Winn Hardin: [00:00:06] Hey everybody, it’s Winn Hardin, co-host of the Manufacturing Matters podcast. Thanks for joining us for our latest episode. I’m here with my good friend and associate John Lewis. John, how are you doing today?

John Lewis: [00:00:16] I’m doing well, Winn. How are you?

Winn Hardin: [00:00:17] I’m doing great. It’s a pleasure today. So today we’re going to talk a little bit about something that’s a very important topic that doesn’t always get a lot of airtime. And that would be geometric dimensioning and tolerancing, also known as GD&T. And to discuss the topic, we’re here with Tom Geiss, who is the founder and president of GD&T Basics, also known as Engineering Essentials LLC. Tom, how are you doing today, sir?

Tom Geiss: [00:00:40] Doing great. How are you two doing?

Winn Hardin: [00:00:42] Really good. Thanks for taking some time out of your busy day to join us. You know, I’m not sure when this episode will air, but it’s right before Automate. Actually, we’re doing live today, so forget that on-the-air thing. So I really can’t believe you carved out some time for us this week. I really, really appreciate that.

Tom Geiss: [00:00:59] Yeah, this is great. I’m excited.

Winn Hardin: [00:01:00] So before we delve into the deep topics related to GD&T, can you tell us a little bit about GD&T Basics? How long you guys have been around, the training service you’re providing? Tell us what you’re doing.

Tom Geiss: [00:01:12] Sure, yeah. GD&T Basics, as a company, we’ve been around for almost 10 years now. We started in 2015. We are a training and consulting company. We train companies in how to properly use everything from the fundamentals of GD&T all the way into advanced design and inspection topics that they need help with. And we have clients ranging all the way from small machine shops up to large clients that we have like Honeywell, SpaceX, and a few others that we work with. So we mainly cover breaking down how geometric dimensioning and tolerancing works on an engineering drawing and how that is communicated through the manufacturing process and especially into the inspection side, where they have to go in and now fulfill those requirements and make sure that they are met.

Winn Hardin: [00:02:11] Okay. Does this tie into an older term like “design for manufacturability”?

Tom Geiss: [00:02:17] Pretty much. Yeah. Really what we have when it comes to GD&T is really you design with the function of the part in mind. And so GD&T is all about designing to make sure that the part obviously functions, because if it doesn’t function, it’s not a very good part. So it’s carrying though that functional requirement into the manufacturing side without restricting or tightening any of the tolerances that are unnecessary. So, really, GD&T is like a language of specifying exactly what the designer wants without tightening other tolerances that may not be as critical, so that then when it gets to the inspection side, it’s inspected in the exact same way that the function has required. So we’re going to be passing a lot more parts and having a lot more scrap that way, if things are designed correctly of course.

John Lewis: [00:03:15] Yeah, that’s really interesting, Tom. On the manufacturing angle, technology adoption is kind of reshaping the industry and manufacturing right now, with companies transitioning from machine-based assembly to these more integrated smart factories with more and more automation. And in the context of increased robotics and automation, machine vision and automated inspections that are combined with data analytics, can you discuss how GD&T plays a crucial role in ensuring the precision and quality of manufactured products as the process becomes more automated.

Tom Geiss: [00:04:07] Yeah. Of course. I got my start in the automotive industry. I worked for ZF and BMW in my career. So the automotive industry definitely was on the forefront of a lot of the machine learning type things and automated inspections, as well as a lot of the vision systems that would check the quality of the parts and things like this. So this is really not new, at least in the background that I’ve come from. We’ve had camera systems set up, spotting defects and things like that for years. But it’s just to the capabilities of these systems that are coming out now. The way GD&T plays a role in this is GD&T is really the language or the specificity of what the exact requirement is. So it’s not necessarily just used on precision things. The goal of GD&T is to be extra clear of exactly what the designer wants in the final assembly, so that we’re not checking things that may not be as important but we’re really lasering in on those things that really do make a difference in the function of a part. And so you could imagine in the automotive industry, microns make a big difference when it comes to bearing surfaces and high RPMs and things like that. So when it comes to GD&T, we’re able to now communicate not only with each other, but also to these inspection tools, exactly what the parts setup needs to look like, as well as what needs to be checked to verify these sometimes extremely tight requirements that we’re going to be looking at. So while GD&T isn’t just used for things that are precise, it is the language to make sure things are extra clear so we know exactly how to and what we are measuring.

Winn Hardin: [00:06:00] Does this also help the automation engineers, the solution providers, to reduce their cost to basically build that solution, whether it’s an assembly solution, a material handling, an inspection solution, just to the right level of resolution, specificity?

