Tensile Strength Testing

[rsterne]

Sometimes too many variables just confuses the issue, even when I change them one at a time!.... Now I am seeing some failures through or around the eye for the quick pin, something I cannot have happen, and didn't see in the first rounds of testing.... All I can assume is that I an succeeding in making the test section of the coupon (5 x 5mm) strong and stiff enough that is is transferring the load too well to the ends (or actually they are breaking before it)....



The top row is the coupons I mentioned in the last post, with 110% flow in the inner walls only.... They failed right at the end of the test (narrow) section, the one with the 30 deg. angle ripping a piece out of the wall leading to the eye, while the faired in one (right) broke right where the 8 layer walls part to wrap around the eye.... It was the strongest one yet, breaking at 8785 psi, even allowing for the wider test section from the extra flow.... However, the increased density due to the extra flow is something we would not do in practice, because of the detrimental effect on the finish, so I am dropping back to 100% flow for all testing going forward.... In addition, because some of the samples are starting to stretch for up to 30 seconds after the force peaks, I am going to time only to when the force peaks (as closely as I can), and we will call that the "Yield Strength" of the plastic, which is effectively the maximum stress you can use without seeing permanent distortion....

The second row are printed with 8 walls of 0.40" wide, at 100% flow.... That works out to only 4.8mm of walls, filling a 5mm wide part, which is why i started playing with the flow in the first place.... The angled coupon was the first one where I saw it stretch for 30 seconds after yielding, and the faired in coupon broke as it should.... Row's 2 through 4 used the "Extra alternate wall" feature to better lock in the infill....

For the third row, I changed the wall width to 0.42", so that it properly filled the 5mm test section.... Failure of the angled coupon was the same, but the faired in one ripped the outer wall from the eye.... When I looked back at the Cura Preview, I found out that increasing the wall thickness dropped the wall count around the eye to only 7 instead of the 8 previously, and I think that is why the eye failed....

The fourth row was the same as the third, but I increased the temperature to 220*C instead of the 205*C I have been using previously.... It performed great, strongest of all the coupons tested at 100% flow.... Why the angled coupons stretch, while the faired in ones break, I have no idea....

The bottom row was the same as the 4th, except instead of using 100% fill (lines at 45 deg.) for the triangles between the hole and test section, I increased the number of (possible) walls to 16 instead of 8 (7), so they were printed with no infill, just walls.... The angled coupon performed exactly the same as the other one printed at 220*C, but the faired one broke the eye!.... Now I'm getting frustrated....  .... Here are the yield strength numbers for these tests....

30 deg. angled, 110% flow.... 7501 psi.... some delamination
30 rounded, 110% flow.... 8785 psi.... OK
30 angled, 100% flow, 0.40" walls, 8 alternated with 9.... 7949 psi.... stretched
30 rounded, 100% flow, 0.40" walls, 8 alternated with 9.... 8379 psi.... OK
30 angled, 0.042" walls, 8/9.... 7855 psi.... stretched
30 rounded, 0.042" walls, 8/9.... 8086 psi.... eye failed
30 angled, 220*C.... 8306 psi.... stretched
30 rounded, 220*C.... 8484 psi.... OK
30 angled, all walls.... 8348 psi.... stretched
30 rounded, all walls.... 8086 psi.... eye failed

I am going to try a different test section that is 6mm wide and the whole coupon only 4mm thick.... This will allow me to have 16 walls in the test section, and have 8 of them wrap around the eyelet on the ends.... Stand by for the results.... 

Bob

[rsterne]

Here are the first tests with the 6mm wide by 4mm thick test section.... The angles were changed to 1 in 2 (26.6 deg.).... The bottom two angled coupons had rounded ends to save material.... and became the only angled coupons to shed part of the eye!.... hmmmmmmmmmmmmmmm....  .... All coupons has the outer wall printed first, to hold the dimensions better, and not have to drill for the pins....



Top row was 0.38mm walls, so that 16 of them (8 per side) are 6mm wide for the test section, and also they flare out to wrap around the eye.... They didn't use the alternate extra wall function.... I used 110% flow for the inner walls....

