In today’s episode, Tom and Tracy Hazzard talk about CAD part size versus going to your actual 3D printed size. They discuss how and why this is a problem, and also provide some great tips and examples of what you can do to avoid this problem or how you can deal with it when such issues with actual 3D printed size arise.
CAD Part Size vs. Actual 3D Printed Size
We have had some requests for some more tech questions, tech topics. We’ve got a good one and it’s something we’ve been experiencing a lot lately as we’ve been doing a lot more 3D prints that are for short run production. We’ve done this for ourselves and we’ve done it for clients. There are really some important considerations, especially if you’ve never done it before.
We’re talking about going from CAD files size to your actual 3D printed size, which is your ultimate output. But there’s also the translation between going from there to your slicing software and then from the slicing software to the printer and then from printer to printer, because we go from printer to outsource sometimes. We do put it on our printer and it looks amazing and then we send the same file, we think it’s great and we send it to an outsource. But also, we have a challenge with a couple of our clients right now where we’re going from a CAD file, we’re doing the test run on our 3D printer ourselves, checking it out, but the reality is it’s also going to be made in a different material when it hits the outsource. It might be made in sandstone type or might be made in metal. We’ve got a material translation issue on top of not just the file translation in the actual 3D printed size.
Here’s what I’m going to say about this from my perspective and my experience as the primary CAD guy and who makes models and is in the trenches printing these things, is that you’ve got to start with correct geometry, correct dimensions that you know what your target is. As your building something in CAD, you’ve got to start with that. For instance, I’ve had some things lately where I’ve had, for a client, an existing printed circuit board that has to go inside of a two part 3D printed housing. You’ve got to know what your holes are and the space that is required for all of the little components attached on to this printed circuit board, but especially the mounting holes for it. You’ve got to then know these dimensions and build your CAD file to it.
But quite often, what’s going to happen is when you 3D print it, just because of the variances from one 3D printer to the next, especially in FFF, extrusion based 3D printers, you’re going to have variations in the actual 3D printed size. Certain printers extrude a little less than others, some under extrude, some tend to over extrude. All those differences can make it so that when you print on one kind of printer to another, you’re not going to get the same dimension or results. You have to adjust for that. That’s where it gets to be a bit of pain in the neck.
I just want to point out a good practice that we have here at our company and the way that we operate with all of our factories. It’s still a coordination when we’re doing this with factories as well. When we’re going from drawings of any kind, CAD, they used to be old school hand drawing, whatever they are, we put in a good practice here of having a criteria sheet that goes along with all of our specification packages. In other words, we don’t just send drawings and leave it at that. We actually put together a specification package that specifies exactly the material, any kind of hard data or hard criteria that we have.
If the critical measurement, the critical factors, and that’s what we call this particular form, is our critical factors form. In our critical factors form, it might be, “This is the exact size of the circuit board and this is its position in which it goes on the drawing. We expect it to fit within this dimensional tolerance.” It’s more of that organization and forms than anything, because I really get so sick of forms. But they are very important when you’re trying to make sure you get the results you expect. If you don’t specify exactly what you’re critical factors are and what you need, you can’t expect to get something back that’s actually going to work for you.
In the case of using 3D print service bureaus, and we did this just recently, remember we had our daughter’s wedding and did the 3D print favors. In that episode, I talked about the bracelet. I designed that. There were two different parts and they had to fit together and snap together. I designed that CAD to what the dimensions that I wanted it to be, number one, and then I tested it out here on our 3D printer. The first time I put them together, it worked just about perfect. Those things snapped together in the actual 3D printed size.
But when I sent those files to an outsource, still using an FFF 3D printer, just a different model of an FFF 3D printer from what we have, when they printed it and they put the two pieces together, they were barely holding together. They were so loose, they were coming apart and they needed to then change the dimensions of one of the two parts so that they would be closer in tolerance, this A part and the B part, so they would snap together.
