Surprised you didn't try nut pockets (what prusa does with their printer parts). That creates a nice solid wall that isn't distorted by brass teeth. I typically use wood screws if it's a one time use, and the inserts if it's a part that may need dis-assembly (also looks more professional). In injection molding it's exactly the same, if it's a one time assembly then self-tapping, if it needs to be serviceable there better be inserts included.
This is my exact design philosophy. If it's going together once it's self tapping(even machine screws self-tap in plastic well if the tolerances are low enough) and if it's going to be taken apart multiple times it's getting threaded inserts.
I design most of my prints with a nut pocket. then I will stop the print in the slicer, insert nut, and continue print. this allows me to use a lot of off-the-shelf hardware store nuts and bolts without having to buy more expensive specialty hardware. But I'm also not printing high volume, so it's easy for me.
I think it could be interesting to see how the number of perimeters around the hole and the type of infill (geometry and density) affects the strength of the insert (using CNC kitchen ones for example)
I did something like this on my printer, but ran a battery two tests: number of perimeters, and diameter of the hole. I incremented the hole in .05mm increments and tested about 20 options, but I didn't have a fancy torque setup so it was all eyeball and hand-feel. Once I did that I was able to save custom hole profiles into my cad tool to make those custom holes.
I was thinking similar thoughts about increased perimeters and/or increased infill density. My spin on this would be to use cylindrical modifiers around the hole to apply the perimeters and/or infill density localized to the hole instead of on the full print (that presumably doesn't need the extra plastic).
It's crazy to me he didn't go with higher infill and more perimeters. I always beef that stuff up anywhere I'm using inserts. Also, it looks like his prints had poor bonding of the top infill layers to the hole perimeters. I have a less "pretty" tuning setting that really overfills that to make sure my hole perimeters are bonded to infill everywhere possible.
Flip the camera inserts to the other side of the printed piece. That way you have a larger flange you are pulling against the print. Essentially acting as a larger backup washer.
Also, doing the same thing with the heat-set inserts. Pull them THROUGH to the far side, rather than pulling them OUT of the side that they were pushed into.
Also, there's the possible benefit of the recess used for tightening the insert mating with the plastic and preventing twist out when tightening and loosing the bolt.
Test it with a regular hex nut countersink in a hexagon hole from the back on the part or a square hole in side the part and push the nut in from the side. I do this all the time and it works perfect and does not require any special nuts.
Friendly tip - for anything starting at M4+, I am not using inserts for plastic, but inserts for wood. They have larger diameter of the metal part. And I am heating them up and screwing them to the plastic - significantly increasing the contact area. Works much better. And the best part - you can then screw in the bolt from the other side, make the insert invisible and even stronger :)
The use of those traditional taps may be why his simple thread an m2.5 hole.did not work. I have a spiral tap and, as long as I have a few extra wall layers, find it works perfectly every time. The only thing I would like to find is a way to simply set extra walls on holes only without a lot of fiddling.
Ah yes, I was a bit surprised that tapped holes worked so poorly considering how often I use it for non load bearing parts. My taps are also spiral fluted, and I don't power tap them so that is probably why I have better luck with it. I wouldn't necessarily recommend it, most of the time I do use a captive nut or insert. But it isn't that bad.
Single-cut taps are the choice when using machines for manufacturing. They require a higher torque because they finish the thread in one pass. They are also more sensitive to torque peaks that occur when cutting a thread by hand with a tap wrench. In this case, a tap set consisting of a pre-cutter, center cutter and finish cutter is used.
Perhaps consider printing a 100% infill 'doughnut' of say 10mm around each insert/tap/thread point to eliminate some of the variability caused by your part flex and infill interface, along with using an omnidirectional infill like Gyroid. It would improve pull-to-pull consistency and more accurately demonstrate the ultimate/average strength of each method.
This is exactly what I do. In the slicer add a cylinder modifier with 100% infill (usually just larger than the diameter of a standard washer you would use for that fastener. I've had no issues with part failures for my applications going this route. I make a lot of sensor mounts so there isn't generally a lot of dynamic forces on them, mainly just static loads.
I typically use self-tapping 3mm screws in 2mm holes with no printed thread. Even for parts that need occasional disassembly and re-assembly they work great. And let's not forget that lots of injection molded cases do the same thing. And in my experience PLA parts survive unscrewing and rescrewing better than the industrially produced ABS parts. As I'm writing this I'm printing 3.5" to 5.25" mounting adapters that I designed today (because the ones on Thingiverse are all just imitations of mass produced parts and these designs work poorly for 3D printing) and the means of fastening the adapter to the PC case is by means of 3mm self-tapping screws. I always use the longest screws that will fit.
yeah, self tapping screws: fast, easy, cheap and often plenty strong. If strength is a question, bolt and nuts. I only rarely print things that I plan to take apart multiple times.
I skip the self-tapping part and just use regular M3 machine screws and force them on the hole. Power tools will melt the plastic, but if hand tightened, they usually work just fine. "Just fine" as in the level that I never opened the insert box I ordered a year ago.
This is the way. Although I design my holes to be 2.5mm for m3 screws. I very rarely have problems with this kind of self-tap and it's always self-inflicted harm. ( i.e. impatient assembly, bad design, faulty print, etc)
I too use regular M2, M2.5, M3 and M4 screws as self tapping screws (by hand). I print a hole smaller than nominal and using a separate slicer modifier solid (prepared in CAD), I add a 2mm thick ring of plastic (perimeters) just around the holes to be later tapped by the screw. If you use long screws as shafts (sticking out of the part) this method will effortlessly give you perfectly straight (perpendicular to surface) shafts. A thing almost impossible to achieve with inserts (that are impossible to sink in perfectly perpendicular to the surface). And you can screw and unscrews many times without ruining the tapped hole. I do not use inserts anymore for pieces not intended to be screwed/unscrewed repeatedly. I use PETG.
In combat robots, we all use plastite screws. They have triangular lobe coarse threads. You just screw them into the plastic, no tapping required. A common failure mode we see are heat set inserts pulling out, but the plastite screws fare much better. We also tend to use a lot of walls since we’re printing for strength. Usually 3mm walls or more. Plastites are also lighter than inserts.
A handy tip for increasing the number of perimeters around just a hole: You can put an infinitesimally thin annular gap spaced out a few perimeters lines from the ID of the hole. This will force the slicer to an extra set of perimeter lines around the hole without having the increase the number of perimeters everywhere. I'll actually put these phantom gaps anywhere I need to add solid perimeter lines in an area that would otherwise just be infill.
@@SixTough More like a cylindrical gap of, say, 0.001mm wall thickness but the height of your screw. The radius of the cylindrical gap should be larger than the original screw hole size by several (4 - 6) line widths.
@@SixTough nope, ctting a 0.0001 wall thick cilynder around the hole full depth, or maybe two concentric at 1 mm distance from eachother (assuming you print 0.4 mm wide layers), that forces the slicer to put 3 walls around the hole, another 3 around the first gap and another 3 around the second gap, but the gaps being so narrow the walls for each circle will melt into ecother giving you a 9 walls print.
Would be more interesting to test screws made for screwing into plastic, like "plastite" instead of screwing ISO screws directly into plastic. Right tool for the right job!
@@shonkydonkeyThey actually use them a lot on cars. Both on plastic and metal parts. I think so they clear any paint or debris out of threads on the way in. You can usually tell because they will have a small slit in the thread on 3 sides. Sometimes the tips are also slimmer than the threads or look slightly triangular.
support this, but the difference is the reuse ability of such threads as the plastic weakens every time you reinstall the screw. The question is how often can you do this with the same load.
Did you test torque-out at all? Because that pull-out test is honestly almost entirely pointless as that is just not the way you use them in, dare I say it, almost any case.
In my experience the screws for wood work much better. The hole in the printed part should be a bit bigger than the core and since these screws have a very deep and sharp thread, they usually connect very nicely with the part. The failure point is typically the connection between the printed hole and the rest of the part. For tapping the thread into PLA i found that a decent lubrication makes the process almost perfect. I typically use some grease since it does not flow away unlike oil. By the way adding a little of grease on woodscrews eases the joint process a lot. Cheers
@@fred-9929this can be true, but straight-shank wood screws are far easier to find in your local hardware store and orders of magnitude cheaper than ordering a small quantity of plastic -specific screws. If you size the hole 1% larger than the nominal minor diameter of a wood screw it wont be any more prone to crack.
@@fred-9929 Drywall screws are available in fine and coarse thread. Fine thread works well, and is even self drilling in thin steel with the good quality ones, and I have used them a lot that way, even on steel a lot thicker than the thin steel used in drywall work, like steel tubing with 1.2mm wall thickness. The coarse ones, or those designed for composite wood, are very good, and almost always stronger than the substrate, but are generally not going to self drill in anything other than composite board, or through 0.6mm steel plate..
I usually embed nuts inside the print by modeling a cavity and pausing the print at the top layer of the cavity. Nuts are usually a lot easier to find for a particular size than threaded inserts, and more broadly useful. If you need even more strength, modeling in space for a steel washer between the nut and the inside wall of the part increases the engagement with the plastic part by quite a bit.
For the past 5-6 years I’ve just self-tapped into a tighter printed hole and had excellent results that way. Yes, occasionally I’ll use a nut on the back and once or twice inserts but for 90% of the loads I put my prints through (including my CNC machine), the self-tapped printed holes seems just fine.
Same, and I have hundreds of products (modified, industrial card printers) out in the wild assembled like this and not a single fastner (mostly M3) failure.
My favorite is the wood thread insert with spikes on a flat washer on the back of the part. Also always add some extra walls around the thread and it won’t pull out but generally I’d print a section of 100% infill there anyways. If you can get 4x diameter on the hole depth, printed threads chased with a forming tap are wicked strong. And like others said, burying a nut in the print in some way is always a super good option. If the nut is near an edge, you can add a keyhole slot on the side to slip the nut in and then screw through the front. I had to make a super tricky assembly once and used a thick 3mm washer tapped for an m4 bolt and slipped it in through a side slot that got plugged with epoxy to hold the threaded washer and plug in.