The 2nd row was the same, but used the alternate extra wall.... Both rows performed great....

The 3rd row was also 8/9 alternating walls, but 100% flow.... The angled coupon (left) shed the eye (which I never found), the faired one was fine....

The bottom row used all walls, with no infill.... Both coupons tore the outer walls from the inner walls of the eye.... Here are the yield strengths....

27 deg. angled, 0.38mm wall, 110% flow.... 8453 psi.... stretched
27 deg. rounded, 0.38mm wall, 110% flow.... 8892 psi.... stretched
27 angled, alternate extra wall, 110% flow.... 8903 psi.... stretched
27 rounded, alternate extra wall, 110% flow.... 9046 psi.... OK
27 angled, 8/9 walls @ 100% flow, 220*C.... 7896 psi.... eye failed
27 rounded, 8/9 walls @ 100% flow, 220*C.... 8353 psi.... OK
27 angled, all walls, 220*C.... 7788 psi.... eye failed
27 rounded, all walls, 220*C.... 6004 psi!.... eye failed

It appears that the squared end on the coupon helps keep the eye from distorting and tearing out.... So, even though the rounded end saves material and looks pretty, I guess I better not use it, and see if that prevents further failures at the eye.... 

The highest yield strengths were once again the coupons with 110% flow on the inner walls, but the top surface of them is very rough, and it is not likely we would use that setting, so I am going to stay with 100% flow.... Using all walls seems to not be quite as strong as using the criss-crossed 100% infill for the triangles.... My guess is that the infill acts like crossbracing and helps prevent movement of the walls of the test section past where the width increases.... The alternate extra wall does indeed help lock in the infill, so I am closing in on a good coupon design....

Bob

[WhatUPSbox?]

Bob, great work. That's a lot of interesting data. You mentioned that you were mostly interested in the relative data between the coupons rather than the absolute numbers. A filament strength number may make a useful benchmark though. Unfortunately the internet numbers for PLA are all over the place, with some below what you are getting with printed parts. It may be useful to pull on just a filament to get a number for your specific batch.
Thanks for sharing your progress.

[rsterne]

OK, so I redesigned the angled coupon so that the slope of the sides is 2:1, which works out to 26.6 deg.... The original ones I downloaded were 45 deg. (1:1), and the tended to pull apart on the slope, I think because the pull exceeded the layer bonding.... I got rid of the radius leading from the slope to the test section because it didn't seem to be doing and good.... I made a second version where I increased the thickness at the eye by 2mm, so there is a very shallow, stepped slope between the eye and the test section, but on the top surface only.... I ran a test at 220*C, and it seemed to work fine, so I decided to try other temperatures.... On the next set, at 210 deg. I ran out of filament, and changed to the same brand (Creality), but in white.... I ran the rest of the tests in white, it didn't really seem to matter.... Here are the spent coupons, 250*C at the top, dropping by 10* each row.... The coupons with the thicker upper surface at on the right....



and here are the results....



The force required to break the constant thickness coupons tended to be higher, and varied more than the thicker ones....

Bob

[rsterne]

I then tested some Overture "Econo" PLA.... I did them at 190, 210, 230 and 250, top to bottom in the photo....



Here are the results....



The overall tensile (yield) strength was MUCH lower than the Creality PLA (like HALF), temperature made virtually no difference, and if you look closely at the photos, you can see the stretched portion is "crazed".... meaning it has a whitish appearance due to microfractures as the material stretches.... Not a good product if you are looking for strength in PLA, in my opinion....

Bob

[WobblyHand]

How does PETG do?

Interesting on the 2:1 difference in strength between the two brands of PLA.  What's the price differential between them?

[TorqueMaster]

I suppose we should give Creality filament some due respect.  I've never heard it mentioned as good nor bad, so it's easy to forget about.  Overture is generally considered pretty good. 

This is from a "printability" standpoint -- most folks don't intentionally break their prints.

