Ensuring Quality with IoT: How RAFI and ifm Digitalize Injection Molding

Hello, dear friends of IoT. In this episode, we dive into the world of production—more precisely, into the injection molding operations at RAFI. Since 1900, RAFI has been manufacturing high-quality HMI components such as displays, touch panels, and push buttons. Today, they show how process data can truly be put to use together with their IoT partner ifm and the IoT solution moneo. We talk about the typical challenges that come with such implementations and share concrete results. You’ll also get a sense of what to look out for in your own projects.
My guests today are Guido Zoll, who works in development and design at RAFI GmbH & Co. KG, part of the RAFI Group, and Christoph Schneider, Vice President Product Management Application at ifm solutions GmbH. You’ll learn which specific use case RAFI has implemented, what savings were achieved, and what business case lies behind it. As always, you can find all details about this and similar implementations at www.iotusecase.com and in the show notes.
Enjoy the episode—let’s go!

A warm welcome to you, Christoph, and to you as well, Guido. Christoph, let’s start with you. How are you today, and where are you joining from?

Christoph

I’m in my office right now, working on some evaluations and project preparations, and I’m looking forward to today’s podcast with you and Guido.

When you say office—where exactly are you based, in which city?

Christoph

I’m located near Baden-Baden. I live there and have my office in Appenweier. I’m officially based at ifm solutions in Siegen and usually move between our headquarters in Essen, the Siegen site, my home office, and our production facility at Lake Constance.

Very nice. And Guido, where are you located? I assume you’re not exactly in the same region. We’re at RAFI in Berg near Ravensburg, about 25 kilometers north of Lake Constance.
Do you also have production directly on site there?

Guido

We’re at RAFI in Berg near Ravensburg, about 25 kilometers north of Lake Constance.

Do you also have production directly on site there?

Guido

Exactly. Around 1,000 people work here, with development and production located at the same site.

Great, I’m glad to have you both here today. Let’s start with a quick introduction for those who don’t know you yet. Guido, you work in development and design at RAFI and bring many years of experience in product development—especially in areas where robust HMI technology and quality assurance are key. Maybe you can tell us what fascinates you personally about IoT and digital projects, and why you enjoy your work.

Guido

I originally trained as a toolmaker, specializing in injection molds. After that, I studied mechanical engineering and have now been with RAFI for 28 years, working in various design and development areas across different industries. Because of my background in injection molding, I started looking for ways to connect with sensor manufacturers—and that’s how I came across ifm. ifm offers a wide range of sensors and, with its software moneo, provides the ability to record and analyze process parameters.

So, the project we’re talking about today comes directly from your own production—right where the injection molding machines are located, and you’re personally overseeing it?

Guido

Exactly. In the injection molding process, hot material is injected into a large steel mold and then cooled down again. The goal is to better understand the process and the mold itself—to identify which parameters, such as cooling times, temperature control, or cycle times, influence quality and efficiency.

Very interesting. Let’s move straight into your project. But first, Christoph, a quick introduction for you. You’re Vice President of Product Management Application at ifm solutions GmbH. Maybe you can give us a quick overview of the company. I’m not even sure how long we’ve known each other—it’s been a few years at least. You focus, among other things, on solutions around moneo. ifm is globally known as a market leader in sensor technology, family-owned, and has been in the market for many years. Now, you’re also heavily engaged in the IIoT space. Could you explain how IoT is anchored at ifm and what kind of importance projects like the one with RAFI have for you?

Christoph

At ifm, IoT started quite early—around 2003, with vibration diagnostics. But we quickly realized that sensor technology alone wouldn’t be enough. We needed solutions. That’s why, starting in 2015, we began building up our IIoT division. Our goal is to evolve from being a world leader in sensor technology to becoming a true digitalization and solutions provider.
With our IIoT platform moneo, we provide analytics, dashboards, and AI functions to bring measurement values from OT into IT. These data can then be processed and analyzed to give customers tangible results—like how long a machine will keep running, whether there are early signs of failure, or when alarms should be triggered if a process drifts out of control.
All of this can be achieved with moneo. Our vision is to integrate the entire automation landscape—controllers, sensors, and all related components. In addition, we’ve built interfaces to ERP systems so that we can create a continuous data flow from the sensor level all the way up to the ERP, enabling truly end-to-end solutions.

You mentioned that you’re recording process parameters and specific timing data. Can you explain in a bit more detail why this project exists in the first place and why it’s so important for you?

