Building spindles with computer modeling

Nov. 1, 2000
SPINDLE PERFORMANCE CAN BE PREDICTED BEFORE THE PROTOTYPE IS BUILT, WHICH HELPS WIDEN THE SPEED RANGE FOR CHATTER-FREE MILLING.

SPINDLE PERFORMANCE CAN BE PREDICTED BEFORE THE PROTOTYPE IS BUILT, WHICH HELPS WIDEN THE SPEED RANGE FOR CHATTER-FREE MILLING.

In manufacturing, there is the ideal world and then there is the real world. The ideal world is the showroom floor or test lab where machines, tools, and spindles operate exactly to specification or even better than promised. Transporting these products into the real world is often a success. But, according to Paul Ramer, President of Dyna Drive Inc., Eastlake, Ohio, end user frustration is too often the norm when built-to-specification spindles are transported from the ideal world to the shop floor.

Ramer's experience with Dyna Drive — a company working in cooperation with SKF/Russell T. Gilman, Grafton, Wis., and SKF Schweinfurt, Germany, to make spindles — has shown him that the problem begins in the spindle design process. The right questions need to be asked of end users to determine their specific application needs. Then the right diagnostic tools need to be used to make sure the spindle is customized accordingly. The idea of spindles being custom designed for a specific family of tools is a fairly new and complex concept. Dyna Drive uses computer modeling for the analysis and optimization of high speed/high power spindle designs, as well as for performance prediction, even before the prototype is built.

Traditional spindle design and manufacturing does not consider tooling or workpiece material used in the final application, says Ramer.

"Typically, the process starts with the end user saying, 'I want a 30kW spindle with a speed range at constant power from 5,000 to 24,000 rpm.' In reality what happens," explains Ramer, "is the end user will stick a tool into his spindle, expecting to get 30kW between 5,000 and 24,000 with all his tools, cutting all his materials. He then may find out that the power he can 'get into the chips' without chatter will change with different speeds and stay, in general, way below the rated 30kW. Frustrated, he calls the spindle manufacturer because he thinks it's a bad spindle. The spindle manufacturer says that, when run with a dynamometer, the spindle shows the characteristics he paid for. But the dynamometer test doesn't use a cutter for load; the dynamometer is the load. So the test sheet says the end user has the power he paid for, and the spindle manufacturer doesn't feel obligated to do anything about the 'problem.'"

In the machining process two basically different types of vibration may occur. 1) Vibrations due to forced excitation, mainly caused by imbalance, at a frequency identical to the rotational speed of the spindle. 2) Vibrations due to existing "natural" resonant frequencies of the system, which get excited by the wide frequency band energy generated by the cutting process. The spindle can be rotating at any rotational speed and produce vibrations at or near a dominant frequency (mode) of the system (tool/toolholder/spindle/machine).

There could be four or five or eight natural frequencies (modes), and, on the test bench, the spindle runs fine — balanced with less than 1 mm/sec vibrational speed. As soon as you add a tool in its toolholder and start cutting, one of several of the natural frequencies get excited, the cutting process gets instable, and chatter occurs.

This situation happens in particular for end mills with a length to diameter ratio above 4:1. According to Ramer, "It is not possible to design a spindle which gets the full rated power into the chips over the whole speed range, at least not for long cutters. There is no technology that can make that happen. That's just physics. The spindle design has to be optimized for a certain well-defined tool family."

Dyna Drive does its optimized spindle design with help from software and corresponding hardware based on scientific work performed by Prof. J. Tlusty and his research associates at the University of Florida, Gainsville. Dyna Drive starts the design process by talking with end users to determine the group of tools, the material, and the ideal cutting speeds that the spindle will be customized for. Today's analytical methods allow spindle optimization and performance prediction during the design process, prior to building a prototype. Ramer says that cutting tests with the prototype verify the analytical results, in general, pretty well. This process allows Dyna Drive to cut down design time and lengthy and costly prototype modifications.

According to Ramer, this is different than how most manufacturers do things.

"Most spindle manufacturers build spindles based on designer experience. This experience is still very important, but since the industry has gone to higher speeds, higher power, and longer tools, the experience of the designer isn't good enough anymore."

Throughout the design process, Dyna Drive works to achieve ideal spindle performance running different tools in particular speed ranges. These ranges of chatter-free milling are referred to as sweet spots, and Dyna Drive strives to widen those sweet spots as much as possible.

"Normally you have a narrow sweet spot. You stay in that narrow range or you're cutting with chatter. After our modifi-cation, you can stay in a comparatively wide speed range and get the full power into the chips," says Ramer.

Once the prototype spindle is created, Dyna Drive suspends it on a thin cord and taps it with a hammer for testing. This provides the energy to the spindle (like cutting will do in the application) and then the frequencies can be re-measured and re-calculated with the software to predict an even more accurate milling performance.

Dyna Drive can develop a mathematical model for each kind of cutter being run on the spindle. The company mostly works with aerospace customers. According to Ramer, aerospace aluminum milling is all about stock removal rate. Most aerospace manufacturers want maximum kW on the upper end of the speed range between 20 to 24,000 rpm or even higher.

"We make the model and optimize the spindle so that it provides the end user with full power on the upper end for a particular family of tools. When faced with 6 tools, we might say, 'For 4 tools we can give you full power in that range, but with the other two tools you'll only get 80%.' Then when they come and say those two tools are not important anyway, it's ok. But if they say that those are the two most critical tools, we modify the spindle to sacrifice power in another tool. All this is done before actually building the spindle. The reality is, you can't put any tool in any spindle and run it at full rpm and power. You can't cheat physics. No one can design that spindle. But now, with this method, we can design the best spindle possible for a given set of tools."

Latest from CAD and CAM