Tom Geiss: [00:06:14] Yeah, of course, and that’s really what the tool GD&T is used for. It’s meant to communicate from the designer to really the inspector and the people making the parts obviously. But parts aren’t just made; they’re also inspected. Carrying it through the whole way. Exactly what the requirements are down to the level of how this part actually needs to be set up in a way to mimic what the designer had the intent of, because there are multiple ways  if you just have, let’s say, a width on a part – there are multiple ways you could actually check that part, and you could get a few different answers, whether you’re checking with calipers or whether you’re putting it up against a gauge block, whether you’re virtually simulating it using a CMM, you might get multiple different results. And so what GD&T is meant to do is specify exactly what the requirements are for that inspection, so that we’re all now measuring it the same way. And so you can imagine that the automation, if we’re adding that now to the factor, that is so much more important now, because if we’re not following the right directions or the right guidelines that the GD&T has set out, we’re now just inspecting things worse, just at a faster rate. So it’s important to make sure that we understand all of these requirements that are coming through and understand and speak the language so that we can automate these machines to do it for us.

Winn Hardin: [00:07:39] Yeah. Obviously, you’re right. Throughput is not the end-all, be-all, right? Yield is the end-all, be-all that we need. Usable parts, as you were saying earlier when we were talking. Now you mentioned automotive. You yourself come from the automotive industry. But I know that GD&T, it sounds like it’s just as critical from the customer side as it is from a solutions provider side, right? Both sides need to be really intimately aware of how all these parts fit together.

Tom Geiss: [00:08:05] Yeah, absolutely. And again, when it’s done correctly, and that’s what keeps us in business is making sure that these companies are aware that this is a standardized language. There are written rules for this, whether you’re using the ASME Y14.5, which is our GD&T standard in the States here, or the ISO standards, which have a bunch of standards that cover a lot of the GD&T requirements. There are rules behind this. And that way when you are communicating from a supplier, whether they’re just down the street or they’re all the way overseas, we all need to make sure that we’re doing things the correct way. Same way. So that way it’ll eliminate disputes. Whereas if I was measuring the part over in China with some of the parts that I have, and they’re setting up the part, I know that when they send it over and I’m setting up the part according to the GD&T rules, I’m doing it the same way that they had to do it, following those directions.

Winn Hardin: [00:09:02] Gotcha. One quick follow-up if I could. Real quick, John, let me follow up on this one. So you mentioned earlier machine shops. We’ve talked about automotive. You work in, is it fair to say all manufacturing industries essentially that are interested in GD&T?

Tom Geiss: [00:09:18] Yeah, yeah. We work aerospace, automotive, heavy industrial. Biomed is a big growing sector. Energy, you name it. If they’re making a physical part, we typically work with them.

Winn Hardin: [00:09:30] And you mentioned a U.S.-specific and the ISO international standard too. What is the sentiment amongst these different industries as they look at GD&T?

Tom Geiss: [00:09:38] When we first started the company, actually one of the reasons I did start the company was there was a big reluctance to start using this language, because it was new. The unstandardized, unofficial way people typically did things was just using plus-minus coordinate dimensions to put everything on a drawing. In fact, that’s the way I learned in college. It was just, if you want to do a width on something, you just say how far it is. Over the course, though, of probably the past decade that we’ve been in business, we’ve seen a big increase in the adoption of using GD&T, where I think it’s safe to say now that it is truly the industry standard way of communicating your tolerances on a print. You have to know how to use it if you’re going to be functional in the modern manufacturing world. Just because, as you mentioned before, with the automation and coming into things, we have to be extra clear of not just what the requirement is but also what the rules are so that we’re all following the same thing, we don’t have disputes, right? Because disputes cost money, whether it’s a good part or a bad part.

Tom Geiss: [00:10:52] Just arguing over it can shut a line down for a while, right? I remember at ZF, if we shut the line down, there were some cases where we were going through $20,000 a minute, because we were arguing over whether one thing was inspected correctly or not. So these decisions have major repercussions, especially in high-volume manufacturing. But you can imagine all over when it comes to quality standards that you’re dealing with. So I think more and more over the years, the more people started to use it, they really feared it less. It’s really not as complicated of a system as people make it out to be. Really, that is what we’ve seen in the more recent years, and I think it’s safe to say that in the past few years, it really is now the industry standard. A lot of the OEMs, in fact, that we’re working with are now making it a rule that their suppliers have to use GD&T on their drawings or they can’t be a supplier anymore, because it just eliminates that need for, hey, we’re arguing over this thing where why don’t we just look to the standard and say, no, it’s this one thing. It’s unambiguous now.

Winn Hardin: [00:12:05] Which is going to save bucks.