Guido

We wanted to understand exactly what happens during an injection molding cycle. We had several questions: How do the temperatures behave inside the mold? What pressures occur? What are the inlet and outlet temperatures of the cooling fluid? We wanted to know our actual process conditions and use these parameters to derive how we could improve both the process and the product quality. For example, can we detect from part deviations or tolerance issues that the cooling fluid has become too warm? We wanted to find out where we can intervene to counteract such effects—and at the same time optimize the cycle time of the entire process, perhaps even shortening it by one or two seconds.

I see. Maybe one more question before we go deeper—what kind of products do you actually manufacture there, and for what types of customers?

Guido

We produce plastic injection-molded components for human-machine interfaces. That includes input devices for medical technology, such as CT scanners or X-ray machines, but also for tractors, snow groomers, and other types of machinery. Essentially, everywhere a human interacts with a machine—through buttons, emergency stops, key switches, or touchscreens like on a coffee machine—you’ll find our components inside.

So it’s all about plastic injection molding, meaning the housings or components needed for those interfaces. You mentioned earlier that your project mainly focuses on process data—around the injection molding process itself, but also upstream and downstream steps, like the cooling circuit. Are those the main data points you’re looking at?

Guido

Exactly. In addition to the cooling circuit, we also monitor the machine’s power consumption to detect when it draws peak loads. We also keep an eye on the compressed air supply for the robot arm to identify possible leaks or loose fittings. Such deviations are visible directly in the trend curves, and we can set up alarms in moneo to detect them in real time. That way, we can immediately see if a problem occurs—for example, if the robot arm drops a part because the air pressure isn’t right.

That’s super interesting. So, when we talk about the business case behind it—the point where you said it’s worth investing in this technology—what direction does that go in? You mentioned quality management earlier. Is it mainly about improving part quality, which is obviously crucial for your customers? Or how do you calculate such a business case?

Guido

It’s about both quality and productivity. At the same time, everyone in the team—from group leaders to shift supervisors—has access to the key figures. Everyone can see in the trend curves whether a process is running stably or if something’s getting out of control.

Got it. You’re already quite far along with this project, while many listening might still be at the beginning of such initiatives. How did you actually come up with the idea to focus on these specific data points—like power peaks, compressed air, or cooling circuit parameters?

Guido

The main idea was to use resources more efficiently and only as much as truly necessary—whether that’s cooling water, electricity, or compressed air. We also wanted to avoid power peaks. When several machines switch on their heating elements at the same time, it causes high load peaks. With the data we’ve collected, we can now control the process so that machines start up in a staggered sequence. This helps reduce those peaks and save energy.

Projects like this can be quite complex to implement. Christoph, a question for you—when you visit customers, do you usually find ifm sensors already in place, or are there often components from other manufacturers as well? How does such an integration typically work, also in this project with RAFI?

Christoph

We always start by talking with the customer to understand exactly where the pain points in production are. What problems exist with the machines? Which aspects are most critical? We bring a lot of experience from previous injection molding projects. Typical bottlenecks include mold cooling and compressed air consumption.
Next, we look at what’s already available on the machine and what needs to be added. Often, components like power meters are already installed and can be easily integrated into our software. Depending on the machine type, control system data can also be very useful. So, we determine which data are needed to solve the customer’s problem and what might need to be retrofitted. Based on that, we make recommendations—some customers handle the implementation themselves, others work with partners. In some cases, we support them together with system integrators.
The overall goal is always the same: to make the machine transparent. As Guido already mentioned, one of the most important steps is to make the relevant parameters visible in the first place. Only with real transparency can you identify where to take action. Often, anomalies appear that no one noticed before. For example, you might assume that water is flowing through the cooling circuit in the right amount— if the finished part looks fine, you think everything’s okay. But maybe the mold is being overcooled, or the cooling performance is uneven. Once you can see this data clearly, you can control cooling more precisely and even shorten the cycle time.
That ultimately means you can produce more parts per hour without sacrificing quality. And that’s the point where new ideas start to emerge: What else can we improve? What exactly happens in the process when the material is injected or when the temperature profile in the mold changes? These insights open the door to the next optimization steps.

You often hear that in projects like this. What I find fascinating is that you analyze these bottlenecks so systematically and feed your existing knowledge into the process. You’ve already implemented hundreds of use cases. Do you have an internal knowledge base that your teams use when working with customers? Something like: “Injection molding machines—we know the typical issues”? And then, as in this case, the customer—like Guido—adds his process expertise to identify where it’s worth looking closer?

Christoph

Yes, at our facility on Lake Constance we have a dedicated team that focuses on application development. Whenever we receive a new project from a customer, we recreate the application in-house, essentially building and testing it ourselves to ensure it works as intended. Once verified, we document the entire setup. This information is then made available to our colleagues worldwide.
We maintain an internal website where all these applications are described in detail—including the specific use case, the customer’s challenge, and the resulting benefits. This way, every sales engineer or application expert doesn’t have to start from scratch. If someone visits a customer and sees an injection molding machine, they can immediately access existing knowledge and reference cases.
This approach allows us to conduct much more targeted discussions with customers and quickly identify which data provide the most value. Only after that do we move into deeper analysis or individual customization.

