Are Our Heat-Set Insert Datasheets Wrong?

Were our datasheets always wrong? Heat-Set inserts for 3D prints have a specific shape with a pilot end that’s very useful for pre-seating them into your part before melting them in one after another. However, if you’ve ever used them in your design and followed the instructions on our packaging, you might have noticed something quite frustrating. With the recommended diameter, the inserts often don’t fit your part, and you might even end up with excess material where you don’t want it. So, you start increasing the hole size. But eventually, there's not enough material left for the knurling to grip properly, and you lose strength. So why does this happen, and is there a sweet-spot diameter where pre-seating is easy without compromising performance? I tested dozends of samples with my tensile tester to answer that question for my own curiosity and so you can make better and stronger parts and the results were pretty surprising. Let’s find out more! Guten Tag everyone, I'm Stefan, and welcome to CNC Kitchen. This video is sponsored by PCBWay!

Over the few years, we have sold millions of these threaded inserts, and to understand what makes them work so well in 3D prints and why hole size is so important, we first need to examine them more closely. The insert, made from lead-free brass, has two main sections. At the top is the milled knurling that later locks into the plastic. It is cut in opposite directions to provide a good balance between pull-out strength and torque resistance. What is less obvious is that this knurled section is also slightly tapered. That makes the insert easier to install and prevents the lower part from pushing away too much of the material the upper section needs to hold onto. At the bottom is the pilot end. It does not contribute much structurally, but it lets you pre-seat the insert in the hole before melting it in. At least in theory

because if the hole is too small, that part quickly becomes annoying because now you have to pick up every insert manually with the hot tip and try to maneuver it to its designated hole. But why don’t your inserts fit even if you designed the holes to size? 3D-printed holes almost always end up smaller than designed, and often by different amounts depending on the printer and material you use. SO if you’ve ever designed a part you might have already slightly increased the hole size to compensate that effect. But can you also go too big? When I built my Positron 3D printer last year, I had to install a ton of threaded inserts into the printed parts. Most of them went in great. The pilot end of the inserts fit nicely into the holes, so I could easily place several of them and then melt them in one after another. But there were also a few parts where I had the feeling that the holes in the plastic were simply too large, and only part of the milled knurling actually engaged with the plastic. I still went ahead with it and the printer works fine, but ever since then, I’ve had this question in the back of my mind: how much do holes that are too big affect the strength of my inserts?

So instead of speculating, I did what I usually do and printed a whole bunch of sample parts to test the real-world effects. If we take a look at the datasheet of our M3 inserts, we currently recommend a hole diameter of 4. 0 millimeters, which is slightly larger than the outer diameter of the pilot end of the insert. It’s obvious that at the pull-out strength will start to decrease if we make the hole bigger and bigger but I was also curious about the opposite direction. I also wanted to know whether we might actually gain strength by melting the insert into a slightly undersized hole, where the plastic can really flow around the part. That’s why I settled on a diameter range from 3. 6 millimeters all the way up to 4. 6 millimeters, which is actually the outer diameter of the knurling on the insert. I printed three test pieces per diameter in one big print job on the CORE One L using Polymaker PolySonic PLA. From previous tests, I already knew that it’s important to print these samples with a lot of walls and infill. Otherwise, during testing, the plastic itself fails before the insert gets pulled out, and that would make the results useless. If you’ve ever printed functional parts, you probably know that holes almost always come out slightly too small. There are several reasons for that, including material shrinkage, segmentation of curves, and simply the way plastic beads are laid down. And this varies depending on the printer and material you’re using, so the results will be slightly different for everyone. I used drill bits as diameter gauges for my samples and measured that my printed holes ended up about 0. 25 millimeters smaller than what I designed in CAD and that fits very well with the typical offset I add when designing parts. But the change in hole size in a print is also due to the shape and roughness of the inner surface — not just the diameter — I also set up a more generic test. I printed a second full set of samples and drilled the holes to their exact nominal diameter. So we’ll have samples that to compare that are just printed and then the same samples with a perfectly smooth, drilled bore. Then it was time to melt the inserts in. For this, I used our new USB soldering iron set to 240°C together with special tips that make good contact with the inserts for optimal heat transfer. Normally, you would place the insert into the hole first and then start melting it in. But when the hole is too small, that simply isn’t possible. In that case, you first need to pick up the brass insert with the soldering iron and then carefully move it over to the hole. When the insert slowly melts into the part, two things happen. Because there’s too much material in the hole, the hot insert pushes some of the plastic out of the way. Part of it bulges upward on the top side, while the rest gets pushed ahead of the insert, creating a burr on the bottom. And that can be really annoying, because it may prevent you from screwing a bolt all the way through the insert. When the diameter is just right, the insert fits into the hole with a small amount of pressure. That’s great, because it holds the insert in place and prevents it from accidentally falling out when you move or bump the part. At the upper end of the range, the hole becomes so large that the insert basically just falls straight in — and could almost be installed without a soldering iron at all. And I was very curious to find out how well those would actually still hold. So over the next half hour, I melted almost 50 inserts into the samples. For better consistency, I used our new insert straightener. This is an aluminum tool with a small tip and a large, flat face that lets the inserts be installed perfectly perpendicular and flush with the surface. You melt the insert in as usual with very little force, but stop just before it reaches its final position. Then you remove the soldering iron and press the insert the rest of the way in with the tool. That left me with very consistent samples because I was sure that all the inserts were installed in the same depth and also straight. The tool even removed the bulge ontop of the parts where the holes was significantly undersized.