That’s really where you go into this communication process saying that that is your expectation that these two pieces are going to snap together so that your outsource performs to that. We didn’t even think to put that in this particular case because they have a standard form. There was a place you could put notes. I did tell them in there. While I couldn’t upload a document, I did tell them, “Guys, these parts are meant to snap together. Please make sure you test it before you go printing all of them.” I was warning them in case they weren’t already going to be thinking about it, don’t just go sending all these things on your print farm to print 75 of these things overnight until you’re sure the first one works. That’s where you really want to make sure that you’ve got an outsource who is going to communicate back with you. This was Voodoo Manufacturing.
That works very well, but we’ve had some other experiences with some other parts lately. I want to talk about your options here, what you can do. You have the CAD file, you have the slicing software and then you have the printer itself. Then you might have material changes that happen along the way. You think you’ve got it dialed in and you get all that right, but then you have the complication of another material change.
Here’s where rapid prototyping can become not so rapid. You need to be aware of this. If you’re not prepared for it, you could end up with not enough time to accomplish the task you’re trying to accomplish, whether that’s a school assignment you’re trying to get done on time or it’s a job for a client or just something of your own and you’ve allocated a certain amount of time or you have a deadline when you want to get it done. You could be really surprised when you end up, you’ve done it and you printed it and the part you’re going to mate with it doesn’t fit, like that printed circuit board. Or I’ve also done a USB flash drive when I need to insert that into something. Those are hard metal and otherwise manufactured parts that have to fit.
What can you do? Let’s say you size it right for you part and you print it out and it’s too small. The part you want to mate with your 3D printer isn’t going to fit. You have a couple options. Your slicing software will allow you to increase the scale of the part by percentage wise and you can choose and you’re going to scale it only in the X dimension, only in the Y dimension or in all three dimensions, or actually it will let you probably do the Z also independently, each of the three axis independently or all of them together.
If it’s a general scaling percentage difference issue, you can do that. Just scale it up. That’s what Voodoo Manufacturing did with the bracelet part. I actually had them scale it only in two dimensions because it didn’t need to be scaled in all three. I didn’t wanted the fit on the wrist to be the same, but I could scale it in the other two dimensions. That’s what they did. I think it ended up being 5% or 8% larger and then it worked. That’s where your critical factors, that’s why understanding that criteria is important because you would be able to specify and say, “This dimension can’t change because if that dimension changed, it’d be falling off people’s wrist,” in this particular case. “Or it’d be too tight to get on,” if they scale it down.
For example, think about the bracelet as a C-shape going over your wrist. In that dimension, I didn’t want it to change in terms of the C-shape relative to people’s wrists. I didn’t want it to change. There was one dimension I wanted fix. Just so happen, the orientation of my mating part, we could scale the other two axes proportionally and it didn’t affect the sizing around the wrist. In this case, it could get wider, slightly wider because the width of the dimensional bracelet portion didn’t matter as much. It could get slightly taller I guess, on the outside of the bracelet, not on the inside of it. As long as I stayed away from that one axis that was parallel with the length of your arm, then the fit around the wrist was going to change the least. That’s what I did.
Another way, some 3D printers will allow you to actually take an already sliced file and just scale it up slightly just by the interface on the 3D printer. You could say, “Increase the scale of this X percentage,” but it’s usually very general. It’s like a total scale up or scaled down, they don’t give you usually the ability to do each axis independently. The other way is you can go back into your CAD file and scale it there. If you want to be sure that you have CAD data that is identically accurate to what you’re printing out on that specific printer, then you may want to do it in your CAD software scaling it up and do some trial and error. Scale up, print it, not quite right, scale up a little more, print it again.
At least that way you have a record. Because sometimes I think it’s easy to forget when you start with the CAD file and then you change the percentage in the slicing. When you want to remember a couple weeks from now, “How much did I increase that percentage in the slicing software?” you don’t always remember or maybe you didn’t save the settings file differently. If you did it in the CAD software, you save another version with a different name and you’d have that.
