Next level software automation and efficiency for fabrication shops

In This Episode

Learn more about Salvagnini America

Learn more about podcast sponsorship opportunities

TRANSCRIPT

Tim Heston: Where's the trajectory for the future metal fabricator out there?

Clay Case: Yeah, I think that's extremely interesting to think about, and I think that's the million-dollar question, right? So where are we going to be 10 years from now?

Sara Spring: This episode of the Fabricator Podcast and Automation Talk is brought to you by Salvagnini America, a provider of sheet metal fabrication equipment and software. The Automation Talk video series and ebook focus on the role of software automation in modern precision sheet metal fabrication. This season takes a deep dive into order processing, nesting, bend simulation, and IIoT for tracking equipment performance. Click the link in the description to learn more.

Tim Heston: Hi, this is Tim Heston of the Fabricator Magazine, and thanks so much for joining us for this special edition of the Fabricator Podcast, sponsored by Salvagnini America, and I'll let my colleagues introduce themselves. So why don't you introduce yourself?

Clay Case: Yeah. Hi, I am Clay Case, head of engineering at Salvagnini America.

Matt Humerick: My name is Matt Humerick. I am CAD cam and OPS coordinator at Salvagnini America.

Tim Heston: All right, and during this podcast, we'll be discussing a new... It's as part of our special series with Salvagnini called Automation Talk, and we're talking about next level fabrication with software. It's bringing fabrication up to a next level of efficiency, mainly getting that flow, all the way from art to part, all the way from initial design to final shipment of that work piece. So first, Matt, I'll start with you, but bring me through the bottleneck of manual preparation of piece parts and any kind of engineering work that goes into an automated system on the floor and what kind of bottleneck that can provide, and you can have very expensive machines sitting idle on the floor.

Matt Humerick: It's correct. So when it comes to bottlenecks in the process, there could be bottlenecks from the design of the part. There could be bottlenecks at the drawings of the part being incorrect. There could be bottlenecks from part process to nesting, so there could be bottlenecks all throughout the process.

Tim Heston: And how do you overcome those bottlenecks? What are the root causes?

Matt Humerick: Some of the root causes that I see are incorrect modeling of parts. That's one big thing that we deal with. Another thing is information exchange. So sometimes, we don't get all the information from the original designer of the part so we don't know exactly what this part entails as far as material, if there's forming features, whatever it might be. There's a lot of lack of data flow.

Tim Heston: So Clay, I'll bring you into the conversation, but where should a shop start to try to get that data flow? To actually, all right, we need this information. How can we build our processes to streamline that flow?

Clay Case: Yeah, I think a lot of it is historical precedence, so you observe your processes and find out where the bottlenecks are, and then you try to alleviate that process, open it up a little bit, maybe try to add some additional resources to it or try to optimize your resources in creative ways.

Tim Heston: Optimize your resources. And also, another big constraint I see out there is throwing information over the wall between design, engineering and manufacturing. So Clay, why don't you get into that a bit as far as what happens when those kind of over the wall challenges happen?

Clay Case: Right. Yeah, whenever there's a challenge, short of having somebody step up and take ownership of it, sometimes it's much easier to pass it off. And you have your operations guys blaming the design guys, blaming the operators, so you get out the blame thrower and try to figure out whose fault it is. But I think the resolution to that is having a strong leadership and having someone that really takes ownership of the process as a whole from a holistic standpoint, rather than just focusing on only their specific tasks, let's say.

Tim Heston: Right. And bring me through the information waste. Maybe Matt, you can chime in here as far as the information waste that comes from throwing things over the walls, things that can get missed in the process, and then we'll talk about how software can meet this need, but I just want to get this information waste message out there.

Matt Humerick: Absolutely. So there's a lot of waste when it comes to design for manufacturing. There's a lot of designing for the way it looks and the way it might feel inside of the software, but actually generating and making it, the program and the part, is usually something a little different. So that is critical where we get the information going back to design. That way the design is correct and we have more of a stream flow all the way to the end process of the part shipping out the door.

