In this episode, Tim Nolan, my friend and fellow University of Pittsburgh alum shares his experience with technology to serve clients at Data Science Automation. We discuss how markets are moving and the ROI from a technology or systems change. We also discuss target clients and ways to market these services.
Sam Schutte 0:00
In today’s show, we have Tim Nolan. Tim is the Technical Director of Automation Engineering at Data Science Automation out of Pittsburgh, Pennsylvania. I met Tim at the University of Pittsburgh, and today we’re going to talk about industrial automation and programming embedded devices. Tim, welcome to the show.
Tim Nolan 0:15
Thanks a lot, Sam. It’s good to be here.
Sam Schutte 0:17
Yeah, absolutely. Great to have you have you on. So maybe a good way to start our conversation is tell me a little bit about how you got into this as a career.
Tim Nolan 0:27
Sure thing. So data science automation is where I’m at right now. But those skills go back to like you mentioned at the University of Pittsburgh, when I was a graduate student in bioengineering. Well, Pitt calls it bio engineering. A lot of other places call it biomedical engineering, but it’s artificial organ design and things like that. My main responsibilities in the lab were lab automation, data collection, data analysis, things that were used, with a number of programs like LabVIEW and MATLAB and associated programs. And so that’s what I ended up doing my thesis work with and some kind of specialized sensor work. And that, of course, was then a big part of my resume that I kind of had up there. And I was contacted directly by DSA for having known LabVIEW. And MATLAB was the kind of thing they were recruiting for. So that’s what got me in as my entry level programming position there. Back in 2006.
Sam Schutte 1:23
Okay, so yeah, so so basically the tools that you used, even though you were studying biomedical engineering and stuff in school, the the, the software tools you were using as why you ended up really cashing in on sort of and, and using and spending most your time on your career since then.
Tim Nolan 1:39
Yeah, absolutely. It’s really been the kind of tools expertise and that’s kind of what DSI presents as a tools expertise, rather than any specific industry as well. So it aligned well from that side of things.
Sam Schutte 1:52
Great. And so when when you kind of started with GSA and in this field, you know, what was one of your kind of first big projects you worked on, or one of them kind of more game changing things you did that. That was interesting to you.
Tim Nolan 2:06
Well, I had a real nice lead in, you know, I had more senior engineers kind of took me under their wing and got me up to training got me up to DSA standards, and actually kind of bittersweet because he literally retired yesterday, Ben Rayner, one of the the big, big names in the LabVIEW world, but um, you know, so I had several years of experience. And then I finally got to be a lead engineer on a big project. And in that project, it was an embedded monitoring system. So a lot of times you can run LabVIEW or what have you on a computer, you know, so you have a Windows PC, and it has hardware connected to it, you can gather data from that. But for long term systems, you want kind of more deterministic operating system. So you need to start going into what we call real time systems. And that’s a separate operating system. So at that time, we were using FarLap which was one of the OS is on what they called the field point system. Crazy name. I think it’s named after a horse from Baltimore. Like, it’s like a thoroughbred racehorse or something. And then those real time systems have moved on to like VX works. And RT Linux VX works is like what’s running the Mars rovers? I think right now. But this was an embedded system for coal plant monitor. So we needed something super robust that could be halfway up a smokestack and wouldn’t care about freezing weather, hot weather. So is this ruggedized industrial computer that had to basically constantly maintain all of the variables that are going on? So what’s the flow rate through the smokestack, what’s the temperature, and everything like that, because it was connected to a sampling system that was kind of pulling out little samples of the smoke. So that went into a device that measured how much mercury was being emitted into the atmosphere. So for the EPA regulations at the time, every smoke stack had to basically have one of these monitors on it. And, you know, we had to run in all kinds of temperatures had to run for a long time without anybody looking at it. And actually didn’t even have, you know, a screen that you can interact with it we had developed at that time, basically a PDA that you could walk up to it, it would connect wirelessly. You could check all the data, you could pull data off, and then walk away. So you didn’t have to have a screen that was there all the time.
Sam Schutte 4:25
Yeah, so Tim, you mentioned real time operating systems for folks that might not be familiar with that. Can you explain, you know why you might need a real time operating system as opposed to using something just like Windows?
Tim Nolan 4:37
Sure thing, Sam. Now in general, a lot of people are familiar with Windows, but as I’m sure we all know, Windows can crash on occasion. So it is not something that you use for long term support and not something that is amazingly reliable. So what we have in a real time operating system is what we call determinism now Windows is a general purpose operating system. And it’s great for what it does, you can do all kinds of things, you can have a network in place, you can be having virus checking, you can have all that kind of stuff independent, but they also interfere with each other. So they will, you know, have the virus checker interfere and lock up your system for a little bit on a real time operating system. You have what’s known as determinism, like I mentioned, you have a bunch of processes, but they operate in a hierarchy. So the top level process that you want to make your most deterministic, nothing can interrupt it, if it needs the processor, it needs the resources of the system, it gets it. And then there are all kinds of descending orders of priority below that of other processes. So you can always predict the order in which things will happen. And certain things won’t preempt each other. And that gives you a level of reliability. Number one, and that your main process is operating in a loop that’s deterministic as far as time has low jitter. And you know, your loop rate and your acquisition Whatever you’re aiming for, is always happening at the same rate, but also, it can’t get interrupted. So it’s memory, managing the memory managing the processor. And that gives you long term stability and helps you prevent system crashes if it’s implemented correctly.
