I upgraded my Pro system to Ex with servos. Everything is dialed in and running very well with simple (plywood, for example) projects while I still learn the new software and system.
Today I did a 3D carving and had disastrous results – I clearly had feeds and speeds WAAAY too aggressive for a Fusion adaptive clearing pass on hard maple. While the parameters were bad, the servo system did not quite live up to my assumptions/expectations.
Using a spiral downcut bit (all I had – I usually use upcut for adaptive clearing) the aggressive speeds clearly caused the Z axis to drop down during what should have been coplanar to the work surface. This caused massively aggressive over cuts and eventually broke the bit when it traversed over remaining material. All because Z was way off. I didn’t see this until too late due to the dust boot.
This was obviously because of bad programming, but my question is why the servos with their feedback loop did not detect this and stop the program as was my expectation. When all was said and done I moved Z to machine coordinate zero (not through homing), and it was exactly where it should be. It seems the machine knew the correct location, but while running the g-code, didn’t care that the Z position was wildly off by as much as an inch.
If this is true this tells me that the servos didn’t actually lose position.
The way the servos work is that they have an encoder in them. If you lose even a little bit on your position the encoder sends a signal to the control and everything is halted.
This sounds to me like an issue with your G code program. If you can post the G code and/or the Fusion file I could take a look for you.
Thanks Eric, that is what I thought. Without restarting or changing anything but z-zero I did do an air run (well above any material) and it appeared to be correct, at least looking at the DRO and eyeballing the machine.
I’ve attached the g-code. I can send a picture of the result in a bit.
One thing I have had happen cutting hard maple too aggressively with a 1/2 inch up cut bit is the bit got pulled out of the collet slowly as the job went on. Not enough attention to torque putting on the bit combined with over aggressive cutting for my setup. I did not miss steps, but the result was a disaster. If your bit broke off, it might have been getting pulled out of the collet along the way.
What post processor/machine combo did you use in Fusion to make this?
This is a copy of the Avid provided Fusion post. Other than the name change, it is exactly the same as Avid’s. I always keep my own copy so that when I do modify the post I can compare the output with the unchanged Avid one, if necessary.
I was referring to machine Z zero, e.g. all the way to the up.
The distance the z axis was off is so much by the end of the job that I would think it would have exited the spindle (e.g. not that much inserted into the collet). I still have the broken bit in the machine, I’ll look closer when I get home to see if the bit might have slipped.
Each of those “rings” should be at the same Z level. Note: it was clearing .375 inches above what you see here. It’s like it progressively got worse the longer it ran. The slot cut in the center circle is where the bit broke – this should have been a move above the workpiece.
Now I am wondering if the bit did slip as suggested above and the machine was actually at the correct Z level the whole time. I should be able to determine this by comparing how much of the bit is still in the collet and using the last measurement off the tool changer plate.
That looks to me like a bit slipping, so I’m curious to see if your measurements support that conclusion. I have found that over time collets can wear and when they do this is more likely to happen. They aren’t that expensive so I have started treating them as a wear item. I always keep spares around and if I have a problem with a bit slipping in goes a new one. In my case I use 1/4" bits more than anything else and I usually replace the collet every 6-12 months depending on how hard I’ve run the machine.
I also think it looks like bit pullout – especially since you were using a downcut. Not fun when it happens! Definitely examine the collet for wear/damage, or just toss it.
Given what you said about not seeing a “Drive fault” and the machine returning exactly where it should have after the cut is done this sounds like what happened.
You can find tightening torque charts for ER20 collets around. Depending on the bit size it can be 50 ft pounds or more. In my book that’s “gutentight” but not “break a sweat or grunt a lot” tight.
That’s all to say don’t be afraid to snug those in there good.
Sure enough, bit slip it is. When I got home I marked the location of the installed bit with electrical tape, tried to tighten the collet (seemed tight), then removed the bit.
The depth difference from the top outer (first) cuts to the bottom is about .75". The width of the electrical tape is about .75", and the bottom of electrical tape is depth I would have inserted the bit. There is no doubt it slipped that amount.
The collet and nut look fine – I rarely use 3/8" bits. Unfortunately it cost me a $100 bit (Amana 46449-K).
Thanks for all the feedback and help! And @Corbin it was actually your beautiful bowls that was the inspiration for this test piece.
I had this happen to me, and learned to clean every new bit (and often the collets) with a degreaser (I prefer acetone). I learned that new bits are often shipped with a thin film of oil on them to prevent rust, but that oil also lets them slip under load.
That looks like it’s specific to the tapers; I would hesitate to put lubricant on the bit itself unless the lubricant said it was for that purpose, because of slippage.
Usual preface: I’m with PreciseBits, so while I try to only post general information take everything I say with the understanding that I have a bias.
This is correct. We don’t recommend using this in the bore of the collet. Regardless, I appreciate the shout out grossmsj.
That being said there’s a few other things here…
Part of the clamping force and slip resistance is based on the mating of the tapered surfaces of the spindle and collet. So cleaning those may help.
While you don’t need MAX torque, as Eric said, the spec for ER20, UM/hex nut, with a 3/8" bore is 59ft-lb/80Nm. Depending on your wrench that could be asking a lot.
Off the top of my head DIN standard for collet engagement is 4x diameter or the full length of the collet. From that picture it doesn’t look like there was enough blank shank for proper engagement or clamping (depending on which end you mean by the bottom). That alone could let this slip or possibly damage the collet.
Hope that’s useful. Let me know if there’s something I can help with.
I think you hit on something important here… While this isn’t scientific, it seems like you with the short wrenches we all likely have “very tight” is probably pretty darn close to the correct torque.
After about 5 years of quite a bit of CNCing that Eric’s ‘gutentight’ is about right. If I have bit slippage it’s always been one or more of a) an old bit, b) a worn collet, or c) a worn or damaged nut.