Tom Geiss: [00:12:08] Exactly. And you can imagine, all of us who have worked in manufacturing know all of those big crucial meetings that  always happen after 5:00 in the evening. The line’s shut down. We’re trying to figure out how to get parts shipped, red tagged, everything, and GD&T isn’t a solution to that. But it’s standardizing the way so that we’re all talking the same language when we are finding that solution. And we are figuring out exactly what’s wrong, and we now know exactly what to inspect and how to do that.

Winn Hardin: [00:12:42] Thanks. Okay, great.

John Lewis: [00:12:43] Yeah, Tom, we talked earlier about the GD&T in the context of increased automation. And one of the things that’s impacting automation in significant ways recently is artificial intelligence and machine learning. They’re disrupting manufacturing in amazing ways. Do you see the integration of AI with GD&T potentially enhancing the precision and efficiency of manufacturing processes?

Winn Hardin: [00:13:12] Or the other way?

Tom Geiss: [00:13:13] Yeah, of course. As I mentioned, GD&T is really the specificity language, really making sure that we’re clear exactly what it is. You can almost think about it as a programming language too, where if we are going to be able to automate things, we have to use this language to be able to specify exactly what the machine is supposed to be doing. We have a couple clients that do laser scanning technology on automated robotics arms. So they basically scan the part, but they go all the way around and they know exactly where to go with that. Well, if we’re not identifying things like what are our datums or the origin our measurements are supposed to come from or we’re not communicating to the machine what exactly are we checking on an engine bore and specifically how round it should be in addition to the size that it should be, things like that, we’re going to be missing a lot of these requirements. And the automation only goes so far as the specificity that we’re entering or putting into it for those requirements. And so I think GD&T is going to become more and more important as we get more automated with things, because it’s going to be the language that we use to tell the machine, hey, this is exactly what we want you to do.

Winn Hardin: [00:14:35] I’m curious, using GD&T standards and language itself to define tolerances, do we see that playing in or does that play into any SPC, statistical process control, or optimization? And I guess this ties in a little bit to John’s last question on artificial intelligence, where you might use some data mining to help determine, well, we’re seeing this reoccurring issue, let’s start looking here as the source of the issue. Does it feed back into a closed loop manufacturing optimization process?

Tom Geiss: [00:15:10] Yeah, it definitely has. And coming from the automotive world, SPC is our bread and butter. We can’t measure every part, right? So getting into the statistical process control, we need to be very specific of what those requirements are, high and low as well. And typically GD&T also can feed right into that same style, where we have an SPC requirement that is a little looser when it’s measured over the course of 1,000 parts. But if you measured an individual part, it then has to be tighter. The same rules apply with GD&T. It’s just another way of doing it versus just using the coordinate dimension itself. It’s also going to help eliminate those disputes that do pop up, because now your supplier has to follow the exact same rule that you are when you find that defect, and you’re all going to be doing it the same way to make sure there’s no dispute in the method of discovering the problems or the root cause analysis that you’re doing.

Winn Hardin: [00:16:12] Gotcha, gotcha. So you may be more familiar with CAD QIF than I am. It’s my basic understanding that we’ve got CAD design files and we’re including some manufacturing tolerances into these design files, essentially, right? So maybe if you’ve got facilities around the world, we’re binding together the actual product dimensions with how it’s manufactured, tolerances at each step. Do you see GD&T coming together, maybe leveraging QIF in the future, which I think is just in its early adoption phases, if I’m not mistaken?

Tom Geiss: [00:16:44] Yeah, I think QIF is probably the next stage in the side of what we call a model-based definition, or MBD, where the first step is putting the tolerances on the 3D model itself, right? Because typically, probably 90% of the industry, we’re still using 2D blueprints to do everything. We have the three views and that’s still the way. Ten years ago when I started the company, we heard rumors that MBD, model-based definition, is going to be the future. We are seeing some OEMs and some places adopt model-based definition, having the 3D model carry everything forward. But still, if your supplier doesn’t have the same CAD package as you do, you’re going to have to somehow communicate that down to your supplier as well. In fact, I have a funny story at BMW, we did a lot of MBD early on, model-based definition. And they also did a little bit I guess you could call it QIF, where they carried forward some of the CAM data also, I believe, from the suppliers on there, just so that they can inspect certain things. I wasn’t involved in this decision, but they wanted to carry the MBD that us in the design world were already using to inspect parts, looking at tolerances and seeing. There was a rumor that they wanted to carry it down to the shop floor.