I like to use “alternate extra walls” and “connect infill lines” to really boost the strength of printed parts. The perimeters become interlaced with the infill structure and resists pullout and deformation. I use grid for faster prints but prefer gyroidal for strength in all directions or for parts susceptible to warping.
FYI, when you tap plastic you should hand tap not use a drill, you will heat up the part and melt it. Also do 1 revolution and clean out the tap, then continue the next revolution and cleaning until you are done. If not you will mess up the threads. Also don’t lube the PLA, oil will weaken it.
I'd be really interested in seeing the results of different infills as the failure mode was always to rip the connecting walls, I think that'd be much more useful for figuring out how to get the strongest output for each usage case!
The vast majority of installation torque (80-90% in steel nuts and bolts) goes into overcoming friction so it's not a good way of measuring pull out force. Once installed the assembly strength is determined by bearing and pullout strengths, not twisting of the insert.
For extremely small screws (M2, M2.5, M3) I tend to just screw directly into an oversized printed hole; basically the tapping method, but skipping the tapping step. I get fine results, but I'm also never designing parts for strain, just assembly.
Two ideas I have are creating a wide ring in the middle of the screw hole that acts like an anchor in the shape of a washer. Or creating a recessed area on the backside to actually install something like a fender washer to spread the load out if you really need the strength.
Please also test the strength of self tapping screws. The "threads" of those, are spread over several layers, and seams to make a quite strong bond. It's a way quicker solution to use, that I have had good succes with.
This video needed to be made, thank you. Please test nut captures and the strength of inserts from the opposite side. Tapping holes with a hand drill is something I never do regardless of material as while it takes a little longer I find that doing it by hand results is less damage to the threads especially for fine pitch threads.
First year Uni manufacturing here in Australia taught me that a bolt has a un-threaded shank portion where as a machine screw (not a wood screw) is threaded all the way to the head. Love the testing videos, keep up the great work.
Actually, the only difference between a screw and a bolt is what you call them. They are literally the same type of fastener. Don't get too caught up on the name.
@jetah50 the problem is that any supposed difference is easily disproven. Bolts have a smooth section? Fully-threaded bolts (often called "tap bolts") are a thing and are commonly used to attach automotive body panels. Bolts used with nuts? Lycoming cylinder bolts are installed into tapped holes in the crankcase. Screws are only designed for tension and torsion loads? Nylon shear screws are commonly used to attach model rocket nose cones to the rocket body in order to prevent unplanned separation. And I could keep going.
The only difference I’ve been taught and that seems to consistently stick is that bolts mate with a tapped thread whereas screws cut their own. Machine screws simply have a tooth profile capable of carving into the metal/plastic. Wood screws cut their thread into wood but I don’t think I’ve ever seen a wood bolt, because the material being held doesn’t matter to a bolt, as it knows it has a matching tapped thread to fasten with. A screw on the other hand needs a different profile, pitch, taper etc depending on the material it’s self-tapping into.
Hi Thomas, first at all thank you for your always excellent videos. If this video, you are in fact testing resistance vs. torque which is fine to know how the attachment will resist to an over torque. But according to me, should you be able to test resistance vs. axial force, this will provide tremendous inputs on how these fixtures resist under load. And we could have some surprise with plastic tapping. And should you be also able to test self-tapping screws at the same time, this will give us a whole overview of attachment resistance.
I find myself wondering about the testing methodology. When turning a screw in PLA, won't the friction heat/weaken the PLA giving the metal on metal solutions an unfair advantage?
This is a fantastic video! You mentioned that you solder lead free. Could you make a video on how you do this and what solder you use. I have been trying to use lead free solder for a while and I just can't get the solder to cooperate.
Few other types of threaded inserts: "thin wall threaded insert" and "helicoil" both would be interesting to have in the comparison. As another comment mentioned, would be interesting to know print settings and do a review on things to improve the joint. I would also love to see how using a "matching" grade of screw would affect the strength of joint (i.e plastic screws). Ace video as always Tom, loving the new clean feel of the production.👍
Thin wall inserts will be about the same as going up a size on the bolt, but the hole will last a lot more insert remove cycles. Probably best is any of the inserts, and make the wall of the hole a full fill to double the diameter, which will allow better transfer of stress to the bulk media. As most of the failures come at the join of the insert plus fill to the bulk void space, a better bet is more material there, either denser for a 2 diameter volume, or better infill that uses more filament, but at least handles load better.
Great video 👍 something that people don't often talk or think about is that in many cases threaded inserts aren't necessarily "stronger" than direct threading...but in applications where you need to often remove a threaded fastener (eg. a thumbscrew) they are a must (for example helicoils in aluminum).
This should be upvoted to the top! The helicoil example @emberprototypes gave is fairly common knowledge / industrial practice. Aluminum is too soft for repeatedly bolted/unbolted joints. They mentioned "often" but more realistically even one removal and reinstallation is too many - bolts into aluminum is often a one-and-done thing and unreliable to remove and re-bolt. The helicoil at its simplest form is just a tiny steel thread to increase the wear resistance right at the joint interface. This is doubly true for plastic.
Large diameter steel washers can spread the load over a wider area of the plastic when used in conjunction with a steel not. A thread for the screw can be included in the hole so that a lockout effect can be achieved with the steel nut.
I personally think, the Torqueout Test won't give completely comparable results, because of different coefficients of friction (galvanized steel bolt - PLA/brass/Nickel plated brass). A pullout test would give indefent results, sadly requires a test machine. But I think the "ranking" would be quite similar. I tend to use woodscrews for simple "once" connections, Bolts in printed threads if the simple connection needs to be a bolt for other design reasons. Sometimes printing only the first few millimeters and let the Bolt cut the rest by itself (friction stops unwanted rotations). Heat inserts only when I need wear resistance, if even more is needed Ensat-Gewindeeinsätze. Really strong threads I get by designing in massively longer threading depth or a nut placed inside during the print.
Fortunately, I'm not the only one who sees that this test is worthless. Twisting force is measured here, not pulling force. It's nice that the coefficient of friction between the different materials is measured here, but that is of no use to us. With a drop of oil on the thread it suddenly becomes a bad result. And a bad thread becomes a good result.
Another thing which is worth noting, is loosening of the screws over time. Having a metal threaded insert allows you to use a more diverse range of threadlockers (e.g. Loctite 222 or 243 etc). I designed a custom hotend shroud and ended up simply under-sizing the holes and forcing machine screws through as "self-tappers" which holds pretty well, but the screws do end up getting loose (I think it's because the bowden tube puts varying pressure on the hotend as it travels across the x-axis). Something else worth testing, is instead of using machine-screws, use actual thread forming screws meant for self-tapping into plastic. They tend to have a sharper, more coarse thread and are often seen in injection molded parts.
I prefer cut threads for everything below m6. I use 3d printing mainly to prototype parts that eventually get machined. 99% of the time some cheap PLA is sufficient for my needs. usually I print with 3-5 perimeters. I usually use alcohol as cutting fluid. few drops of cutting oil dissolved in the alcohol works even better. also aluminium taps work way better then normal ones (the ones that have every other tooth missing to reduce friction) through hole whenever possible, bottoming taps that push the swarf up only when really necessary. ...in 3d printing I can make the hole super deep or all the way through the part even if I'm not going to do that on the machined part for obvious reasons. a final note: for embedded nuts: get some square ones - they work better in every regard I can think of compared to hex nuts when it comes to incorporating them into 3d prints. (don't rotate as easy, easier to design the pocket/slot, usually better bridging if pocket has to be upside down...)
Yeah, I bet hand tapping vs power tapping would reduce the friction and heat generation of the tap in the plastic...thus preventing the ripping of the threads out of the part However hand tapping a buncha holes suuuuuuuck I wonder if a better lubricant such as silicon grease would work better than denatured alcohol (I use Danco food grade silicon grease from lowes on my parts and bearings on my 3d printers/cnc machines, not only is it slick as greased owl crap...but it doesn't attract dust like oils and "dino" grease)
I use a tapping arm. For my cordless drill to tap. Having the tap straight also reduces the friction quite a bit compared to just cordless drill. Alcohole does some minor lubrication but is great at cooling.
Tom, I doubt you'll even see this but, we need a machining channel from you. I also see you for your blondiehacks reference! Been watching for a decade or so. I can't thank you enough for the knowledge you've shared.
I would love for you to test the direction of pulling out, AFAIK, Voron parts sometimes have you thread in the direction of installation and sometimes opposite. Taking your thread adapter into consideration, the large flange on the end would take a LOT of force to pull out in the opposite side of the install. Also it would be nice to know how much the different techniques of linking the perimeters to the infill would work! Great video!
Also they sometimes put inserts in a way that the other part will made contact with insert ( all force will be on metal ) and plastic part where insert is will be on that insert.
I mostly use the modeled threads, but if I need the strength I will use a normal nut pressed in the back of the part. Where I think the inserts shine is with parts that need to be disassembled many times. The modeled threads are only good for a couple assembly cycles and a nut can sometimes get out of place if the press fit is not great.
my favorite by far is taking one of the screws I'm going to use for the project and then cutting relief slots at the tip with a Dremel and cutting wheel. I then use that one screw to tap and clean all the screw holes leaving perfectly matched treads. The end result is better than printed threads, not quite as good as using wood screws but there are more head shape options and definitely not as strong as bolt and nut but definitely less fussy.
Its also worth noting that one should appropriately break the chips as you tap a tread depending on the style of tap used. Most cant handle or do a very poor job if you simply screw in all the way with out backing back out every couple of turns. Additionally, PLA melts, shocking, so you do have to go slow even by hand enough heat can be generated to destroy your threads during the tapping process.