Tim Heston: So dig it deep into that stream flow. For me, I find this fascinating, especially when you're saying you've got a press brake and you've got a panel bender. They both have different manufacturability requirements. Bring me through those differentiations, and when you get that granular, how efficiency can flow upstream from there.

Matt Humerick: So with software, when you make a model, you make a model based upon your needs and what you're going to run it on most of the time. There's a difference between machines, so we have a difference between a panel bender and a press brake, for example, the developments. We can speak about the development. Instead of having to have a different development for a panel bender versus a press brake, with the advent of the software, what we can do is we can take your original 3D model, we can create the correct developed flat pattern for a press brake for example, but also the correct one for a panel bender. They are slightly different from each other so it's best to have a software that can do both developments for you.

Tim Heston: Both developments. Why are they slightly different? Explain the why for folks.

Matt Humerick: So it's the way that the metal is formed on a panel bender. We actually stretch material, so our K-factor is slightly less than what you would have on a standard press brake.

Tim Heston: Right. So taking that information, how can you build in, say, design for manufacturability so you can really streamline flow? It's like, all right, you've got that information. Now let's bring that up to the design element. I don't know if you want to join there.

Matt Humerick: When it comes to that, basically what it is that the person that designs it needs to understand the process on the floor also. So a lot of the development that I've seen has been someone knows how to make a 3D model but maybe they don't understand the processing of the part on the machine level. It's very, very important that that data gets transferred back, that knowledge gets transferred back. That way, they make the 3D model for manufacturing.

Tim Heston: Make the 3D model for manufacturing to get all the material idiosyncrasies in, to make sure you have tool access as well, right? So you have the clamp down tool for panel. So Clay, step in here regarding how a shop can get started on their software journey. So say you visit a shop and it's got a disjointed software system, and you say, "All right, we need to start a transformation." Where can a shop start?

Clay Case: Yeah, I think the biggest first step is identifying that there is a problem, right? Because I feel like a lot of shops maybe that do have disjointed processes and software, that they don't maybe necessarily realize it. So sometimes it's consultant, sometimes it's a sales opportunity. It just really depends. But bringing that issue to light I think is really the first step, and then depending on where the biggest disjointedness is, I don't know what the right word for that is, but figuring out where that is and then that's where you start to chip away at the problem. And as you start to resolve one or two issues, then that brings out other issues that then you resolve and you resolve and you resolve.

Tim Heston: Right, right, right. Yeah, I've been to so many shops where... In fact, I visited one shop where they actually redesigned their front office to have a literal flow down the engineering and design and manufacturing to have everybody talking with each other, so have that physical connection. Now, they didn't have a software platform that actually allowed for a stream like flow so I wanted to dovetail into that. Describe the Salvagnini approach to software and how it's evolved over the years. You've had OPS for few years, now you have Stream, and so bring me through the different levels of the platform for me.

Clay Case: Probably a better question for Matt.

Tim Heston: Matt. Go for it. Matt, you've been testing it for a while, right?

Matt Humerick: Yeah, since 2017 when Stream first came out. So there's been the evolution of modeling inside of the office down to the shop floor, and the way we looked at it as, instead of an individual creation of a program for a machine, we want to look at it as a production flow, a process from beginning to end. So we want to be able to take your 3D rendering or even your 2D with bin lines, let's say for panel bending, and we want to be able to take that and we'll be able to create a panel bender program, but then we want to be able to automatically create the correct developed flat pattern and hand it off to a laser machine or a punching machine. So the thought process was, here's our final product. We want to make a streamline to that final product without a bunch of offsets in between. Makes sense?

Tim Heston: Right. And how does that flow information back to OPS as far as scheduling and as far as programming and automating that front office operation?

