AMT Tech Trends: The Force
Release date: 11 September 2020
Episode 33: Now with extra rambling! The boys start by talking some smack about 5G. Stephen is learning about atomic force microscopy and then how COVID test swabs are being reinvented. Ben goes on about General Motors getting something right and linear algebra. Steve introduces a new robot OS that can turn conventional robot arms into cobots! Ben says the U.S. Army has a new additive policy. Steve tries to explain Toshiba’s robot development. Ben wraps up with how Tesla is making a car part less complicated but probably more expensive.
- simple.wikipedia.org/wiki/Atomic_force_microscope - www.expressandstar.com/news/health/c…st-for-virus/ - www.sme.org/aboutsme/awards/wu-…d/wu-award-winners/ - roboticsandautomationnews.com/2020/09/09…ots/36158/ - www.army.mil/article/238868/new…escalating%20costs - us.acrofan.com/detail.php?number=337160 - www.core77.com/posts/101742/Tesl…medium=from_title
Benjamin Moses: Hello everybody and welcome to the Tech Trends podcast where we discuss the latest manufacturing technology research and news. I am Benjamin Moses, the director of manufacturing technology and I'm here with ...
Stephen LaMarca: Stephen LaMarca, the technology analyst.
Benjamin Moses: Steve, how are you doing today?
Stephen LaMarca: Doing all right, man.
Benjamin Moses: Great. It's been a busy week. I've been watching a lot of TV for some reason and-
Stephen LaMarca: Good for you.
Benjamin Moses: It's the holidays, man, what are you going to do?
Stephen LaMarca: Yes.
Benjamin Moses: I've seen a lot of commercials for 5G. Is 5G important to you?
Stephen LaMarca: Oh, geez, no. No.
Benjamin Moses: It seems like it's the new hot thing, but ...
Stephen LaMarca: I mean, we've got to find somebody who will actually utilize it. But then again, no, we don't because we still need to find people that not only need to utilize the high speeds of an actual hardwired connection, but people who are willing to upgrade their entire internal network. When somebody needs the speeds enough to the point where they say, "Okay, let's try to upgrade our internal wired network to the speeds that 5G claims to offer" and then they see how expensive it is to rewire everything. I mean, my computer I think if I wanted to run 5G speeds on a wired internet connection not only would I need an internet connection fast enough for those speeds, but even if you bought the internet connection that would be of that speed I personally would need a new motherboard too because the motherboard's ethernet jack doesn't support those speeds. I want to say it's kind of like 4K TVs or to go back to computers, ray tracing. You don't need a ray tracing graphics card, because there's no ray tracing capable ... There's not enough ray tracing capable video games right now. And then there was the whole hubbub a while back that everybody thought when the RTX series of graphics cards by Nvidia came out that were the supposedly, allegedly, the first ray tracing capable GPUs. A year later after the RTX series came out they found out if you have a GTX 10 series version of graphics card Nvidia just released an under the radar software update, so if you have one of the most recent GTX cards, like the generation of graphics card before the RTX card those are now ray tracing capable. It's going to be that, but on a much more expensive level. If you need fast internet speeds, slow down and find out what's going to be the most economical for you.
Benjamin Moses: Yeah. I find it really interesting that you're chasing the bottleneck, especially for on the consumer side mobile applications. I remember the first time I got a data-enabled phone, a really, really long time ago when I first started working. It was great. I could get emails. I had apps. It was a Windows-based phone-
Stephen LaMarca: How nice.
Benjamin Moses: ... which in retrospect back then when I first used that phone, it's just as capable as my latest Android devices. I could copy and paste. I could run multiple applications and this was ages ago.
Stephen LaMarca: Oh, you could do split screen.
Benjamin Moses: Yeah. You could do split screen. It was amazing. But, I actually turned off the data after about a year realizing one, the cost of it. Back then, it was fairly expensive for unlimited data, but then I was doing basic stuff where if I'm sitting in an office or even at home I could use my desktop or other applications or other devices. I didn't need to be mobile all the time or if I did I could turn it on for a month if I was traveling, which I have done. When they extend to what I'm doing now, I still use it just for phone, just for email. I use it for Google Maps, trying to make sure I get home at a decent time or Waze if you're that kind of person and social media. I'm struggling to find where 5Gs really going to make my life better as I'm walking around the city or if I'm driving around or traveling. If I'm streaming videos or if I'm doing anything that requires any decent bandwidth I'm going to be on a WiFi network.
Stephen LaMarca: Yeah. I think if you're streaming 4K video to your ... Wait a minute, before I even get onto that and I do want to touch back on that, what you've said about chasing the bottleneck is absolutely correct. You need to prioritize your bottlenecks because if you're going after 5G and you need 5G speeds you have other bottlenecks.
Benjamin Moses: Exactly.
Stephen LaMarca: Sure. I'm not saying that internet speed or network connection speed is an important bottleneck for you, but you probably have other bottlenecks at that point. To go back to what you were saying with your mobile device, if you need 5G to stream 4K video to your mobile device, you really need to consider other bottlenecks. Does your body physically have irises capable of registering 4K video quality on a screen that's five inches wide in your hand on a moving train? Are your eyes capable enough for that?
