Cracking in PC/ABS (Chi Mei PC540) injection molded parts

[IgorPev]

Hi guys,

I would like to consult with the group regarding a failure I am experiencing in one of my injection molded parts.

The part is plastic shell enclosure made with injection molded PC/ABS (Chi Mei PC540), the screw socket is cracking in both of the bottom bosses. The cracks appear between 1-3 month after assembly, they may occur in storage or in the field.

I have tried to simulate the cracks myself without any successes, I have identified that there is a small gap (~0.2mm) between the two enclosures so some tensile stress may be applied.

The screw is a self-tapping type, the bosses does not show any singes of cracking, only the socket does.

Does anyone know how to test injection molded parts (not transparent) for stress / how to expedite cracking?
Does anyone experienced PC/ABS cracking this way before? (No oily / other solvents are present in the assembly line).

Thanks in advanced.

 

[Nereth1]

ABS may or may not be susceptible to chemical attack from hand lotions etc, but polycarbonate sure is. I have seen mass failures of parts with similar symptoms to yours made of a PC based material due to the wrong additives for chemical resistance being used in the blend. For the record, investigation found the material was susceptible to a lot of chemcials used in the production line, including the hand soap. This being a hand tool, it could be exposed to the same. Some pointers that may help you diagnose:

1) Is it a constant stress or constant strain part? Constant strain parts are difficult to make crack in a delayed manner (i.e. 2 months away) without chemical attack, and thus indicate chemical attack. Constant stress parts can of course creep.
2) I would suggest opening up a crack on a part that has failed, and looking at the profile of the crack. A brittle looking crack or crazing in the area may be indicative of chemical attack. A ductile failure not so much.
3) A lab/the material supplier should be able to look at the crack profile and get at least an indication of if chemical attack has taken place
4) If there isn't any tensile stress in the area, chemical attack won't really be a big issue - but even moulded in stresses can provide the tensile stress necessary for chemical attack to proceed.
5) Get some translucent samples of the material made up if you can (remove the coloured master batch if possible, for example), you will be able to see crazing and so on much easier in them.

 

[chicopee]

Taking another look at the first picture, it appears that there are a couple of strands, since normally ABS has no reinforcement, is this plastic fiber reinforced?; also it appears by the appearance of the width of the crack that the crack originated somewhere else but not in the boss. What is the application of this part and is it exposed to low temperatures?

 

[IgorPev]

Hi guys thanks for your help, it is highly appreciated.

These are enclosures of a HHD and its is (not in storage nor in the field) not subjected to any temp extremes.

PC/ABS is wide spread bland (mostly in consumer electronics) so i find it hard to believe that it can be so easily attacked by hand lotion.

I am still investigating the issue and i have found that the shells are in constant tensile strength due to 0.2-0.3 mm gap between the bosses so the part may creep, hard to imagine that such a gap may couse failures in plastic molded parts.

The crack area is brittle so i'm thinking about bad batch of material, i'm molding now parts from Sabic C6600 which will be later compared to the current Chi Mei PC540, plus i am supposed to get to transparent samples for internal stress check using two polarized films.

 

[Nereth1]

It might not be hand lotion, that's just an example - many chemicals can cause this issue. Chemicals (paints, thinners, cutting fluid and oil, etc) can literally evaporate and deposit themselves from airborne onto PC and crack it. Although that would be an extreme case.

The 0.2-0.3mm gap is closed by the screws I assume, in which case since it's constant strain it's difficult for it to cause an eventual crack, since stress relaxation will take care of it.

 

[3DDave]

Nereth1,

Is the reference to constant stress vs. constant strain language peculiar to plastics? If so, what is the way that stress is divorced from strain? I prefer linear solutions where strain = K*stress, but somehow this does not apply here, so I'm guessing it's a domain specific reference. I did try to look, but don't know how to differentiate the search terms to get what's described here.

 

[Nereth1]

3DDave,

It's the result of viscoelasticity in plastics. Somewhat similar to creep at higher temperatures in metals. Constant stress and constant strain loading cases are ideals but most cases are a mix.

In any case, plastic slowly stretches over time.

Constant strain: If you take a 100mm rod of plastic and stretch it to 105mm and clamp it there, in addition to any classical plastic deformation from overstress, it will also slowly molecularity reorganise itself and take a permanent set somewhere between 100 and 105mm. The longer you leave it at 105mm the closer to 105mm its' length will be when it relaxes after you remove the clamps holding it at 105mm. While it is clamped, if you strain gauge up the clamps, you will see the force on them exhibited by the plastic slowly drop over time. It won't likely break in this situation without something like chemical attack happening simultaneously, because neither stress or strain are rising.

Constant stress: Hang a weight on another 100mm rod of equal weight to the force your clamps that were holding at 105mm, and the plastic will immediately stretch to 105mm (as would any material). But wait a while, and the plastic will slowly flow and stretch beyond 105mm. The mechanism is the same as the relaxation in the constant strain situation, except the force never decreases so the effect we observe is a bit different. This will more often cause an actual fracture, as the plastic just keeps stretching until it goes, although the time period can be extreme in some cases if load is low enough.

If he is screwing two plastic parts together to a set position (gap=0 from gap=0.2mm) then that is mostly a constant strain situation, we wouldn't expect to see a delayed failure in that case without something like chemical attack or other service loads (e.g. dropping it one too many times).

 

[3DDave]

Nereth1,

Thanks. Makes sense.

When I looked at the OP pictures it looked to me as if the crack terminates in the bore of the screw recess and starts elsewhere. I usually prefer an overall picture, in this case of both sides of the part to see where the fractures begin and end. For certain the fractures look very brittle.

I see from the IDES material page
Tensile Elongation (Break, 73°F) 100% ASTM D638
Tensile Strain (Break) 95% ISO 527-2/50

So a little deflection should not be hurting this unless there is a significant stress riser or these specs define elongation differently than for metals. Also, IDES copy protects their page.

http://catalog.ides.com/Datasheet.aspx?I=62673&E=69709

The original resin maker page,

http://fjt.chimeicorp.com/gate/gb/

http://www.chimeicorp.com/en/products/product.asp?pid={510F54DB-0F81-4AE4-9F78-EF28BEF26551}

has different values than IDES, but not very different.

 

[Nereth1]

That's exactly right, 3DDave, and that's why I'm thinking about chemical attack or similar rather than pure mechanical failure. Chemical attack makes things fail below normal breaking stresses and strains, and tends makes them fail after a significant time delay, in a brittle manner.

 

[stick1]

There is a standardized method for testing for chemical compatibility using test coupons. See the attachment. Also, sharp corners should be avoided as they cause stress concentrations that lead to cracking. The mold maker should be able to go back and add slight radii to the sharp corners that are on or near the load path.