Inside View - Eduardo Torrealba, Lumafield and X‑Ray Vision of Parts

Eduardo Torrealba, founder and CEO of Lumafield, on how industrial CT gives engineers a safe, fast way to see inside products and make better decisions. He discusses LumaField’s approach to trials, safety, resolution and GD&T from scans.

• Value of non‑destructive 3D inspection for complex assemblies
• Differences between industrial and medical CT and how shielding works
• Try‑before‑you‑buy approach and what it costs
• Main manufacturing use cases across batteries, plastics, and electronics
• Limits and trade‑offs for resolution, contrast, and dense materials
• CAD‑to‑scan comparison and ethical lines around reverse engineering
• GD&T from scans as a faster alternative to extensive CMM programs
• Company footprint, manufacturing in Massachusetts and software in San Francisco

Transcript

Meet LumaField And CT Scanning

Roopinder 0:19

Hello, and welcome to FoDES, the future of design and engineering software podcast. My name is Roopinder Tara. On the show, we will have guests that will discuss tools and technology that engineers will find interesting and useful. Hi, Eduardo. Good to meet you. I want to welcome Eduardo Torrealba to the show. Yes for having me. Eduardo is the CEO and founder of Lumafield. Lumafield is a scanning company selling hardware and software. Solutions are used for industrial CT scanning. Curious about CT-scanning because that is mostly used as far as I know. I've been a patient of CT scans and MRIs. They're always used in the medical field. So how does it relate to manufacturing?

Eduardo 1:03

Industrial CT scanning is as old as medical CT scanning. The technology was invented at the same time. And the value proposition for human beings or cell phones or drones or batteries, it's all the same, right? It's the desire to see inside of something without having to cut it open. And with human beings, the price of cutting a person open is very high, both in direct costs and in risk. And in industrial CT scanning, product development, destructive testing could also be very expensive. We bring not just the 2D X-ray of an object, like if you get your arm X-rayed, if you have a sprain or your word that you have a broken bone, we get the full CT the 360, 3D data of the entire inside of an object. It's invaluable for complex internal geometries, inspecting complex assemblies, multiple parts that are nested together, looking at things like intermittent electrical contacts. But at the end of the day, the thing we give our customers more than a scan or a couture is confidence to go faster. If you're sure that everything you're making is good, you're able to operate closer to the limit of your materials, your processes, your suppliers, you can get more efficiency out of every process that you run. That's a huge concern for every one of our customers today, whether they're a Fortune 10 mega corp or a small startup that's just getting started in production.

Roopinder 2:25

How should I say this? Why should I believe you when you say you have a really good solution? What do you know about CT scanning? What's your background? I see here. You're at Form Labs and you were so you were a director of engineering. Then you went on Yeah. Then you went on to freelance. Consulted with a lot of people, right? A lot of companies and advised them. And then you decided to go off on your own. You could do this. Why tell other people how to do it when you could do it yourself?

Eduardo 2:51

Why should you believe me? And my answer to that is like, you shouldn't. There's no reason that you should have to just trust me. We actually engage with all of our customers through what we call the try-by motion. We're the only CT scanning company in the world that I'm aware of that'll give you a trial of our unit. We'll deploy a system to your facility for 60 days, scan anything you want, look at the data, examine it, give it a torture test. We'll let you experience this technology before you have to pay us money every year to access our software, hardware, or services. Do that? You'll approve it. But we do that in the United States. We don't do trials in other countries right now. If you're a company in the United States, we'll put a system in your facility. We don't do it for free. You have to pay a little bit to try it. We believe there's a lot more value in actually experiencing. You don't know if it'll work. Let's try it out. And if it doesn't work, we're totally fine with that. I don't want my customers to have my products if it's not going to work for them. There's no reason you should trust us. You should go and use it yourself before you commit to anything at Lumaffield.

Roopinder 3:46

Or do I have to give you my first board to hold as a hostage?