Benjamin Moses: Yeah.
Stephen LaMarca: Can your eyes actually utilize the speed of that internet connection for the 4K video? Will the screen of the cellphone or mobile device turn into a hotplate that you can boil water on?
Benjamin Moses: Right.
Stephen LaMarca: Will the battery last more than 15 minutes?
Benjamin Moses: Yup.
Stephen LaMarca: Internet, sure, it's the next step and I want to support ... I do support faster internet. Everybody loves faster internet and it's not so much that they love fast internet it's that they hate slow internet.
Benjamin Moses: Right. Now, I will give you the ... I don't know how big the use case, so on the East Coast I would say this is probably a really small population of users that are going to rely on this, but in real applications, especially when you look at the global usage. Back in India, wired connections are few and far between unless you're actually in the office, so most people ... The big thing was to have a landline, so you were successful if you had a landline at your house. That jumped from everyone get a landline to two percent of the population that has landlines to 80% of the population has cellphones because it's just easier to activate a SIM card than running a physical landline. I think that's in the US also. I wouldn't say it's easy to get landlines or running fiber line to people's homes is practical in all applications. Also, I mean, I do see that in the city too because as the increase in congestion and population densities in cities it's getting more difficult to run utilities in some locations, but you still have to run power, so if you're running power, other utilities, a single cable probably doesn't hurt, but it's just interesting.
Stephen LaMarca: You're absolutely right.
Benjamin Moses: We'll see.
Stephen LaMarca: I think you're absolutely right on that. When I was Vermont in New England shooting outside, not Outside the Shop, Road Trippin' with Steve for IMTS Network, as much as I love Vermont and it was so awesome getting to go back there and visit there and just smelling that fresh Vermont air and seeing how beautiful it was, I just stop. I forgot the ugliness that was it's hard to get a good signal out there.
Benjamin Moses: Yeah, that's true.
Stephen LaMarca: I think 5G, I'm sure Vermont would be happy if they get 4G LTE right now, you know?
Benjamin Moses: Yeah, yup.
Stephen LaMarca: Everywhere that is.
Benjamin Moses: I see.
Stephen LaMarca: I was at a gas station and struggling to get Google Maps running properly.
Benjamin Moses: There's one spot when I drop my daughter off at daycare, when I park in front of their house, my phone drops to 3G for some reason. I don't know why.
Stephen LaMarca: New development?
Benjamin Moses: No, no, no, no. It's probably just some low spot in the signal. One thing that does hurt you as you increase your frequency speed or your frequency is you lose capabilities to penetrate through buildings. That's one drawback about 5G is the frequencies are so high that the ability to penetrate walls or buildings are actually pretty poor so you lose 5G strength in the early implementations that I've seen. That's where 4G and 3G are better because the frequencies are lower, but whatever.
Stephen LaMarca: I think I saw a video. No, I did see a video last year, that one tech guy on YouTube, Marques Brownlee. I probably totally mispronounced his name.
Benjamin Moses: Great production value. I love watching his videos.
Stephen LaMarca: He's really good production value. He did a video on 5G on the first neighborhood in California or LA that had a 5G network up and running. He showed his phone that had this blistering high speed on speedtest.net or whatever the website is. I know I use fast.com now or fast I don't know, man. I haven't had to check the internet in a while, first world problems. He showed a speed test on his phone directly underneath the 5G tower that everybody thinks is spreading coronavirus. That's fake news. He was directly under the transmitter, if you would, transceiver, and it showed crazy high speed on his phone, like 15,000 megs or something crazy like that. Then he was like, "Okay, I'm right underneath it right now and we're getting this high speed and I'm on 5G, let me walk across the street."
Benjamin Moses: Just across the street.
Stephen LaMarca: Just across the street the signal totally dropped. I mean, it didn't drop. He was still getting like 3K meg, but ...
Benjamin Moses: You're still in line of sight, just a little bit longer distance.
Stephen LaMarca: Yeah. We're talking still line of sight, across the street. Forget going through a wall. That was cool.
Benjamin Moses: Let's shift gears. You mentioned something about AFM?
Stephen LaMarca: Yeah. Since Monday, two days ago, I've been seeing a lot of stuff pop up on Tech Trends about atomic force microscopy. Got lucky right there, I said it right.
Benjamin Moses: Don't say it again.
Stephen LaMarca: Atomic force microscopes. I'm not going to try saying it again. AFM, atomic force microscopes, microscopy, a form of metrology, because we're measuring something here.
Benjamin Moses: Sure.
Stephen LaMarca: AFMs came up a lot on Tech Trends. I've seen a handful of articles talking about stuff like how a new method to reduce noise in AFM and I'm like ... I love metrology news and I wanted to read all about it and I started and I was like, "Dude, I don't even know what an atomic force microscope is," so I hate to break it to you and let you down to say that I don't have a particular article, of the articles that I've seen, to talk to about because before I even get into any of that news I wanted to learn, dude, what is an atomic force microscope. I went to one of my favorite websites to get me learned on things. Wikipedia, simple English, and I typed in atomic force microscope. Low and behold, they had an entry, which was awesome. Basically, what it told me is an AFM, an atomic force microscope, is very similar to a ... Well, it's similar in one respect to a scanning electron microscope or an electron microscope in that it can see and visualize and thus measure something as small as onto the submolecular level. With an electron microscope, you can actually see the alignment of electrons in your whatever it is that you're looking at. An atomic force microscope can actually see even closer to the atomic level, which is ... An electronic is actually subatomic, so I don't know why they're comparing them like that. Anyway, the difference is they're both similar in that we're talking the atomic level-
Benjamin Moses: Very small.