Sam Schutte 6:16
Yeah, exactly. So I mean, I would imagine, obviously, if you’re running a nuclear power plant or something, you can’t have the thing pause to install virus updates and the entire system drags, right? Yeah, exactly. All those kind of applications. And so you’d mentioned you know, and your time their data science automation. can tell me a little bit about the company and sort of its history and how it got started. I think it’s, it had been going for quite some time when you joined to believe.
Tim Nolan 6:47
Yeah, very true. We go back to 1993. Our founder and still president Rich Bregman had actually kind of been doing this job internally in Pittsburgh Plate Glass. So the local company in Pittsburg, his job was internally to go around to the other departments and get people to kind of pay his funding to automate their systems. And they kind of realized, well, you know, I could do this on my own. Yeah. And so he founded the company back then. And, you know, we work in kind of the research, manufacturing, government business operations, it’s all across the board. And even from an early time, he realized that he could kind of spread out and do a number of things. So you know, we acquire we analyze present manage data, we design, simulate, test and validate products, we monitor, predict control and optimize processes. And even we invent draft prototype and build machines. We just delivered a like 10 foot tall surgical light test machine that was kind of fully integrated and built in house so we do the full full turnkey systems, however, whatever level you need, and early on as well, is very dedicated to teaching and education. So he was teaching LabVIEW as a course As well as implementing it as a consultant. So he was actually the first certified LabVIEW certified instructor outside of National Instruments, the the company that makes LabVIEW back in the day, and we were the first certified Training Center outside of National Instruments as well. So two halves of our business have been in there since the beginning, both the consulting and the training side.
Sam Schutte 8:23
Right. And so, you mentioned LabVIEW, a lot of folks may not be familiar with LabVIEW, and kind of why it is different as a programming environment. So you’re talking a little bit about what it is and how it’s different.
Tim Nolan 8:37
Sure thing. So yeah, LabVIEW, as you said, not too familiar. I was rather amused in the past couple of years that’s actually rank programming languages by use. And LabVIEW, I think is 36 or 37th at the moment at the moment, so we’re beating Hadoop but not many others. So what LabVIEW is, it’s the graphical programming language. So if you look at somebody’s programming in LabVIEW, it might not look like programming or your traditional type line by line, as you’d see in Python or C sharp or anything like that. You have functional blocks that you connect by wires. And ultimately, it’s a different philosophy of programming because it’s data flow rather than control flow. So as data is available to each node, like you can imagine, like water in a pipe or electricity in a circuit, it operates on that item and then gives output on the other side. It was developed by essentially a bunch of PhDs, standard Silicon Valley. I mean, only now we’re in Austin, Texas, rather than California, but in some guy’s garage, namely Dr. Trueshard and Jeff Kadosky. Jeff Kadosky was the programming side of ni back then, they just said this is something we need to control our instruments, and they developed it as a accessory to the heart. Were production of National Instruments where they were making things like GPIV cards and data acquisition cards. And they just created out of whole cloth, really. So it has a bunch of interesting things in place like it’s inherently a parallel language. If I have parallel wires on a system, it’ll operate in a parallel level. I don’t have to write any extra code for parallel operations. But yeah, it’s its main strength is instrument control. So automation, acquisition, getting a bunch of different kinds of hardware to talk to each other is kind of what it was built around and where it’s wheelhouse really lies.
Sam Schutte 10:39
It sounds like out of box, it can run on some of these real time operating systems you mentioned to just sort of, it’s built handle those as well.
Tim Nolan 10:46
It is, I mean, they’re different like toolkits like, you get LabVIEW and then you get LabVIEW real time or LabVIEW FPGA or LabVIEW emotion that are kind of specialized, that just kind of give you an extra set of pallets. And you’re working with them in the development environment.
Sam Schutte 11:03
We’ve talked about this before a little bit, I think but you know, is there anything that, you know, that you could do in save C sharp that you can’t do in LabVIEW? I mean, can you build entire, you know, enterprise systems in LabVIEW? Or is there any restriction there? I mean, I guess there’s, there’s best use, like you kind of said
Right, like,I would never go and try to build a word processor in LabVIEW. For example, it would not be its use case. But there are very large enterprise systems that are done, you know, you can do factory level systems and factory level automation. They’re using it to control the Large Hadron Collider. So a lot of the fine magnetic adjustments that need to be made at a fast speed, they again use the compact Rio system, which is another embedded real time controller to do very fast adjustments of operations at that level. So and people are using it, like we mentioned in very critical applications, such as on naval vessel And in nuclear engineering, and aerospace and so on.
it seems to me like, we’ve done some work recently, for instance, pulling data off of, you know, say PLCs, or devices and stuff. And, you know, if you’re doing that at a sort of, you know, base programming language like C sharp or something, you know, you’re opening the connection, you’re pulling it down, and then running a SQL query against the data, whatever. It’s all very atomic, right? For line by line. And so I could see where if you just kind of had a block, you just plug in and say, pull the data done. Yeah, exactly. And it can be a tremendous time saver. And probably, you know, even a lot of errors, especially if it also it can be parallel.
Tim Nolan 12:41
Yeah, because you also get a certain amount of the advantage that a lot of Macintosh users do, because the hurdle for the hardware and the hardware is built for the programming because you a large section of LabVIEW stuff is from National Instruments, hardware. So that gives you an understanding and a and a synergy. That is kind of built in from the ground up.
Sam Schutte 13:01
So when you started a DSA, you were a developer there, I guess, correct. And now you’re the technical director of automation and engineering. So how is that role change? I mean, do you manage a staff of people, they’re kind of in charge of all the projects and programming side or what what is your role kind of there now.
