Modern CNCs Make Easy Work Out of Tough Stuff
Manufacturing Engineering Magazine July 2018
By Ed Sinkora

CNCs are getting faster, smarter and easier to use nowadays. CNC stands for “Computer Numerical Control” which could help us when we’re struggling with high-speed machining, need better surface finishes or higher accuracy, have training and retention problems, or want a better handle on your production efficiency.

Who Needs the Speed?

Like everything computer related, CNCs are constantly getting faster. To take a few examples from the dizzying array of specs you’ll see, Heidenhain controls process data blocks within half a millisecond, the Siemens SINUMERIK 840D sl can control up to 31 axes and 10 machining channels with just a single processor, and the pulse count is up to 32 million per revolution in FANUC’s latest drives and motors. Who needs this kind of power?

First, anyone trying to move multiple axes simultaneously, especially if they need to do so rapidly. A senior application engineer in USA pointed specifically to the increased use of composites in aerospace where “faster high-speed five-axis routers are commonly requiring more processing power. We also see a need for high-speed machining of aluminum in five axes on high-performance machining centers.”

High-speed machining requires fast feed rates and constant chip loads. You want to transfer the heat into the chip and not into the tool. But if the control is too slow to manage the data coming from your CAD/CAM system or post-processor, you won’t be able to guarantee a constant chip load, which burns out the tool very quickly. You also don’t get the same surface finish or accuracy. Faster controls are also needed to run spindles as high as 80 – 100,000 rpm, again to maintain constant chip load. The CNC is a big component of success in these areas.

Machines that combine machining methods, like mill-turn machines or machines that both mill and grind are also driving the need for ultrafast processors. For example, if the part is not centered on a milling machine’s table and you want to do a turning operation, you would have movement in all axes, not just the rotational axes.

Modern CNCs include a function generally called “adaptive feed control”, which uses measured spindle load to adjust the cutting speed. For example, if you’re cutting “air” the control automatically speeds the feed to the maximum you’ve set. When the tool is buried in the cut the control slows the feed to maintain constant, safe torque. All on the fly. The feature is particularly good for “unattended machining and trochoidal milling. It even further optimizes CAD/CAM packages that adjust machining speeds based on predicted material removal volume.”

Higher Accuracy, Better Finishes

Advanced controls offer much more than the ability to handle kinematic complexity and multiple processes, or the ability to machine faster. Perhaps most importantly, they offer higher machining accuracy and improved surface finishes for mold and die and other applications where these factors are critical. One common approach is dual feedback that uses both the digital motor’s encoder and a separate absolute linear encoder to inform the control.

The mention of “algorithms” brings us back to the common perception that modern CNCs are a black box. Indeed, there are often sophisticated programs running in the background that go well beyond simply executing the moves defined by your machining program.

Forcing a machine to go from point to point along such a contour can create vibration and gaging marks on your workpiece. On the Heidenhain control, you can establish a bandwidth of a few microns and allow the machine to move smoothly within this tolerance band around the programmed contour. The control also optimizes the speed and feed depending on the shape of the contour, while maintaining the accuracy.

There are also a features that correct for inherent machine errors, like Heidenhain’s Adaptive Chatter Control (ACC). We work with the machine tool builder to determine how vibrations can occur in their machines, as this differs from machine to machine and we determine what action the control should take to eliminate such harmonics when they occur, which again differs from machine to machine. The machine tool builder puts this data in the PLC, which feeds it back to the control and it runs automatically.

Heidenhain, FANUC, and others also have functions that compensate for acceleration-dependent position errors at the tool center point. The errors depend in part on the stiffness of the guideways, the distance between the feed force application point and the center of the mass, as well as the distance between the center of mass and the tool center point. So, Heidenhain partners with the machine builder to understand the mechanics of the machine and makes calculations based on that understanding to correct for these acceleration errors.

Advances in HMIs

As you might expect, recent advancements in the human/machine interface (HMI) include icon-driven designs and the kind of fingertip control popularized by smartphones (pith to shrink an image, spread to zoom, etc.). A home screen that doesn’t look anything like a FANUC screen, with rows of different icons for tool data, editing your program, maintenance, etc. You can also customize it, for example adding a programming app from another vendor. Most importantly, this makes running out control much more accepted by millennials. CNC vendors have managed to not only give the machine operator more programming capability independent of off-line CAD/CA, they’ve made it relatively easy.

Other Setup Aids

Besides intuitive interfaces, today’s CNCs also help the operator with guidance for each function. Heidenhain has a new system called VSC (Visual Setup Control) that takes a picture of your setup after you’ve optimized the orientation of the part in the workholding, the tightening of the bolts, and so forth. The control then compares that image to subsequent setups in the same production run and alerts the operator to any anomalies, such as a wrench being left on the part or a missing hole, giving him the option of proceeding, switching to the next pallet, or stopping for corrective action. VSC is so sensitive that the camera even recognizes a bolt head that’s not perfectly flat due to being over torqued.

Although all controls are tuned at the factory based on the expected workholding, part size, and so forth, changers in the field (intentional or otherwise) sometimes require adjustments. Automatic tuning of the axes using Fagor Finetune software is now a standard feature. This software tunes the parameters of the CNC and drives using Bode diagrams to optimize the high-speed cutting features. Fagot and others also provide a built-in oscilloscope to the CNC for further advanced tuning.

Whose Control Is It Anyway?

Before leaving the subject of HMIs it’s worth noting that the interface actually presented to the operator is often partly, if not entirely, the creation of the machine tool builder and not the CNC vendor. By the same token, you often have several control options from the same builder, so it pays to educate yourself on what’s out there.

Harnessing the Power of Data

Industry 4.0 and the Industrial Internet of Things are arguably the most important driving forces in today’s industry. Information is the key to making smarter decisions and today’s CNC machines are expected to publish productivity and production-relevant information and to network with other intelligenct devices to reduce setup time and costs.

As part of this effort, Hurco freely provides an open-source interface to its control on GitHub and partners with a number of robotics and productivity monitoring companies to broaden the interconnectivity capabilities of its control.

While the Heidenhain control can tell you virtually everything that happens in great detail makes it easy to get a relatively constrained set of data on things like spindle on, spindle off, and error codes that provide “very clear and decisive analytics without all the fuss and bother”. It is also allows the operator to provide input on why the machine was not running, which can be critical data point a fully automated system would not catch.

For example, a sensor detects an abnormal vibration coming from the Y axis of a certain type of machine tool, it can determine (based on algorithms derived from large amounts of smart data gathered from other users) that the vibration is due to a bearing wearing out on the Y-axis ballscrew. That part can then be ordered and shipped to the customer before they are even aware of the problem. This goes beyond preventative maintenance to predictive maintenance. A major step forward.