Stephen LaMarca: ... or even subatomic visualization and measurement. The cool thing that I found when reading the Wiki article was an atomic force microscope the differences are an atomic force microscope is actually really similar to ... It's on a quantum level in that it's a really, really, really small probing measurement arm. Originally I wanted to say it's an CMM and then it's like that's inaccurate. It's more like just a probe arm, really small. Because you can't have sensors because sensors aren't small enough to physically measure the movement of that kind of probe while the probe is really small and can measure the smoothness of an atom, if you would. What they use instead is, and this is the microscope part because when you think microscope you think optics and when we're talking about a physical probe arm that's not optical. The optical part is the arm has a reflective surface on it that goes up and down with the surface that it's feeling and a laser is shined on that surface and reflected back and as the arm goes up and down and fluctuates the distance that the photons travel and that laser beam is either elongated or shortened and that's how they quantitatively measure the smoothness or whatever it is they're trying to measure with the AFM. I just thought it was wild because while I'm not sure what all these articles have to do with the manufacturing industry because it's way more science-y than it is manufacturing right now, it was cool because it's just like a manufacturing probe arm. It's just like manufacturing metrology, but on a quantum level. The manufacturing industry is constantly in search of that almighty micron. AFMs laugh at the micron. Micron, we're talking subatomic.
Benjamin Moses: It is-
Stephen LaMarca: You get out of here with that micron.
Benjamin Moses: It is funny you mentioned that because I have a couple articles on getting to an improved, accurate machine component and the trend of subtractive manufacturing getting to better surface finishes and more precise manufacturing, we're just a bunch of years away from being light years away where we used to be 20 years ago.
Stephen LaMarca: Oh, yeah.
Benjamin Moses: Where we're seeing that is increased horsepower, increased fuel efficiency. We're seeing moving components last longer. While this is super interesting to me being able to drag a arm and then using a laser to get a topographical image on this thing that you ... It's incredibly small.
Stephen LaMarca: You're using all the right words.
Benjamin Moses: It's super small, but in 10, 15 years something along those lines is going to be required because of how precise manufacturing is going to become.
Stephen LaMarca: Oh, yeah. Even though right now this is the expensive new technology, in the future the older technologies that are simpler and less accurate are actually going to be more expensive because the technology is having to back peddle to keep up with those old products.
Benjamin Moses: Sure, yep.
Stephen LaMarca: That's something I've recently seen with the firearms industry. When you look at the 1911, that 100-plus-year-old pistol design, which I don't want to offend any Boomers, I'm not going to smack about. It's one of the guns that won the world wars, but those guns are so expensive compared to more modern pistols and handguns because they're not manufactured. They can't be manufactured using modern machines, modern means rather. Those guns were made through hand fitting and stuff like that. They were certainly more modernized in terms of manufacturing than something like a musket, a flintlock or just a lock action in general, sidelock. Compared to a Glock, something like that, which I don't think sees any human interaction in the manufacturing process other than assembly. I bet you robots even assemble those before they go in the box.
Benjamin Moses: We should investigate that.
Stephen LaMarca: That would be great. Once we're allowed to fly overseas or even-
Benjamin Moses: If we ever.
Stephen LaMarca: ... travel at all. I'll live in New England.
Benjamin Moses: Awesome, Stephen, let's get into some articles, man. I'm excited to learn about-
Stephen LaMarca: All right.
Benjamin Moses: ... some stuff you found.
Stephen LaMarca: Yeah. Switching gears again. It was funny seeing coronavirus test swabs coming up in manufacturing news again because it's like, "Come on man, that's like so three months ago, five months ago even." But then again, at the end of the day, we're still in the midst of a pandemic and what did I just say, "We can't travel anywhere because this is still a thing." A university in the UK, let me try to pronounce it properly. I bet you I'm going to butcher it even though it's English. Wolverhampton University or the University of Wolverhampton in the UK has designed a 3D printed swab that more effectively and more efficiently tests for COVID.
Benjamin Moses: That's cool.
Stephen LaMarca: There's really not much in the article words, but in the picture of it the swab is really cool looking and when you see the design of the swab the end of the swab that actually collects all of the material that is to be tested for COVID is like a lattice and when you see this swab it's like that swab can only be made by additive means, so-
Benjamin Moses: That's cool.
Stephen LaMarca: ... it was really cool. It's awesome to see that the pandemic is still here. That's not the good thing, but the awesome thing is we're still innovating amidst the pandemic. So much has been done. Technology has advanced so far just in the past five months to help beat this virus, which is really cool.