Trust, Trials, And Buying Model

Eduardo 3:49

No, I'd say the cost of the trial depends on the size of the organization and the length of time and shipping. It's something like the equivalent of paying for a few scans from a third-party lab. But it's very reasonable in terms of the cost to do a trial. You're not putting a million dollars down. Our scanners don't cost that much. They're much more affordable than anything else in the space. That's part of the reason we have such an open and transparent process that we put our technology in front of our potential customers. Now we make sure in here like a company and we're not putting this in somebody's garage. But it is X-ray equipment. We've got to be careful with it. But yeah, we've done it for tons of organizations. We deployed these things in conference rooms and factory floors and production facilities in Asia.

Roopinder 4:28

They're all over the world at this point. So it is X-rays, and I'm reminded of the medical use of it. In medicine, I'm going to get zapped with X-rays, but they're going to be of lower energy than an industrial machine, correct? I would totally get fried if I went under one of your machines.

Eduardo 4:44

Wouldn't get fried. There's a huge number of factors to skip over all the biological and radiation safety and physics, which we take very seriously at low seals. Our systems are in closed cabinets. It's a sealed lead box. So you can't turn the X-ray system on unless the door is closed. There's a special interlock where if you were to run into it with a forklift and jam the door open, it would cut the electricity of the X-ray source. So it's extremely safe. There's no radiation leakage. We hold ourselves to a higher standard for shielding than the federal standard. There's no detectable x-ray emissions, even a tiny amount of distance away from the system. You could sit next to it all day long and never have any kind of additional exposure. Certainly less than you would get on a short flight. Our systems don't pose any of those risks because they're sealed systems to keep all the x-rays inside the box pointed at the point. Frankly, 3D printing is not a production technology. Our customers do things like medical devices, batteries, consumer package goods. Like I talked about the shampoo bottle situation here. We have a lot of customers that work in aerospace and defense, and they're making things like defensive flares or energetic systems. These are items that are made in huge quantities. And being able to inspect them and make sure that they're operating properly, that's the predominant use case. Ultimately, is someone who works in 3D printing for a long time. It's really much more of a prototyping and short-run manufacturing technology than it is a mass manufacturing technology. Most of our customers are inspecting injection-molded parts, assemblies, electronics, things that are made in huge quantities.

Safety, Shielding, And Radiation Basics

Roopinder 6:27

Batteries, I'm reminded of battery technology. So I'm a mechanical engineer, and you are too. Congratulations on learning all that physics. You should have a secret handshake. We'll do it. Yeah, there we go. So you had to learn a lot of physics, right? Well beyond what an ME normally has to learn to get into this business. You can study electronics, but the battery technology, batteries are suffer from this thing called dendrification. Is that the right term? And get the dendrites. The dendrites tend to grow and as the charges, discharges, and they're microscopic, I understand. Now, is that something that's a use case? Something like what you have?

Eduardo 7:01

Yeah. That is a failure mode for batteries. That is not usually the primary place where our customers are using us. Our customers use us to inspect freshly manufactured cells. They usually haven't had enough charge-discharge cycles on them to find that. There are use cases where our customers are doing research and development versus production. A lot that goes into batteries. I've come to really appreciate a lot of the failure modes in the development, the production, and use of batteries. We have a deep dive into 18650 lithium-ion battery cells called the battery report. We've got over a thousand scans of batteries that we ran to do statistical analysis of various manufacturers and different products that you can buy online for these applications. What we found in that report is that a significant percentage, almost 10% of the third-party, refurbished, non-name brand cells that we inspected had critical failures that could actually lead to a short or even a fire with those batteries. The report that we did on batteries, Samsung, Panasonic, they build really amazing pieces of technology. And part of that is the reliance on really high quality information about everything that they're manufacturing. Whether that's T-scanning or other processes, that level of knowledge is difficult to bring to a small or medium-sized manufacturer. We've tried to make T-scanning affordable enough that you can operate your 300-person defense contractor in Nebraska the same way that Apple or Samsung or Philips or Boeing operates their quality control program by making this technology more accessible.

Roopinder 8:26

Are your tools, your systems used for nefarious purpose? I'm talking about industrial espionage in the case of batteries. How are my competitors' batteries made? Rather than take them apart, I just reveal every all the secrets of my competitors. Sure, there's a lot to that question.

