit for an hr, and slowly cool down. That did squat. Or didn't do squat. :D

I do know that something can happen near 190 because if I do the same thing with the piece unclamped, I can make the bow much worse. :rolleyes:

Is annealing plastic the same as getting the molecules to slowly flow around until the stress is relieved? In other word, suppose you take a thick ring and cut a slit in it. Can this C shape be "annealed" back into a flat bar? What happens to all the molecules that were at the ID and OD?

My google experience with polypropylene has been poor. What are these other phase change temperatures? I only found the glass transition temp which is very cold.

So far, my best result has been to heat the Al clamping plate on the convex side on the stovetop to 250f, leave it there for 2 minutes, (so the heat does not have time to evenly distribute) and then dunk it in cold water. About 70% of the bow can be removed this way. I don't know whether it is the 250f, the uneven heating, or the quenching that makes it work.

Evan
07-15-2012, 03:06 PM
At 190 all of the material will be into the phase change zone. Full conversion to amorphous doesn't take place until 234 and it takes a great deal of heat input without a rise in temperature. This is typical of phase change materials including water ice. Annealing does allow the molecules to reorient without fully decrystallizing which would cause large distortions. I misspoke somewhat when I said "total phase change". The material does not fully decrystallize at 190 but all of it is much more mobile that it is at 180 which is fully crystallized.

I'm not certain about other phase change temperatures but many crystalline plastics have more than one. The one that matters is at 185 to 234. The reason the material doesn't crystallize directly below melting is because it supercools as do most thermoplastics. If the material is above 234 it is technically melted although not in a fully fluid state.


... suppose you take a thick ring and cut a slit in it. Can this C shape be "annealed" back into a flat bar? What happens to all the molecules that were at the ID and OD?

Never tried it but I suspect that it isn't possible without fully melting the plastic.

When I say to cool slowly I mean slower than the natural cooling rate of a pot of hot water. That isn't slow enough for all parts to cool at the same rate.


BTW, there are many formulations of polyprop. There are a range of additives that change the properties a little or a lot. Perhaps you have a slightly different formulation than the last time.

beanbag
07-17-2012, 06:07 PM
Evan, thank you for your info. Where did you get the values of 190 and 234, etc?

I got some info from the manufacturer, and their procedure is heat to 275 for a long time, and then cool 10 degrees per hour. No wonder aiming for 185 didn't work.

But I am too lazy to do a 15 hr anneal process and came up with this instead: I made a fixture that has a slight convex surface, and clamped the part to that. There is a little groove to fit in a thermocouple probe. Heat to 250F, let it sit there for 4 minutes (bottom of part gets hot, top gets barely warm) and then dunk in cold water. There is some room for tweaking, but hey, it works, and the piece doesn't seem to shift around after a day. Maybe there are some residual stresses inside. Oh well, at least the piece is flat.

Evan
07-17-2012, 06:39 PM
I don't recall where those numbers came from. I have a very large collection of materials properties sheets and pages that I save whenever I come across anything interesting. I work with just about anything I can get my hands on in my continuing experimenting with materials. Plastics in particular are very useful and interesting. I'm pretty good at remembering numbers too.


I got some info from the manufacturer, and their procedure is heat to 275 for a long time, and then cool 10 degrees per hour. No wonder aiming for 185 didn't work.

Nice to know. That will most certainly take all the material into the full phase change zone meaning no crystallization remaining. Then by cooling it so slowly it ensures that all of the material crystallizes at the same rate and time ensuring no differential strains.

Incidentally, with most plastics long term aging in a cool place will do the same thing. With nylons aging for ten years will ensure absolutely strain free material.

beanbag
07-17-2012, 10:11 PM
Incidentally, with most plastics long term aging in a cool place will do the same thing. With nylons aging for ten years will ensure absolutely strain free material.

One of the first things I tried was clamping the piece bent the opposite way overnight. The next morning, it looked straight, but returned to its original bent shape after 3 hours. I know that PP has creep properties, but I think it also has "memory" of its previous state, and the only way to erase this memory is via heating.