Excellent. And to everyone listening—I’d be curious to know what use cases you’re currently working on. Feel free to share them in the comments or connect with us on LinkedIn. Every use case is unique, but this one is a classic example of quality monitoring combined with energy management. And of course, an open invitation to join our IoT Use Case Community to exchange ideas and experiences.
Guido, Christoph—if it’s okay with you, I’ll link your LinkedIn profiles in the show notes so listeners can connect with you directly and share insights.

[15:59] Challenges, potentials and status quo – This is what the use case looks like in practice

What were some of the typical challenges you encountered during implementation? Were you able to access all the necessary data right away, or were there difficulties during integration?

Guido

The first challenge was to get everyone on board. Once the system was installed and team leaders could see the tangible benefits, the right discussions started happening. They now recognize the current status of the tool and the injection molding cycle directly in the trend curves. The next hurdle is to translate insights into actions. For example, if I realize that a cooling pump needs to run slower or faster, I need both the authorization and the technical capability to access the interfaces of external devices. That means reaching out to the manufacturer of the cooling or temperature control unit to get access to their interface. The same applies to robots or injection molding machines. These interfaces are crucial— our goal is to turn the insights gained from moneo into real adjustments, such as changing pump speeds, robot movements, or other control parameters.

So that means you’re accessing the manufacturers’ device data, using the existing interfaces, and avoiding new data silos instead of adding extra measurement points?

Guido

Exactly. On one of our machines, for instance, we currently use two temperature control units. Each has four sensors for the inlet temperature and four for the outlet temperature, plus sensors for coolant flow in liters per minute. We monitor both the ejector side and the nozzle side of the mold. These trend curves have already revealed key optimization opportunities. For example, if we operate Cooling Circuit 1 with a specific flow rate and expect a temperature difference of two degrees Celsius, but only achieve 0.8 degrees, that means we’re pushing too much water through the system— essentially wasting energy. We can reduce the flow rate to reach the desired temperature delta while keeping the cycle stable.

Fantastic. That really makes the business case behind it clear—and how this kind of project pays off over time. One thing that comes to mind, Christoph: does the EU Data Act play a role for your customers? Manufacturers are increasingly required to make machine data available. Maybe that’s even an enabler for creating interfaces and making data easier to use. Do you encounter this topic often?

Christoph

Absolutely. It’s coming up more and more. You have to keep in mind that these machines contain a lot of manufacturer know-how. Many are hesitant to share too much about what’s happening inside their systems. The more machines you have from the same manufacturer, the easier it becomes to access certain information. Still, getting to the control system data is often difficult. You know the data exist in the PLC, you know they’re being measured—but you simply can’t access them easily. In that regard, the EU Data Act could definitely help.

A huge topic, no doubt. I’m actually considering doing a special episode on it. I think we last covered it in detail back in Episode 107, so feel free to check that out if you’d like to learn more. The topic is also being discussed extensively in our community. Many manufacturers are asking how to enable data access and build proper interfaces. Some have never heard of the regulation, while others are already working on export functions. I believe this kind of collaboration and openness is essential if IoT projects are to succeed at scale.

Christoph

We’re seeing this trend very strongly, for example in the beverage industry. Producers are increasingly demanding transparency over their equipment. We’ve had many discussions in this space because we can read and process control and sensor data from almost all major manufacturers. That’s not only helping beverage producers—it’s benefiting many other industries as well.

[21:03] Solutions, offerings and services – A look at the technologies used

Christoph, earlier you mentioned that you offer different solutions for customers, made up of various product modules, depending on their requirements. Guido, maybe let’s start with you—what exactly are you using from ifm, and what does the solution look like for your users, like team leaders or group supervisors?

Guido

From ifm, we’re using compressed air sensors and flow sensors, such as the SBT633 and the SD8500 flow meter. These sensors capture key process parameters and provide the basis for us to control our processes more precisely.

And then, Christoph, those data flow into your moneo software, right? You call them moneo IIoT devices. These could be sensors from your entire product portfolio. So how does it continue from there—from the sensor level up to the application layer?