Use Cases Beyond Prototyping

Eduardo 8:43

We could talk about what is industrial espionage, what is reverse engineering. I could tell you that the vast majority of our customers, that's not the reason they buy our technology. They're buying it to look at the things that they make. A lot of our customers use a tool that compares the as manufactured object with the CAD model that they make. They're the ones that design the product. They're really looking at the things they make themselves. You can take apart anything you buy. It might avoid the warranty, but you can take it apart. You can T-scan anything that you buy. If you decide to use that information to violate another company's patents, that would be illegal. There's nothing illegal about buying something and taking it apart. If I work at Samsung, I can buy an iPhone and take it apart. That's nothing illegal about that. The problem is when you're violating their intellectual property or their patents in new products that you're making. But you can always look inside of things. I mean, teardowns, right? iFixit. There's a partner of ours now, actually, they do a lot of scanning on this. They take apart every piece of electronics that's released and talk about the way that it's manufactured and assembled and how repairable it is. I would personally classify that as very different from industrial espionage. I would think of industrial espionage as gaining access to trade secrets that you're not supposed to have access to. And our technology is never used for those kinds of purposes.

Roopinder 9:48

It might be industrial espionage. If you broke into Apple's headquarters to look at their phones, but if you bought that $130 cable scan it, that's totally fine. Just answer that question. Why is it worth $130? That's right.

Eduardo 10:00

I've come to appreciate the complexity of the USB C cables quite a bit more since starting to stamp it.

Roopinder 10:05

Engineers, it's really important to see what's inside the object. And it's frustrating when you can't do that. When you have to shatter something, use some strange means that take it apart. Why not CT scan it to see how it goes? If I want to detect voids or very small objects, by the way, you have to tell me how small this can go. How is that differentiated by color? For example, if I want to see a void in a casting, will it show up red?

Eduardo 10:29

The visualization of the data in our software, you could actually apply a number of different color schemes to that. So I could show a void as red or blue or green or whatever it might be. What's really happening is, as I said before, the x-rays are passing through the object where we build our own X-ray detectors from scratch that we invented here at Lumafield to be more affordable, longer lasting, more robust, better fit for the use cases our customers have. Through some really incredible algorithms, we reconstruct that into a 3D representation. It shows up with different levels of density. If you scan a plastic product, it'll be one color. If you had plastic and the metal absorbs the x-rays differently than the plastic, and so you get some contrast between those two things. Those can show up as different colors or different shades or whatever visualization you want to apply. And the voids would show up as whatever color you have empty space set to in your visualization.

Roopinder 11:18

It's hard to discern what is broken or damaged in an X-ray because of that lack of color differentiation. Yeah, the contrast, right? Yeah, exactly.

Battery Insights And The Battery Report

Eduardo 11:26

So we have we have pretty good dynamic range. We're able to visualize these things. Just went resolution earlier and say we can usually resolve things in our high-resolution system down to about which is very, very small. It's not small enough to inspect, like, you know, the transistors on a NVIDIA chip or something crazy like that. But you can see one-tenth of the diameter of a human hair. So you can see a lot inside of that to that level of resolution. How is it? There's a number of factors that influence resolvable feature size. The resolution of a CT scan is a function of the X-ray source. So think about how like finely, this is an imperfect analogy, but like how tight is the tip of your pen, right? If you have a very small ink pen versus a Sharpie versus the big paint marker, you get a different level of resolution when you're drawing. So think of that as sort of the light source that we have that we're shining. And then there's the resolution of the detector technology. And you could think of that as like you have this pen and you're coloring over different squares and they turn on or off depending on whether they're getting hit with the ink. The resolution of the system, the resolution of the detector, and then another one that's the, you know, so you've ever done shadow puppets with your hands. If you take a flashlight and you put your hand on the wall, the shadow is the same size as your hand. But if you walk away from the wall and you hold your hand close to the light, you get this thing called geometric magnification where you're able to get like a very large shadow. And so that distance and the cone angle, there's a lot that goes into the model of what determines the resolution of one of these systems. But we figure all that out and just give you the number and you don't have to worry about all those things. We just tell you the resolution whenever you're setting up a scan. And we've actually automated many of those things. If you buy a traditional C scanner, you're buying a scientific instrument. And when you're a scientist, you want access to everything, like all the raw values. You want to be the one that creates the algorithm or the process that you use to achieve your result. It's like taking a camera and setting it to fully manual and individually tweaking every single setting. Get the raw image to get exactly what you want, right? I mostly just want to take my phone out and hit the button and get a picture, right? That's what our engineers want. That's what our customers want. They want an auto feature for CT scanning, and that's what we built for them in our software. You don't have to know about everything I just described to understand if you're getting a good scan or a bad scan with the messy else.