Hi, I use the best combination of materials. Let me explain: if I'm working with plastic, I also use a plastic screw. But too small sizes often do not allow this. So I place the empty space under the female screw and then pause the slicer on the last layer to place the nut there. As a result, I have a part that fits the hole perfectly, has the correct threads and the right material hardness. I would be glad to see similar tests with my method of making threads
Super interesting. Personnaly I usually extrude the shape of the nut from the other side of the part. thats my prefered solution as you often get the nut provided with the bolt, you dont have to buy plenty of insert or spend the time inserting them. Maybe you can do an addon video and compare to these results to this method, that would interest me a lot. Thanks !
I am surprised you tried to tap the holes... what i do is print a perfect hole (lets say 2.9mm in diameter for m3) and then slowly let the screw make its own "threads" while going in, I found its a really nice and strong solution... probably stronger than modeled or tap holes...
You should be able to test for lead content of the Alibaba inserts using a swipe test. These are marketed for use in testing house paint and toys, but would work for this purpose too.
Most brass is leaded brass (2-3%), its not really an issue unless you: use it for a childs toy, use it for food storage, or grind it after the fact. You shouldn't be using 3D prints for these purposes anyway.
@@jaro6985 My point was focused on whether you should buy CNC Kitchen's brass inserts because of being lead free brass. Or alternatively, since there is a lot of plumbing fixtures that are in-fact lead free, is it possible to get cheap inserts that are still lead free. And I'm sorry, but using 3d printing for toys is a pretty standard use case.
@@richdobbs6595 3D printing toys for young children is just as unsafe as using leaded inserts, most people do not have food grade setups. I'm telling you all china inserts should have lead in them, so no need to test.
Lead free bass literally means only 2% lead. This is also true for lead free brass that is used in water supplies. Lead free is a legal term not a normal english understanding. It is very misleading.
A few things spring to mind here: 1) If the limiting factor is the perimeters around the hole, beefing that up to the point where the insert turns in or pulls out seems necessary to get meaningful test data on the insert or thread. 2) I don't advise tapping under power for this sort of thing - use a manual hand-operated tap wrench. It'll keep the heat down and you'll be able to feel what you're doing. I've actually used threaded inserts and stainless machine screws to replace self tappers on bits of plastic under-bonnet trim on my car - they are a huge improvement! 19 years of taking those old self-tappers in and out doing maintenance and they didn't really tighten up properly anymore. A great quality of life modification.
I've found that friction threading fasteners into unthreaded holes works quite well.... What you do is you rapidly turn the fastener in about a 1/4 turn then back out a little less than 1/4 then back in a 1/4 until the fastener is seated. The friction from this rapid back and forth heats the fastener so it doesn't cut rather it melts it's thread into the plastic. If you size the hole correctly it will take a little bit of muscle to seat the fastener and you will be able to feel the resistance slightly lower when the fastener starts melting it's way in.... The result should be a tight thread with it's layer lines fused together. Something you will not get from a printed tread. Also this method of threading does not clog the hole or remove any plastic. The plastic is pushed outward into the walls, if anything leaving the wall thickness slightly higher.....I've used the word fastener in this description over bolt because I've used this method with screws as well as bolts up to 1/4-20. Other than that if I just need a thread in my print where the bolt will be repeatedly adjusted I heat set an ordinary nut into the print. Which is cheap strong wears good and I don't have to order anything online to complete my project. With heat set nuts I print a hexagonal recess in the hole slightly smaller than the nut and push the nut in with soldering iron set to as low a temp as I can where the nut will still push in. Like with fiction threaded fasteners no plastic is pushed down into the hole. The plastic is instead also pushed outward increasing the wall thickness. So far I have had zero issues with pullout of friction threaded fasteners and heat set ordinary nuts have only failed in bad designs and when over torqued.
Yea, dont power tap plastic dude. Especially fine threads. This is the second video of this channel I've seen today where your process is severely flawed because you have no idea what you're doing. How about doing some research first. Blind leading the blind.
I switched from stocking machine screws to self tapping plastic-threading screws a little while ago and have had a fantastic time with them! Most of my parts aren't particularly load-bearing, but need to be held together... They're just so simple to add to a design!
I am just saying this before watching... I print all my threads on my designs and don't have an issue, so I am very interested in the results. EDIT: now having watched it I can see how the inserts would have a benefit but unless there is a need for more strength I'll stock with printed
I see you teased the rivnuts. I've been using rivnuts as inserts for a while. I design for them to be inserted from the back side so that the flange can act as an anchor. The resulting connection is more robust than typical inserts.
Great idea to run this test. I did 3 years ago a similar test for my machine designs and found out Prusa Pla is the best PLA for this application (the strongest, less ductile) and actually don‘t put a insert in and don‘t tap the hole. Just put around 4 perimiters and let the screw tap itself while screwing it in. With PLA this generates that much heat from friction that a perfect new thread forms around the screw, which additionally acts like loctite and is self locking. Takes some time to find the right hole diameter for each screw, but worth it.
Thomas, in tru holes you are blessed with having the possibility of using the insert from the backside of your workpiece. which creates the fact you have to pull the insert not only out but also tru the complete printing. And a bit of a hassle but you can also demonstrate the "print in place" possibility, if someone might be interested. possible with an ordinary nut.
When I played with this situation (using PETG) I found that unless you are constantly assembling/disassembling, that just drilling an undersized hole in the fastener-area that was 100% infill, that the raw plastic was just as strong as the inserts. In a couple of cases, it seemed like maybe the insert was better but not consistently. Finally, the only times I use inserts is if I want ease of assembly/disassembly cycles or for a customer that is under the belief that they just gotta have them! Thanks for doing the tests. One last thing: I also found that after a couple of screw-unscrew cycles that adding cyanoacrylate glue worked well to harden the plastic screw-hole. I also found that the longer the screw remains in the hole the more the plastic confirms around and makes it tighter. I do feel that if you would just drill undersized holes instead of printing them, you would get better results not using inserts.
Thank you for putting in the research work! Since you asked what to check next: Self-tapping screws, wood screws, inserting nuts, use of washers (larger area might not pull out as easily?), using the inserts from the other side (pulling through the entire part instead of pulling them back out where they inserted) and of course more walls / more infill / both...
What if you push in the insers into the opposing side of the hole of where the screw goes into? I imagine the additional plastic the insert will press against when you try to pull out the screw will add a substantial amount of resistance.
Whenever the design allows it, I press hex nuts into the back side of the part with a clearance through hole to screw in from the other side. Make the nut recess compliant to avoid splitting the 3D print. PrusaSlicer has a couple of tricks that help. Modifiers allow more perimeter layers and/or more infill to strengthen the 3D part around the threads. PrusaSlicer also has negative volumes that can increase part strength in a particular direction that can be useful to support loads attached with threaded fasteners. There are slicer tricks for bridging over the top of the inserted nut pocket, with RUclips demonstrations. For smaller inserted nuts, typically M4 / #8 or smaller, using square nuts instead of hex nuts will help prevent the nut from spinning in the 3D printed part.
for low stress/torque, for M3-5, makign the thread directly with the screw, works, it will heat up due to friction, but it will work. That's how i mounted a adaptor plate for a shifter in my sim rig
I've also tried a two stage hole, which does the same thing CNC is trying to accomplish with his section on the insert before the grips. This seems to work better with inserts that don't have that feature, when using an iron to put in the insert. My main problem with iron installed inserts is they don't get in straight or in the right location, unless just pressed in from the other side. I think that STL files provided online should allow the insert to be pushed in from the other side or at least have a two stage hole to help start the insert in with an iron.
This is why I print my wall loops with 6 lines thick if I'm planning inserts. I have found that this adds significant strength increase in pull-out strength. Where I think the opposite threading would help, is if your object is expected to operate under pressure in a warm environment. That should help decrease the failure pull-out point. To me though, the plastic will always be the more likely failure point before the insert fails. The benefit of the insert, is if you plan to frequently remove the screws for maintenance/access reasons (like case access screws for example).
Pro tip I recently discovered and have been using for my 3d printed items: R-Type Nylon Rivets. You can put things together pretty easily with them, and unlike heat set inserts, or screws or bolts, you dont need nuts, you dont need to model any threads or places to insert a heat insert, when you prototype you dont have to throw the hardware out with each failure, and they're cheap as chips. All you need are holes for these, so its dead easy to use them to prototype and for final assembly due to them being cheap and reusable. You can even toss in a few extra if you are selling a product with how cheap they are.
Taping into the plastic works MUCH better if you just let the screw tap the whole, especially if it's a "screw in once and leave it there" sort of deal. (You can unscrew, and reapply the screw, you need to be very careful to align your threads correctly, but if you're going to do this more than a few times, I'd go with an insert at that point.) As all of the material that was there, is still there, just displaced. In my experienced tapped holes into 3D printed parts are FAR weaker than self tapped holes. You also need to understand, you're going to be fighting a LOT of friction, so it's not going to be the smooth and easy engagement you're used to from say metal on metal engagement. No, this does not generally split the plastic, even when it's extremely tight. Though I generally print with a minimum of three walls, and if I'm going to be doing screws, I may step it up to four, depending on application. You do want the hole as small as you can get it, and the bolt/screw actually still be able to fit in enough to tap in. Also there is utterly no need to tap a plastic hole for a metal screw, I can't believe Tom even bothered with it. It's not like the plastic is going to deform the treads on the screw!!! Further, I can't believe he was tapping the hole with a drill!!! I tap metal holes by hand, and here he's using a power tool on plastic, and wondering why it's melting! 🤦♂
The best thing about the inserts is that they allow for the screws to removed and replaced many times without messing the plastic up. Holding force on most screws into plastic is usually fine for screw it together once jobs.
I usually use square nuts that are inserted during a pause in the print. The thing is that you are then not pulling the nut out of the material, you are compressing the perimeter material, so it's *much* stronger than inserts. You can regulate the strength by the depth of the insertion and the width of the solid infill around the hole.
USSA woodworking channel uses threaded inserts for wood in their designs. They also insert it from the back of the part so the chamfer helps spread the load. Would be interested to see how they perform. I only use inserts for parts that need to disassembled fairly often. Normally I just undersize the hole by the thread thickness, and let it self tap.
