Origins of chatter: General Industry Coverage

Dear Doc: I cylindrical grind hardened-steel shafts and sometimes get chatter. There’s a debate on the shop floor, with some saying the wheel is dull and the material-removal rate is too high, and others claiming the wheel is not true. Is there a way to tell? Our 16 “-dia. wheel runs at 12,000 sfm and the workpiece at 40 rpm. I have counted about 70 waves of chatter on the workpiece.

The Doc Replies: There are two types of chatter. The first is forced chatter, where the chatter comes from some specific frequency, such as an imbalanced wheel, an out-of-true wheel or a bad bearing. The second is self-excited chatter, where “things just start bouncing around” at the natural frequency of the system. This is exacerbated by a dull wheel or a high mrr.

Here’s a quick way to determine the type of chatter you’re experiencing. Divide wheel rpm by workpiece rpm. In your case, the wheel is running at 2,866 rpm [(12,000 sfm × 12 in./ft.) ÷ (π × 16 in./wheel rev.)]. Therefore, the wheel-revolutions-to-workpiece-rotation ratio (2,866 ÷ 40) is 71.65. That’s awfully close to your count of 70 chatter marks per revolution, indicating an out-of-true wheel.

But be careful! It might just be self-excited chatter at 47 Hz (70 marks/rev. × 40 rev./min. × 1 minute ÷ 1 min./60 sec.). If it’s self-excited chatter, then it will always be at 47 Hz regardless of the grinding parameters. That’s because self-excited chatter occurs at the natural frequency of the system, and the natural frequency is constant—it doesn’t change when the grinding parameters change. If it’s forced chatter because of an out-of-true wheel, it will follow the wheel rpm.

Conduct a simple test. Decrease the wheel speed to 9,000 sfm (2,150 rpm). Now the wheel-revolutions-to-workpiece-revolution ratio is 53.75 (2,150/40). If you count about 54 chatter marks per revolution, it indicates an out-of-true wheel. But if it stays at 70 chatter marks, it indicates self-excited chatter.