Roopinder 13:36

That's very good. Definitely helps me. I thought of another use case that would be really good for e scanning. I don't know. I saw this watchmaker. It was a 60 Minutes episore a watchmaker making extremely intricate watch movements in Swiss watches. They get down to very small. Okay. I thought I really want to know how that works, but I don't want to take it apart because it'll never come back together. Are watchmakers a customer? And that's the same.

Eduardo 14:00

We have not really gotten into that market. There are levels to being a watch person versus not a watch person, and at the very beginning of that. But yeah, I do love the mechanical design pieces. I'm a mechanical engineer. It's not a surprising sort of overlap in interests. I've done some scans of watch movements. I'll send you one that you can take a look at in Voyager. When we have scanned high-end mechanical wristwatches, it's kind of two limiting factors that make that a difficult application for CT scanning in general, whether it's a million-dollar system or one of our systems. One is there's a lot of precious metals and high-end wristwatches. And things like gold, silver, and platinum, they have a very high atomic number and they absorb a lot of x-rays. You need a very powerful X-ray source to punch through a gold Patek Philippe watch, for instance. The other thing that's interesting is that a lot of the glass on the crystal for a lot of modern high-end watches, there's a polymer anti-reflective coating. If you take a cheap piece of plastic and you put it on the sun, it'll turn yellow. It'll color. The same thing can happen with anti-reflective coating on the crystal on a watch. And that doesn't happen under normal conditions. You could leave it outside for years and it wouldn't happen. But if you have a super powerful X-ray that's punching through platinum and hitting your anti-reflective coating, it can yellow that coating. So you end up with this combination of moderately fragile polymers that are very thin with super absorbent gold, silver, and platinum. Nobody wants a yellowed watch face. Exactly. So you're right. So and then in the lower cost systems, there could be some applications for that. I would wager that the high-end Swiss watch world is probably a place where they are at the absolute cutting limit of what quality looks like. And that's very manual. It's a lot of by hand operations. Right. If you're watching and you work at Omega or Rolex, I'd love to get you a system and come visit your facility.

Roopinder 15:48

They have to take care of that problem somehow.

Eduardo 15:50

They could examine the movement without the crystal or a case. The area that we make the biggest difference in Lumafield is putting inspection technology that our customers have never had in their hands. We are not the tool. We don't make Formula One cars. If you're a Formula One car driver, you're shopping at a very different place than a person who needs to buy a fleet of F-150 pickup trucks for their entire field services organization for their utility company. We make the F-150 pickup truck. We don't make the Ferrari Formula One car. And so we sell to these people who want to go from zero to 100. They don't need to go from 100 to 110 or from 1,000 to 1,000. We typically sell to a different type of organization than maybe the highest end luxury manufacturers. It's really a different design philosophy. We're more about accessibility than we are about highest end performance.

Ethics, Reverse Engineering, And IP

Roopinder 16:38

I think we've covered everything. I do want to give you a chance to talk about what you do. I read the press announcement about the GDT Act. Are you able to now put GDT on parts that you have scanned? I thought, wow, this makes inspection of parts really much more useful.