Tim Nolan 13:18
So I have done all of those things. I was in the middle of that, a program manager and a project manager. So I would manage projects and control a team. And it’s not like I have a set number of direct reports because there’s another director and another project manager to other project managers actually at DSA. So we kind of formed teams for the length of a project and then release those teams and they might be working at the end if those things now that’s the project management level, which I’d kind of started about seven years ago. And you know, I learned that on you know, how to focus more on the communication and the registration and the requirements coverage and budgetary and, and things like that. But as we’re growing as a company, we realized, we want to make sure that we have kind of internal processes set up, and that we also have things that are making it. So our president doesn’t have to do everything. You know, that’s always the trap of being a founder, right? Because, you know, for so many years, you are HR and you are sales and you aren’t marketing anymore, you know, Operations Wing. And you know, it’s not what you want your your CEO doing. So, Jeff, the other director there and I have taken on more roles, such as, for example, I’m in charge of our internal training. So I kind of meet with a lot of the engineers and make sure they can get their certifications. And then I’m also in charge of our internal testing standards. So when we make our unit tests, and we make our acceptance testing for factory acceptance or site acceptance, I make sure that we’re adhering to industry standards. And that our customer is satisfied with the kind of tests that we’re putting together. So I’m, that’s why I’m in charge of engineering and automation. For example, Jeff is in charge of operations as his director position. So he’s more of what is the staffing? How are we billing? How ours? How’s the balance going out for that? Or is all of our data being managed internally? so on so we kind of divide and conquer from that perspective.
Sam Schutte 15:26
Awesome. Cool. Yeah. It’s interesting. It’s a pretty neat role to move in to, you know, having been a developer to move into that role. That’s kind of what everybody wants to do eventually is move into sort of running the show more so it’s great. So and when you mentioned testing and some of these things a standards and that you’re having to develop for your customers that kind of reminded me of or maybe want to ask, you know, what are what are some of the key needs that are out there, that your customers have an industry that’s really kind of driving change on their side and really kind of pushing them into working with you guys on projects? Well,
Tim Nolan 16:04
I know it sounds kind of trite, but a lot of our customers are looking for efficiency. Because it’s from everything from right now, a user is writing something on a clipboard and be much faster to just pull that into a computer and get it in place to their analysis. Data is somebody staring at an Excel sheet for five days, and they can’t, you know, they can’t make it happen any faster. But it I know, you’re asking about what is the industry driving here, but we are kind of interesting as well, because depending on where we’re going, we’re seeing different technological levels. I mean, we’re in Pittsburgh, of course. So there’s a lot of steel applications around here. And we’ll go into a steel application, and it’s the finest technology 1970 could give you and we’re needing to update that. Yeah. Or even just, frankly, a lot of other government institutions. You know, like I mentioned, I started in 2006 at DSA, and my first Job was we had to migrate from Windows XP to Windows 98, where the windows migrate something out of Windows 98 into Windows XP because they weren’t supporting windows 98 anymore. So yeah, quite often you have to hit a lot of those things. But we have a lot of like I mentioned research and cutting edge technology people. And there are several things that are becoming really big, is funny back in 1993, when rich founded the company, he was like, okay, we’re all about data. We’re all about science. And we’re all about automation. Great, we have a name. But now data science is coming into its own as a thing. And you know, we’ve actually over the years got a lot of experience with that we participate with the Pittsburgh tech Council and we actually present at the Data Science Conference locally in town, and I bring that up just because it’s called I love it when you call me big data. And the interesting thing is that those conferences, basically the first step in everybody’s presentation, Okay, so say you have your giant block of data? Well, we’re the step right before that. We’re real experts at getting machines to report their current status and get more and more information back to the user, as it’s going on. And that’s reflected in things like operations management. So for example, somebody wants to be in their office and say, What is every machine on my floor doing right now? Because that tells them how much output they’re getting, you know, which machines breaking down all the time? And who’s, you know, is it getting fixed in enough time, you know, more kind of information that gets fed into like the industrial engineering side of things. And then even beyond that, we’re getting into condition monitoring. So it’s to the point where, you know, in a factory, beyond, you know, my machines that are making anything I have air conditioners and fans and so on. And normally you just run them and then they break down and you fix them. But if you put things like accelerometers on there, you can start seeing the bearing going out of alignment. You can start to see things wear out. So you can do preventative maintenance earlier and then fix them earlier. And to that end, we’ve expanded that into a number of our customers our like power companies and so on where they’re wanting to monitor their, you know, these giant breakers, you know, you have the little breakers in your house, but do we need these giant eight foot long ones that they put on the power poles? Or even the transformers to make sure that Hey, is this about to fail? And can I fix it now rather than blacking out a city block or whatever the case is?
Sam Schutte 19:32
And and I imagine Are you guys getting into like, you know, we’ve talked about prediction, applying artificial intelligence to some of that as well as machine learning.
Tim Nolan 19:40
We would like to, I mean, I’m sure a number of people are like, yes, we’re doing that immediately. And it’s everybody,
Sam Schutte 19:47
nobody really is
Tim Nolan 19:47
nobody really is.
