Re: CNCShark HD3 Randomly digs into wood - HELP!
Posted: Wed May 10, 2017 8:50 am
It almost sounds as if it's having an issue with a flat-bottomed v-bit, where the toolpath imagines it goes to a point. Does your v-bit have a true, sharp point on the end? Or is it a wood-carving bit with carbide cutting surfaces welded onto a steel body, that has a flat tip. Those are normally chamferring bits, not true v-bits, which are expected to reach a true point. After all, you can't get a true V carved with a real point in the bottom of the V, if the bottom of the bit is in fact flat. Okay, maybe a picture will help?
When there's a flat-bottom v-bit being used, and the toolpath wants a cut at a certain XY location, it calculates the Z ordinate based on its understanding of where the cutting faces are with respect to the very tip of the bit. That is, it assumes the cutting at a particular Z ordinate goes all the way to the center of the bit. But, when there's an "undeclared" flat spot on the tip of the v-bit, the touchoff tells it that the flat end of the bit is that point, and so the actual cutting edge is "lower" than what the software thinks it is.
Try this: instead of a v-bit for the failing cuts, try creating that exact same tool, but as an engraving bit. Specify the diameter of the v-bit's flat spot as the TIP diameter of the engraving bit. Remember that, like with the v-bit, the Diameter parameter is NOT the shank, but the OVERALL DIAMETER of the cutting edges. If, as I suspect, your v-bit doesn't actually reach a true point, measure that tip diameter, and enter that in the engraving bit.
As you start the run in the part that is failing, if you notice it's not doing it right, STOP THE MACHINE. If you can tell it's not running right, DO NOT let it run to completion. Why? Because the "bad" cuts are removing too much material, which means that if you re-ran a correct cut on that same part, it wouldn't touch anything, so you can't tell if it's right without redoing the entire thing. Stop wasting your wood. The moment you see it not cutting correctly, stop the bad cut. Then just re-run that one toolpath as you adjust it, letting it get a little farther with each adjustment. But again, only so much farther that you can tell whether it's working properly. That way you can use the same failed part to test a whole bunch of ways to fix it. Makes the whole debugging process faster as well. And, slow the machine down so you can watch it carefully. Pause the cut when it's doing a bad thing and check what the system says the Z is. See if that value really makes sense (yes, requires math with the V-bit). Being a very careful and close observer (eye and hearing protection required!) will get you to a solution faster than running it "for reals".
Also, the amount of current used to sense the touchoff is on the order of 2-3 milliAmp. The likelihood that it's a "conductance" issue and gives you that much incorrect depth is essentially zero. You'd only see that if the v-bit visibly pressed itself into the touchplate, which would definitely leave a mark.
Getting closer...maybe
Regards,
Thom
When there's a flat-bottom v-bit being used, and the toolpath wants a cut at a certain XY location, it calculates the Z ordinate based on its understanding of where the cutting faces are with respect to the very tip of the bit. That is, it assumes the cutting at a particular Z ordinate goes all the way to the center of the bit. But, when there's an "undeclared" flat spot on the tip of the v-bit, the touchoff tells it that the flat end of the bit is that point, and so the actual cutting edge is "lower" than what the software thinks it is.
Try this: instead of a v-bit for the failing cuts, try creating that exact same tool, but as an engraving bit. Specify the diameter of the v-bit's flat spot as the TIP diameter of the engraving bit. Remember that, like with the v-bit, the Diameter parameter is NOT the shank, but the OVERALL DIAMETER of the cutting edges. If, as I suspect, your v-bit doesn't actually reach a true point, measure that tip diameter, and enter that in the engraving bit.
As you start the run in the part that is failing, if you notice it's not doing it right, STOP THE MACHINE. If you can tell it's not running right, DO NOT let it run to completion. Why? Because the "bad" cuts are removing too much material, which means that if you re-ran a correct cut on that same part, it wouldn't touch anything, so you can't tell if it's right without redoing the entire thing. Stop wasting your wood. The moment you see it not cutting correctly, stop the bad cut. Then just re-run that one toolpath as you adjust it, letting it get a little farther with each adjustment. But again, only so much farther that you can tell whether it's working properly. That way you can use the same failed part to test a whole bunch of ways to fix it. Makes the whole debugging process faster as well. And, slow the machine down so you can watch it carefully. Pause the cut when it's doing a bad thing and check what the system says the Z is. See if that value really makes sense (yes, requires math with the V-bit). Being a very careful and close observer (eye and hearing protection required!) will get you to a solution faster than running it "for reals".
Also, the amount of current used to sense the touchoff is on the order of 2-3 milliAmp. The likelihood that it's a "conductance" issue and gives you that much incorrect depth is essentially zero. You'd only see that if the v-bit visibly pressed itself into the touchplate, which would definitely leave a mark.
Getting closer...maybe
Regards,
Thom