Reading time ( words)
Andy Shaughnessy speaks with Roland Valentini, Gardien Group COO, about the company’s new OnTrack2 ERP and process management software. He explains how OnTrack2 is purpose-built for customers around the world and why it’s much more than a replacement for OnTrack1, which was being stretched beyond its original intent.
Gardien’s North American Group recently began using the OnTrack2 system. OnTrack is an internal tool, and the vision is to use it to improve performance, be transparent, and overall, give the customer—the PCB manufacturer—the best quality of service.
Andy Shaughnessy: I'm speaking today with Gardien Group COO Roland Valentini. How are you doing, Roland?
Roland Valentini: I’m doing all right. How about yourself?
Shaughnessy: Very good. Before we start, why don’t you just give everyone a kind of a brief background about Gardien.
Valentini: It’s my pleasure. Gardien, as you know, is a global testing and inspection service company. We operate about 18 facilities in five countries. Simply put, we provide a range of electrical test services like the standard short test, 4-wire Kelvin test, hi-pot, micro-short detection, and these kinds of typical services, as well as optical inspection services, which range from AOI and AVI to final inspection. And we provide all these services either as an in-house solution to the PCB fabricator in their factory or in our own offices.
Shaughnessy: All right. We recently saw you released OnTrack 2 ERP software, following on the heels of OnTrack 1. Tell us about OnTrack 2.
Valentini: All right. OnTrack is our in-house purpose-written process management and administration tool. ERP doesn’t really cover it fully. The tool helps us provide the services in an efficient, transparent, and traceable way. With the original OnTrack 1, we achieved several of the goals that we had. We made available the expertise and the know-how of the individual people to the entire organization. We built the best practice processes, so we looked at what we’ve learned in different countries, interacting with the different customers and different technology levels. Then, we put that all together and said, “Okay, that is the way we want to operate in the process,” and OnTrack reflects that by having such a tool that minimizes the chance of an employee making a mistake and deviating from the standards that we’ve set as the best practice.
Why is that important? We at Gardien very much focus on the process approach of quality control and not so much on the machine itself. The machine does not guarantee the integrity of the results. It’s the process of data handling, product handling, understanding and utilizing the different machines’ capabilities, and proper documentation and resilience that makes a stable process. OnTrack helps us to do that. It handles all the incoming data from the PCB manufacturers and together with the instruction information. We have that in our engineering department, and they will then create suitable data for all of our machines that can potentially run the job. Like if it’s a fixture test job and a certain size, it will then provide the necessary information for all the machines that could run it.
OnTrack will push the created information to the individual location. The operator, when the job arrives, doesn’t have to search for it. They open it up and say, “This is the job I want to test,” for instance, and then it shows them to the location. They don’t have to look for the right machine and the right fixture. It’s there, so it eliminates the chance of making a mistake by picking something wrong.
Then, the process screen guides the operator through the individual process steps, including how to set up, what to set up, and how to run the individual task on the machine, reports the results, and have some counter boards.
There’s also some integrity checking in there. If you’re counting the boards, the total number is 50, and you have 49 good and two bad, the number is not correct. It will help the people avoiding those kinds of mistakes, and all of these process steps are barcoded. We always know who, when, where, and what, and it gives us the ability to have traceability, trackability, and transparency. We can always look into what happened, where it happened, and so forth.
OnTrack provides stuff that’s within required quality documentation. The people on the floor don’t have to mess with it or write the shipping and commercial documentation, and you would be surprised how many different flavors are out there in terms of what people want in terms of documentation and how they should look. Japan traditionally has a different view on it than America or China. You don’t have to do that individually.
Shaughnessy: I saw on the release it said that each country’s requirements were fed into the new system. That’s pretty interesting.
Valentini: With the original OnTrack 1, a lot of these things were hardcoded. Whenever you had the chance that it changed, then you were running the risk of breaking your legs somewhere. You’re making an adjustment here to make a customer happy, and all of a sudden, something else didn’t work anymore, and OnTrack 2 is more modular. That makes it a lot simpler to adjust that and add a customer. If they have a different idea of what they want to see, then it’s easier to create the necessary information for them, and then they just see it.
