Dr. Olsen's Lab

Many people assume that the “accuracy” of an abrasive waterjet (AWJ) cutting system can be determined by simply reviewing the manufacturer’s specification sheet.  That assumption then leads them to the conclusion that the machine with the most precise table specifications must also be the one that produces the most accurate part.  Unfortunately neither the assumption nor the conclusion is true, and the individual who buys an AWJ system based on table accuracy specifications may be very disappointed when he measures the parts being produced.  It is not that AWJ system suppliers are intentionally trying to mislead potential buyers, but rather that simple X-Y table specs do not tell the complete story behind the accuracy of a final AWJ-produced part.

In the first place, the listed X-Y accuracy of a cutting table is very often based on static measurements of X and Y displacement—the actual change in location of X or Y carriage along the X or Y axis compared to a change directed by the machine controller.  This may be an actual comparison made using a precise measuring device such as a laser to determine the true distance moved, or it may merely be the stated linear accuracy of the ball screw or other linear motion device used to move the cutting nozzle.  In either case, this approach fails to take into account any variation in location of the nozzle due to squareness, twist or straightness errors in the machine.  Other measurements may track the actual nozzle position vs. the commanded position for point to point moves. Such measurements of accuracy based on static conditions really do not give a good indication of the accuracy of the system while it is actually cutting and thus do not really predict the accuracy of the part being produced.  Additional errors occur when the machine is moving due to servo following errors, vibration and other motion related sources. A far better way to evaluate X-Y accuracy is to measure it using a dynamic test, such as the Renishaw Ball Bar ™.  This approach electronically measures the deviation of actual nozzle motion from that of a perfect programmed circle while the nozzle is moving at a rapid speed typical of an actual cut.  By performing such tests in a series of locations around the cutting table and system manufacturer (and the potential purchaser) can get a much more realistic evaluation of the true accuracy of the X-Y motion system in actual operation.

However it must be realized that even an accurate specification of X-Y positioning accuracy while in motion will not fully predict the accuracy of a part produced on an AWJ system.  The reason for this lies in the nature of the cutting jet itself.  Simply put, it is not a rigid inflexible cutting tool.  The jet bends as it moves.  The lower part of the jet does not behave exactly like the upper part.  The width of the jet changes and deforms.  The jet tapers as it cuts.  And all of these things change with different materials, different thicknesses, different nozzle parameters, different cutting speeds, different rates of acceleration and a host of other factors.  Thus the accuracy of the motion system only sets the lower limit on the error seen in a part.  By moving very slowly everywhere along the part path a user can approach this accuracy. However, no user wants to cut that slowly.

When cutting quickly, the speed must be accurately controlled to manage the jet deflections so as to minimize part error.  This task is very difficult to do with manual programming. It is the control software that sets speeds, accelerations, rates of change of acceleration, actual cutting path, taper compensation angles and other variables based on its own embedded algorithms and cutting strategies.  The most precise cutting table in the world will not make an accurate and repeatable part quickly without the right software.

We at OMAX are proud of the design of our cutting tables and their inherent precision and accuracy of motion.  However, it is our control software that allows them to make accurate and repeatable parts.  That is why we tell potential OMAX users not to judge our system (or any other AWJ system, for that matter) on the specifications of its X-Y accuracy.  Make some actual test parts and measure them and compare.  The measured accuracy of actual parts produced on an AWJ system is the one and only true measure of accuracy.