Matt Humerick: Actually, so when it comes to OPS, OPS, what it is is a software that will take your order that you give it in either a CSV or XML format and it'll create a work order, some instructions to run on the machine, some nesting for example. So as it's running on the machine, OPS is always connected to the machine, so it knows when this part that you asked to be produced, when it's been produced. It will give a report back to your ERP system and say, "Hey, you asked for this particular part. I ran that particular part this time and date." More information can be had where it says, "This is the weight of the part. This is how much material you used. This is your scrap." That goes all back to of course the ERP, then you can control your accounting and understand how much utilization I'm having, how much scrap I'm having so I'm running my machine as efficient as possible.

Tim Heston: Can it get that granular? So it's feeding back into the ERP, but can you get that granular as far as, all right, what was the reason behind the scrap? Was there a common cut that didn't work? Was there a slat position issue because you're cutting over the slats or something to that effect? Can it get that granular?

Matt Humerick: OPS does not, but we have another software. Yeah, we have software that will tell you there was some error and what was the error on the machine while it was producing.

Tim Heston: And that could communicate that data back to-

Matt Humerick: You can get some of that information and you'll be able to see, I'm having the same common errors, let's say for example, when I run this common part or something.

Tim Heston: Now, also then describe to me after the fact software, the tracking software. Is it Links, right?

Matt Humerick: Links, that's right.

Tim Heston: So how does that fit into the software puzzle?

Matt Humerick: So Links is our IoT connection. There's always information going out to the cloud, whether it's production data, whether it's real-time events, so that anybody that has access to that can see what's going on. So it's important for the people that actually keep track of the part production, the people that keep track of the machines, whether it's maintenance, whether it's a production manager, whatever it might be. All of that data is there and readable to them so they can see, for example, I have a message come up on the screen. I could be at home and I get a notification on my phone that says, "The machine stopped at this time for this reason."

Tim Heston: Right. Okay. Yeah.

Matt Humerick: So it helps with preventative maintenance, it helps with understanding your production, it helps with understanding your OEE capability, where you're at and how everything is flowing inside of your shop even when you're not there.

Tim Heston: So see, then how does this affect, say... I'm thinking about how operators grow up in this industry and how it's like, all right, you've had an operator who's been here 20 years, but then consider, how does this shape a metal fabrication career as far as how you learn the business and how you improve in the business? Burn me through that as far as how... Because you were on the shop floor so you remember all the headaches.

Matt Humerick: Yep. I've done a little bit of everything. So it's a huge gain for the company because the people that you have that are running the machines, that have more interaction with the machines also have more interactions with the programmer, the guy that made the actual part, the designer of the part. So there is a lot more understanding of the complete production instead of, "I just go to work every day, I make this one part." You understand a lot more of the process.

Tim Heston: Right. Yeah. I saw there's one part, this great DFM example where I believe what it was, they had a weld line that was moved, a weld seam that was moved slightly up, but they took advantage of the very short flanges you can make with a panel bender. Say, "Hey, we could bump this and have a short flange and have a spot weld on the side versus a perfectly mated corner weld, which was extraordinarily challenged to weld." And they made that simple change and it flowed all through the system and then cut days off the production side. So Clay, I want to get some historical aspects of the software as far as how far it's come, and are we at an inflection point as far as, all right, things are really... For me, it just seems, especially when I'm walking the show floors, things are really starting to gel as far as software goes. Everything's connecting. So bring me through your experience as far as where software is coming from and where it's going.

Clay Case: Sure. Yeah. So many, many, many years ago when I started at Salvagnini, we had very different software packages than what we have now obviously. Technology has come a long, long way. So much so that at one point, that it was almost faster to write the programs by hand than it was to try to chug everything through the computers because they just didn't have the processing power that we do now. So as the processing power starts getting more and more stronger, we can start implementing new algorithms and adding more complexity to the process in order to find the most efficient outputs. So as we've evolved through that, three or four generations of software that I've personally have witnessed, everything from our old stuff to our more modern Stream, it's really just made the process a whole lot more efficient and it's really based on just taking advantage of the computational power that we have these days. Looking forward, I think we're going to pull a lot more from AI and ML. I think we're going to start bringing in aspects of that into the software world, not just Salvagnini but just the industry in general. As that technology becomes more readily available and more affordable for any company to implement, I think we'll see that really start to drive a lot of things and find new efficiencies in areas that we didn't even know were an issue at this point.