Sam Schutte 19:49
everybody would like to basically
Tim Nolan 19:50
yes. Now, we do do a lot of learning but at a very lower level. So a big thing that we were doing, I mentioned is testing and inspection. Something that we’ve really gained as a skill over the past few years is visual inspection systems. So you take pictures of cameras, and you can measure, you know, alignments. And you can measure even, you know, sizes, you know, you can, the camera becomes your ruler, the camera becomes your caliper saying, hey, are these two flanges far enough apart? Has this been put in the right location. And to that end, we’re starting to put learning algorithms in place where it can be like, okay, these are good ones. And these are bad ones. And I can train you on what a passing and failing system is. And even to, we found that a lot of our customers they want OCR, you know, they want it to Hey, look at this and read what is printed on it, you know, even to check that their printers working. And every single font is different. So you have to retrain their computer to their font quite often in order to get that information off.
Sam Schutte 20:53
Yeah. Interesting. So when companies are putting these kind of systems in, how are you what kind of ROI are they getting, and how do you measure that
Tim Nolan 21:00
That is a big important part. And it even goes back to our sales process. You know, it’s an early thing that we ask our customers and we’re like, Hey, you know, we can make what you’re asking us to do. And we can accomplish it. But if it costs you $2 million, and you save $100,000, I don’t think we’re going to go forward in the business. So why don’t we visit the ROI is one of the first things you know, when I get a sales call, or make a sales call, that’s what I’ll I’ll say, It’s okay, we’re trying to improve this thing, what what saving will that get you and over what period of time? And you know, what, what is it? When do you have to realize that as well, because these are things that accompany you know, I can’t decide for a company do they need a two year return on investment to need a five year didn’t need a two quarter return on investment, you know, depending on where they’re at, and the savings coming in a lot of places, right? So, you know, I can speed things up. So if I could produce more widgets per hour, then I make more money and the return on investment is there. But then there’s also a lot of costs. Just have quality, right? So if I have a system that tests things efficiently and I can get to non destructive testing, then you know, I’m not at the highest level, you know, if I crash a car to crash test it, I can then sell that car, right? And you’ll see that in a lot of other smaller applications, or even cost a failure. You know, if I have a system that is more precise than somebody put it together with their fingers, then maybe those systems those devices that come off Are you have a lower failure rate. Or especially it’s like, if I can test to the level that you never have an escapement from the factory, you know, I don’t want something going out into the world and being like, oh, it failed there. Because that can be very, very costly, because at that point, they have to question everything and pull everything back, you know, the, you know, stop a recall kind of thing. Exactly. And we try to take them through that whole line of reasoning across all of those because, you know, savings come from a lot of places and You know, it’s not just, you know, hey, you have to pay one less person to put these things together. It’s the return comes from a lot of places.
Sam Schutte 23:06
Yeah, I think I find you probably do as well, there’s a, you know, a broad spectrum of value and savings and gain, right? And the hardest ones to pull in and assess with the intangible ones. I mean, I’ve had people that, you know, want to rebuild a system because they, that makes them feel good, makes their employees happy, right, stuff like that. And that is a value. And there is there is but it’s very intangible. What is the dollar value of that a year? Hard to say? Obviously, the on the other end of saying, well, this saves us 30 cents per per minute or something. You can do that math, right. Yeah. Do you think when do you think companies are happy to work with you on that sort of doing that sort of ROI assessment or is it you know, I think a lot of times when when people call engineering and technology companies, they kind of want to just ask like, Well, do you have these skills? Do you have skills X, Y, and Z? And then are you really good at them? Like that’s their assessment? Right?
Tim Nolan 24:03
Right. And we just said in a couple different ways. Number one will tell customers off the bat, if there’s a solution that you can just go buy and screw together and it will work. We’re not the people to hire. You know, we are the people who are making the custom situations where there is not a ready solution for you. And you know, we’re fine to let everybody else have that kind of business. And yes, of course, they do want to know, do we have the skills and that is another reason why our kind of internal certification drive is so high, you know, we want to be able to say I have this many certified LabVIEW architects, I have this many certified LabVIEW embedded embedded developers. And while there are a certain amount of architects out there in the world, you know that they have gone through a minimum level of training you’re not getting some guy who had LabVIEW for a semester in college and are now is now calling themselves a consultant. But it is kind of a sale. Question, you know, we have to make sure that they’re qualified and that, you know, we’re qualified to help them out. We don’t want to waste their time. They don’t want to waste our time. So we’ll often take them down through what we call the pain funnel, which sounds a lot more heavy metal than it actually is. Because we need to figure these things out, like, Oh, well, we’re getting rejections on the line. Well, why does that matter? Oh, well, it it slows down the process. Okay, well, I’m sure that’s okay. Right? No, no, if it slows down the process, then I have to hire another guy. Okay, well, what happens if it well, if I have to hire another guy, I’m gonna get fired because my Okay, now we’re at where it affects you, and where it’s gonna affect the company. And then therefore, it’s a lot easier to say, Okay, I this is where I want it because I can’t sell somebody I can’t make them decide to buy from us. I need to find out from them. Is this a good fit? And can we solve their problems because we’re here to make them more efficient. We’re here to make them more successful at their job. And there has to be a gap there. There has to be a what can we improve and how We’re gonna make your life better before we can proceed forward.
Sam Schutte 26:03
When you can, you can’t want it more than they do. They have, they have to want to improve. And if you show them how to improve, and there’s, we’re just you know, we’re just looking for a quick fix. Again, it’s like you said, it’s not a fit.