Thanks for another great test. This was basically already the conclusion that I have come to over the years. If I only ever intend to screw something together once, I may just size the hole so the screw bites in and cuts threads, or use a nut. Same for quick one-off prototypes. However anything I intend to unscrew or use multiple times, gets inserts. I made myself a custom profile in openscad for the inserts that I use so that the hole is a little wider where the insert goes, so that a little less plastic is pushed down but there is still plenty to melt around the brass and make a tight hold.
I print a slightly undersized round hole and tap with a screw with a tiny bit of lithium grease on it. This keeps the plastic from melting too much and shapes it into a perfect fit. Since it’s not a cutting tap bit, none of the plastic is removed. I also design parts with “nut pockets”, if there’s enough space.
I printed some tool heads for my Dillon 650 Reloading Press. I did it just for low-stress depriming only, so very light loads. I dabbled with printing threads, but what worked best was printing 20 walls in the holes, and running a 7/8-14 tap through the holes. Yes, it was messy, but it worked, and continues to work.
One thing, you didn't mention would be a rivnut. Slipped into the hole from the other side. And since it doesn't have the rotation ribs, you can use a bit of glue if you don't want it to rotate. You can match the rivnut length with the part thickness, or recess the back side to match the rivnut length.
For the injection moulding inserts, to make the plastic pushed away less of a problem I insert a bold into the insert and then put the soldering iron on the bold...like that the area that needs to remain free for the bolt is clear
I have had good success by adding more walls around the hole and then taping threads by hand. NOT a power tool. It takes a little longer but it also eliminates the heat issue with the tap.
This test left out the OG method of screwing into printer parts - print slightly undersized and use the screw to tap the threads as you put it in. You don't compromise the wall thickness at all in this way, and an added bonus is that the printed part now acts as a bit of a threadlocker.
I have done a lot of tapped M3 holes in plastic. I print the holes at 2.9mm and run a tap through them. All you need is a screwdriver handle for a TAP. You can print this. You can tap them by hand. No coolant necessary. No T handle. Just screw it in like you're screwing in a self-tapping screw. Run it in far enough that it's coming through the other side - even if that's down in a 5mm diameter hole - and blow the swarf out of the flutes with a quick puff of air. Then run the tap back out by hand. So many RUclipsrs who try to do threaded plastic holes over complicate it with T handles or power tools. That stuff is not necessary. A 3D printed knob/handle and 15 seconds of elbow grease is enough. Of course it's not as strong as inserts. But for things like raspberry pi cases and phone stands, it's more than good enough. You can use superglue as threadlocker if you have trouble with the screws coming loose, I don't really find that to be a problem. Also a 2.9mm hole is good for a #4 self tapping screw. So it opens up other hardware options without having to change anything in the CAD - if you're sharing your STLs or whatever.
An idea i just jad while watching this was to redesign the insert to be a truncated cone that is pressed into the opposite side of the material. This would greatly increase the contact area of the insert. It also spreads the load out over more material and gets rid of the aligned perimeters that pull out so easily. No blind holes and slightly more limiting but would also be more hidden. Or maybe just putting the bolts in from the opposite side of the inserts. The inserts do seem to move a lot of material out of the way, which may be creating a shoulder for the insert to push on. You're likely still only contacting the perimeters though, which would still be the weak point.
As you hinted at I'd like to see how much better it performs if you modify the 3D printed parts to improve the strength of the connection to the rest of the part.
Great informative & practical segment. I designed a hole with fins (negative cut-outs) extending into the infill area, increasing the hole perimeter material and the engaged surface between perimeter & infill. They are arranged radially around the hole. In my example for a 5mm hole, the fins are 0.5mm wide by 3mm long, with 8 in the radial array. I start & end the fins 0.5mm inside the object surface so that they're not visible from the outside. Since the area around the hole is where the failures are located, this should increase strength substantially. This also would provide space for the excess material to go with thermal insertion. I haven't tested or even printed this design yet - wish I could send a screenshot of how it looks in the slicer.
I use M3 screws as self tappers. The holes have to be slightly oversized to push the excess material into threads. This is technique used in RC model kits along with true self tapping screws. There are also thread forming taps available - these are better for 3D printed parts than thread cutting ones. They are strong enough to pull two warped PETG parts together.
Tom, another significant aspect of threaded inserts is the ability to torque a fastener tight enough such that they don't back out. If your fastening application compresses plastic at any point, you will never be able to tighten your fastener enough such that it cannot back out. Another option you could try is a Threaded Flanged Insert Nut. They are designed for use in wood, but I find they work just as well in 3D prints.
Very comprehensive test! From a Design For Assembly standpoint, threaded connections can be achieved with stainless DIN 7981/7982 screws. Faster, easier and cheaper.
I have not been able to put inserts in with an iron consistently. What I have done is pushed in the insert from the other side, and made the hole smaller than the insert hole on the bolt side. This has made the insert stay in, when torqued to secure the parts. I have also looked into designing parts to snap together, which can eliminate gluing, that can be messy. Designing features that snap together is easier than I thought. It's also faster assembly, but screws are still stronger.
One thing to remember with modelled threads is that they require a small layer height. No point modelling threads when you intend to print at 0.4mm. And remember that small layers = longer print time.
I did 6 screws into PLA, each with a drop of CA when I screwed them in. After some months I needed to take the thing apart and ALL 6 #6 screws broke off at the head without removing the screws with no deformation of the plastic at all. Sometimes the clamping force is less important than the shear force and in those cases direct screwing into plastic can work pretty good.
Alternatively, create a pocket for a hexagonal nut, pause the print and drop the nut into the pocket before continuing. You can also use a square nut slid in from the side or a nut from the back if you have acces to the back.
Thinking about it, for parts where you have enough depth to use a nut+washer and have it covered up by the top layers that might be the strongest combination of things. Alternating extra walls makes a pretty big difference in my experience, and so does making sure there is enough material there by adding an extra wall or two. Most slicers also have a wall overlap setting that gets them welded together more.
I was wondering why captive nuts weren’t used. im guessing Tom can show these would have blown the others out of the water due to the failure state (assuming parts are “apples to apples”). I’ll hold judgment until these are in the follow up…
I've used internal nuts (kind of like internal magnets) as threads with really good success. You design a hex cavity inside the volume of the part. Pause the print when the depth is such that the nut fits flush with the top layer. Then continue the print, locking the nut inside the print.
As for tapping M3 or such - here is, how I do it: I print a 3mm hole, which due to tolerances will end up shy of 3mm. Then I heat up the screw, screw it in, wait for it to cool down, and unscrew itagain. Works a treat. Should you use stainless screws or dont have a magnetic bit, blue tak or such is your friend for holding the hot screw to the bit :-)
I've always used printed threads and they are fine from M5 up. Pro tip is to tap them after printing but not with a drill and machine taps but with a proper set of hand taps. For anything smaller than M5 I typically use hexagonal holes and nuts. Great video BTW!
I have not needed to print many parts or projects needing a fixing like this but I think one thing that has not been mentioned is, How Tight Does Your Fixing Need To Be? clamping down the parts to destruction does tell you their ultimate stregnth but what if you do not need that stregnth, say just for a lid on a box but need to be removed often and look professional, I think most of these fixings would do that quite nicely. Having said that, you went to great legnths to bring this topic to our attention and did a fine and thorough job too. Regards, Geoff.
Great video. I appreciated the quantified testing and also your willingness to toss out bad data as needed. My personal wishes would be: 1) Holes with thicker walls. I usually design my parts to final size and don't have Cura do any hollowing, so I can make any desired wall thickness. 2) Test with PETG. I don't trust PLA for my boxes, which can get pretty warm in use. 3) Small bolt sizes. M3.5 is a maximum for me, and M3 is probably the sweet spot. 4) I'm mainly interesting in the brass inserts, but a comparison with a self-threading wood screw would be useful, especially if it includes multiple cycles of insertion and removal.
Surprised you didn't try nut pockets (what prusa does with their printer parts). That creates a nice solid wall that isn't distorted by brass teeth. I typically use wood screws if it's a one time use, and the inserts if it's a part that may need dis-assembly (also looks more professional). In injection molding it's exactly the same, if it's a one time assembly then self-tapping, if it needs to be serviceable there better be inserts included.
This is my exact design philosophy. If it's going together once it's self tapping(even machine screws self-tap in plastic well if the tolerances are low enough) and if it's going to be taken apart multiple times it's getting threaded inserts.
I design most of my prints with a nut pocket. then I will stop the print in the slicer, insert nut, and continue print. this allows me to use a lot of off-the-shelf hardware store nuts and bolts without having to buy more expensive specialty hardware.
But I'm also not printing high volume, so it's easy for me.
That'd certainly be stronger threads, but it seems like that would just make the plastic fail faster since there's even less of it at that point.
Lol. Nut pockets.
Nut pockets are great, and if you design it well you can have lots of plastic between the nut and flange to make it very strong.
I think it could be interesting to see how the number of perimeters around the hole and the type of infill (geometry and density) affects the strength of the insert (using CNC kitchen ones for example)
true or one print with 100% infill
I did something like this on my printer, but ran a battery two tests: number of perimeters, and diameter of the hole. I incremented the hole in .05mm increments and tested about 20 options, but I didn't have a fancy torque setup so it was all eyeball and hand-feel.
Once I did that I was able to save custom hole profiles into my cad tool to make those custom holes.
I was thinking similar thoughts about increased perimeters and/or increased infill density. My spin on this would be to use cylindrical modifiers around the hole to apply the perimeters and/or infill density localized to the hole instead of on the full print (that presumably doesn't need the extra plastic).
I second that.
Voron settings for their printed parts are 5 or 6 perimeters. Likely to increase the hold of inserts
It's crazy to me he didn't go with higher infill and more perimeters. I always beef that stuff up anywhere I'm using inserts. Also, it looks like his prints had poor bonding of the top infill layers to the hole perimeters. I have a less "pretty" tuning setting that really overfills that to make sure my hole perimeters are bonded to infill everywhere possible.
