We want to control the machine

    Interview with Reiner Schmid, Tebis AG

    This is a principle that Tebis has followed for more than a decade, and Reiner Schmid, Tebis AG’s head of Product Management, still views it as a unique approach. In an interview with Christopher Detke from NCFertigung, he explains what makes a good CAM system, where the challenges lie and what machines he currently regards as the most complex.

    Hello, Mr. Schmid. 14 years after Tebis introduced its virtual machine shop, programming of lathes with simultaneous machining will now be supported based on the same principle, starting this summer. Why has this taken so long?

    This timeframe shouldn't be misunderstood. We have gradually expanded the spectrum – for milling machines, laser machines, robots and turning/milling machines. We have now reached the point where we can support what I consider to be the most complex machines, while still maintaining the principles of our CAM system. However, the principle of our CAM system has not changed.

    You mentioned several different machine types. In your opinion, which of these is the most complex?

    I think this would be multi-channel lathes or turning/milling machines with simultaneous machining. The market trend is clearly moving toward these types. Many users have recognized the benefits of multi-channel or simultaneous machining technology.

    What does multi-channel technology mean?

    This term comes from the field of controls. Controls manage the components moving through channels. Every channel is essentially a tool holder. Channel 1 can be a melding head and channel 2 a tool turret with a turning tool. But the part, in other words the workpiece holder, can also be a channel, or the secondary spindle that takes over the part at some point.

    What are the benefits of this?

    It reduces logistical efforts significantly. You can perform all machining steps on a single machine, which has substantial advantages in terms of precision. Of course, this has substantial advantages in terms of precision. It also eliminates the need for additional equipment such as fixtures or clamping devices. This makes machining extremely effective, especially in terms of planning and control.

    So, lathes with simultaneous machining are not exactly a novelty. Why are they still the current trend?

    This has taken some time in die and mold manufacturing, where we have many customers. The benefits were recognized earlier in the general machine manufacturing industry. But the need for lathes with simultaneous machining has been very clear for some time. Considering this, we have put significant effort into this area.

    Who are your primary customers for your turning/milling machine modules?

    It is primarily manufacturers of small series and single parts, rather than those in serial production. Although we can still provide support for them as well, the goal of series manufacturers is to reduce time. This is not possible with a CAM system. We must be honest about that. A CAM system or developer claiming the ability to determine the exact times will actually always be off.


    It has to do with the precision of the simulation. For example, the cycle time of the turret, positioning movements and measurement of the tools. I simply cannot simulate these times precisely enough. But this is what the series manufacturer wants, because for them, every second counts. We can program the sequence for them, but we cannot provide a simulation down to the second.

    So, I can't explicitly determine machine run time in advance using Tebis?

    Yes, you can. We have addressed the machine run time despite the challenges this faces. Because of course it is very important to know exactly how the machining operation will proceed and how long it will take in order to optimize the internal processes in the machine. We can already do this relatively well, based on the feedback from our customers.

    Relatively well?

    We can achieve an accuracy of 90 to 95 percent, which we consider to be very high, considering the fact that we are offline.

    How do you achieve this precise calculation?

    We achieve this because we not only account for normal and rapid feed rates in the calculation, but also acceleration and jerk.

    But in this case, you must have the most accurate possible data for every user and for every machine. Who stores this data?

    We call this a hybrid situation. The user can enter jerk data for the machine. But in practice, the user often will not know these values. Our service technicians can provide support in this area. They have a great deal of experience in setting up different machine types. At the same time, our algorithm converts the entered data which brings us even closer to the actual machine behavior. In other words, this is a complex topic.

    Normal feed rate, rapid feed rate, acceleration and jerk data. That’s enough information for 95 percent accuracy?

    No, which is why we also account for all traversing movements. We don't just simulate the calculated toolpath. We have taken an approach with our simulator technology that I believe is unique.

    What is unique about it?

    We handle every position of the machine. For example, once the machining operation is complete, it is time to move to the tool change position. This position can be approached more or less intelligently. We also calculate the path.

    So, Tebis finds the fastest direct route, given the available space?

    Yes, because the problem with the machine is that it doesn't know where the part is and where there is still space in the operating area. This depends on the clamping situation. We want to tell the control specifically and precisely what it must do, and then the machine does that as well. If you generate the NC code simulation starting from the control, you can only simulate what the control specifies. We approach this from the other side. We address the machine with very simple commands so it does exactly what we want.

    Let me summarize here. You basically program from the virtual machine, while the usual procedure is to output the NC code from the control logic?