Benjamin Moses: I do find that is a fascinating approach to get rid of the cotton swab basically that they would use to probe the entire nose to a plastic lattice that will collect basically the same type of material. If you look at getting or improving supply chains, things like that, you get rid of the need for cotton and the entire ecosystem of that to just growing a single growth part.
Stephen LaMarca: Mm-hmm (affirmative). A single material.
Benjamin Moses: Yeah, single material. That's awesome. That's really cool article.
Stephen LaMarca: What did you have for us?
Benjamin Moses: The next one I got is a little more techy, it's about multi-axis compensation.
Stephen LaMarca: Nice.
Benjamin Moses: This is an award from SME and NAMRC. I have attended their technical conference for a bunch of years since I joined AMT and they have an award, the SM Wu Research Implementation Award. It's a fascinating award when they track a bunch of the research projects that are discussed in their conference and then they see if it has made it to any type of commercial application. In this particular case, a couple of GM Motors engineers developed a way to adjust their compensation for multi-axis subtractive manufacturing process lines. A few men in General Motors, their machining all types of different widgets, both in house and of course they have their entire supply chain that's doing it also, but they are processing a lot of parts themselves. The biggest use case are their own motors, their own engines.
Stephen LaMarca: Right.
Benjamin Moses: Imagine if you've got a machine, this cast engine block and it's fairly complex. In this case, they have a multi-axis machine that's doing it. And then once you have your engine block you've got your tooling, you've got your rotary tables and you've got your spindle. You have all these places of errors potential, based on compensation that needs to be accounted for. If you imagine the simple three axis, everyone does a tooling comp the size of the diameter of your cutter. You can adjust that compensation. There's probably one or two other axes, offsets, that you can incorporate. But what they've done is they took a step back and looked at the whole process of calibrating the machine in trying to figure out where are the errors that they need to compensate for, the global compensations that they need to account for. What they did is there's a couple of techniques that you can do now. You could use bull bar tests, you can use lasers and a bunch of things, but those are kind of independent of the part and it takes away from the process because you have to do that every once in a while and you're trying to constantly run production parts. What they decided to do is that, let's take one of the production parts that we machined, do a ton of inspection on it so we get data rich analysis of what the part looks like and then use that data to feed back into the global compensation. What they're doing is they added probably a little bit more inspection data to some of the production parts, but they continued running the parts and using that data to feed back into the global compensation for that specific machine. Now it's not continuous data ongoing, so what they do is they run it for a while and then maybe the next week they'll run it again and do this test, but it's a fairly complex process.
Stephen LaMarca: It's not for every part.
Benjamin Moses: It's not.
Stephen LaMarca: It's like batch test. It's a form of batch testing-
Benjamin Moses: Right, exactly.
Stephen LaMarca: ... in compensation.
Benjamin Moses: Yeah, it's not-
Stephen LaMarca: When you say compensation, do you mean like before they even have a part they design whatever they're trying to make like an engine block. They design the engine block because it's a GM and because it's GM I imagine it's a 350, but they design the engine block in CAD and they have their CAD design. They move it to G-code, they run the program in the machine. They spit out their first part and then they scan the part every which way and essentially scan it until they have a 3D model of the part they just made and then they compare the part they just made to the original model. By seeing what changed, what may or may not be in and out of spec, they use those differences to compensate for the next part being made or in this case because it's not ongoing the next batch of parts being made.
Benjamin Moses: Right. Right. There's a couple things that are semistatic, so they're compensating for a couple things. You have your different tables. If there's an A and B table on your machine you can compensate for those. Also you have to compensate for your fixture. If you have a static fixture that's in your machine that's compensated for and then any other additional rotary offsets. As long as none of that changes, those compensations should be consistent, but in manufacturing things wear, you may bump it or things will wear down, so realistically-
Stephen LaMarca: It's about wearing parts.
Benjamin Moses: Yeah, exactly.
Stephen LaMarca: How many times is a part being fixtured.
Benjamin Moses: Yeah, exactly. As soon you take that fixture off and remount it somewhere else, you're probably going to have to run-
Stephen LaMarca: It's totally different.
Benjamin Moses: Right. Right. If you look at setting up a machine, I would say it's part of the setup process. As long as that one fixture, one assembly, one type of process is constantly being run that data is probably still valid. Now you will want to run validation checks to make sure things haven't been ... That the third shift guy didn't drop coffee on it or bump it or something.
Stephen LaMarca: Nothing's out of spec.
Benjamin Moses: Nothing's out of spec, right. It's a really interesting process using the existing data so you're not shutting down the machine to verify compensation using the data that you have on hand to drive the machinery. Now, in looking at the research paper, it's a really complex math. It's funny because [crosstalk 00:26:07] they pull a lot of data and they're doing a lot of arrays and matrices and linear algebra, which is the most complex algebra of all time. But then once they distill it into the end goal of improving accuracy and therefore improving throughput, this is really tangible results that they're seeing on the floor.
Stephen LaMarca: Wow, man, I just got a flashback. I remember taking linear algebra. Well, first off, the hardest math that I had ever taken. When I say hardest, it was the difference between the last math course that I had taken, which was Calc III and there was the huge jump to differential equations. I struggled with differential equations. Linear algebra was definitely harder, but I did better in the class, just because I feel like the jump wasn't as drastic. Oh, man, linear algebra is intense. I don't want to go back.