Flip the camera inserts to the other side of the printed piece. That way you have a larger flange you are pulling against the print. Essentially acting as a larger backup washer.
Also, doing the same thing with the heat-set inserts. Pull them THROUGH to the far side, rather than pulling them OUT of the side that they were pushed into.
Also, there's the possible benefit of the recess used for tightening the insert mating with the plastic and preventing twist out when tightening and loosing the bolt.
But if you can access the rear, you'd just put a nut pocket in there generally.
My thoughts too!
@@thingswelike a heat set insert will fit in places where a nut pocket can't
Test it with a regular hex nut countersink in a hexagon hole from the back on the part or a square hole in side the part and push the nut in from the side. I do this all the time and it works perfect and does not require any special nuts.
This was exactly what I meant with my comment, but better writen :)
Yes this is the standard (and correct) way of doing it, unless you want to save weight.
This. It's so much stronger than heat inserts that can pull out or strip the hole and spin, and uses dirt cheap off the shelf parts.
@@daliasprints9798 well you can use the heated inserts from the back side too but yeah those could potentially still strip the plastic
Was surprised he didnt do that when he was showing a nut in the beginning.
Friendly tip - for anything starting at M4+, I am not using inserts for plastic, but inserts for wood. They have larger diameter of the metal part. And I am heating them up and screwing them to the plastic - significantly increasing the contact area. Works much better. And the best part - you can then screw in the bolt from the other side, make the insert invisible and even stronger :)
That is a realy good idea. I will try that out the next time :)
I definitely want to see this idea tried in the next video. Great idea
Just a FYI, spiral fluted taps are much better suited to tapping plastics and soft materials like aluminum as they evacuate the chips while tapping.
The use of those traditional taps may be why his simple thread an m2.5 hole.did not work. I have a spiral tap and, as long as I have a few extra wall layers, find it works perfectly every time.
The only thing I would like to find is a way to simply set extra walls on holes only without a lot of fiddling.
Ah yes, I was a bit surprised that tapped holes worked so poorly considering how often I use it for non load bearing parts. My taps are also spiral fluted, and I don't power tap them so that is probably why I have better luck with it.
I wouldn't necessarily recommend it, most of the time I do use a captive nut or insert. But it isn't that bad.
Cool didn't know that
Single-cut taps are the choice when using machines for manufacturing. They require a higher torque because they finish the thread in one pass. They are also more sensitive to torque peaks that occur when cutting a thread by hand with a tap wrench. In this case, a tap set consisting of a pre-cutter, center cutter and finish cutter is used.
Exactly! Plus use a super sharp tap which you only use on plastic parts and not on metal parts ..
Perhaps consider printing a 100% infill 'doughnut' of say 10mm around each insert/tap/thread point to eliminate some of the variability caused by your part flex and infill interface, along with using an omnidirectional infill like Gyroid. It would improve pull-to-pull consistency and more accurately demonstrate the ultimate/average strength of each method.
This is exactly what I do. In the slicer add a cylinder modifier with 100% infill (usually just larger than the diameter of a standard washer you would use for that fastener. I've had no issues with part failures for my applications going this route. I make a lot of sensor mounts so there isn't generally a lot of dynamic forces on them, mainly just static loads.
I typically use self-tapping 3mm screws in 2mm holes with no printed thread. Even for parts that need occasional disassembly and re-assembly they work great. And let's not forget that lots of injection molded cases do the same thing. And in my experience PLA parts survive unscrewing and rescrewing better than the industrially produced ABS parts. As I'm writing this I'm printing 3.5" to 5.25" mounting adapters that I designed today (because the ones on Thingiverse are all just imitations of mass produced parts and these designs work poorly for 3D printing) and the means of fastening the adapter to the PC case is by means of 3mm self-tapping screws. I always use the longest screws that will fit.
yeah, self tapping screws: fast, easy, cheap and often plenty strong. If strength is a question, bolt and nuts. I only rarely print things that I plan to take apart multiple
times.
I skip the self-tapping part and just use regular M3 machine screws and force them on the hole. Power tools will melt the plastic, but if hand tightened, they usually work just fine. "Just fine" as in the level that I never opened the insert box I ordered a year ago.
This is the way.
Although I design my holes to be 2.5mm for m3 screws. I very rarely have problems with this kind of self-tap and it's always self-inflicted harm.
( i.e. impatient assembly, bad design, faulty print, etc)
@@bogusF and also if the hole got to small, a 2.5 drill by hand and all ist good^^
I too use regular M2, M2.5, M3 and M4 screws as self tapping screws (by hand). I print a hole smaller than nominal and using a separate slicer modifier solid (prepared in CAD), I add a 2mm thick ring of plastic (perimeters) just around the holes to be later tapped by the screw. If you use long screws as shafts (sticking out of the part) this method will effortlessly give you perfectly straight (perpendicular to surface) shafts. A thing almost impossible to achieve with inserts (that are impossible to sink in perfectly perpendicular to the surface). And you can screw and unscrews many times without ruining the tapped hole. I do not use inserts anymore for pieces not intended to be screwed/unscrewed repeatedly. I use PETG.
In combat robots, we all use plastite screws. They have triangular lobe coarse threads. You just screw them into the plastic, no tapping required. A common failure mode we see are heat set inserts pulling out, but the plastite screws fare much better. We also tend to use a lot of walls since we’re printing for strength. Usually 3mm walls or more. Plastites are also lighter than inserts.
A handy tip for increasing the number of perimeters around just a hole: You can put an infinitesimally thin annular gap spaced out a few perimeters lines from the ID of the hole. This will force the slicer to an extra set of perimeter lines around the hole without having the increase the number of perimeters everywhere. I'll actually put these phantom gaps anywhere I need to add solid perimeter lines in an area that would otherwise just be infill.
Ill have to remember this one. Could have used this trick the other day
Nice tip - learning to make the printer do what you want is so important in this craft.
Annular gap? Like cutting a circle into the existing hole with ~0 height?
@@SixTough More like a cylindrical gap of, say, 0.001mm wall thickness but the height of your screw. The radius of the cylindrical gap should be larger than the original screw hole size by several (4 - 6) line widths.
@@SixTough nope, ctting a 0.0001 wall thick cilynder around the hole full depth, or maybe two concentric at 1 mm distance from eachother (assuming you print 0.4 mm wide layers), that forces the slicer to put 3 walls around the hole, another 3 around the first gap and another 3 around the second gap, but the gaps being so narrow the walls for each circle will melt into ecother giving you a 9 walls print.
Would be more interesting to test screws made for screwing into plastic, like "plastite" instead of screwing ISO screws directly into plastic. Right tool for the right job!
This is the best way in my experience. They are also called trilobular or trilobe thread forming screws for plastic.
Yes, they work quite well too in my testing. Also things like 'helicoils' do a good job to just add extra strenght.
screws for wood work as well. nothing is cheaper than drywall screws too
@@shonkydonkeyThey actually use them a lot on cars. Both on plastic and metal parts. I think so they clear any paint or debris out of threads on the way in. You can usually tell because they will have a small slit in the thread on 3 sides. Sometimes the tips are also slimmer than the threads or look slightly triangular.
support this, but the difference is the reuse ability of such threads as the plastic weakens every time you reinstall the screw.
The question is how often can you do this with the same load.
Did you test torque-out at all? Because that pull-out test is honestly almost entirely pointless as that is just not the way you use them in, dare I say it, almost any case.
LOVED the Blondihacks "Yahtzee" nod with the lathe part-off 😂
In my experience the screws for wood work much better. The hole in the printed part should be a bit bigger than the core and since these screws have a very deep and sharp thread, they usually connect very nicely with the part. The failure point is typically the connection between the printed hole and the rest of the part.
For tapping the thread into PLA i found that a decent lubrication makes the process almost perfect. I typically use some grease since it does not flow away unlike oil. By the way adding a little of grease on woodscrews eases the joint process a lot. Cheers
Wood screws are often conical, and can crack the plastic parts. Better use screws specific for plastic.
@@fred-9929this can be true, but straight-shank wood screws are far easier to find in your local hardware store and orders of magnitude cheaper than ordering a small quantity of plastic -specific screws. If you size the hole 1% larger than the nominal minor diameter of a wood screw it wont be any more prone to crack.
@@fred-9929 Drywall screws are available in fine and coarse thread. Fine thread works well, and is even self drilling in thin steel with the good quality ones, and I have used them a lot that way, even on steel a lot thicker than the thin steel used in drywall work, like steel tubing with 1.2mm wall thickness. The coarse ones, or those designed for composite wood, are very good, and almost always stronger than the substrate, but are generally not going to self drill in anything other than composite board, or through 0.6mm steel plate..
I usually embed nuts inside the print by modeling a cavity and pausing the print at the top layer of the cavity. Nuts are usually a lot easier to find for a particular size than threaded inserts, and more broadly useful. If you need even more strength, modeling in space for a steel washer between the nut and the inside wall of the part increases the engagement with the plastic part by quite a bit.
That's an absolutely brilliant idea, thank you for the tip!
Since you have no need to use a spanner, you could use square nuts, which have a bigger area (and a simpler cavity shape).
@@fatroberto3012 True, but hexagonal nuts tend to be more readily available, especially when compared to threaded inserts.
For the past 5-6 years I’ve just self-tapped into a tighter printed hole and had excellent results that way. Yes, occasionally I’ll use a nut on the back and once or twice inserts but for 90% of the loads I put my prints through (including my CNC machine), the self-tapped printed holes seems just fine.
Same, and I have hundreds of products (modified, industrial card printers) out in the wild assembled like this and not a single fastner (mostly M3) failure.
Hear hear! Same here. This video seem more like a hidden commercial
Love the machining work and the Quinn reference, yahtzee! Would like to see your mill and lathe setups more!