I found a similar disparity between same-brand regular black PETG and a matte black PETG.  The matte was way weaker, only having ~60% the strength.  Must be the additives.  It looks fantastic, though.

Keep on bustin', Bob!

[subscriber]

Bob, this is very useful data, and I do not even run a printer.  But I do designs for printing, based on assumed printed material strength.

A related field that would provide useful info, is the pull strength of 3D printed threads.  The male thread would be metal, while the female thread would be printed in a "coupon" form compatible with your test rig.  The sample should be able to self-align axially, similar to what your test print and pin arrangement does now.

I think many folks here would like to know how strong a 1/2" deep 1/2 - 20 printed thread is; comparing these strategies:

Thread designed to be used directly off the printer;
Designed with extra material to be chased by a tap;
Designed as an untapped hole, with thread to be fully tapped.

Certainly, the first version listed could be done the most easily, without the need to purchase a tap; assuming an appropriate threaded bolt could be sourced.

Any interest in this would be appreciated.  How can we assist?

[rsterne]

Sub, I can do that, and when I have finished running my strength vs temp tests I will see what I can do to make up the fixtures.... I would assume you want to run two sets of tests, one with the threaded hole printed vertically (Z) and the other with it printed horizontally (X and or Y).... I only have 3/8" x 1" horizontal slots in my machine to accept the coupons, with a 3/8" hole vertically and smooth 3/8" pin (shortened bolt)....

We would need a male thread adapted to fit into the moving slot.... I would think a piece of aluminum milled to fit into the slot (and pivot loosely), drilled for the pin, with the end tapped 1/2"-20 and a piece of thread rod inserted and loctited into place should do the trick.... We might need a similar adapter for the fixed end to accept a thicker coupon that would contain the female threads.... My concern is that the strength of 1/2" depth of threads might exceed 400 lbs, as the maximum I can test (without making a new lever) is 2000N - 200 kg - 440 lbs.... and if I go more than that, I might have to brace the vertical fixed attachment to the lathe.... The leverage across the holes is already 3630 lbs, spread across two 5/16" bolts an inch apart and the fixture leaning against the cast cross brace of the bed, about an inch above that.... I can install a stay and turnbuckle to the rear of the bed, with most of the stress in that stay, but if we stay less than 400 lbs force to strip out the threads I should be OK.... Whether the lead screw thrust bearings are strong enough is another question, and I really can't risk damaging the lathe.... 

You would need to tell me how thick the coupon will be, and the width (I assume you would print a cube with a threaded hole).... so that I can make the adapter to suit....

Bob

[rsterne]

Bruce, the price difference is not proportional to the difference in strength.... I will be doing PETG shortly....

Bob

[WobblyHand]

Bruce, the price difference is not proportional to the difference in strength.... I will be doing PETG shortly....

Bob

Bob, I didn't think there would be a strong relationship between price and strength.  Interesting, however, that there's such a big difference though.  Could even be lot dependent.  ("Lot" as in from lot to lot of the production of the filament.) If I had something "critical", I'd not be relying on previous tests, but would be testing the filament that I planned to use.

PETG would be interesting.  I do most of my printing on PETG now.  It seems more pliant than PLA, but it is strong and it doesn't sag in the heat.  PETG also isn't brittle, like PLA can be. 

I printed my AXA tool holders in both PLA and PETG, and I think the PETG ones will hold up better.  Think I will replace the PLA units over time.

[rsterne]

I did the testing on regular Overture PLA today.... Here are the results.... Basically it is about twice as strong as their Econo PLA.... 



Maximum strength is reached with a nozzle temperature of 230*C, and once again the coupons with the 5mm thick ends were not as sensitive as the ones with the 4mm ends.... All coupons had a test section that is 5.6mm wide and 4mm thick.... Tests are showing great consistency....

Bob

[subscriber]

Thanks Bob,

My interest in printed threads at 1/2-20 would all be printed with the threads facing up on the printer platen.  Other orientations would have general interest, but due to the asymmetrical printing induced threadform distortion, such threads should all be chased with a tap.  Printed a little tight everywhere and then tapped.

