Determining aggressive values: Testing & Quality Assurance

Dear Doc: I grind 200mm-long hardened steel with a resin-bond 300mm-dia. CBN wheel using the parameters in Table 1. After dressing, the imparted surface finish is good but the grinding power is high. After grinding about 10 pieces, the surface finish deteriorates but the power is lower. After about 20 pieces, the surface finish is rough and I have to redress.

I adjust the speeds and feeds to find a material-removal rate that’ll extend the time between dresses while maintaining a reasonable cycle time. But there doesn’t seem to be any rhyme or reason to it. Why?

The Doc Replies: The phenomenon you describe is of a wheel “opening up” during grinding. This is normal when grinding hardened steel with a resin-bond CBN wheel, providing better cutting but a rougher surface finish.

However, you can control how fast the wheel opens up. Pushing the wheel hard opens it up quickly, whereas grinding timidly opens it up slowly.

Most people think that pushing the wheel hard or not depends only on the mrr. This is absolutely not true. You can grind at a high mrr and have a wheel open up slowly, such as when creep-feed grinding at high wheel speeds. Conversely, you can grind at a low mrr and chew up a wheel quickly, such as when shallow-cut grinding at a high table speed and a low wheel speed.

So let’s forget about mrr. Sure, it’s useful for seeing how much material you’re removing and how much heat you’re generating. But in terms of how quickly a wheel opens up, let’s think in terms of the aggressiveness number. The equation is:

When looking at your parameters, I see you’re grinding at a lower and lower mrr each pass, presumably to reduce the risk of grinding burn. That’s good! I’ve added one more column for the aggressiveness number. The aggressiveness values are consistently in the 20s except for one whopper at 51.1 in pass No. 4, which is the pass that’s chewing up your wheel and shortening the time between dresses, even though the mrr is not high and the DOC is not large.

Table 1: Standard parameters.