My favorite is the wood thread insert with spikes on a flat washer on the back of the part. Also always add some extra walls around the thread and it won’t pull out but generally I’d print a section of 100% infill there anyways. If you can get 4x diameter on the hole depth, printed threads chased with a forming tap are wicked strong. And like others said, burying a nut in the print in some way is always a super good option. If the nut is near an edge, you can add a keyhole slot on the side to slip the nut in and then screw through the front. I had to make a super tricky assembly once and used a thick 3mm washer tapped for an m4 bolt and slipped it in through a side slot that got plugged with epoxy to hold the threaded washer and plug in.
I like to use “alternate extra walls” and “connect infill lines” to really boost the strength of printed parts. The perimeters become interlaced with the infill structure and resists pullout and deformation. I use grid for faster prints but prefer gyroidal for strength in all directions or for parts susceptible to warping.
I would think bicubic is better for overall strength than gyroidal, no?
FYI, when you tap plastic you should hand tap not use a drill, you will heat up the part and melt it. Also do 1 revolution and clean out the tap, then continue the next revolution and cleaning until you are done. If not you will mess up the threads. Also don’t lube the PLA, oil will weaken it.
I'd be really interested in seeing the results of different infills as the failure mode was always to rip the connecting walls, I think that'd be much more useful for figuring out how to get the strongest output for each usage case!
The vast majority of installation torque (80-90% in steel nuts and bolts) goes into overcoming friction so it's not a good way of measuring pull out force. Once installed the assembly strength is determined by bearing and pullout strengths, not twisting of the insert.
For extremely small screws (M2, M2.5, M3) I tend to just screw directly into an oversized printed hole; basically the tapping method, but skipping the tapping step. I get fine results, but I'm also never designing parts for strain, just assembly.
Two ideas I have are creating a wide ring in the middle of the screw hole that acts like an anchor in the shape of a washer. Or creating a recessed area on the backside to actually install something like a fender washer to spread the load out if you really need the strength.
Please also test the strength of self tapping screws. The "threads" of those, are spread over several layers, and seams to make a quite strong bond.
It's a way quicker solution to use, that I have had good succes with.
This video needed to be made, thank you.
Please test nut captures and the strength of inserts from the opposite side.
Tapping holes with a hand drill is something I never do regardless of material as while it takes a little longer I find that doing it by hand results is less damage to the threads especially for fine pitch threads.
First year Uni manufacturing here in Australia taught me that a bolt has a un-threaded shank portion where as a machine screw (not a wood screw) is threaded all the way to the head. Love the testing videos, keep up the great work.
Actually, the only difference between a screw and a bolt is what you call them. They are literally the same type of fastener. Don't get too caught up on the name.
@@GrubbyZebra It seems there's a difference in naming/function when engineering something vs home building.
@jetah50 the problem is that any supposed difference is easily disproven.
Bolts have a smooth section? Fully-threaded bolts (often called "tap bolts") are a thing and are commonly used to attach automotive body panels.
Bolts used with nuts? Lycoming cylinder bolts are installed into tapped holes in the crankcase.
Screws are only designed for tension and torsion loads? Nylon shear screws are commonly used to attach model rocket nose cones to the rocket body in order to prevent unplanned separation.
And I could keep going.
As someone who sells these for a living and deals with technical documents constantly, the terms are 100% interchangeable. @@jetah50
The only difference I’ve been taught and that seems to consistently stick is that bolts mate with a tapped thread whereas screws cut their own. Machine screws simply have a tooth profile capable of carving into the metal/plastic. Wood screws cut their thread into wood but I don’t think I’ve ever seen a wood bolt, because the material being held doesn’t matter to a bolt, as it knows it has a matching tapped thread to fasten with. A screw on the other hand needs a different profile, pitch, taper etc depending on the material it’s self-tapping into.
Hi Thomas, first at all thank you for your always excellent videos.
If this video, you are in fact testing resistance vs. torque which is fine to know how the attachment will resist to an over torque.
But according to me, should you be able to test resistance vs. axial force, this will provide tremendous inputs on how these fixtures resist under load. And we could have some surprise with plastic tapping. And should you be also able to test self-tapping screws at the same time, this will give us a whole overview of attachment resistance.
I find myself wondering about the testing methodology. When turning a screw in PLA, won't the friction heat/weaken the PLA giving the metal on metal solutions an unfair advantage?
This is a fantastic video! You mentioned that you solder lead free. Could you make a video on how you do this and what solder you use. I have been trying to use lead free solder for a while and I just can't get the solder to cooperate.
Few other types of threaded inserts: "thin wall threaded insert" and "helicoil" both would be interesting to have in the comparison. As another comment mentioned, would be interesting to know print settings and do a review on things to improve the joint. I would also love to see how using a "matching" grade of screw would affect the strength of joint (i.e plastic screws). Ace video as always Tom, loving the new clean feel of the production.👍
Thin wall inserts will be about the same as going up a size on the bolt, but the hole will last a lot more insert remove cycles. Probably best is any of the inserts, and make the wall of the hole a full fill to double the diameter, which will allow better transfer of stress to the bulk media. As most of the failures come at the join of the insert plus fill to the bulk void space, a better bet is more material there, either denser for a 2 diameter volume, or better infill that uses more filament, but at least handles load better.
My solution (not applicable everywhere but..) is adding a little pocket for a square nut.. it's not going anywhere and it's fairly cheap.
Great video 👍 something that people don't often talk or think about is that in many cases threaded inserts aren't necessarily "stronger" than direct threading...but in applications where you need to often remove a threaded fastener (eg. a thumbscrew) they are a must (for example helicoils in aluminum).
Exactly. "Strength" is the wrong question which is being asked. If you need strength, you need to bond the plastic together permanently.
This should be upvoted to the top! The helicoil example @emberprototypes gave is fairly common knowledge / industrial practice. Aluminum is too soft for repeatedly bolted/unbolted joints. They mentioned "often" but more realistically even one removal and reinstallation is too many - bolts into aluminum is often a one-and-done thing and unreliable to remove and re-bolt. The helicoil at its simplest form is just a tiny steel thread to increase the wear resistance right at the joint interface. This is doubly true for plastic.
Absolutely correct. You use an insert if you are constantly removing the part otherwise the plastic will just wear away.
Large diameter steel washers can spread the load over a wider area of the plastic when used in conjunction with a steel not.
A thread for the screw can be included in the hole so that a lockout effect can be achieved with the steel nut.
I personally think, the Torqueout Test won't give completely comparable results, because of different coefficients of friction (galvanized steel bolt - PLA/brass/Nickel plated brass).
A pullout test would give indefent results, sadly requires a test machine.
But I think the "ranking" would be quite similar.
I tend to use woodscrews for simple "once" connections, Bolts in printed threads if the simple connection needs to be a bolt for other design reasons. Sometimes printing only the first few millimeters and let the Bolt cut the rest by itself (friction stops unwanted rotations).
Heat inserts only when I need wear resistance, if even more is needed Ensat-Gewindeeinsätze.
Really strong threads I get by designing in massively longer threading depth or a nut placed inside during the print.
Fortunately, I'm not the only one who sees that this test is worthless. Twisting force is measured here, not pulling force. It's nice that the coefficient of friction between the different materials is measured here, but that is of no use to us. With a drop of oil on the thread it suddenly becomes a bad result. And a bad thread becomes a good result.
Another thing which is worth noting, is loosening of the screws over time. Having a metal threaded insert allows you to use a more diverse range of threadlockers (e.g. Loctite 222 or 243 etc). I designed a custom hotend shroud and ended up simply under-sizing the holes and forcing machine screws through as "self-tappers" which holds pretty well, but the screws do end up getting loose (I think it's because the bowden tube puts varying pressure on the hotend as it travels across the x-axis). Something else worth testing, is instead of using machine-screws, use actual thread forming screws meant for self-tapping into plastic. They tend to have a sharper, more coarse thread and are often seen in injection molded parts.
I prefer cut threads for everything below m6. I use 3d printing mainly to prototype parts that eventually get machined. 99% of the time some cheap PLA is sufficient for my needs. usually I print with 3-5 perimeters. I usually use alcohol as cutting fluid. few drops of cutting oil dissolved in the alcohol works even better. also aluminium taps work way better then normal ones (the ones that have every other tooth missing to reduce friction) through hole whenever possible, bottoming taps that push the swarf up only when really necessary. ...in 3d printing I can make the hole super deep or all the way through the part even if I'm not going to do that on the machined part for obvious reasons.
a final note: for embedded nuts: get some square ones - they work better in every regard I can think of compared to hex nuts when it comes to incorporating them into 3d prints. (don't rotate as easy, easier to design the pocket/slot, usually better bridging if pocket has to be upside down...)
Yeah, I bet hand tapping vs power tapping would reduce the friction and heat generation of the tap in the plastic...thus preventing the ripping of the threads out of the part
However hand tapping a buncha holes suuuuuuuck
I wonder if a better lubricant such as silicon grease would work better than denatured alcohol (I use Danco food grade silicon grease from lowes on my parts and bearings on my 3d printers/cnc machines, not only is it slick as greased owl crap...but it doesn't attract dust like oils and "dino" grease)
Another way is to use a good tap-holder instead of putting a tap in the drillhead ... I use a small adapter that I made in 5 minutes. @@haydenc2742
I use a tapping arm. For my cordless drill to tap. Having the tap straight also reduces the friction quite a bit compared to just cordless drill. Alcohole does some minor lubrication but is great at cooling.
Tom, I doubt you'll even see this but, we need a machining channel from you. I also see you for your blondiehacks reference! Been watching for a decade or so. I can't thank you enough for the knowledge you've shared.
I would love for you to test the direction of pulling out, AFAIK, Voron parts sometimes have you thread in the direction of installation and sometimes opposite. Taking your thread adapter into consideration, the large flange on the end would take a LOT of force to pull out in the opposite side of the install. Also it would be nice to know how much the different techniques of linking the perimeters to the infill would work! Great video!
Also they sometimes put inserts in a way that the other part will made contact with insert ( all force will be on metal ) and plastic part where insert is will be on that insert.
I mostly use the modeled threads, but if I need the strength I will use a normal nut pressed in the back of the part. Where I think the inserts shine is with parts that need to be disassembled many times. The modeled threads are only good for a couple assembly cycles and a nut can sometimes get out of place if the press fit is not great.