Benjamin Moses: Definitely flashbacks of sitting in some small classrooms in university. It was tough. What article do you have?
Stephen LaMarca: Oh man, it's going to be tough to follow that one. I saw this really cool article and I figured we can have a nice little discussion about this one. From Robotics & Automation News, they had a headline that's new universal operating system makes robots interact and learn from humans and other robots. That was the headline, but forget the headline, the news is that there's this company called Qobotix and the spelling of Qobotix is really cool. It's Q-O-B-O-T-I-X, so very trendy and youthful, but great name. I think it's an awesome name, Qobotix. What Qobotix is doing is ... What they've done is they've created a robot arm operating system that can be used on whatever robot you have. It's kind of like immediately I thought of Android versus iOS. An Android phone and Windows computers, Windows operating system doesn't run on Windows brand computers. No, they run on Dell, they run on HP, they run on insert brand here. Apple is different, but they're different. I think it's really cool that Qobotix has made a universal operating system that can be used on whatever robot you have so as long as-
Benjamin Moses: That's cool.
Stephen LaMarca: ... the manufacturer supports it. I think this comes up with two positives. One being that it gives you the freedom to experiment with different brands and models of robots. If you're used to one robot at your facility and your guys are buying a second one, but it has to be different than the one you currently have, the learning curve won't be as difficult because you can use this new operating system and both robots can run on the same operating system also. Also too, the big topic that everybody's blabbing about today, supply chain shifting and production lines. Yesterday you may have had two different manufacturing cells, one using a KUKA robot and the other one using a FANUC and now you need to do a production line shift and maybe you need both of those two different robots in the same cell. What's the quickest and easiest way to get them working in sync and working together and communicating properly?
Benjamin Moses: Right.
Stephen LaMarca: It would certainly help if they were using the same software that operates them, the same operating system. That's the other one. I don't know enough about it and I'm so interested to hear more about Qobotix and I'd love to learn more about this operating system. I'm also itching a little bit to get back into the office so we can play with the test bed again because I really miss our cobot.
Benjamin Moses: Yeah, absolutely. You mentioned-
Stephen LaMarca: I remember unboxing our cobot and seeing how easy it was to use. I can't imagine how cool it would be using that thing and then getting another robot. That's made by UFACTORY. If we got a FANUC and getting it up and running was just as easy as the first one because they're both on the same operating system now I think that would be really cool.
Benjamin Moses: Yeah, I think you mentioned a couple of good trends that I just want to recap.
Stephen LaMarca: Sure.
Benjamin Moses: Being able to shift production lines, so getting to a more flexible high mixed, low volume type of environment. I think that tremendously helps you if you have multiple robots. The idea of not going with one brand across your factory, but being able to choose the entire spectrum of robots based on the individual need. You could have a higher model on one side versus something that's less expensive on the other side. Completely different brands. One could be a cobot, one not an industrial, just the traditional industrial robot. I think that's very interesting and I'm definitely interested to learn how it works with ROS, the robotic operating system, right?
Stephen LaMarca: Mm-hmm (affirmative).
Benjamin Moses: This sounds like it could be a layer on top of that where it might add some intelligence or some analytic power to the platform. I definitely agree that being able to have one way to operate the machinery, multiple machines, is definitely a really great approach.
Stephen LaMarca: I mean, we're talking about robots right here, but how great would it be to as a machinist to say, "Oh, if you've ran one 5-axis CNC mill, you've ran them all." I think this universal operating system will give robot programmers the opportunity to say, "Well, if you've run a seven joint universal robot then you can run any seven joint robot."
Benjamin Moses: Yup, awesome. The next article I have is about additive in the Army, big Army. This is not the little T-handle thing that they did a while ago. That old useless [crosstalk 00:32:45] tool that they used for the M249, but this is when there's subdivisions, the US Army Aviation and Missile Command. They've been doing actually quite a bit. They partnered with the Wichita State University. The university has a branch called the National Institute for Aviation Research. We toured their facility with one of their committees last year, I think, and it's impressive. It's really, really impressive at Wichita State. They do not just defense stuff, but they do all aviation type research. One of the interesting things that came out of there was they helped ... When you go on the plane the size of the liquid bottles that you're only allowed to take, they help determine that size based on what's the worst chemical that they can put in there and if it explodes how much damage could it do the fuselage.
Stephen LaMarca: Whoa!
Benjamin Moses: I mean, that's the type of fundamental research-
Stephen LaMarca: They came up with that figure.
Benjamin Moses: Well, I don't want to point all the fingers at them, but they help support the FAA in getting to that figure, right? [crosstalk 00:33:44] And of course layers and layers of safety on top of that, so it's not just that's the maximum that's required, but that's something they're comfortable with. That's the type of stuff that they're doing at Wichita State there. One of the interesting things about this article that I found was the aviation side of the US Army is running into platform problems where their aircraft is lasting longer than the engineering tools that they have that designed it. One example is the Black Hawk. Steve, who's older? The Black Hawk or me?