But here become the tricky aspects, and we actually created a database to keep track of exactly this type of thing. I created this dynamic database in which you could enter that in and keep track of it by every single file from the source of what you are creating. Way back when we were doing it with angels, you would start with the angel design and you had the design, but you would have all the iterations and all the variations would all be tracked to that original design, but you would have all the different settings and the dates you were in those settings and all of that.
I even made it so I could store the slice-setting file in the database. I put pictures of the print in the database. This is when we were working with some more open source 3D printers and especially ones that had more than one extruder on them. There were a lot of complex variables in prints. In order to really dial things in to get it to print exactly the way you wanted, you had to change slight things here and there. We need a way to document that so we could say, “We’re making progress here, but we’re going off the rails on these aspects of the print. Let’s take a step back, go back four, five generations. What were those settings? Maybe we need to start there and then go forward again.”
This is good practices, design of experiment. Design of experiment rules say change one thing at a time, don’t change multiples because it gets complicated and then it muddies your results. You don’t know what’s going on. You don’t know which one worked. You don’t know which one produced the better results. We have another client where, the one with the printed circuit board where we put that in there, we were doing a 3D printed part that is a unique design and they’re testing it out. They’re actually going to sell the 3D print manufactured part as a part of their product and do a short run, prove the market before they tool for it to be really be injection molded.
But there’s another part that’s this back half of the housing that is existing, it’s injection molded, there’s nothing wrong with it and there’s no reason to not use it. But now I’ve got to make a 3D printed plastic part mate with an injection molded plastic part, which has its own issues of shrink variances and other things. Sometimes, especially when you’re in that type of situation, trying to get a 3D print to match an injection molded part, I engineer it and I get it right in a sample sense on our FFF 3D printer.
Now I’m going to go out and have it printed let’s say at Shapeways or Sculpteo in full color sandstone. For a couple hundred pieces or something like that. But you can’t just go and order that couple hundred pieces, you got to do one. You got to wait for it to get back and then try it and make sure it fits and then adjust it and send it back out. Because they don’t do that for you. It’s not a part of their services.
I want to mention two things. One is that a really good way of testing that is to break it up into pieces. You’ve done this super detailed and a various part. Sometimes we had issues with the head of the angel for instance. You would just print the head portion or the top portion, one section of it. Being able to break up and say, “If this is my critical factor section of it, if it’s just the top part of the bracelet or if it’s this top part of the angel, where can I break off and just only adjust that section of the design so that it prints really quickly, so I can run through rapid iterations really fast or spend a lot less money having it made out and so it comes back to me quicker and I didn’t spend a fortune making the whole piece and having the whole piece be trashed at the end of the day?”
You can save money and you can save time that way in doing that and just breaking off the most critical portion of it. Remembering that as long as it’s the only part that’s critical in dimensional tolerance or whatever you’re trying to accomplish. That saved me a lot of time only printing a portion of it, not having it print the entire thing every time, for sure. I think that’s a rookie error if you do that. Maybe most people who’ve got some mileage under their belts 3D printing wouldn’t do that. But definitely, especially if you’re a student or somebody just starting out, you definitely want to consider that. Printing small pieces is really important.
I want to talk also about the flip side of that, which is being like Hubs, providing services and costing it out. This is where it goes so wrong. I think people don’t understand that until you’ve gone through some really horrible projects that went really bad and then you realize how burned you got on your time and your effort. You have to build in to your cost, to your prices, to what you quote someone to work on a project, the ability to take the time and the cost to you of time to do these types of iterations and to be able to build that into the project.
If they’re not willing to be patient with that or if they’re like, “That’s way too much cost,” they’re probably not a good client for you. Don’t undersell your services or underestimate your time. That’s a path to being upside down on the cost of a project. It’s not just being upside down on the cost, if at the end of the day, they didn’t leave you enough time, they weren’t tolerant or flexible in that and understanding that that’s what it was going to take to get it right, then at the end of the day, you’re going to deliver something that isn’t going to fit, it’s not going to be right because you’re going to run out of time and then you’re not going to have a satisfied client anyway. It’s really a slippery slope for you to be underestimating in all ways. In our experience, it’s just all kinds of bad and we refuse clients who aren’t flexible like that.