What about wood screws? They are made to hold on to the wood fibers that are weaker than metal. They are also very cheap
My thoughts exactly.
my favorite by far is taking one of the screws I'm going to use for the project and then cutting relief slots at the tip with a Dremel and cutting wheel. I then use that one screw to tap and clean all the screw holes leaving perfectly matched treads. The end result is better than printed threads, not quite as good as using wood screws but there are more head shape options and definitely not as strong as bolt and nut but definitely less fussy.
Its also worth noting that one should appropriately break the chips as you tap a tread depending on the style of tap used. Most cant handle or do a very poor job if you simply screw in all the way with out backing back out every couple of turns.
Additionally, PLA melts, shocking, so you do have to go slow even by hand enough heat can be generated to destroy your threads during the tapping process.
Hi, I use the best combination of materials. Let me explain: if I'm working with plastic, I also use a plastic screw. But too small sizes often do not allow this. So I place the empty space under the female screw and then pause the slicer on the last layer to place the nut there. As a result, I have a part that fits the hole perfectly, has the correct threads and the right material hardness. I would be glad to see similar tests with my method of making threads
Super interesting. Personnaly I usually extrude the shape of the nut from the other side of the part. thats my prefered solution as you often get the nut provided with the bolt, you dont have to buy plenty of insert or spend the time inserting them. Maybe you can do an addon video and compare to these results to this method, that would interest me a lot. Thanks !
That's what I thought too! 👍🏼
I am surprised you tried to tap the holes... what i do is print a perfect hole (lets say 2.9mm in diameter for m3) and then slowly let the screw make its own "threads" while going in, I found its a really nice and strong solution... probably stronger than modeled or tap holes...
You should be able to test for lead content of the Alibaba inserts using a swipe test. These are marketed for use in testing house paint and toys, but would work for this purpose too.
Most brass is leaded brass (2-3%), its not really an issue unless you: use it for a childs toy, use it for food storage, or grind it after the fact. You shouldn't be using 3D prints for these purposes anyway.
@@jaro6985
Oh. So you're saying I shouldn't 3D print a children's toy lunchbox with brass shavings on it for decoration?
@@jaro6985 My point was focused on whether you should buy CNC Kitchen's brass inserts because of being lead free brass. Or alternatively, since there is a lot of plumbing fixtures that are in-fact lead free, is it possible to get cheap inserts that are still lead free. And I'm sorry, but using 3d printing for toys is a pretty standard use case.
@@richdobbs6595 3D printing toys for young children is just as unsafe as using leaded inserts, most people do not have food grade setups. I'm telling you all china inserts should have lead in them, so no need to test.
Lead free bass literally means only 2% lead. This is also true for lead free brass that is used in water supplies. Lead free is a legal term not a normal english understanding. It is very misleading.
A few things spring to mind here:
1) If the limiting factor is the perimeters around the hole, beefing that up to the point where the insert turns in or pulls out seems necessary to get meaningful test data on the insert or thread.
2) I don't advise tapping under power for this sort of thing - use a manual hand-operated tap wrench. It'll keep the heat down and you'll be able to feel what you're doing.
I've actually used threaded inserts and stainless machine screws to replace self tappers on bits of plastic under-bonnet trim on my car - they are a huge improvement! 19 years of taking those old self-tappers in and out doing maintenance and they didn't really tighten up properly anymore. A great quality of life modification.
This is a valuable contribution to 3D print engineering, Tom!
I've found that friction threading fasteners into unthreaded holes works quite well.... What you do is you rapidly turn the fastener in about a 1/4 turn then back out a little less than 1/4 then back in a 1/4 until the fastener is seated. The friction from this rapid back and forth heats the fastener so it doesn't cut rather it melts it's thread into the plastic. If you size the hole correctly it will take a little bit of muscle to seat the fastener and you will be able to feel the resistance slightly lower when the fastener starts melting it's way in.... The result should be a tight thread with it's layer lines fused together. Something you will not get from a printed tread. Also this method of threading does not clog the hole or remove any plastic. The plastic is pushed outward into the walls, if anything leaving the wall thickness slightly higher.....I've used the word fastener in this description over bolt because I've used this method with screws as well as bolts up to 1/4-20.
Other than that if I just need a thread in my print where the bolt will be repeatedly adjusted I heat set an ordinary nut into the print. Which is cheap strong wears good and I don't have to order anything online to complete my project. With heat set nuts I print a hexagonal recess in the hole slightly smaller than the nut and push the nut in with soldering iron set to as low a temp as I can where the nut will still push in. Like with fiction threaded fasteners no plastic is pushed down into the hole. The plastic is instead also pushed outward increasing the wall thickness. So far I have had zero issues with pullout of friction threaded fasteners and heat set ordinary nuts have only failed in bad designs and when over torqued.
Yea, dont power tap plastic dude. Especially fine threads. This is the second video of this channel I've seen today where your process is severely flawed because you have no idea what you're doing. How about doing some research first. Blind leading the blind.
I switched from stocking machine screws to self tapping plastic-threading screws a little while ago and have had a fantastic time with them! Most of my parts aren't particularly load-bearing, but need to be held together... They're just so simple to add to a design!
I am just saying this before watching... I print all my threads on my designs and don't have an issue, so I am very interested in the results.
EDIT: now having watched it I can see how the inserts would have a benefit but unless there is a need for more strength I'll stock with printed
I see you teased the rivnuts. I've been using rivnuts as inserts for a while. I design for them to be inserted from the back side so that the flange can act as an anchor. The resulting connection is more robust than typical inserts.
you are washing your hands too much with harsh alcohol hand sanitizer ? you need to moisturize !
Great idea to run this test. I did 3 years ago a similar test for my machine designs and found out Prusa Pla is the best PLA for this application (the strongest, less ductile) and actually don‘t put a insert in and don‘t tap the hole. Just put around 4 perimiters and let the screw tap itself while screwing it in. With PLA this generates that much heat from friction that a perfect new thread forms around the screw, which additionally acts like loctite and is self locking. Takes some time to find the right hole diameter for each screw, but worth it.
For the love of god please moisture your hands 🫣🥲
If that's what his hands look like imagine how dry his lips are😭
Thomas, in tru holes you are blessed with having the possibility of using the insert from the backside of your workpiece. which creates the fact you have to pull the insert not only out but also tru the complete printing.
And a bit of a hassle but you can also demonstrate the "print in place" possibility, if someone might be interested. possible with an ordinary nut.
Usually, a screw is tapered to a point at the end, and a bolt is the same diameter for the entire shaft.
When I played with this situation (using PETG) I found that unless you are constantly assembling/disassembling, that just drilling an undersized hole in the fastener-area that was 100% infill, that the raw plastic was just as strong as the inserts. In a couple of cases, it seemed like maybe the insert was better but not consistently.
Finally, the only times I use inserts is if I want ease of assembly/disassembly cycles or for a customer that is under the belief that they just gotta have them!
Thanks for doing the tests. One last thing: I also found that after a couple of screw-unscrew cycles that adding cyanoacrylate glue worked well to harden the plastic screw-hole.
I also found that the longer the screw remains in the hole the more the plastic confirms around and makes it tighter.
I do feel that if you would just drill undersized holes instead of printing them, you would get better results not using inserts.
Thank you for putting in the research work! Since you asked what to check next: Self-tapping screws, wood screws, inserting nuts, use of washers (larger area might not pull out as easily?), using the inserts from the other side (pulling through the entire part instead of pulling them back out where they inserted) and of course more walls / more infill / both...
What if you push in the insers into the opposing side of the hole of where the screw goes into? I imagine the additional plastic the insert will press against when you try to pull out the screw will add a substantial amount of resistance.
Whenever the design allows it, I press hex nuts into the back side of the part with a clearance through hole to screw in from the other side. Make the nut recess compliant to avoid splitting the 3D print. PrusaSlicer has a couple of tricks that help. Modifiers allow more perimeter layers and/or more infill to strengthen the 3D part around the threads. PrusaSlicer also has negative volumes that can increase part strength in a particular direction that can be useful to support loads attached with threaded fasteners. There are slicer tricks for bridging over the top of the inserted nut pocket, with RUclips demonstrations. For smaller inserted nuts, typically M4 / #8 or smaller, using square nuts instead of hex nuts will help prevent the nut from spinning in the 3D printed part.
for low stress/torque, for M3-5, makign the thread directly with the screw, works, it will heat up due to friction, but it will work. That's how i mounted a adaptor plate for a shifter in my sim rig
I've also tried a two stage hole, which does the same thing CNC is trying to accomplish with his section on the insert before the grips. This seems to work better with inserts that don't have that feature, when using an iron to put in the insert. My main problem with iron installed inserts is they don't get in straight or in the right location, unless just pressed in from the other side. I think that STL files provided online should allow the insert to be pushed in from the other side or at least have a two stage hole to help start the insert in with an iron.
This is why I print my wall loops with 6 lines thick if I'm planning inserts. I have found that this adds significant strength increase in pull-out strength. Where I think the opposite threading would help, is if your object is expected to operate under pressure in a warm environment. That should help decrease the failure pull-out point. To me though, the plastic will always be the more likely failure point before the insert fails. The benefit of the insert, is if you plan to frequently remove the screws for maintenance/access reasons (like case access screws for example).
Pro tip I recently discovered and have been using for my 3d printed items: R-Type Nylon Rivets. You can put things together pretty easily with them, and unlike heat set inserts, or screws or bolts, you dont need nuts, you dont need to model any threads or places to insert a heat insert, when you prototype you dont have to throw the hardware out with each failure, and they're cheap as chips.
All you need are holes for these, so its dead easy to use them to prototype and for final assembly due to them being cheap and reusable. You can even toss in a few extra if you are selling a product with how cheap they are.