Tim Nolan 26:15
Yeah, exactly. Like, we don’t do quick and dirty. We don’t cut corners, that’s our fundamental thing is like, you’re going to get the well architected, robust, reliable system from us. And we’re not going to give you something cheap that will break. Yes, I
Sam Schutte 26:31
Well, I think being picky like that with the projects you select, it means you don’t end up with things that you end up in trouble on because you’re overextending or the value is not there. And now, you know what I mean? you’re charging somebody for something that they feel doesn’t have the value and it just it just means you get better customers.
Tim Nolan 26:47
Absolutely. I mean, you know, qualification is a two way street. You know, I qualify the customer, they qualify me back. So,
Sam Schutte 26:53
Exactly. Yeah. So once you’ve got a qualified customer, let’s say so what kind of relationship. Do you know working relationship do you create with that client? I mean, are you sending an army of engineers out on their site for six months? Do you do everything remote and virtual? There’s a lot of travel involved? How do you kind of communicate with them and work with them? What’s that look like?
Tim Nolan 27:17
Sure. Now in the consulting field, it’s almost impossible to get away from some amount of travel. And you know, there there have been times where I’ve been traveling like 25%. So one week out of every month or so on. But you know, sometimes we go to two ends of that, you know, we have a full fabrication shop a full fabrication lab locally. So quite often, we’ll try to get as much done here as we can before delivering to a customer but the magic always happens at the customer site. So we’re often be there kind of deploying and testing and refining and making sure that’s in place even if it takes several intermediate trips. Now, we did have a customer even earlier this year. Whether we’re setting up a gigantic assembly line with a whole bunch of modules that had to talk to each other, there’s no way we’re going to pull that back and set it up in our office. So we went to Iowa for eight weeks. And we had two dedicated engineers and a project manager there for that time period, trying to get it done, because they had to finish it, and they had to send it to China. So we were certainly, you know, a part of the kind of supply chain there for them. So having embedded engineers, there is something that we also offer, and it’s completely scalable. You know, we can come for an afternoon and walk you through your LabVIEW code and give you some advice and shake your hand and walk away. And that happens quite often. I you know, we have a training mystique about us that you know, it’s something we’ve always done, or you can have an embedded engineer that’s there for six months to get your project done.
Sam Schutte 28:50
Yeah, it’s funny, isn’t it? How you know the deployment of machinery and sort of systems and stuff so forth, you know, in the physical world is so different than the software world right. I mean, you If it’s just coming in installing, like you said, A LabVIEW program, well, as long as you have VPN or something, you can do that from anywhere potentially. Right. But then in some of these factories, I mean, I’ve seen plenty of customers where, like you said, Some parts are made in a small fabrication shop, shipped to somewhere in the US where it’s all put together and then shipped to China, you know, in a much, much bigger chunk, right? And just the logistics of that can get incredibly complicated and interesting.
Tim Nolan 29:27
Yeah, we, we try to be flexible, we’re there to make the customers like life easier. So we’re not gonna force them to, you know, do something that’s not efficient for them. But you know, we’ve gone to UK, we’ve gone to Israel, we’ve gone to Malaysia to deploy things. It’s not like we only operate out of our local city here, although that is certainly nicer for commute periods. But yeah, quite often with a customer, the Find the level of effort to a certain level, and it’s kind of something that we address in the kickoff meeting. We have procedures in place for all these sort of things. We’ll say pay, we can test these things and give you intermediate reports, we can give you intermediate deliveries. And then we can deliver on site. And that’s quite often what they’ll go for. But you know, they’ll also say, we’ll make a desk for you. Now, you mentioned the difference between software and hardware engineering. But we’re starting to actually like to pick up a number of things from say, you know, app developers and so on like that. You know, for right now we have this system, again, with that customer from the UK that I mentioned, where they have a whole bunch of distributed embedded systems that are controlling things. Some of them are local here in Pittsburgh, some of them in the UK, and they wanted to, you know, I can just open up an iPad and see what all of them are doing in every country. Now, they still want updates to that. So in that we can’t go to every oven that they’re running in every country and deploy like that. So we had to set up a deployment system like you have with Excel or with Twitter or what have you where the app on your phone will be like, Oh, I just received an update push. And I need to update locally. So we’ve had to set up that kind of infrastructure as well to support our customers. Because another big thing we try to emphasize is that we’re not going to, you know, deliver and then wash our hands and walk away. We really want to educate and support you throughout the entire cycle of everything that goes on. And our best customers have really reacted well to that where we’ve gone through three, four or five revisions of their product because, you know, they’re updating their product. So they need more tests, they need more fixtures, and so everything kind of gets cycled through that way.
Sam Schutte 31:38
You know, that seems like that’s a real big pain that I’ve heard folks talk about when it comes to the type of machinery we’re kind of talking about here is basically the fact that there’s so much travel involved on their part to go to these facilities just to plug into a serial port and update a robot for instance, right folks have told me they have you know, these are other partners of mine and companies. I know that They’ve been in places they’ll have 100 robots, but there none of them have any kind of network connection or not connected at all. And I mean, you physically have to plug into every single one and patch it or give it a new program. Right. And so kind of supplying that, I guess you might call it middleware, right? That does, you know, that says, hey, there’s an update that needs to be applied to 12 out of these 50 robots. Right. You know, that that is a custom thing. It doesn’t always exist sometimes. And it and it saves evidence, you know, if the guy considered his desk in Tennessee, and install that in Japan, or something, I mean, you know, rather than flying to Japan, that is a personal pain.