Christoph

In projects like this, we consistently rely on IO-Link as the communication standard between the sensor and the IO-Link master. That allows us to get much more information out of each device. With a conventional connection, you might get two values, but via IO-Link, you can access five or six.
For example, a compressed air sensor provides the current pressure, flow rate, temperature, and—through an integrated totalizer—the cumulative air consumption. These values are collected by the IO-Link master, which converts them to an Ethernet protocol and integrates them into the customer’s internal network.
From there, the data either flow to a local computer running moneo or through an edge gateway into the cloud. The trend is clearly moving toward cloud-based solutions, since many customers want to reduce their IT infrastructure. Instead of maintaining local servers, they prefer to securely transfer their data directly to the cloud for analysis.

Okay. And for you, Guido, as an end user, there are probably different roles involved, right? It’s mainly about the visualization of data. I’m not sure how far along you are—these projects are often still evolving— but you mentioned measuring things like cooling water flow, temperatures in mold halves, and similar parameters. Are those data visualized in moneo, or on a local computer?

Guido

Exactly. And then it’s mainly the production planners who come into play. They design these systems and draw the right conclusions from them. The goal is to shorten cycle times while reducing energy consumption, so that we can—just as Christoph mentioned—produce more parts per hour.

Right, that closes the loop from the beginning. You mentioned earlier that the real goal is to draw the right conclusions from the data. I think that’s the real discipline—actually looking at the data and deciding what to do with it.

Christoph

Exactly. The key is to work with the data. A system like this isn’t meant to just produce pretty dashboards. The real value only comes when you interpret the measurements and take action based on them. That’s when you get tangible benefits.
Shortly after commissioning, we noticed that air consumption at one machine was steadily increasing. RAFI was able to correct it immediately, saving a significant amount of compressed air, as well as energy and CO₂ emissions. Without that insight, the machine might eventually have failed—or at least driven energy costs unnecessarily high.

Excellent. And that’s what you do with your moneo IIoT Insights, right? That’s the product group for this kind of application. You help generate these insights, or the customer can do it themselves depending on the project. So the platform basically serves as the enabler for dashboards, quality metrics, and other evaluations.

Christoph

Exactly. moneo not only prepares the data but also enables further processing. With moneo Insights, we can run various AI tools that perform calculations, detect anomalies, and similar tasks directly on site. But the typical first step is always transparency. Once the data are visible, the next step is automated anomaly detection—to make processes even more targeted and efficient.

[25:59] Transferability, scaling and next steps – Here’s how you can use this use case

Guido, what’s currently in the pipeline for you?

Guido

Right now, we’re planning to expand the project to additional machines. As Christoph mentioned, we also want to integrate more sensors—for example, to measure humidity in the dryer, ambient temperature in the hall, or data from the air conditioning system. The goal is to record all parameters in moneo that could potentially affect the quality of the injection-molded parts.
When you look at the resulting characteristic curves, you can quickly see which factors have the greatest impact—whether it’s hall temperature, humidity, or even an open hall door causing fluctuations. There are many parameters that need to be monitored to ensure consistently high-quality production over time.

Excellent. And Christoph, how do you view the project from your perspective? Are there any new product features or developments you’re currently working on? Can you give us a little peek behind the curtain at what’s coming next?

Christoph

The project has started off really well. As Guido mentioned, we’ll continue to add more elements— that’s exactly what’s on the roadmap for upcoming projects. From our side, we’re planning to integrate a new tool into moneo called Asset Health. It allows us to derive direct action recommendations from existing measurement data. So instead of having to interpret on their own that a high pressure reading might indicate a clogged filter, the software will automatically notify the user: “Filter clogged – please replace.” Or it might flag that “Motor vibration levels are too high – please check the bearings.” These kinds of features are extremely helpful for understanding data and turning insights into immediate actions.
Another focus will be Remote Connect. Through the moneo Cloud, we’ll soon be able to remotely access machines and interact with all systems connected to the cloud, such as machine controllers. We’re also developing a text- and voice-based query feature, essentially a kind of chatbot. Users will be able to ask, for example: “What was the hydraulic oil temperature over the past 24 hours?” The system will then respond with average and maximum values, and if desired, display a trend curve for the past 14 days. These features are planned for integration into moneo next year.

Excellent. Thank you both for the glimpse into the future. Maybe we’ll catch up again in a year or two for an update on how things have evolved. First of all, thanks to both of you—especially you, Guido—for giving us such a practical look into how RAFI approaches these kinds of projects and the real benefits they create. It was fascinating to see the production process laid out so concretely. And of course, thank you, Christoph, for your perspective as a solution provider— it became very clear today how you support customers from sensor level to IT integration.
With that, I’ll hand it over to you for a final word.

Christoph

Thanks from my side for the great conversation. I think we were able to provide a good insight into what we’ve achieved together with RAFI. I’m looking forward to follow-up discussions and diving even deeper into the topic.

Guido

Same here—thank you very much.

Thanks for having us, and have a great rest of the week! Take care.