Taping into the plastic works MUCH better if you just let the screw tap the whole, especially if it's a "screw in once and leave it there" sort of deal. (You can unscrew, and reapply the screw, you need to be very careful to align your threads correctly, but if you're going to do this more than a few times, I'd go with an insert at that point.) As all of the material that was there, is still there, just displaced. In my experienced tapped holes into 3D printed parts are FAR weaker than self tapped holes. You also need to understand, you're going to be fighting a LOT of friction, so it's not going to be the smooth and easy engagement you're used to from say metal on metal engagement.
No, this does not generally split the plastic, even when it's extremely tight. Though I generally print with a minimum of three walls, and if I'm going to be doing screws, I may step it up to four, depending on application. You do want the hole as small as you can get it, and the bolt/screw actually still be able to fit in enough to tap in.
Also there is utterly no need to tap a plastic hole for a metal screw, I can't believe Tom even bothered with it. It's not like the plastic is going to deform the treads on the screw!!! Further, I can't believe he was tapping the hole with a drill!!! I tap metal holes by hand, and here he's using a power tool on plastic, and wondering why it's melting! 🤦♂
It would also be cool to see a resin version of this test, with different resins, of course.
The best thing about the inserts is that they allow for the screws to removed and replaced many times without messing the plastic up. Holding force on most screws into plastic is usually fine for screw it together once jobs.
I usually use square nuts that are inserted during a pause in the print. The thing is that you are then not pulling the nut out of the material, you are compressing the perimeter material, so it's *much* stronger than inserts. You can regulate the strength by the depth of the insertion and the width of the solid infill around the hole.
USSA woodworking channel uses threaded inserts for wood in their designs. They also insert it from the back of the part so the chamfer helps spread the load. Would be interested to see how they perform.
I only use inserts for parts that need to disassembled fairly often. Normally I just undersize the hole by the thread thickness, and let it self tap.
Thanks for another great test. This was basically already the conclusion that I have come to over the years. If I only ever intend to screw something together once, I may just size the hole so the screw bites in and cuts threads, or use a nut. Same for quick one-off prototypes. However anything I intend to unscrew or use multiple times, gets inserts. I made myself a custom profile in openscad for the inserts that I use so that the hole is a little wider where the insert goes, so that a little less plastic is pushed down but there is still plenty to melt around the brass and make a tight hold.
I print a slightly undersized round hole and tap with a screw with a tiny bit of lithium grease on it. This keeps the plastic from melting too much and shapes it into a perfect fit. Since it’s not a cutting tap bit, none of the plastic is removed. I also design parts with “nut pockets”, if there’s enough space.
I printed some tool heads for my Dillon 650 Reloading Press. I did it just for low-stress depriming only, so very light loads. I dabbled with printing threads, but what worked best was printing 20 walls in the holes, and running a 7/8-14 tap through the holes. Yes, it was messy, but it worked, and continues to work.
One thing, you didn't mention would be a rivnut. Slipped into the hole from the other side. And since it doesn't have the rotation ribs, you can use a bit of glue if you don't want it to rotate. You can match the rivnut length with the part thickness, or recess the back side to match the rivnut length.
rivnut was at the end next to the tee nut
For the injection moulding inserts, to make the plastic pushed away less of a problem I insert a bold into the insert and then put the soldering iron on the bold...like that the area that needs to remain free for the bolt is clear
@MadeWithLayers I could watch you 'make chips' all day! So glad you finally have a workshop you deserve. Can't wait for more chips to come!
I have threaded inserts (somewhere) & have never used them. Screwing into the bare holes in the print has worked just fine.
I have had good success by adding more walls around the hole and then taping threads by hand. NOT a power tool. It takes a little longer but it also eliminates the heat issue with the tap.
This test left out the OG method of screwing into printer parts - print slightly undersized and use the screw to tap the threads as you put it in. You don't compromise the wall thickness at all in this way, and an added bonus is that the printed part now acts as a bit of a threadlocker.
I have done a lot of tapped M3 holes in plastic. I print the holes at 2.9mm and run a tap through them.
All you need is a screwdriver handle for a TAP. You can print this. You can tap them by hand. No coolant necessary. No T handle. Just screw it in like you're screwing in a self-tapping screw.
Run it in far enough that it's coming through the other side - even if that's down in a 5mm diameter hole - and blow the swarf out of the flutes with a quick puff of air. Then run the tap back out by hand.
So many RUclipsrs who try to do threaded plastic holes over complicate it with T handles or power tools. That stuff is not necessary. A 3D printed knob/handle and 15 seconds of elbow grease is enough.
Of course it's not as strong as inserts. But for things like raspberry pi cases and phone stands, it's more than good enough. You can use superglue as threadlocker if you have trouble with the screws coming loose, I don't really find that to be a problem.
Also a 2.9mm hole is good for a #4 self tapping screw. So it opens up other hardware options without having to change anything in the CAD - if you're sharing your STLs or whatever.
An idea i just jad while watching this was to redesign the insert to be a truncated cone that is pressed into the opposite side of the material. This would greatly increase the contact area of the insert. It also spreads the load out over more material and gets rid of the aligned perimeters that pull out so easily. No blind holes and slightly more limiting but would also be more hidden.
Or maybe just putting the bolts in from the opposite side of the inserts. The inserts do seem to move a lot of material out of the way, which may be creating a shoulder for the insert to push on. You're likely still only contacting the perimeters though, which would still be the weak point.
As you hinted at I'd like to see how much better it performs if you modify the 3D printed parts to improve the strength of the connection to the rest of the part.
Great informative & practical segment. I designed a hole with fins (negative cut-outs) extending into the infill area, increasing the hole perimeter material and the engaged surface between perimeter & infill. They are arranged radially around the hole. In my example for a 5mm hole, the fins are 0.5mm wide by 3mm long, with 8 in the radial array. I start & end the fins 0.5mm inside the object surface so that they're not visible from the outside. Since the area around the hole is where the failures are located, this should increase strength substantially. This also would provide space for the excess material to go with thermal insertion. I haven't tested or even printed this design yet - wish I could send a screenshot of how it looks in the slicer.
I use M3 screws as self tappers. The holes have to be slightly oversized to push the excess material into threads. This is technique used in RC model kits along with true self tapping screws. There are also thread forming taps available - these are better for 3D printed parts than thread cutting ones.
They are strong enough to pull two warped PETG parts together.
I once tapped E-SUN cf-nylon and was amazed by how well it worked. It almost felt like tapping aluminum the way the chips came out!
Please test with 100% infill
Tom, another significant aspect of threaded inserts is the ability to torque a fastener tight enough such that they don't back out.
If your fastening application compresses plastic at any point, you will never be able to tighten your fastener enough such that it cannot back out.
Another option you could try is a Threaded Flanged Insert Nut. They are designed for use in wood, but I find they work just as well in 3D prints.
Very comprehensive test! From a Design For Assembly standpoint, threaded connections can be achieved with stainless DIN 7981/7982 screws. Faster, easier and cheaper.
I have not been able to put inserts in with an iron consistently. What I have done is pushed in the insert from the other side, and made the hole smaller than the insert hole on the bolt side. This has made the insert stay in, when torqued to secure the parts. I have also looked into designing parts to snap together, which can eliminate gluing, that can be messy. Designing features that snap together is easier than I thought. It's also faster assembly, but screws are still stronger.
One thing to remember with modelled threads is that they require a small layer height. No point modelling threads when you intend to print at 0.4mm. And remember that small layers = longer print time.
I did 6 screws into PLA, each with a drop of CA when I screwed them in. After some months I needed to take the thing apart and ALL 6 #6 screws broke off at the head without removing the screws with no deformation of the plastic at all. Sometimes the clamping force is less important than the shear force and in those cases direct screwing into plastic can work pretty good.
Alternatively, create a pocket for a hexagonal nut, pause the print and drop the nut into the pocket before continuing. You can also use a square nut slid in from the side or a nut from the back if you have acces to the back.
Thinking about it, for parts where you have enough depth to use a nut+washer and have it covered up by the top layers that might be the strongest combination of things. Alternating extra walls makes a pretty big difference in my experience, and so does making sure there is enough material there by adding an extra wall or two. Most slicers also have a wall overlap setting that gets them welded together more.
I was wondering why captive nuts weren’t used. im guessing Tom can show these would have blown the others out of the water due to the failure state (assuming parts are “apples to apples”). I’ll hold judgment until these are in the follow up…
I've used internal nuts (kind of like internal magnets) as threads with really good success. You design a hex cavity inside the volume of the part. Pause the print when the depth is such that the nut fits flush with the top layer. Then continue the print, locking the nut inside the print.
As for tapping M3 or such - here is, how I do it: I print a 3mm hole, which due to tolerances will end up shy of 3mm. Then I heat up the screw, screw it in, wait for it to cool down, and unscrew itagain. Works a treat. Should you use stainless screws or dont have a magnetic bit, blue tak or such is your friend for holding the hot screw to the bit :-)
I've always used printed threads and they are fine from M5 up. Pro tip is to tap them after printing but not with a drill and machine taps but with a proper set of hand taps. For anything smaller than M5 I typically use hexagonal holes and nuts. Great video BTW!
I have not needed to print many parts or projects needing a fixing like this but I think one thing that has not been mentioned is, How Tight Does Your Fixing Need To Be? clamping down the parts to destruction does tell you their ultimate stregnth but what if you do not need that stregnth, say just for a lid on a box but need to be removed often and look professional, I think most of these fixings would do that quite nicely. Having said that, you went to great legnths to bring this topic to our attention and did a fine and thorough job too. Regards, Geoff.
Great video. I appreciated the quantified testing and also your willingness to toss out bad data as needed. My personal wishes would be: 1) Holes with thicker walls. I usually design my parts to final size and don't have Cura do any hollowing, so I can make any desired wall thickness. 2) Test with PETG. I don't trust PLA for my boxes, which can get pretty warm in use. 3) Small bolt sizes. M3.5 is a maximum for me, and M3 is probably the sweet spot. 4) I'm mainly interesting in the brass inserts, but a comparison with a self-threading wood screw would be useful, especially if it includes multiple cycles of insertion and removal.