Tim Nolan 32:37
Yeah. I mean, and it’s an efficiency thing, too. I mean, it’s saving everybody money, saving everybody time. And it can really help out. There are some tools that seem to be coming out right now. That we’re not going to have to roll around quite as much as we have in the past. Now, have we been burnt by accepting a new tool before it’s ready for primetime? Yes, but you know, there’s certain ones that are getting tested out and that are being rolled out by National Instruments who we’ve always kind of trusted from a software perspective, that might help us out quite a bit with that. So it’s an exciting time for remote access and things like that.
Sam Schutte 33:14
Well, and speaking of business kind of all over the world, I’m curious what kind of what have changes into sort of the global economy, how those changes have affected your clients? What kind of impact Have you seen from that?
Tim Nolan 33:25
Well, I’ve certainly seen everything from trade to regulation affect us. I mean, the only reason I got to do that mercury powerplant thing was because of EPA regulation that got passed, right. And also, you know, over the past 10 years, you know, even though Pittsburgh is kind of this, well, at least used to be the Center for steel, and so on. You go to a lot of these steel plants, and they don’t have any money for internal infrastructure, you know, so they’re, they’re in such big trade wars that they just don’t have capital expense things. Now, that’s changed recently. We’ve certainly seen a lot more opened up from that perspective. But you know, even just, you know, things that happened in the news like, you know, the whole Boeing issue with their planes getting grounded and having to be redone. You know, we’re four stages removed from that, but we’re definitely seeing that you know, we have clients that make testers that do the torque meters that go on to the, onto those planes. So there are several stages in the supply chain removed but Boeing is basically shutting their processes down. So it’s interesting to see how the economy is moving for all that
Sam Schutte 34:31
Whenever that you know that 800 pound gorilla in an industry makes a move it to the ripple effect throughout so many, you know, things you wouldn’t even think would be affected, you know, some lubricant manufacturer in Ohio somewhere that you didn’t even know does anything with Boeing right? I mean, there’s so many companies and jobs it can it can get affected when that you know, Boeing and size that they are makes it makes any choice. You know, so many things are dependent on some of these very large companies.
Tim Nolan 35:02
Yeah, absolutely. Now, a nice thing as well, we’re a member of what we call the control systems integrator Association. I mean, I guess we don’t call it we join them, they name themselves. But every year at their conference, they have kind of an industry analysis and what where’s the economy going in the future that we’ve always found very useful? You know, so there are a bunch of implementations about you know, what jobs robotics are creating, you know, so often in the news, you hear of robots and automation or destroying jobs or, or so on, but it’s like, Okay, well, if you actually analyze it, it’s making 120% of the jobs that were lost. So you know, because of like, we’re saying we need more infrastructure. We need more it. We need more people running these web services. We need more people maintaining them. So it’s not the working on the plant line job, but it is a bunch of people supporting this automation process. That is amazing.
Sam Schutte 36:00
But arguably, those are higher paying jobs. I mean, the the software developers building a control systems for the robots get paid more than the guy who would have hand assembled the chair seat. Yep. Right. You know, and now there’s maybe there’s fewer of those jobs needed, because, you know, you only need free developers to write the software for millions of robots. Right? Better, you know. And that’s not necessarily true, either. I mean, you might need hundreds
Tim Nolan 36:26
And even like I mentioned, you know, you we have people who are now laying internet cable, and they’re doing the power plants that are dedicated to these server farms out there, you know, they there, there are all kinds of levels of jobs that are being supported by just the fact that there’s automation and cloud computing.
Sam Schutte 36:44
Exactly. Yeah. When you look at the tools and technologies we’re working with, right now, you know, if we were to talk, say, three years from now, what sort of technologies and and sort of solutions do you want to get into in that timeframe that you think you’ll be doing sort of three years from now? Let’s say
Tim Nolan 36:59
I see I think the big thing that we want to get into is the cloud based computing. Now, it’s a buzzword right now, of course, we both know some smart people who are into Microsoft Azure and so on. But it is really driven from our customers they want. They don’t want to have to go plug into a computer and get a report or have some guy walk out to something and find out what the data is they, they want the central headquarters to know what all of their plants are doing at any one time. They want the analysis to be happening independently. So once you’ve trained to the system, so you can just say, hey, it’s sending me a flag that there’s going to be an inventory shortage in three weeks. If we continue on the way we’re going right now, because we have an imbalance in our shipping. It’s it really seems like the automation is going to be another level. It’s not going to be the robot on the assembly line. It’s going to be automation, the whole process of where the parts are coming from when they’re getting assembled when they’re coming out when they’re getting packaged. So it’s going to be an issue. See in that line that we’re really going to see, in addition, the cost of these embedded computers and so on and is becoming lower and lower. I mean, you know, eight years ago, you know, who had heard of Arduino or Raspberry Pi or any of these computers on a chip. And you know, nowadays we’re getting up to where we’re having industrial processes being around things like the Intel Galileo platform and even some of the more advanced Arduino. And you know, right now, one of the most competitive imaging solutions on the market is built on the Raspberry Pi platform. So I think the hardware is becoming reliable you know, it’s no longer hobby level cheapness. It’s becoming reliable and embedded and I’m gonna have, you know, everything talking to everything else. And I wanted to say it that way first, before I started dropping, oh, Internet of Things, right, you know, the, you know, the thing gets thrown around all the place, but I’m literally seeing more and more things talking to each other, more and more things reporting their status. Back to central location and people doing very good decision making, and very good business planning based off that, that I think is really driving the industry and the Internet of the, you know, maybe the industrial Internet of Things, IoT rather than IoT would be the more appropriate thing to say.