Stephen LaMarca: The Black Hawk.
Benjamin Moses: The Black Hawk is definitely older than me. The first flight for the Black Hawk was in 1974. Imagine if you're an engineer in 1974, you're probably using a piece of Mylar and doing a drawing on top of that with ink and a pencil. You may transition in the 80s to some type of CAD or maybe doing some finite element analysis at that point also. But, we're in 2020 and we have parts that don't technically exist anymore in the engineering world, right?
Stephen LaMarca: Yeah.
Benjamin Moses: If they've got an older aircraft that's towards the end of life and they've got a casing that has a crack in it that can be fixed somehow, but there's no engineering data for it. There's no drawings that exist. How do they continue that platform by just doing repairs? So they've-
Stephen LaMarca: Yeah, there's no CAD file.
Benjamin Moses: Yeah, there's no CAD file.
Stephen LaMarca: You got to find some pencil drawings.
Benjamin Moses: Exactly. Or the supplier may be gone. Suppliers change all the time. They turn down bids, they go out of business. The whole ecosystem of engineering data and suppliers is really, really tough to manage when you've got something that's been around for a really, really long time. What they've been working on is a way to create a digital twin of the entire Black Hawk so that in the future if something breaks they can figure out the best way to re-manufacture that part. It's an interesting dilemma of a couple of things. One is the engineering data support of that, but also how do they remanufacture something that doesn't exist or can't be manufactured or is not being manufactured right now.
Stephen LaMarca: So I just want-
Benjamin Moses: I have a quote here from ... oh, go ahead.
Stephen LaMarca: I think that's awesome. I can't wait to hear the quote, but this is really great because this immediately makes me think of when you're going from hand done drawings and designs to what's now the digital age there's a lot lost. What makes me really proud about our military is they seemingly have learned from the mistakes of the Swiss watch industry when quartz watches came about and almost killed the mechanical watch industry. A similar thing happened to this and the Army's getting on top of this before it happens to their Black Hawks which is really great. To some degree you could say that the Army, and I'm sure a lot of soldiers would disagree with this, but the Army works more efficiently than the Swiss watch industry. Works better than a Swiss clock. It's cool.
Benjamin Moses: Now, so-
Stephen LaMarca: What quote?
Benjamin Moses: The quote here is from someone in the military here. It's, "A key focus of IMCOM's AM policy is on inserting evolving technologies into enduring designs that have relied on traditional manufacturing processes throughout their acquisition lifecycle, however, future aviation are benefiting as well from advanced manufacturing." In the end, I agree with you that they're evolving their previous platforms, but I'm really curious to see what they're doing with their current platform, so this isn't the Air Force. The Army doesn't have an F-35 ferry do they? The Air Force, Navy, and Marines do, right?
Stephen LaMarca: Air Force maybe it, right? Yeah, I think so.
Benjamin Moses: I'm just curious what they're going to do with the new platform that they come out with, what are they going to do with the data rights associated with that platform. Example of this Black Hawk that they're redesigning, they're taking all the CAD data and putting it into their own vaults. They're using the [inaudible 00:37:54] facility at Wichita as a means to achieve that data, but something that's going to be designed in the future I think we're going to run into the similar problem that consumers have right now with data rights. When I rent a movie or when I purchase a digital movie, if you ever read the end user license agreement you don't own that movie, you paid for the right to use their platform for as long as they hold it. They could revoke your access to those movies at any time without telling. They'll probably tell you, of course. For example, you pay for a subscription of Netflix. Movies come and go on Netflix all the time. If you buy something on Amazon Prime and their license that they have with the studio gets pulled you no longer have license or access to that movie through Prime. I'm curious to see for future platforms how they will handle digital data rights and kind of who's going to own and pay for what in the future, but we'll see.
Stephen LaMarca: Well, if there's one American organization, if you would, to figure that out, it would definitely be the military because there's a bunch of not necessarily copyright, but definitely IP protection and concern when it just comes to any sort of military TDP, technical data package. Talk to [Colton FN 00:39:13] about that.
Benjamin Moses: All right. Steve, let's talk about some ... You got some more info on robots, don't you?
Stephen LaMarca: Oh, yeah. Toshiba has been doing a lot of work with mobile robot arms-
Benjamin Moses: That's cool.
Stephen LaMarca: ... and it's not so much that ... It doesn't feel like it's that new because I feel like this is a pretty old concept as have been like a couple of things we've talked about today. What Toshiba's basically experimenting with, but actually getting a lot more progress done with it. They're not experimenting with it on an academic level. They actually want to produce these and make them, which is awesome. Toshiba's essentially taken a warehouse robot, which is basically a mobile platform. It's a robot that ... It's not a forklift. It looks like a coffee table that can hold a lot of weight and it's on wheels and it just moves from one end of the warehouse. Somebody puts stuff on it and the robot knows where to go. Once it's told it has its payload it goes somewhere else. Toshiba's basically taken that, made them more accurate and then instead of carrying a payload, instead on top they've put a robot arm, so they're making-
Benjamin Moses: Cool.