Tim Heston: Right, right. And I see AI having an amazing impact on not only efficiencies with flow, but with DFM, where if you have say Links. Now, does Links take into consideration, all right, here it was the flow and here was the part and job that was associated with that flow. Does it do that?

Matt Humerick: Links just gives the data on what was actually ran.

Tim Heston: What was run.

Matt Humerick: What was physically run.

Tim Heston: But you can technically correlate that with the jobs, like the job number, and then go back to a design and say, "All right, this design improved flow by this amount."

Matt Humerick: You can, for sure.

Tim Heston: If we tweak this, how will this change it or et cetera. And then you could correlate data with, say, maybe back to OPS. OPS goes back to ERP. If we switch material suppliers, is this material better or worse than this material, et cetera, or our maintenance schedule, predictive maintenance and slat maintenance. It's going through because you have minutia with your... You guys have ceramic slats I know, but also you have intelligent software to actually not cut over the slats like everybody else, and so it's like you have all this integration. So Matt, describe how you learned this process initially. Bring me through the history of you in Salvagnini and in sheet metal fab. Where was this industry when you entered it?

Matt Humerick: Yeah, that's interesting. So before I started working at Salvagnini, I was not in sheet metal. I was in the automotive industry.

Tim Heston: Okay.

Matt Humerick: So when I started at Salvagnini, I started first in mechanical installations, so I understood the mechanics of the machine. Then I moved over into the electrical aspect, so I understand the electrical connections, the communication between machines. From there, I moved into what we call our startup, where we turn the machines on, we calibrate them, we get the machines dialed in. We help with programming, we train the operators of the machine, and we also train maintenance on how to fix the machine and how to read problems on the machine. From there, I moved into applications, which applications is a feasibility study of parts, so for a punching machine or for a panel bender. And then from there, I was gifted with the software, so I've seen the full aspect of the machine from the mechanics all the way up. And when I started to now, the way that we actually produce is much more streamlined, much more easier to use, much more easier to understand.

Tim Heston: Right. And it's become an integral part of the whole solution I think, where I see if you have silos of information... Years ago, I used to visit a lot of fab shops, and of course me being an editor, they say, "Oh, we bought this new machine. Let's go see it." And so I would see this massive FMS, a massive unit. "Oh, isn't this wonderful?" And then you go to the other area of the shop and, "Oh, that's the assembly area, so the machine is forwarding to the assembly area." Well, okay. So you bought this FMS and now you have this amazing amounts of whipping just sitting there. Are you shipping any more product? And so it's changed. Software's broadened the aspect of, all right, are we shipping more products out the door? Can Links help with that as far as downstream processes beyond the final bend, as far as that's concerned.

Matt Humerick: So when it comes to links, no, but OPS, what OPS can do is it can send information to our machines and the information can go from our machines to a secondary process like a welding machine, so an inline welding machine. We can send data as far as here's the part that's coming, here's the parameters that you need to weld this part. So it's really a complete communication flow from beginning to end. When it comes to lean manufacturing, you only produce what you need to ship. You don't have the room to store material, you don't have the room to store parts that are ready to be shipped. It has to be a lean production flow from beginning to end. So that's where the whole OPS thing comes in, where an order comes out from their ERP system and says, "I need these parts. I need these parts. Within the next few hours." OPS can control that and say, "Oh, okay, you want these within the next four or five hours. Let me make some nesting, send it to the machine for production." But you could also make OPS to where it sends it to the next available machine. So just because you send it to a machine A doesn't mean that it can't go to machine B. So if you have that balancing act where this machine gets too much production, it can automatically direct it over to a secondary machine. That way, your flow is continuing and moving.

