Tensile Strength Testing

[rsterne]

Yes, if I move the pivot on the moving bar so that it is fixed (relative to the existing fixed end), and simply have the carriage move the force gauge, you are correct.... Not sure what you mean by "in between".... However, the bar will tilt by the multiple of the leverage, so if the coupon gets 5mm longer, at 4:1, the end the gauge is mounted to will move 20mm, and therefore be many degrees out of square, which could affect the measurement.... Right now, there is no angle change at all.... 

It might be a good design for measuring thread strength, however.... but realistically, for a threaded connection, isn't the maximum torque you can apply before failure more important?....

Bob

[rsterne]

OK, so yesterday and today I ran some iSanmate Wood PLA+ that I had, and the 20% wood filler certainly weakened it.... Here are the results....



Interestingly, the minimum temperature recommended on the reel of 190C produced the strongest horizontal (X) coupon, but at that temperature the layers were not at all bonded well, and the vertical coupon failed at only 1260 psi, and when it did, it failed in two spots!.... The strongest layer bonding occurred at about 210C, and at that temperature this Wood PLA+ was less than half the strength of Overture PLA.... It's great for appearance (but strings at over 210C anyways), but not very strong....

Bob

[WhatUPSbox?]

Not suggesting you change your current test approach. It sounded like you and Subscriber were coming up with a coupon for 1/2 inch thread that might impose high loads on your lathe.

[subscriber]

realistically, for a threaded connection, isn't the maximum torque you can apply before failure more important?....

Not for the applications I have in mind:  A 1 to 2" OD cylinder is screwed onto the barrel muzzle by hand, until it is tight enough to square up on the mounting shoulder, and not move or loosen during the shot cycle - or after multiple shot cycles.  Tighter than that just uses up thread strength, that would otherwise prevent the "cylinder" from blowing off.

[subscriber]

Stan,

We certainly do not want to risk damage to Bob's lathe. 

I suspect that for decent prints, the thread strength will be enough.  So, if a typical depth of 1/2-20 thread can stand 400 lb, then I would consider that a proof test. 

A typical thread depth is 10 to 13 mm based on barrel stud lengths I have seen recently.  To predict the failure load for a 1/2" thread depth or 10 mm for that matter, reducing the engagement to 1/4" or 5 mm and then testing to thread failure would be very reasonable.  If 1/4" or 5 mm thread engagement exceeds 400 lb, then use 1/8" or 2.5 mm thread engagement, until the test part fails.

If we have to rely on too few threads to avoid exceeding 400 lb, then the test repeatability may become poor. Then perhaps using a full 1/2" or 10 mm long thread would be better, but the thread form would be configured to be shallower to reduce it strength that way.  The exact plan can be refined, based on early results.

How much force does a muffler apply to the barrel stud during firing?  One way of measuring that indirectly is  to design a series of short test mufflers that have thread depths at 10, 5, 2.5 and 1.25 mm, or equivalent inch measurements, starting at 0.5".  These mufflers would be test fired in turn until the thread stripped and the muffler flew off.  Or, a similar series of muffler where the thread axial depth was kept at 10 mm (or 0.5"), and the thread form reduced for less engagement, till the muffler blows off.

These blow off test results could be correlated with Bob's pull force measurements.  So we know what we need; and what we got.

[rsterne]

Yes, Stan, thanks for that.... I certainly would not want to overstress the lead screw thrust bearings.... or the vertical support that holds the fixed end now....

sub, if I get the time I'll try and get the adapters done in the next week or so.... I would start with 3 threads (0.150") of engagement and see how close that gets to (or exceeds) 400 lbs.... I assume PETG?.... Has anyone done any checking into the hoop stress of said muffler.... Having one grenade would not be desireable.... 

BTW, I have read that increasing the number of walls until the object is solid (and using no infill) produces the strongest prints.... That is what I am doing on my test coupons, the walls touch throughout the (horizontal) test section....I would think that using the round threaded portion would be an excellent candidate for that approach, it would take about 9 walls (@ 0.4mm) to do that....