Sam Schutte 39:16
Absolutely. And, you know, I think it’s interesting to me too. And with regards to cloud and cloud infrastructure, the sort of strategic decisions that that people are kind of inventing, or it’s allowing them to do that. I’ve heard recently that, you know, you wouldn’t necessarily first think of so for instance, there’s a very large company that that I’m familiar with, that they sort of in the next five years would like to spin off a lot of divisions. Let’s say they have they have thousands of divisions, and they’ll probably spin off. I don’t know x number of them. So they are trying to go everything to cloud everything. Right. Well, why? Well, because they currently have maybe 12/15/20 data centers across the world. With just a myriad of servers stacked and intermixed together for all the various divisions, right? So if you sell a large, you know, let’s say a billion dollar division to your to your closest competitor, they don’t really want your their servers in your data center anymore. Right? Well, what does it cost to go through? Figure out which servers are where? Who owns what, and then take them and move them all out? Oh, and plus, each server might have multiple departments or division software running on
Tim Nolan 40:30
and maybe you can’t find it at the server level? It’s, you know, it’s split.
Sam Schutte 40:34
Yeah. And then well, exactly. And then who has access to it, what users what passwords, I mean, hundreds of I mean, well, a lot of money 10s of millions of dollars easily to go through and split up and, and then re condense these data centers, right. Whereas if the entire company is running on, say, Azure, it’s a command line, you know, you say, because, you know, you have all your servers grouped. All these virtual servers belong to one division, and you just run a simple script that takes away all these users access and gives it to a different set.
Tim Nolan 41:05
You’re increasing the liquidity of your servers
Sam Schutte 41:10
Exactly. So it’s almost like a, it’s almost like a real estate opportunity, like in the same way that, you know, maybe, I don’t know, if you’re a trucking company or something. It’s better to have just a couple main depots than to have hundreds of depots because who wants to buy a company that has that, like too many locations, right? So it’s interesting because it’s like they’re making they’re, they’re making they’re sort of acquisition, merger, agility hire by moving into Cloud has nothing to do with cost or technology or anything like that, you know? That was kind of interesting. So definitely a lot of changes in that space recently. What do you what do you think is kind of the most personally rewarding project you’ve worked on recently that you’ve kind of got the most out of?
Tim Nolan 41:54
So like we mentioned, kind of customer interaction is a is a big thing for us, and it Nice to have a nice loyal customer. And maybe Laura’s not the way of rewarding customer. We have one they make the make dry eye treatment. But it’s it’s it’s a division of allegan. So allegan is a giant pharmaceutical company, these, this group was acquired by them, where it’s actually a device, you stick up your nose, and it makes your eyes water. All right, much more technologically efficient than pulling a nose hair. Let’s just put it that way. And, and you know, I’ve had personal friends who have lost corneas and everything like that to dry eye. It’s a very serious condition. So it’s nice number one to be able to have something that’s personally affecting people. But number two, we’ve really become trapped. You know, we were brought in just to do a regression tester. But at this point, where do we make all of their fixtures in software that test their new software in their r&d department. We’re part of their entire fabrication process. So we test the boards before they’re assembled. We test their batteries, we test the charging system for it. We tested after assembled, we serialize it, package it. So we were doing the whole cradle to grave for their entire process. And that was the one I was mentioning. We had their version one, we have their version 1.5. We’re now just delivered their version 2.0 that they’ve been so satisfied with our returning our efficiency and our product that we’ve given them that kind of reward. And that’s just being the project manager on something like that is great.But,DSA is also involved in the community. So I had kind of brought up the idea of getting more involved in the first robotics program. So this is the robotics program that is intended to teach all kinds of levels. There’s the high school level, and then there’s like effects and then there’s even a Lego division essentially for elementary students. And it was founded by Dean Kaman. All right, so for those who don’t know, the segue to the whole bunch of other kind of amazing invention stuff, and I met him at ni week, so I got I got when I was winning an award For, actually, we’ve had mercury monitoring systems here, this comes full circle, he had just don’t convinced me to donate the embedded brains that are going in all of these projects across every FIRST robotics team in the country. So the compact Rio platform, and LabVIEW was the new product that was these robots were being built upon. So, to that end, I’m like, well, we should be teaching these people we teach LabVIEW. So let’s go and teach our local high school. So we would do a seminar at the beginning of every season and teach all the local Pittsburg high schools, how to use LabVIEW, and what the environments like and how to get your robot working. So they don’t have to worry about that technical side of things. And then I would go around to different schools to train them on that side. And even at the event, you know, I basically would wander around the pits where they’re putting these robots together and help them sort out control systems issues before they went out to compete out there. And that was really, really rewarding just to see kind of the future robotics engineering. His training and how excited they got about it, and everything like that. So that was a big thing that I’m really glad that we got involved in.
Sam Schutte 45:07
Yeah, it’s awesome. Because I, you know, in my experience, I recently recently was asked to speak at a local high school. And I kind of thought I’d be talking about computer science and But really, the kids just want to talk about video games, right? Which is fine. I mean, nothing wrong with that. But that’s, that’s not exactly what most people in the industry do. Right. And so it sounds like you were doing a little bit more maybe a little bit more practical, maybe real world. I mean, people do actually do robotics stuff a lot more maybe than they write
Tim Nolan 45:38
Well, that’s actually been a problem. In the control systems realm. We’re getting a lot of talented engineers who are graduating, who want to write apps for the phone or want to build the next bluetooth headset or stuff like that. And these are all valuable things. These are all things that I use, but I we’re kind of losing the the skill set To the the new hot thing, and we’re not getting controls engineers and so on, you know, I’m doing, hopefully a little bit of my part to get them excited about, hey, it’s cool to make motors spin and capture images and make decisions and things like that at a low level or at a non fit in your pocket level.