Eduardo 16:54

We talked to our customers and we asked them about how they use our tools. What other tools do they have in their quality lab and their factory? How do they use those tools? Is there something they wish we could do? Coordinate measuring machines are pretty incredible. They've been around for a very long time. They're a little bit older than industrial sea scanning. They haven't changed a whole lot in that time period. I talked about injection-molded plastic. A lot of high precision injection molded plastic for things like continuous glucose monitors, drug delivery systems, lure locks. There's a lot of medical devices that are plastic. If there's a problem with the geometry, you don't get function and then you can have an issue with patient care. And so a lot of those parts end up being inspected with high precision ordinary measuring machines to verify things like the injection mold before they make 100,000 of something or a million of something. They'll check kind of first article or maybe even first and last objects in a shift to make sure there hasn't been any drift in the quality metrics. In programming a CMM takes a long time. It's a full-time job. It's a very slow process to move the probe tip around and touch in a bunch of different places. The interpretation of that data set can be somewhat complicated. With Limafield, you get the full data set through one of our scans. You can intuitively go into the geometry set your measurements, the things you care about for parallelism, perpendicularity, angles, dimensions, all of that you can do in our software program. One, press a button and apply that to scans that you don't have to meticulously fixture and then wait several hours. You can throw 20 of something in a scanner, get a bunch of data, and then have these measurements very quickly. It's just a much easier, faster way to gather that same decision that maybe you had to make in the building field. You can do 10 or 20 times as many. You could have more confidence to operate on the edge of what's possible with your designs, your materials, your partners, your production processes.

Roopinder 18:35

Do you actually making and producing systems and selling them already? This is an ongoing business. This isn't just your normal startup, right?

Eduardo 18:43

We do between 10 and 100 million dollars of annual revenue. We have multiple customers in the Fortune 500. We have systems in North America, Europe, and Asia. We're almost 200 employees at this point, offices on both coasts. We've raised over 160 million dollars in venture capital. Long way to go, but we're definitely a real business at this point. Now you're in Cambridge, but your house offices in my neighborhood in San Francisco, correct? We have about 50, 55-person office in San Francisco. I'm out there about one week a month.

Roopinder 19:09

We're in San Francisco.

Eduardo 19:10

Oh, we're in Soma, right beside Oracle Park.

Roopinder 19:12

Where are the machines actually manufactured?

Eduardo 19:14

They're manufacturers here in Massachusetts. We have a 20,000 square foot facility where we do all of our manufacturing for our hardware products in San Francisco. We focus on software engineering, browser-based tools, the back end, our DevOps organization, some of our software RD, the construction algorithms, some product team members, a lot of our marketing team is in San Francisco. And then our sales team is distributed around the country. We think of manufacturing as a face-to-face business. So our sales team is in California to access customers on the west coast.

Roopinder 19:41

Makes sense. Yeah. Come here for the software, go there for the hardware. Makes perfect sense. Oh, they're in it for just the venture capitalists. You did have great success with venture capital. Were they not from the Silicon Valley?

Eduardo 19:51

Yeah. All of our VC money is from Silicon Valley for sure. So we're definitely a traditional venture bag like Lux Capital, Kleiner Perkins, Spark Capital. IVP, you know, Sandhill Road, or I guess most of them are up in San Francisco now.

Roopinder 20:04

Well, congratulations on that.

Eduardo 20:05

You're gonna have to pay it all back eventually, right? So it's not about raising the money, it's about building the company. And that's the thing I'm much more proud of than the fundraising. But it helps you don't really celebrate stopping at the gas station on a road trip. It's part of the process, but it's not the point in the trip.

Resolution, Contrast, And Visualization

Roopinder 20:18

That's a good way to put it. Well, thank you very much. That's been great talking to you and great to learn more about the technology and the products. I wish you well. CT-scanning definitely has a place. Like I said, love to look inside things without having to put them back together. Me too. I think it will be great. Thank you so much.

Eduardo 20:33

Appreciate you having me.

Roopinder 20:34

Thank you for listening to FoDES, the Future of Design and Engineering Software Show, brought to you by ENGtechnica. I hope you have learned of a new application or technology that will help you with your job. If you have an application you think would be of interest to other engineers, please let me know by emailing me at roopinder at engtechnica.com or message me on LinkedIn.