Let’s talk about material again for a minute. We talked about it in a few episodes where we printed an outsourced 3D printed part in stainless steel. It was the closest functional material we could 3D print something and that would perform the way we need it to test in an actual product. Even though the end part in production was not going to be made of stainless steel, we 3D printed it in stainless steel. I was quite pleased and actually surprised, and I think we used Shapeways to print this part, when I got it back, dimensionally, it was dead on to my CAD file. I measured it with my digital calipers in every which way and put it in to the product and was able to use it and it functioned just great. I was really pleased that at least the stainless steel process seems to be tremendously accurate to the CAD file that I created anyway.
I haven’t always found that to be the case in different materials. I’m also not going to say that that’s going to be always the same across different service bureaus. I actually uploaded the part to Sculpteo and Shapeways. For some reason, Sculpteo’s stainless steel cost was a lot higher. I don’t know why. I ended up using Shapeways for speed and timing. Also I think it took too long. That’s not always the case. This is not a knock to Sculpteo, because actually in some more recent prints that we’ve been doing in the full color sandstone, Sculpteo has been less expensive and faster turnaround than Shapeways. Even Sculpteo, when I looked at them producing 50 of an item instead of one, actually gave a discounted price for a larger volume whereas Shapeways didn’t.
I think that each of these different service bureaus out there, it’s just a little recommendation and a note to self for you. Make sure you shop around between a few different service bureaus when you’re going to have something printed out and even checking Hubs, any different place that might be available to print your part for you. You got to really shop around because I find different service bureaus specialize more in some different areas and have more experience in different materials and different machines. I think they all pretty much print most of the same materials. There are some that have ones that others don’t, but most of the same materials. If you look into their pricing and their quantity and all that, you’ll find some differences that may meet your needs. Don’t think that every service bureau is just going to give you the same price and go with the first one that you try.
Here’s is the case to be thinking about the complexity of your overall project and your ultimate goal. While we were going to print this at Shapeways or we were going to do some of those sandstones products at Sculpteo, ultimately, when we go into full production, meaning we’ve ran a couple hundred pieces, we’re probably going to go to a different outsource, one that does an entire volume or we can have an entire bed dedicated to our client’s product. It’s one where you could lease out space on someone’s machine basically. Excess space that’s going to be running overnight anyway.
You might go with somebody else to make it, but here again lies the problem. Complexity of doing that, you still really need to make a final design sign off sample, making sure that the tolerance is right and are they even capable of doing that for you? Again, while it sounds like really great to be able to lease the space, you might lose time in trying to get to that outsource because you have to do yet another design sign off sample. Considering all of that and both cost and time and scheduling is so critically important to handling to complexity of these translation issues and these fit and form issues.
It’s really hard to go and be able to use that excess capacity on 3D printers that are out there. If you can’t just do one and test your dimensions, that’s a risky proposition if your part needs to have a critical fit with some other material. I think that’s going to be a very, very difficult situation. For us, we were considering doing it with the angels because we might make 50 to 100 at the holiday time or something like that, but the reality is that they didn’t have to match to each other, there was no fit between that. As long as they didn’t come out distorted or something and they look like angels at the end of the day, it was going to be fine. It just really depends on what you’re printing, whether or not you can allow something like that, you have the flexibility to do something like that.
Anyway, we hope this topic has been really helpful for you thinking about dealing with all of these complexities and translation issues as you’re trying to achieve your ultimate thing, whatever it might be, what the FFF are you printing. We would love to hear from you if you’ve got some tips or translation tools that you guys use that have been successful for you, we’d love to hear from you. Please send us a message on 3DStartPoint.com or anywhere on social media @3DStartPoint. Thanks for listening, everybody. This has been Tom and Tracy on the WTFFF 3D Printing Podcast.
Listen | Download | View
Hear the episode of the WTFFF?! Podcast by using the player above OR click to download any episode.
Help Us Help You!
Have some feedback? Leave a comment below. We will read and respond
Please also review us on iTunes and share via the social media of your choice.