Tim Heston: And so it doesn't matter the cycle time, so in other words, say it's a slower machine or you're choosing between a press brake and a panel bender. Even though a panel bender is an extraordinarily fast machine, if your press brake has capacity...

Matt Humerick: Where it has capacity.

Tim Heston: But then again, you have to change the flat layout. The software allows you to do that and allows that continual flow. Bring me through, and maybe Clay, you can chime in here, how you can... Because whenever I see a Salvagnini line, I always say, "All right, are you running batch? Are you running flow? Are you running a combination?" Because I've seen absolute single piece kit flow through these systems that's pretty impressive, but then I've also seen certain levels of batch and I've always asked fabricators, "How do you choose?" "Well, it's flexible. I can do various levels depending," but how does software on top of that say, "Hey, you can really tweak and dial this in." Maybe talk through that for me.

Clay Case: Sure, yeah, and I think it really depends on the customer. So Tim, the customer's going to decide, do they want to focus on kit or do they want to focus on batch? And focusing on batch, that makes it a little easier maybe to tune programs because you're going to run the same program 200 times every Monday, every Tuesday, and so you can really analyze it heavily, physically and via the software, and really start tuning some things. When you start running kits, especially if it's a dynamic kit, maybe some parametrics are involved so you're never going to run the same nest twice, it makes it a little more challenging. But you can really pull a lot of data from the software, from Links, from OPS, and really start tooting the families of parts. So that's where you can use some intelligence and try to design that way.

Tim Heston: And you can totally optimize that, bring it back to OPS which brings it back to your ERP to say, "All right, this is the financial results of actually producing this." So shared data throughout.

Clay Case: Sure. Yeah, you can track the timings and throughput, scrap and all of that stuff, and really kind trace it back to maybe you have a certain family that's eating a lot of scraps, something like that. But otherwise, you can focus on just tuning everything based on cycle times and throughputs. Yeah.

Tim Heston: All. So Matt, bring me through, again, you've been on the front lines as far as the operation is concerned and you've been in all aspects of the industry. When you talk to operators, you've talked to a lot of operators over the years, if they're new, and boy, do we have a lot of new operators now. My gosh, I went to a press brake seminar here at the show and the moderator said, "How many folks have operated a bending machine for more than two years?" Nobody raised their hand. I mean, literally. I looked like an old fogey, [inaudible 00:22:37] God. So if they're fresh to the industry, where should they start learning to be the most successful, and to not only know the basics but also to know the power of software so they can take their skills to the next level?

Matt Humerick: That's a very good question. So it depends upon the manufacturing. If you're using a panel bender, it's much more easier for someone to understand. If you're using a press brake, it takes I guess more knowledge, more hands-on to get to that point, right?

Tim Heston: Right.

Matt Humerick: So it's really driven upon the person. If the person really cares and really wants to do the best they can, they could be a great press brake operator. Like you said before, we don't have the press brake operators like we used to. It doesn't exist, so you have to automate. So when you come to automation, everybody is like, "Well, automation is complicated." Automation is not complicated when you break it down to what it is. When it comes to software, automation is key. If you can automate your process of creating the part, of doing the tooling setup, of doing everything that is necessary in a visual type of way in the software, the guy that sits there and runs the machine can understand how it's going to run before he runs it. Does that make a little bit of sense?

Tim Heston: Right, right.

Matt Humerick: So there's not a whole lot of in-depth, crazy thought and base knowledge that you have when it comes to panel bending because it's all graphical and it's all in your face and you can see it and understand it easily.

Tim Heston: But also the creativity behind it. How can an operator that's just starting out in the industry over time actually get into the DFM process to say, "Hey, listen, I've seen this form 15 million times, and Hey, you could actually move that joint up here. It'd be so much easier."

Matt Humerick: So DFM to me comes from frustration.

Tim Heston: Yeah, I like that.