Stephen LaMarca: ... mobile robotic arms on top of a ... Well, they're making robot arms mobile by putting them on warehouse robots. It's really cool because you know when you go into IMTS, you see some really advanced robot arms knowing exactly where something is. They have a vision system and they know exactly where it is and they know how to grab it regardless of how fragile the part is. They grab it wherever it is, even if it's moving. They grab it, puts it somewhere else and it does its task. Those robot arms that we've seen do a task like that at ITMS have always been on a pedestal. Have always been on ... What's the term for it, is it a pedestal?
Benjamin Moses: Sure. Sure.
Stephen LaMarca: Okay. They've always been in a little cell and they've not been mobile. They've been bolted down and the robot itself is always in the same place every time. The parts that it's grabbing may be in a different location, but now Toshiba's like, "Why does the robot have to be stationary," and they're really doing a lot with that.
Benjamin Moses: That's interesting. I think back to the article that I mentioned previously, have you ever been to a factory where they are using robotic arms for either machine tending or packaging?
Stephen LaMarca: Yes. Yeah.
Benjamin Moses: I would say 50% of the cases is the robot moving all the time?
Stephen LaMarca: No.
Benjamin Moses: Yeah. If the robot arm is tending a machine just loading raw material, taking off, which is great, passing on the next cell, its got a lot of downtime. It's waiting for those cycles to finish. If your machine cycle's five minutes, say you got five machines in your cell in a circular and the one robot arm serves all of them, which is great. It's waiting five minutes every time to do a 30-second operation. I think one potential application here I see for the Toshiba robot is having one arm servicing multiple cells. You can time your cells or you can go in between each of these cells. The total cost of your robot plus the platform that it's driving on is obviously going to be higher than just one robot, but it's cheaper than five robots that you're going to have stationary and I think this is an interesting approach on a couple layers. One, of course, for warehouse robots just being able to have increased capability to pick up stuff from bins and shelves, but on the factory floor if the load capability is high enough and it truly does have, let's say, enough self awareness that it could position itself and say, "Hey, this thing's done, let me move it to over here and then let me to go the next cell when this is ready." I think being able to amortize your cost of that platform over five different cells is pretty awesome.
Stephen LaMarca: Yeah. Because a robot I think is limited and I'm repeating you a little bit, but it's limited by the stuff that's around it.
Benjamin Moses: Exactly.
Stephen LaMarca: It's limited by the fact that things need to come to the robot before it can do the things that it needs to do and then move on. As Toshiba is making ... This is no new concept, it's just Toshiba's finally taking it from an academic project and turning it into a product. It's great to see that Toshiba wants to make the robot go to its work and then do its work and then go to more work. I mean, it kind of makes me think of when I worked retail, when I worked at Total Wine & More. Everybody would have their list of jobs that they got to do today for the day and you'd go do it and maybe you worked really fast so you could get a little bit of downtime. If a manager walked by and saw that you were just standing there not doing anything they would be like, "Go find some work to do." Just like imagine telling a robot to do-
Benjamin Moses: Go find some work.
Stephen LaMarca: Seeing a robot just standing there. It's like, "He can't help it man, the thing can't move."
Benjamin Moses: Yeah, exactly.
Stephen LaMarca: But now robots have no excuse.
Benjamin Moses: Yup, yup. That's awesome. The last article I have actually goes against robots, boo for robots. It's an article from Core77 and it talks about one of the Tesla [crosstalk 00:45:08]. Core77.
Stephen LaMarca: Oh, oh, yeah.
Benjamin Moses: It's a article about one of the Tesla factories that's producing the model Y. Elon Musk outlines a problem statement where they have a throughput issue it sounds like. They have one of the rear assemblies for the rear crash zone has quite a few pieces. It's about 70 some components that gets glued, riveted, joined in many different ways. It's not an uncommon problem in automotive manufacturing where they probably have a casting, some extrusions, some other sheet metal parts and they chose those processes based on the economics of those parts individually. Then everything gets put together somehow, which is great. I'm a big fan of multi-piece, multi-component assemblies, but the problem that he's running into is one, I think he's got a footprint issue and the article kind of alludes to that later that bringing all these components into one place and having an assembly line put them together they're running out of space. What the team over in Tesla has decided to do is let's convert this entire assembly into a single casting. High pressure casting that they could basically reduce 70 pieces down to one or two it sounds like. In this case, it looks like they're pushing about 180 pounds of aluminum for their casting, so it's a pretty big assembly. The machines that they're bringing in are fairly large. Let's see, the machine weighs 410 tons for this high-pressure casting. It's interesting. In the end, they talk about potentially throughput savings of not having to process it, but in the end the quote is that they're saving footprint. They can bring in basically one machine. I'm not sure if they're doing additional processing to it. It didn't sound like they were, but who knows?
Stephen LaMarca: I just hope that saving reflects to the customer in some way.
Benjamin Moses: No, no, no, they won't do that.
Stephen LaMarca: Because I really like Tesla and-
Benjamin Moses: Do you?
Stephen LaMarca: I do like Tesla and I like Elon Musk-
Benjamin Moses: Good for you.
