Viewpoints: Let's Set a Standard for Volumetric Accuracy
In all machine tools, mass is driven along the way system.The mass needs to move straight, without pitch, yaw, and roll, because any variation of these three movements is amplified as travel goes further and further along the machine tool's way system to the outer edges of the work envelope.Volumetric accuracy is the sum of positioning, straightness, wobble—all of those characteristics that influence how accurately the tool tip moves through the travels of the machine.Without high volumetric accuracy, the machine tool will produce unacceptable parts. Typically, the machine's dynamic volumetric capability should be at least 50–80% tighter than the part tolerance. In high-precision machining, most users need an 80% capability.
I believe that volumetric accuracy should have a number associated with it, so buyers can compare machines. And that number should be a value—for example 20 µm, 12 µm, 50 µm—that reflects the machine's true positioning accuracy. It should be based upon a world standard, not simply a standard established by machine tool builders. I don't care who comes up with the standard, but it ought to boil down to a number, so that people can say: "I need to buy a machine tool with a volumetric accuracy of XX," as in the old days they could say, "I need a machine that can repeat to 0.001", or position to 0.001"."
A complete definition of volumetric accuracy requires measuring many elements, and is difficult to achieve. Some systems use four-body diagonal-displacement errors to define the volumetric positioning error, but this approach isn't 100% accurate. Today, standards committees are working on a definition of volumetric positioning accuracy. This discussion has been going on for far too long, and it needs to be resolved.
As a manufacturer of what are arguably the most precise machine tools in the world, such as "mother" jig-boring machines and thread grinders, along with custom CNC machining centers, our company has been able to define volumetric accuracy by taking 20 geometric measurements on our machines. That procedure reflects our culture, and meets our customers' requirements, but most other builders are not geared to perform such tests. The standard needs to be less complicated, more understandable, and something that everyone accepts.
We can simplify the concept by establishing a specification number by which to compare equipment. For example, the difference between machines A, B, and C could be a value such as 1, 2, 3, 4, or 5, with 1 having the greatest volumetric accuracy (true position in the cube), and 5 having the least. To make it more palatable, those numbers could represent a range. For instance a 1, which would be in the Mitsui Seiki realm, could be from 5 to 15 µm.
The next logical question, however, is: "What is the methodology for proving that range?" One way to do it is to measure every micron position in the cube. I don't think a builder needs to do that. It really comes down to straightness of travel. In a HMC with 900 mm of travel, if straightness of travel is under 2 arc-sec in X and you square up the three axes, positioning accuracy is at 12 µm. If it's 2.5 arc-sec, it's 15 µm. If straightness is 3 arc-sec, positioning accuracy is 20 µm. That has been our company's experience.
An autocollimator is used to analyze pitch and yaw. Movement of the instrument's cross hairs up and down indicates pitch error, whereas movement left or right indicates yaw error in the way surface. Roll error is measured with a precise spirit level. Typically, the autocollimeter is moved along the way surface in 100-mm increments, and pitch and yaw error are measured at each interval. Then the spirit level is moved along the way system to measure roll error at the same intervals.
Anyone can argue that there are other variables. There's table wobble, thermal distortion, bend, spindle runout, and a host of other issues that affect this value. That's why the issue has been impossible to resolve. I say, simplify it. Express volumetric accuracy as a measurement of travel straightness, and be done with it. As long as everyone knows the other factors that can affect accuracy, this approach is still a fair way for a user to convey the precision of a machine tool. It's much better than the way we describe accuracy now.
This article was first published in the July 2007 edition of Manufacturing Engineering magazine.
Published Date : 7/1/2007