Hoping someone can help me, been banging my head against a wall now for 2 weeks on this.
Trying to do my first cuts in Aluminum on my R7 CNC. I’ve got the settings dialled in pretty good: 0.7DOC, 750mm/min, 20 plunge, 40% stepover, Climb cut, 15000rpm, 4mm bit, single flute, coated upcut.
The cut quality is amazing mirror finish every time and it doesn’t sound like I have much if any chatter. EXCEPT…
On the very last perimeter of the pocket on the very last pass, it chatters like crazy in two spots, top right and bottom left, resulting in an out of round cut.
Checkout the video:
The bed has been levelled, I’ve measured the straightness of my spindle and it varys by less than 0.001" over its course of travel in both ZX and ZY. I’ve checked the runout of my bit and am using a Maritool precise collet, the runout is so low I can’t measure it with my less than stellar dial indicator.
I’ve tried going slower, faster, deeper, harder, all to no avail and every single time with the same exact result in the same places on the cut. I’ve tried doing it as a profile and as a pocket, same result.
For I can understand, the Z axis guides are based on some roller on alu profile channels. This is not a very rigid solution. Even worse if the guides are not with steel-on-steel contact.
So, I am also thinking it’s a backlash PLUS a lack of rigidity. Usualy the milling of a ‘pocket’ has a final finishing pass with a very small depth of cut. When the depth of cut become comparable with the total deformation+backlash of the whole system the mill won’t wont cut but more like ram. So, I think, both commenters are right.
Try not doing the final shallow pass and see what happens. You can experiment which is the minimal thikness of cut for not appearing that bad chatter.
@Alex_Paverman so your saying if I make the final pass thicker it could help? Can you explain a bit more what you mean about the total deformation + backlash? Is it that the machine is ‘relaxing’ during the final pass because there’s less material pushing back on the bit and so the bit becomes more parallel which on a curve like this results in it ramming into the wall and then as a result pulling down into the bed?
My ‘tech’ english is not so performant but I will try.
As a result of cutting forces the whole system is defforming. The less rigidity, the more deformation. The ‘system’ is a whole: the tool + collet + the spindle (the bearing system, and the spindle himself) + the guiding system (all three axis) + all the rest of structural parts. When the total deformation of the ‘system’ is comparable with the depth of cut, the cutting edge of the mill may be rejected out of the cut and very ample chatter can appear.
The more strength of cut material, the more rigidity is needed.
Just try to eliminate that final pass and tell us what hapend. My oppinion it’s just an educated guess so I am also curios if I am right.
@Ben_Delarre If it were me I’d use something else to generate the G Code and see how that worked. Though it might be a setting in your motion control software too. Some command that you might need, yet is not being put into the file?
Sometimes I have to set the blend tolerance or my machine won’t stay on paths. CNC can be awfully involved and arcane at times.
@Alex_Paverman that ‘last pass’ is a full depth of 0.7mm same as every other pass in the pocket. So I can’t really eliminate it. The rest of the cut is flawless, and if I measure the hole before doing that very last pass its also pretty good in terms of size and roundness.
@Paul_Frederick the gcode is about as simple as it could be, just straight up G3 arcs. The arcs perform perfectly on the previous layers so I don’t think its a GCode issue. I have also ‘air cut’ this by setting z0 to the top of my spoilboard then running the cut down to z0 without going through any aluminum and it works perfectly and is dead level the entire time. This means its something to do with how the tool is engaging with the alu.
I’m intrigued by this backlash idea, I will read up on it. Can anyone explain why it would come into play right on the very last perimeter of the pocket at the very last pass, i.e. when the tool is minimally engaged with the stock.
I suspect it’s something about the fact that the tool is not also clearing aluminum at the bottom. This is the first part of cut where the bit is engaging maximally on the side (maximum deformation force) and not at all on the bottom (almost certainly adding some sort of damping/stabilizing to the motion of the bit, possibly even constraining the bit from rejection out of the cut).
You’re using climb milling. Climb milling works great on rigid machines, but creates too much sideways force on this style machine when cutting aluminum. Try conventional. Even that may not work though; it looks like you have too much flex for aluminum.
Hmm yeah that sounds like it could be along the right lines @Ross_Bagley . But then I thought about it some more, during the earlier stages of this last pass its still doing the same thing. The bit is 40% engaged (the stepover) as it takes each perimeter around the cut. When it gets to the wall its still 40% engaged, but I guess there is some rubbing going on against the wall and its this which is causing the vibration and I’m assuming I’m hitting resonance.
What I can’t understand is I got someone to do the cut on their R7 (different machine obviously, same build, same settings, same bit - from the same stock, same gcode) and it cut perfectly.
@Todd_Fleming I’d ordinarily agree with you, but I know this is definitely not the case as the parts for this machine were cut on one of these machines, and I have had this part reproduced on one of these machines without issue. Its definitely a setup issue in my build.
FYI I have also done this same pocket such that I left 0.2mm of stock on the outside edge and took it off in a final pass at each depth. This too caused chatter in the exact same way as seen in this video.
It really is very weird. I’m going to try and do another run tomorrow with different chipload using settings from fiddling with GWizard. Maybe it will help. It’s so bizarre though seeing it run on the same style of machine and run flawless and seeing the cut on mine look great and then fail so agressively right at the last. So frustrating.