Stephen LaMarca: ... but it sounds like ... This happens all the time in the auto industry. Just look at door panels for crying out loud. You've got a switchboard to roll down your windows. You've got the lock. You've got all those things and you think that they're individual components, but let's say you go to whatever manufacturer car you have, you go to a wholesaler of parts to replace something and you come to find out because your passenger front window switch breaks. You think I just need to replace the switch for the front passenger window and you come to find that nope, that's an entire assembly that requires replacement of the entire interior door panel and what you thought was maybe a $50 switch with markup is actually a $1,500 assembly. Going from 70 ... I don't know if you said 70, but going through 10s of parts that make up a crumple zone to one to three parts to make a crumple zone, to me that sounds like what normally was a $500 fender bender requires a $3,000 new crumple zone. In 10 years, will that car be totaled if you get in the slightest accident.
Benjamin Moses: True, yeah. That's the dilemma I faced. I had one of my wife's first car was a Nissan Xterra that we bought when we first got married and it was towards the end of its life recently and one of the license plate lights went out. I looked up through the traditional means of going straight to the dealership website and looking up the part. It's a license plate light that's through the bumper. A little hole through the bumper and some lights, [crosstalk 00:49:18] a wire harness. No, no, I tried all that, that didn't work.
Stephen LaMarca: Oh, no.
Benjamin Moses: It fell at one point and the wire got ripped out on one side so I failed the inspection. I needed to replace the wire harness basically with the bulb. I go to the website and I say, "Okay, I need this" ... I see the part number for the wire harness and said, "Guys, do you have this?" I called around. No, it's sold with the bumper.
Stephen LaMarca: Oh, geez.
Benjamin Moses: I got to replace the bumper for this light? I almost threw my phone across the room. I was like, "All right, I guess I got to go to the junkyard." I found some third party that kind of fit so in the end-
Stephen LaMarca: That's good.
Benjamin Moses: ... I agree that the aggregation of these parts is kind of an interesting dilemma.
Stephen LaMarca: I know the answer to this question, but where's that car today?
Benjamin Moses: Oh, it is not close by. I had it towed away when it couldn't pass inspection.
Stephen LaMarca: Yeah. [inaudible 00:50:04] it's as cool as much as I love lane keep assist and the, what is it, the automated cruise control that detects the car in front of you and slows down automatically so you could be reading a book at the wheel, which you are not supposed to do at all. Back in the day, you'd get into a small fender bender and you break your taillight or your headlight, whatever. It's a couple hundred bucks. If you don't have halogens, maybe it's a couple hundred bucks. If you're still running halogens, it's probably 50 bucks for a new headlight assembly for all that plastic and whatnot. Today, with all these LEDs ... It's not even LEDs that are the culprit. Sure, it's great that you don't have to replace bulbs in modern headlights and taillights, but with lane keep assist and all that stuff, all of that junk is packaged in the head and taillights.
Benjamin Moses: Oh, is it really? Oh, no.
Stephen LaMarca: The parking sensors, the stuff that allows the car to parallel park for you, all of that technology, those military grade lidar and radar systems are in the head and taillights.
Benjamin Moses: Oh, no, that's terrible.
Stephen LaMarca: You look at ... I think it was the Volvo, the modern Volvo cars with the Thor's hammer style-
Benjamin Moses: Oh, sure.
Stephen LaMarca: ... headlight [crosstalk 00:51:36] that's what it's called. That's what they call it, the Thor's hammer. That broke a world record for being the first $1,500-
Benjamin Moses: Headlight.
Stephen LaMarca: OEM headlight.
Benjamin Moses: Geez.
Stephen LaMarca: Yeah, you break a headlight in a Volvo, good luck. [crosstalk 00:51:51] That could total a car.
Benjamin Moses: That could total-
Stephen LaMarca: In 10 years, that could total the car.
Benjamin Moses: One headlight, done. Oh man.
Stephen LaMarca: We can't do this.
Benjamin Moses: It's unsustainable. All right, Steve, that was a great episode. Where can they find more info about us?
Stephen LaMarca: Sorry, there's a car alarm [crosstalk 00:52:05]. I think somebody just got it.
Benjamin Moses: Working from home, man.
Stephen LaMarca: Our listeners can find more about us by going to amtnews.org. If you want to keep track of us through the Tech Trends weekly update, what is it called? I don't even know what it's called now, man.
Benjamin Moses: Our podcast? AMT Tech Trends.
Stephen LaMarca: You can find the podcast at amtnews.org. You can find the weekly technology update.
Benjamin Moses: The Tech Report?
Stephen LaMarca: By going to amtnews.org/subscribe. That car alarm really threw me off. Man, talk about other technology that's expensive.
Benjamin Moses: Am I going to have to edit this now?
Stephen LaMarca: Yes. I'm sorry, Ben.
Benjamin Moses: That's all right. It's a good try, Steve. You can find us on any podcasting app by searching for AMT Tech Trends.
Stephen LaMarca: AMT Tech Trends. Search on your favorite podcast app.
Benjamin Moses: Awesome. Thanks, Steve.
Stephen LaMarca: Or you can go to amtnews.org.
Benjamin Moses: Great. Bye everybody.
Stephen LaMarca: Bye.