Matt Humerick: If you have to take 14 parts and assemble them together and weld them, your welders are going to be mad, your guy running the press brake is going to be mad, so it comes out of necessity. So if I don't want to take a process and make it a four step process, I'm going to try to make it a one step process that it can be done automatically on a machine, that's the way I'm going to do it. I guess kind of lazy, but also smart, right? Work smarter, not harder. That's the whole thought process behind it. So it really comes back to out of necessity, what is necessary for you? What makes it easier for you to do your job?

Tim Heston: Right. All right, well, some final thoughts. If each of you could chime in, how do you think software will affect the industry's trajectory and the future metal fabrication career? If I'm just starting today, where will my career be in 30 years from now when I'm 42 years old? No, but I mean, where's the trajectory for the future metal fabricator out there?

Clay Case: Yeah, I think that's extremely interesting to think about and I think that's the million-dollar question, right? So where are we going to be 10 years from now? I think it's going to be much more hands-off. Personally, that's how I see it. I think through utilizing ML and AI, I think we're going to really drive a lot with that. But of course, you're still going to have somebody have to keep an eye on things, so there will still be interventions but I think it'll be some of the more repetitive tasks and things like that will be more automated, more and more so just really squeezing out every drop of efficiency that they can get out of that process.

Tim Heston: All right, any final thoughts?

Matt Humerick: I couldn't agree more with Clay. So software is from the time you wake up and look at your phone to the time you go to bed. Software is in every aspect of your life, so therefore, it's going to be continuing to grow because it makes our life easier in the manufacturing world, but also in your day-to-day life. So software is for sure going to continue to grow. How far does it grow before we get into AI? I'm not sure, but for sure we're going to get to that point eventually.

Tim Heston: All right. Well, thank you guys so much for your time. It's been really insightful. And yeah, thank you so much for joining us for this special sponsored edition of the Fabricator Podcast, brought to you by Salvagnini America, and we'll see you next time. And be sure to click on special videos we have, video cast we have detailing getting a deeper dive into what we discussed specifically on one fabricator in Canada, one fabricator in the United States, trained technologies, Canadian customers, CMC, and they'll have both unique stories to tell and we'll see you then. Thank you so much.

Sara Spring: The Fabricator Podcast is a production of Fabricators and Manufacturers Association, located in Elgin, Illinois. The show is hosted by Dan Davis and the staff of FMA Communications. The podcast is produced by Gareth Sleger and recorded and edited with the help of Brandon Geier. Sales support provided by Andy Flando. Additional production support by Elizabeth Gavin, Dana Wiker, Mary Diamond, Mike Owens, and me, Sarah Spring. Thank you for listening.

ABOUT THE FABRICATOR PODCAST

The Fabricator Podcast brings you conversations with people in manufacturing who make things out of metal. We speak with manufacturers, metal fabricators, welders, job shop owners, small business entrepreneurs, artists, marketers, educators, and more. Host Dan Davis also goes beyond discussing just manufacturing and the skilled trades, and chats about pop culture, current events, food, music, movies, comedy, and, of course, robots. The Fabricator Podcast is presented by the Fabricators and Manufacturers Association.

We shape the conversation around metal.

Host: Dan Davis

Producer/Editor: Gareth Sleger

Video Producer/Editor: Brandon Geier

Ad writer/spokesperson/social media: Sara Spring

Additional video editing: Dana Wiker

Graphics: Billy Kulpa

Marketing support: Elizabeth Gavin, Mary Diamond, Pat Simon

Sales support: Andy Flando, Amy Hudson

Web support: Mike Owens, Jared Carlow

Additional support: Ed Youdell, Maurine Semevolos, Lincoln Brunner, Tim Heston, Rafael Guerrero, Josh Welton, Darla Welton, Amanda Carlson-Hicks, Callie Check, Rick Lehnhardt, Judy Steinbach.

Where to listen to The Fabricator Podcast:

• Spotify
• Apple
• Google Podcasts
• Amazon
• Audible
• Pandora

• YouTube
• Facebook
• Instagram
• LinkedIn
• Twitter