Tom Geiss: [00:18:14] Unfortunately, in order to do that, you need some way, if a machine is going to be looking at the model, they need some way to manipulate the model now, right? Because they don’t have a blueprint in front of them to see everything. They have to actually be able to look around, see the different tolerances, see what the requirements are and all of that, because there’s still the human element to everything. So this was going to cost a lot of money. They had to add a bunch of monitors around. And I remember someone mentioning that they were thinking about doing it, and someone suggested, “Why don’t we just take maybe a picture of the front of the part and maybe the side of the part and maybe like the top of the part, and we’ll have that.” And everyone was laughing because they said, “Well, that’s how 2D prints have been working for over 100 years.”

Winn Hardin: [00:19:05] There’s another problem we need to lay at the feet of Google Glass. If they would have just gotten it together.

Tom Geiss: [00:19:11] And I’ve seen some with the augmented reality too, being able to actually see the part and holding the part in your hand and be able to do that, maybe that’s where it is. But all these things have costs. And I know a lot of machinists, and I’m not sure if they’re going to be sticking those goggles on their face anytime soon with some of the technology coming through. So we still have an old school industry that adapts to new technology. It doesn’t mean that the machining industries and the quality industries aren’t adapting to this new technology. It’s just they have to also be able to retroactively go back and work with the existing way that, again, probably over 90% of the industry is working, to slowly carry that transition forward as they do it. But the good thing is, there is some progress. The 2018 version of the recent GD&T standard, also outlined with the 2D blueprint version of the GD&T requirements, they had side by side with the digital definition model version of the requirements that you also would include on your drawing. So now at least the ASME committee is also committed to doing it in tandem. The two-dimensional way and the three-dimensional way as well. So we are seeing progress on that side, carrying this information forward. We are also seeing a lot of very high-tech industries starting to add their CAM data to things, starting to add this quality data moving forward. But if the knowledge isn’t there of what the requirements are actually telling you, if the knowledge isn’t there about, even when they’re programming these QIF files, again, you’re just going to be doing things automated but not exactly the right way, the standard way.

Winn Hardin: [00:21:10] Garbage in, garbage out. Which speaks to the importance of having a full, in-depth skill set on your design team, both on the customer and on the solutions provider side of GD&T and how all these elements come together. How long is a standard course? Can you tell us a little bit about what you guys do and how much is hands-on and how much is book work?

Tom Geiss: [00:21:32] We typically offer three versions of our course. We have a full video-on-demand version of the course, where you can take it at your own time. It’s anywhere from six to eight hours for a fundamentals course that we have. And it walks you through all main 14 symbols of GD&T but also the more important concepts such as datums and parts and just the overall rules that a lot of people don’t consider when they are looking at GD&T, kind of the infrastructure of a part. So it’s typically about six to eight hours for that course. We also do virtual training over the web, just like this. We have public training, where we have a bunch of people from all different companies. We also do custom training, where we can work with just one company. That way we can open up your drawings. We can actually talk about those and have more privacy when it comes to things. And we still have our number one thing that we do is our on-site training, where we actually go to your facility and do the training with your drawings, with your parts. We love taking little field trips to the quality lab and/or the shop floor to actually talk about the real way that these are implemented into them.

Winn Hardin: [00:22:51] Gotcha. So the model-based design that we were talking about earlier How is the adoption going? Where are we at on the adoption curve?

Tom Geiss: [00:23:03] You know, the OEMs are obviously going to be guiding it and implementing it first. That’s typically what we see, the larger companies, and they carry it down to the lower levels. I’ll tell you what the biggest hurdle, though, at least that we see with the adoption of the model-based definition is being included as standard in a CAD package? Seems to be the biggest hurdle holding people back, right? If you buy a CAD package, you get the 2D blueprint thing for free, right? It’s included. You get the assembly modeling for free, but you’re going to be adding on this model-based definition thing. You’re going to have to add it on, and now you’re going to have to hope that not only you added it on but your supplier added it on and your customer has the add-on. And now that they’re all talking the same thing. So where an add-on like a finite element analysis tool is helpful for you to design a part, the model-based definition is how you’re actually exporting and communicating that part to everyone else. It doesn’t help if you just have it. You need everyone else to have it too. So I think that’s where we’re seeing a lot of the big hurdles, at least in terms of the adoption, is just making sure everyone has a universal CAD package that can read those, anywhere from MBD all the way up to those QIF files, to be able to open them up and visualize them and see them.

Winn Hardin: [00:24:29] So are SolidWorks and all the rest of the crew, are they offering this as an option? Are these modules available?