    That's exactly what we want to do. We call this "programming with the virtual machine." We deliberately keep the machine under control. This results in technically correct and collision-free data right from the start. It's supported by sophisticated technology, but we work with very simple NC commands. This is sometimes a different approach for the customer, because the program will contain one or more additional NCSets.

    So, your target group is primarily turning/milling machines?

    Yes, although other applications are also conceivable, such as the use of robots. A robot could also be integrated as a separate channel. This would then take a part from the machine at the perfect time for deburring after roughing and finishing which frees up the machine for high-quality operations.

    What is the exact process in Tebis?

    We always program with the machine as the starting point. This automatically fulfills all the necessary prerequisites. Units such as jaw chucks or steady rests are also directly configured. The user immediately has an overview of what they can do. They set up the machine, clamp the part and specify the manufacturing steps. All the programs are integrated in the Job Manager and then synchronized. That's the big difference. We make our calculations using this information, in other words, with the specific prerequisites of the machine and the situation in the background.

    So, synchronization is the last step?

    Yes, because we have seen that the user must repeatedly switch machines if the intended machine is unexpectedly blocked. I can quickly make this switch with our Job Manager. I move turning operations to the lathe and milling to the appropriate milling machine. If we were to synchronize and calculate immediately, this procedure would be significantly more complex. We already generate a synchronization list in the background while building the Job Manager. This is followed by calculation and output of the NC programs. The users are actually finished once the Job Manager is created, and they don't have to worry about anything else. If further optimization is required, they can take a closer look at the synchronization marks.

    That sounds very simple.

    What's interesting at this point is that we have a complete handle on everything else that is going on. There are many small and large steps between the processes. I would say that more happens in toolpaths and movements around roughing than in the actual machining operation. We check, simulate and postprocess all movements and are therefore reliable.

    So, you account for the entire situation, all of the limitations, reductions and the traversing paths you mentioned…

    Exactly. We also like to refer to this as the intelligent toolpath. By using this, we create a much more reliable process, while other systems first need an NC code simulation and are later almost "desperate" to simulate the control process. We tell the control what to do.

    What makes this so complex for turning/milling machines?

    The actual machining operation – turning or milling – is not the problem. The real art lies in activating the right element at the right time. The interaction of the different channels is far more complex. For example, when is the steady rest positioned, when will the secondary spindle move over, when does the bar loader move in – all must be done without collisions. This has little to do with the actual toolpath, but it is difficult to synchronize and coordinate. It is thus indirectly old technology, but with entirely new challenges.

    And the machines are becoming increasingly flexible.

    This can't be overlooked: The machine manufacturers integrate a large number of options. It is not difficult to say, "Today I'll clamp the tip in turret space 11 and tomorrow in the secondary spindle and hold there." The user has this flexibility and wants to fully exploit it. The difficulty for the CAM system is to maintain control of this flexible implementation. I can do this simply by generating an NC output and telling the machine, "Now the tip is in the turret." But then I'm not controlling it, and what I do depends on what the control and the machine do. We want to program this so reliably that the optimal NC output results for all process steps, positioning every channel correctly at all times.

    The user's flexibility will be even greater when more dynamic tools are implemented in the future. The tool manufacturers are becoming increasingly innovative in this area.

    Yes, we have often experienced this. However, the prerequisites of the machine, tool and part spectrum have to fit together. We regularly ask our customers if they want new tools or production processes, and we adapt accordingly. We are always ready to develop dynamic processes. But we are currently focusing on virtual machine technology.

    So, the high-end applications don't always make good in manufacturing facilities?

    These figures are usually manageable. We know this because we are in constant contact with our customers, and I see a continuing need for this in the future. More intensive interaction is needed between all participants: the users, the tool and machine manufacturers and the CAM manufacturers. It's always a bundle that somehow has to be brought together. The potential is there. But everybody is waiting carefully for the others to take action.

    Is there too little willingness to cooperate?

    I think it's more about avoiding risk. The companies wait, because there are only a few examples of applications at first. Nobody switches until a process has been established. Take the example of robotic milling: We have supported this since 2014, but we are only now noticing a gradual increase in demand, such as in model manufacturing.

    Is this especially the case in Germany?

    I believe it is a global phenomenon. Usually it takes some time until a new approach becomes routine. It also has something to do with the structures. We have customers who work with only one tool manufacturer. If they don't have a specific tool, the customer has to change suppliers. These aren't technical problems but rather organizational ones. Innovation is definitely important, but we can see that implementation times are often one to two years. It is therefore important for our customers to have a reliable long-term partner with a high level of technological competence as well as reliable service to accompany them on this path.

    Reiner Schmid, Head of Product Management, Tebis AG