That’s also a perfect moment to thank today’s video sponsor, PCBWay. Whenever we develop new products, there are always parts where it simply makes more sense to outsource a prototype instead of trying to make everything ourselves. And for that, we’ve been using PCBWay quite a bit, especially for CNC-machined parts. PCBWay can manufacture a huge range of custom parts, from circuit boards to 3D-printed parts, sheet metal parts, and CNC machining, which is what we most often use. Getting a quote is very straightforward. You just upload your design to their website, choose the material and surface finish, and you immediately see the price, whether you only need a single prototype or are already thinking about a larger production run. What I especially like is the turnaround time and their support. Our parts already arrive in the following week, which means we can iterate on designs much faster instead of spending a lot of time trying to make them ourselves. Most of you probably don’t have a fully equipped workshop, but you still have ideas you want to bring to life. PCBWay helps you turn those ideas into real projects with high-quality custom parts made in pretty much any material and process you might need. So whether you’re building electronics, mechanical prototypes, or complete assemblies, PCBWay is your one-stop shop and with the coupon code “PCBWay-CNC Kitchen10” the first 200 customers get $10 off their order. Thanks to PCBWay for sponsoring this video.

Now let’s look at the parts, because we don’t only want to find the optimal diameter for strength — we also want to know at which point installation becomes convenient without the excess material. And ideally, I was hoping those two would overlap at some point. As a quick reminder, for the M3 inserts I used in this test we recommend a hole diameter of 4. 0 mm. On the as-printed samples, I observed a severe burr on the bottom of the insert all the way up to a 4. 0 millimeter hole. The 4. 1 millimeter sample still showed slightly too much material, while 4. 2 millimeters looked just about perfect. On the drilled samples, the results were shifted slightly. Since the holes were now a bit larger matching the nominal diameter, I only saw a small burr below 4. 0 millimeters and anything bigger than that looked perfect. This means that from an installation perspective, a actual hole size of 4. 0 mm, as shown in our data sheet, allows you to preplace the parts easily, and after installation, there are no issues with excess material. Since my holes shrank during printing, I have to increase the hole size in CAD to 4. 2 mm to achieve similar ease of installation.

But the most important question is how the strength changes. So it was finally time to head over to my tensile tester. I have a special fixture in this machine where I install the sample plate on the bottom, while the top holds a dummy plate through which a screw goes into our test sample. The testing machine then pulls on the sample at a constant speed until the insert is pulled out. The force is continuously recorded, so I can easily identify the maximum. As much as I enjoy testing things, it can get pretty tedious when you have to test 50 samples in a row. But when the first results came in, they were already looking really interesting. Let’s start with the as-printed samples, since this is what most of us will typically work with. At the smaller diameters — where the plastic is packed tightly around the insert — the strength is the highest. On average, the M3 inserts pulled out at around 1400 newtons, which is almost 150 kilograms, or about 325 pounds. Keep in mind that this is with one material and my setup, your results will vary so don’t take these numbers as face value but just to see trends. If we look at the average values, we can see them slowly decreasing as the hole size increases, until they eventually drop off quite dramatically. For the samples that came directly off the printer without drilling the holes, we earlier found that from an installation perspective, 4. 2 millimeters is the ideal hole size, because the insert can be easily pre-placed and no burr forms underneath. At this diameter, the strength is still about 90% of the theoretical maximum, which in my opinion is still plenty. So this seems to be the sweet spot between convenient installation and maximum performance but we definitely souldn’t go much larger.