Sam Schutte 46:18
Exactly. Yeah. Cuz that’s the thing is not everybody. I mean, there’s, you know, without a doubt, there’s always a shortage of technical talent, I suppose. But not everybody is going to work on Call of Duty, you know, and, and if, if, and I think not everybody would really want to once they got into it, I mean, that is a as a rough. I mean, those game studios are rather tough places to work. So that’s awesome. How do clients sort of find you You think I know you’ve, you know, we’ve talked a little bit about, you do a lot of workshops and training and how that, you know, sounds like that might be a pretty good way for you to market your services or you do achieve a lot through that
Tim Nolan 46:59
It’s a big thing. We teach a lot of courses. So a number of customers will have a LabVIEW or an automation challenge ahead of them. And so they’ll come take a course with us, and whether it’s just your kind of intro to LabVIEW, or one of the more kind of advanced courses and acquisition or vision analysis or embedded development, like I mentioned, they get to interact with us for a whole week. And the nice thing about taking a course from us is that we’ve been in the trenches as it were, right? You know, I’m not, I’m not teaching just theory and reading off of slides, I can give you examples of things that I’ve had in place. I can, you know, tell you about my experiences, I can quote, my computer science roommate and the things he told me about state machine theory back in the day, you know, so I can really give that kind of idea and you know, quite often We’ll train them to the level where they can do it themselves, which makes our competitors completely confused. Like, why are you training these people so that they can put you out of business? It’s like, no we want them to trust us they want, we want them to know that we’re never going to hide anything from them, we’re always going to educate them on what they got, even to the end that when you finish with us, you get unlocked source code from us, you know, we don’t remain retain a password protection or anything like that. It’s, we never keep you hostage at the end. But sometimes they’re like, wow, okay, this is cool. And I’ve gotten a little bit up the slope, but this is way too much time for me to handle. Maybe these guys who I just saw demonstrate their expertise can help us out instead. And that has been a really large lead generator for us in general. You know, we had a project that was two, two years of a five engineer team. So you know, 10 man years of effort that came out of a guy recommending us from a training class. So and we have a we have a full time marketing group as well. being members of a number of industry associations, like control systems integrator Association. Like the Pittsburgh tech Council, like the PA Life Sciences Association, we have kind of register exists on registers that will have some inquiries come into. But we also go out and find people we have kind of our wish list of people would like to work with that we kind of update we have outside sales men dedicated to going and kind of finding customers convincing them, you know, that we’re the best solution to their problems. So there’s not really one answer to your question, Sam, but those are the different places we approach it
Sam Schutte 49:30
Yeah,with omni channel, you know, which is what everybody wants to do nowadays, you know, go on podcasts go on trade groups, everything, you know, and you have to because no one channel will produce.
Tim Nolan 49:40
No, yeah. And like I said, the two while there’s kind of my three biggest accounts, one was a call came in to us because they found us on a web search, that’s that Itreatment group that has turned into a three year project. The big one that we said came out of a class and then you know, our latest one we just were kind of in Cleveland, we’re like, hey, those guys look kind of big. Let’s talk to them. And we’ve had 15 projects with them over the last five years. So
Wow, excellent. So what are the you know, for folks out there listening that may work at companies that are having some of these types of problems, or at least running machinery and doing some of this stuff, what are the kind of the three main problems that that they might be having that they should reach out to your form for your firm or you
A big thing that we help people with is getting from your prototype to your product. So quite often, you get something you work it and you need to get it to, you can start ramping up manufacturing. And we can help you get that so you can design for tests you can design for manufacturing. So that’s our we can help you with your product side of things. The next thing is you are confused about something that is going wrong in your in your process. And so we can give you the extra analysis tools taken to help you sort out things so that’s kind of the second area that can really help Be out. And then third, frankly, we’re kind of guns for hire, you’re like you have smart people, they can do this job, if only you had five more of them. And so you can bring us in, we can work with your people, we can supplement or you know, help our processes and just get a project done that you just haven’t been able to get over the finish line. I would say those are the three main areas that we could really help people out with.
Sam Schutte 51:23
Right. So if folks want to reach out to to ask your questions about any of those things.
Tim Nolan 51:28
Sure. Our website is dsautomation.com we’d love to have DSI COMM But I think that’s owned by some Chinese shipping company. And you can reach me at tdn ,as in Tim David Nolan, @dsautomation.com or our central number which is 724-942-6330.
Sam Schutte 51:49
Great. Well, Tim, thanks so much for coming on the show and talk to me about what you’ve been working on. You and I have known each other for a very long time. I wouldn’t even want to admit probably almost 20- 25 years perhaps. So it’s really awesome to kind of remembering where you started in college and such and just seeing the expertise you’ve gained that, obviously, you know, you really have a lot of mastery over helping customers through these, like really complicated problems. So it’s very cool to see.
Tim Nolan 52:19
Sam, I’m glad we no longer have to use our modem to dial up in to get our email at college like we used to have to. So things have come a long way.
Sam Schutte 52:28
Yeah, yeah, exactly. Yeah. Long distance, high quality audio. It’s amazing.
Tim Nolan 52:34
Thank you.
Sam Schutte 52:34
All right. Thank you.