Bob

[rsterne]

Just doing a strength calculation on the round section.... It will be printed in "Z", so the layers will be trying to pull apart.... The area between the recess ID (14mm) and the OD (25.4mm) is PI (25.4^2 - 14^2) / 4 = 353 mm^2 = 0.54in^2.... My PETG "Z" Coupons fail at ~3600 psi, best case (240C), so that should be good for (0.54 x 3600) = 1944 lbf, or almost a ton.... In theory, over a 4:1 safety margin higher than what I will pull on my setup.... So, a 1" OD it will be!....

Bob

[subscriber]

Sounds good to me, Bob

[rsterne]

I downloaded the square top .zip file.... I see you have 45 deg. chamfers top and bottom of the thread.... How many "good" threads do you have in the "use as is"?.... If I drop it below the platen to shorten the threads, that chamfer makes it difficult to know what I get.... BTW, printing time, all walls in the threaded section, is 2 hrs. 1 min.... I'm thinking we should just have straight thread, no chamfers, with 10 threads (1/2") total.... What are you using for the OD of the "use as is" threads?....

Bob

[subscriber]

Bob,

I removed the chamfers from both ends of the threaded sections, on all three test pieces. 


CAD Dimensions:

The threaded sections are 12.7 mm long

The thread root diameter is 13 mm

The unthreaded cylindrical section is 14 mm in diameter


Three STLs in attached ZIP file

[TorqueMaster]

OK, so yesterday and today I ran some iSanmate Wood PLA+ that I had, and the 20% wood filler certainly weakened it.... Here are the results....

Interestingly, the minimum temperature recommended on the reel of 190C produced the strongest horizontal (X) coupon, but at that temperature the layers were not at all bonded well, and the vertical coupon failed at only 1260 psi, and when it did, it failed in two spots!.... The strongest layer bonding occurred at about 210C, and at that temperature this Wood PLA+ was less than half the strength of Overture PLA.... It's great for appearance (but strings at over 210C anyways), but not very strong....

Wood filaments are notorious for being more for looks than durability.  Some of them have real wood, and will darken more or less depending on temperature, and how burnt the wood gets.  I remember reading about someone who post-processed the gcode file to pseudo-randomly change the temperature, within a specified range, to simulate "wood grain."  The prints will kinda-sorta take stain, too.

Tests like these are great -- some filament vendors seem to cut-and-paste the same temperature range on all their similar products, without even testing, apparently.  Trust, but verify.

[rsterne]

I will test any filaments that anyone wishes to send me.... I need a minimum of 100 grams (33 metres) of filament, assuming no failures, and it takes me a couple of days to run the tests in both the X and Z axis.... If you have half/part rolls you wish to donate, I'll run them for you and publish the results here....

Bob

[subscriber]

Bob,

I don't have partial roles to donate, because I don't have full rolls.  Being as I like to watch people who know more about printing do their thing.  But, I get that expecting you to buy a lot of material to amuse us, is not as much fun for you.  So, the only way I can send you material is to have it shipped directly to you, via amazon or your preferred source.

[WhatUPSbox?]

If the motivation for the 1/2-20 tests is the pressure in an LDC, perhaps that can be explored. Instrumenting the LDC with a pressure gage or strain gage would be great but would take some resources. What about a sacrificial wall section in a special test LDC. The image show the basic configuration. The top section would continue on with the rest of the LDC features. If a one wall (~.4mm) thickness stays intact, one could castle it and tape the insides to seal. Once a configuration fails, it could be compared to a pull test.

[subscriber]

Stan,

I think that making test LDC with thin section to see what they can stand has merits that go beyond correlating back to thread strength.  Stain gauge readings could be interesting.

The problem with LDCs (or the first expansion chamber in particular) is that they are not a known closed volume, that can be relied on to reach a calculated pressure, when a barrel full of air at a known pressure dumped into them.  The bore is a massive leak, so pressure does not build up like a closed vessel.  On the other hand the way the air enters the LDC is more like a car crash than using the brakes.

Again, any clever ideas we can come up with for test purposes would enable lighter LDCs to be designed, so they have enough strength where they need it, and no extra material where they do not.