And to better understand why and especially how they fail, I thought that this was also a great opportunity to finally use my new high-speed camera. If you have something 3D printing related in mind you’d like to see in slow-motion, please let me know! The first thing I learned is that these failures happen very quickly. Initially, I started with a frame rate of 3000 frames per second, but the moment the insert failed was still just a blurry mess. So I increased the frame rate to almost 6000 frames per second. As the load increases, we can see small stress marks forming around the hole, but the insert stays firmly in place. Then at some point, it starts to move, slowly being dragged out of the hole — until it suddenly fails quite violently. It’s also interesting to see how the insert, just before failing starts to rotate due to the angled knurling and this is something you just see in shots like this! They really help us understand how parts behave and interact, and maybe even show us ways to improve their performance. One thing that became very clear here is how important the mechanical engagement between the plastic and the knurling of the insert really is. The plastic doesn’t truly adhere to the metal. Instead, it mechanically locks into the knurled geometry. And this is also why, once the holes become larger, the pull-out force decreases because there is material missing where the brass can grip. Conversely, if the hole is smaller—especially too small—the plastic gets pressed into every tiny detail of the knurling, holding it securely. Additionally, applying enough heat during installation is probably helpful because it allows the plastic to flow properly into the grooves instead of just being pushed away. And improving that groove geometry might actually make for a stronger insert. It also made me wonder whether there are ways to improve how well plastics bond to inserts. I probably need to sandblast them at some point. If you have any ideas, let me know in the comments.

But back to the tests. I also tested the samples where the holes were drilled to the exact diameter. The shape of the curve looks very similar and also the maximum stength is basically the same, but everything is shifted slightly because the holes are larger. In this case, the major drop in strength happens between 4. 1 and 4. 2 millimeters, all the way to the point where the inserts can almost be pulled out by hand. This again shows that our nominally suggested diameter of 4. 0 millimeters on the packaging is actually a very good fit. But what does all of this mean when designing parts that use heat-set inserts? I think the tests showed that we have this sweet spot where easy installation and no burrs overlap with good strength. Ideally, you want your hole size to land within this green range. If the holes are too small, installation becomes a bit more tedious because you can’t pre-place the inserts. But you might slightly improve strength although you may also have to deal with excess material being pushed underneath the insert. If the holes are too large, installation becomes easier, but the strength drops quickly. In some applications, ultimate strength might not actually be required, so going slightly larger can make installation easier and make your design more robust against variations in printer calibration or material. This might actually be what the Positron team was aiming for, although in my opinion, their holes were just a bit too large. But measuring holes by hand or using drill bits as gauges can be tedious and error-prone. And in my practical tests, the hole diameter where the insert just fit also turned out to be safely within the range where the insert still had a strong connection with the part. And I mean, that’s really what these inserts were designed for. So, to make a practical test

I created calibration parts for all of our insert sizes that you can simply print and then use an insert to find the size where it perfectly fits with just a bit of resistance. Since horizontal, vertical, and angled holes can all turn out differently, these test pieces include a range of hole sizes in all of those orientations. Ideally, you would do this for every printer and every material. But in practice, you’ll quickly find that the offset usually doesn’t vary all that much. Typically, for smaller inserts, I simply add about 0. 2 to 0. 3 mm to my diameters. If you’re a real perfectionist, you could even drill the holes to size before installing the inserts. This way you can be sure that you always have a properly sized hole for easy installation and even though you don’t gain strength, you can make sure that you don’t lose any and that regardless of the printer and material you’re using. I’ve heard that some people do this, but in my opinion, that’s a time-consuming and unnecessary step if you already know your machine and materials. Still we are considering putting such a drill set into our online store, and I would be curious if this is something you might be interested in?

So in order to answer the initial question: Should we change the labels on our packages. I think we shouldn’t because with every printer and material holes will come out slightly different and we can’t account for all of this variation. Our numbers define the nominal hole size for best results and you as a user needs to make sure you get as close as you can eiter by measuring holes or using our test parts. But do you agree here or if I already know that holes typically shrink a few millimeters, why not add this right away in the datasheet? These tests were done with M3 inserts, but I’m confident that the behavior will be similar for other sizes as well. Again, try to size your holes so that the insert can just be pre-seated. That will make installation easy while still offering excellent strength. This was a pretty technical video, but I hope you enjoyed this look into the development work we do here to make the products we sell as reliable and easy to use as possible. If you have ideas for other videos and test, don’t hesitate to let me know!