My interest is how strong the threads are VS layer to layer adhesion of the support material; and of how the threads are generated:  What percentage thread needs to be removed off a thread that screws together easily, without slop, VS one that has some play.  Also, compared to a fully taped thread, and one that has been printed overly tight, then chased with a tap.  What wall layer count one uses may also affect the strength of a tapped thread.


Smaller thread sized are more likely to be printed in any "cross-plane" direction or angle.  I think that anything at or under 1/4" should probably always be tapped.  Certainly any thread under 1 mm pitch is asking a lot of a as-printed use thread.

I appreciate that a full half inch deep thread may be too strong, so the way to handle that is testing fewer threads; until they fail at a force level your setup can handle.  Perhaps starting with two threads, then three and so on.  If 3 threads fail at 390 lb, then 9 could be deemed to hold 3 times as much force.

I do not think it necessary to burden you with all possible configurations, but some info on printed thread strength would be useful. 

One variable that one might overlook is tread strength where clamping force is generated by tightening the screw or bolt.  That making it "tight" can use up a lot of the plastic thread strength; especially for small diameters.   The way plastic threads tighten up is not like metal, where there is a "wall"; that if ignored you stretch the threads or the bolt.  I think plastic loads up more gradually, with more drag early on; then a much less obvious "wall".   

I don't know if this aspect is outside the scope of force testing, but for all of us who stripped plastic threads because they were not "tight" yet, it might be a facet of the study.  Hence my suggestion that thread engagement depth be set for tensile testing, but not torqued against anything, other the the friction in the thread generated while screwing the parts together.

[subscriber]

I would imagine that the wall thickness around the female thread would affect pull strength.  If the wall is too thin the part will stretch in diameter and allow the thread engagement depth to go to zero.  Then the test piece will slip off the metal bolt as if it were only friction coupled.  Yes, the thread tips would probably get sheared off, but not the whole thread.

How thick that printed coupon wall is, will form part of the system strength, that the part designer needs to take into account.  So, we have thread shear strength and we have resistance to bulging and slipping off, as system failure modes. 

This suggests two types of test coupon that test different aspects.  The same principles apply to metal, but the bulging and slipping off failures only occur at much thinner walls than plastic.  The softer the material, the thicker the wall needs to be so it does not bulge on the OD around the threaded hole.

The above becomes obvious if you have used thread forming taps.  A wall thickness that allows one to use a thread forming tap, where there is no plastic deformation of the OD (or at the outside wall), should be thick enough for tensile testing for thread shear strength.  If a full thread form cannot be achieved when using a thread forming tap, because of bulging at the part OD, then shearing the thread is much less likely to be the tensile test failure mode; seen when generating a thread by means of a cutting tap or printing; assuming the same nominal thread size and material.

[WhatUPSbox?]

This may give some idea of the pull forces that might be needed.

[rsterne]

2000N = 200 kg. = 440 lbs is just about the maximum I can handle.... and that was for M5 threads and the infill looks like it failed first.... The threaded "cylinder" pulled out of the infill!....

Bob

[subscriber]

Bob,

That is impressively strong for an M5 thread.  And as you say, the thread did not fail.



Stan,

Using inserts would be the smart thing to do.  Assuming they are placed with appropriate alignment.

[WhatUPSbox?]

I use heat installed inserts for the smaller threads for repetitive assembly. Probably not the right application for a 1/2-20 LDC type application that is tilt sensitive. Buckrail site says they use an e-z lock insert. Sounds like a good approach.

[subscriber]

From Stan's video, another one comparing types of insert:

[rsterne]

Threads create a very high axial tension when tightened....

https://www.engineersedge.com/calculators/torque_calc.htm

For steel threads, unlubricated (coef. friction 0.2) a 1/2" thread (pitch does not matter) tightened to just 10 in.lbs has 100 lbs. of axial (clamping) force.... At 10 ft.lbs that increases to 1600 lbf.... 

Bob