Tom Geiss: [00:24:34] It’s an option. I believe it’s an option on all major CAD packages: Inventor CATIA, Pro/E, Creo, all of those. You can have it, but it’s just making sure that, especially if you look at automotive, for every OEM you have tens of thousands of tier one, tier two, tier three suppliers that are all integrating into that final assembly. The OEMs have been probably adopting it for years. It’s just as it trickles down into the lower-level suppliers, that’s where the 2D print, everyone has a printer and can open PDFs, right? And everyone also can open STEP files, which are  kind of the dumb models, right, which don’t have that tolerancing data. But what we’re seeing as the way in between is you open up your STEP file, which is your model, and now you have your 2D print.  It has all the tolerances. The machinists, that’s all they really need to to get rolling on that.

Winn Hardin: [00:25:37] Maybe cloud-based computing will somehow level the playing field and allow every step in the supply chain, manufacturing chain to actually be part of that collaborative group.

Tom Geiss: [00:25:48] Absolutely. Yeah. That would be a major, major thing. I know the Fusion 360 side of things, they’re doing a lot with the cloud computing. Also SolidWorks is doing a lot with that, making the universal formats that can be carried forward. And I think that’s going to have to be the big change that is done. It’s just going to have to be a standard package as opposed to an add-on.

Winn Hardin: [00:26:11] Beautiful.

John Lewis: [00:26:13] Yeah, Tom, we’ve talked a lot about automation and AI and digital, digitization. I’m wondering, is there anything else that we haven’t talked about yet today that you think would be meaningful for our audience?

Tom Geiss: [00:26:25] Well, at least when it comes to this geometric dimensioning and tolerancing stuff, it seems pretty advanced for some people sometimes. And it can be when you’re first getting into it. You know, everyone sees true position, these datums, and max material condition, all these symbols on the drawing, and it seems overwhelming at first. And that’s honestly why we developed our website, gdandtbasics.com.  The reason I originally developed it was I became a local GD&T expert at my company I was working at. And in order to communicate at least what these things were – there was a little bit of confusion on certain requirements – so I created specific web pages for all these symbols and things like that just to help out with communication. Instead of emailing them a long explanation, I just emailed them a URL to something I created. And we realized just how many people that was helping. And so even though, yes, we do have paid training and we do paid consulting, 90% of the stuff that we give away is free on our website. We have wall charts that you can download for free just by going to the site. It benefits all of my clients if everyone gets better at this. So overall I want the entire industry to get better at understanding GD&T, because it really elevates us in the manufacturing industry if we’re all speaking the same language, right?

Winn Hardin: [00:28:03] Absolutely.

Tom Geiss: [00:28:04] So that’s our goal as our company. Beyond just helping our clients, we want to really elevate the entire industry. And so I wanted to send people to our website. Just download. We have a ton of stuff. We have information on every symbol you could possibly have in GD&T. All for free. So I want everyone to take advantage of that.

Winn Hardin: [00:28:26] Beautiful. Sounds like we all need to bookmark gdandtbasics.com and go there and pilfer as much as we possibly can. So thanks again for all the work, time that you’ve put into that solution. And I look forward to seeing more about the training elements. Are you going to be in Chicago next week?

Tom Geiss: [00:28:43] Not next week. We’re going to be there for IMTS though in September.

Winn Hardin: [00:28:48] Beautiful. Do you happen to know the booth number off the top of your head, Tom?

Tom Geiss: [00:28:54] We can put it in the show notes. We’re in the Quality Wing. I think the west end of the building there. But come look at us. We’re in the manual. I gotta memorize that one.

Winn Hardin: [00:29:09] I don’t blame you. That was not fair calling you out on the fly like that. But if anybody does have any questions in the meantime, you can always shoot them here in the message display during this live session. Or if you send us a question, we’ll make sure it gets to Tom after the show. Of course, you can go straight to his website I’m sure and submit any queries you have on their contact page and see what else is on their website. So I really, really want to thank you, Tom, for taking a little bit of time out of this super-busy week and sharing with us why we should all care just a little bit more about GD&T. I didn’t even get to ask you whether this fits into CEU credits or any other qualification. I gotta think this is going to make you a little more valuable to your organization. If you’re the one guy or lady who knows these tolerancing issues, so job security, maybe a little bit of a raise. If you’ve got any anecdotes you want to share there. But until then, guys, send the questions in. If you want to be in a future episode of Manufacturing Matters, check us out on our old shows, you can do Manufacturing-Matters.com. You can reach us at Tech B2B Marketing or Manufacturing Matters on LinkedIn or anywhere you get your podcast. We’re distributing across all major platforms. Until the next time, just remember that manufacturing continues to matter in this world, and I can’t wait to see you next week in Chicago at Automate 2024. Thanks again, Tom.

Tom Geiss: [00:30:29] Thanks so much.