These customers have access to the Gardien floor. A customer can log in, and then we would create a temporary database, and they could see everything on this product that is on our floor. You would see if it’s incoming, what machine it’s on, and how many have already been processed. You can access this historical and statistical information. This way, they don’t always have to call. They can look at the data themselves. As I said, transparency is important to us. This way, it’s completely open.
Shaughnessy: That’s cool. That seems to be what it’s about lately is, I mean, we have all this data floating around, but it’s being able to get it to the people that need it in a format where they can access it. It seems like a lot of what this is doing.
Valentini: Yes, it would predominantly be driven internally because we were dealing with the customers in a different way, and we wanted to make sure that customers small and big also get the same quality of response. Whether you are one customer that has one board to test, or you have a million boards to test, you should get the same documentation and access, so we made it rather simple.
Shaughnessy: And from what I understand, OnTrack 1 was really good, but people were stretching it like far beyond what it was supposed to do. Is this purpose-built from the ground up?
Valentini: Yes. That’s why we decided that we can’t just keep on bolting on new functions to it. It would be much better just to go back and recreate it from scratch, using a different database structure to make it a little bit more resilient, and most importantly, a lot more user friendly. If you have the original software, which is 10–12 years old, the programming has changed. The expectation of what people expect to see in the GUI has changed, and it makes life a lot easier if the handling of such a system is intuitive rather than you have to spend a lot of training on it.
We’re doing it now with tabs instead of different windows. We are following more the standard web browser kind of feeling, so that makes it relatively easy to manage that for a new employee or operator. And we overall built it a lot more modular, so we can make changes more easily. We can build, and we can put the modules together in a slightly different way so that we can reflect on what in the individual countries or individual customers, what they want to have, and how they want to have it without always writing something completely new code.
The OnTrack 2, which was also a part of the process, is a lot more secure. The communication between the computers is all encrypted, and barcoding is now encrypted. You can’t just copy your barcode and use it. This way, we make sure that everything is handled in a secure way, and we do have several data repositories. We do not depend on one location—where if something goes down, we are strapped—so we now have individual sites running completely independent. But we then have collection centers where the data is uploaded, and it’s secure. In case something happens, like a catastrophic event, we’re still able to provide customers with the information, historical information, and data.
The other thing that we have done, which is easier now with the new structure, is we have built in a direct connection and connectivity to equipment. Whilst we have OnTrack terminals on every floor, and we usually have several of them, we also thought it might make life a lot easier if the machine can directly communicate with the system. And to go a step further, we’re still building our own flying probes. We have our own flying probe software. We built a direct interface on that software. When the operator is working on the machine, they don’t have to redo the task on the terminal. It automatically gets all uploaded into the system.
Because a lot of the information is redundant—like entering your name, the order number, the number of boards, and so forth—why should you put it into the machine to run and then putting it into a terminal? This way, we should get more data points, and the quality of the data points should be significantly better than before. And what we hope to get from this is to start learning and getting statistics to see where we can improve, and where we may be wasting time. Is the setup process too complicated? Is the machine idling? We want to look into all these things that make up a successful floor and see how we can make it better.
Shaughnessy: Right. And I think it’s pretty handy that your company is using OnTrack 2. How’s that going?
Valentini: We started with OnTrack in North America with the OnTrack 2 rollout, and that went rather well. The reason why we picked North America is that the programming team also sits in North America. They speak the same language. They’re in the same time zone, and, inevitably, if you put something out there, it’s never going to work entirely as projected. You can’t really test everything to the last bit. That worked well. We had a couple of hiccups that we were able to fix. Also, the number of data points that we have in North America is not quite as high as in other locations. North America makes about 10% of the data that we are creating. For 2019, we had in total about 210,000 orders that we were running and about 300 million images.