All that said, if the first baffle has an effective area of only 1 square inch, then if the residual bulk pressure is 200 PSI after expanding into that chamber, the baffle sees a very rapidly applied 200 lb trying to rip that baffle forwards, off from the base.  Providing the structure survives, I think it is the flow resistance out of the bore that is trying to rip the LDC off the threads.  Because a closed pressurized container would have all the the fore/aft forces balanced in the walls, with no pull on the threads.  Then there is the radial burst pressure that is superimposed on the axial forces...

I have suggested some single known expansion chamber volume LDCs, with a pin hole vent, dry fired to see what wall thickness is required to stand that.  That idea was not immediately popular with the other Bob posting in this thread.  A proposal the other Bob made and already implemented was an Avenger LDC inserts where the section ahead of the shroud had a 1 mm wall.  Another with 0.5 mm wall was also tested.  These were surprisingly robust; but, those sections had the benefit of the reflex expansion volume in the Avenger shroud, and over 3" of shroud ahead of the muzzle to drop the bulk pressure.

Keep the ideas coming; although I think Bob's thread should be more about tensile testing samples, strategies and results.  Even though my thread testing interest is obviously linked to retaining LDCs.

[WhatUPSbox?]

I think Bob was going to make some custom tooling for the thread pull test. If he has not done so already, perhaps the test can be simplified to a double sided pull just using cut down bolts or all-thread for the tooling.

[rsterne]

Muzzle pressures can vary from a few hundred psi to nearly full tank pressure, if the reservoir is large and the dwell longer than the time the projectile stays in the barrel.... The volume of the barrel can vary from 5cc for a 12" .177 barrel to 116cc for a 36" of .50 cal.... If we look at a .25 cal 25" barrel, the volume is 20 cc, and I doubt the first chamber in most LDCs is larger than that.... If you have a 200 cc reservoir filled to 3000 psi, and the valve is still open when the projectile exits the muzzle, the residual muzzle pressure is (3000 x 200 / (200+20)) = 2700 psi, and with an equal volume LDC chamber, that would drop to (3000 x 200 / (200+40)) = 2400 psi.... Think about that when you pull the trigger with a 3D printed LDC....

Now I realize you probably won't be running an LDC on a gun with maximum residual pressure, but if the valve closes when the projectile is halfway to the muzzle (10 cc of barrel volume) the residual muzzle pressure is (3000 x 10 / 20) = 1500 psi, and with a 20cc LDC chamber it would be (3000 x 10 / (20+20) = 750 psi.... not insignifcant....

Stan, that is a brilliant way to simplify the problem.... Make the sleeve with 1.5" of threads, and thread one piece of rod in halfway (a full inch).... On the other end, count turns, starting from 2-3 turns.... If you never exceed 1/2" on the "test end", the other end should not affect the results enough to worry about.... and each test sleeve is disposable anyways....  .... I wonder how strong the shank of a 1/2" bolt is with a 3/8" hole drilled crossways through it.... hmmmmmmm....

Bob

[WhatUPSbox?]

I think the retaining pin can be undersized. More slop helps self alignment. I'm guessing that the shear strength of a 1/4" rod is higher than you want to apply to your lathe.

[rsterne]

Area of bolt shank = PI/4 x (0.5^2) - (0.375^2) = 0.785 x (0.25 - 0.141) = 0.085 sq.in.... A grade 5 bolt is 92Ksi yield strength, so that is a minimum of 7872 lbf.... Should be more than enough!.... I just need some 2" or longer shouldered bolts!.... Here is a design, 25 x 25 x 40mm with 13mm (1/2-20) threads.... All walls, prints in 1 hr. 12 min....



The .stl file is below.... I'll just drill a couple of 1/2-20 bolts and machine the shanks flat to thickness to fit my setup....

Bob

[subscriber]

What is good about Stan's simplified thread specimen is that one will have two data points in one test piece.  The load at failure, and what the thread looks like the end that did not fail, but that took the same load.  I am assuming only one side will fail, most of the time, like a wishbone.