We have a lot less data in NA. It makes it easier to upload, make changes, and modify. The second step was we brought it into China, where we’re doing significantly more in terms of volume. Immediately, whenever something was a hiccup, it became a bigger problem because, all of a sudden, we had a lot more data to deal with. Certainly, it was the right approach starting in North America, but we’re now in China, and we are planning on getting Taiwan and Japan ready before year-end. We have the group covered, and then we’re starting with the reporting processes and KPIs that we use internally.
Shaughnessy: Good. Apparently, the GUI is a lot more user-friendly, and it sounds like you are on your way with this.
Valentini: Thank you. The other thing that we are planning on doing once we have every group company running is scheduling. OnTrack 1 already had a scheduling function, so wherever possible, we are connecting our ERP system with the ERP system of the PCB manufacturer, that helps us to estimate the amount of work, the volume arriving at a certain point in time, and we compare this to the capacity that we have available. The system would then tell the supervisor there’s a problem. You have lot more boards coming than you can handle, or you have more boards scheduled for a certain machine than machine time is available, and then the people can proactively try to manage the situation instead of just waiting for the boards to arrive. They can look at it and say, “Oh, what I’m going to do now?”
Now, in OnTrack 2, we want to be a little bit more sophisticated. What we realized over time is that there are different reasons for how to schedule, why to schedule, what to optimize for, and whether you optimize for time or costs or whether there are certain boards that have a priority. You need to come up with different ways, and currently, OnTrack 1 could only do it in one way. OnTrack 2 will then have the abilities where people can pick up and say, “I have a customer that is purely focused on this. This is their prototype work that needs to go back ASAP. This particular product is purely measured on time, and we’ll do whatever we can to get it back to them.” And this other product where they say, “It’s volume. It doesn’t really matter. I’m a little bit more flexible.” You can use a different approach, and this way makes it easier for both sides to cooperate with each other and get the perfect results out.
Shaughnessy: Wow. That’s pretty good. What’s next? Are you still tweaking OnTrack 2 as it goes here?
Valentini: Yes. The plan was we have copied all the functionalities that we have in OnTrack 1, and we’re rolling it out right now. Once it’s running everywhere sufficiently stable and smoothly, that’s where we want to start building additional functions into it. As I said, we want to improve scheduling, and we want to have direct machine communication, but we also want to add other processes. We want to make a couple of things simpler to handle for us as a company internally, but we also want to look into statistical approaches. Where do we spend time? How do we spend time? Where do we spend costs to try to optimize what we can do here to cater to the individual customer requirements?
These are the future projects I would assume that’s going to go on for a couple of years, but since we have our own programming team, we will keep that as a continuous project. It’s not going to ever end. It’s always going to be something where we have new ideas and a new set of customers, where we are looking into new processes, can maybe add to our portfolio, and then have the ability to just bolt that on to the existing system.
Shaughnessy: That sounds really good. Is there anything else you’d like to add? Anything that we haven’t mentioned?
Valentini: No. I think we have covered quite a lot of what we have in mind. It’s a difficult time for everyone. The overall market is slow. We see it picking up in a couple of areas. We see some changes in technology, and we believe that with the growth of the automotive industry into PCB, that’s going to be some significant push towards traceability and transparency to make sure that you are in total control of your shop floor. One of the other projects we’re focusing on is serialization to be able to pick it up and enter it into the records. When we show what’s going on, we are able to identify on an individual port level rather than just on a batch level.
Shaughnessy: All right. Best of luck with this. Maybe we’ll see you at one of the trade shows. Apparently, the trade shows are still officially going on next year. There’s still IPC APEX EXPO and DesignCon. They’re trying to have in-person shows next year.
Valentini: I hope it’s going to happen. I mean, it’s very unusual for me that it has been 9–10 months into the year, and over the last nine months, I haven’t traveled. Having a global organization, you usually go out every three or four weeks and meet people, whether it’s customers or your own people, and it makes it easier to exchange ideas. I’m looking forward to these abilities.
Shaughnessy: I’m just looking forward to next year. It has to be better than this year (laughs).
Valentini: That’s going to be relatively easy (laughs).