(lots to learn there – there should be a way in G+ to register to the thread without spamming it, sorry guys)
It looks like you get spindle vibrations (better to say micro-vibrations) only on 2 sides of the circle, where Y move is larger than X move. And this is the X axis move which is causing tool vibrating.
I would check two things:
- Is it possible that the machine has different rigidity in X-Z and Y-Z planes? Is the machine cutting fine when larger circle diameter holes? Check machine rigidity and that whole spindle is fixed correctly. Possibly you may need to replace spindle mount with a bigger/thicker/more rigid one.
- How do you fix the workpiece to the table? If the fixture point is far from the tool, you might also get slight material vibration. Try to fix a small piece of aluminum from all 4 sides or 4 corners not too far from the hole being cut and try it.
For both tests you may try larger hole diameters first and then getting down to see when the problem begins.
Just a novice thought here but it seems like the chatter begins just as the mill breaks through the stock material. Could it be possible that your eating into your spoil board underneath and loading up the flutes in your endmill?
chips getting squeezed between spoil board & stock with end mill??? (it’s so fast, but look like it’s about at the length of the 1st chip of that layer, the opposite side being mirrored after the 1st issue)
(like Jeremie, keeping the thread link, sorry for ‘just a guess’)
Lots of great input here! He is getting it as soon as he breaks the lower surface, relieving the bottom of the end mill transfers all load to the side and this is where the issue is.
Rigidity on certain plans - it happens at 2 and 8 o’clock on the circle. Both are after the transmission of push to pull power. And doesn’t happen at 10 and 4.
I’ve got the exact same machine, ran his code in 5052, no issue. What varies is the spoiler board, which has since been modified somewhat.
It’s kind of an odd non repeatable issue for me. Glad you guys are throwing different ideas out there, maybe we can get to the bottom of this.
@Brandon_Satterfield I think you get us all closer to the truth. It is clear a lack of stiffness. But, maybe it is not about the mechanical stiffness. Could be from the stepper. You said both situation (yours an Ben’s) are identical. May be not! The mpp’s drivers could be set different. I don’t know what type of drivers are but, for sure, they have max current setting. Please check. You said something about changing from ‘push to pull’ and this gave me the idea… Can be a momentary lack of torque in a transition phase…
@sszafran the machine is screw driven cbeam in all axes. I have also checked for play in every axis and tuned up the v wheels on all. I have tried larger circles with exact same issue in the same places. Also tried a square and had same issue along bottom and right edges. The sheet is held down with multiple screws into the spoilboard and I have tried cutting between 4 screws within 1inch distance, pretty sure it’s not flexing.
@Chaotic_Logic it’s definitely not excessive load from the spoilboard, only going 0.1mm into it and it’s fine at the earlier part of the cut on the same level.
@Alex_Paverman my drivers on the TinyG are capable of putting out 2amp per winding and 2.5amp with cooling which it has (big fan blowing at it). I have the current set to max on all axes in the config (digital control on the v9 board). The motors I believe are the same Nema23s @Brandon_Satterfield is using.
@Brandon_Satterfield running the desired settings of 600mm/min at 0.6mm DOC through GWizard is showing a 0.0016" chipload for this bit, I think thats a little over what is correct for a 4mm single flute. Gwizardseems to think it should be 0.0013". I might try its desired settings tonight and see if that helps. Though it does think that deflection at these settings is just 0.000022 which is crazy small.
I’ve never found Gwizard to be correct on any machine I have. But that’s just my experience.
@Brandon_Satterfield agreed, though at this point I will try anything. It does have settings for the Shapeoko 3 now though which is what i used as I figure the R7 is about on a par with that.
@Alex_Paverman yes, that is exactly what the translation of motion thought was but it should be closer to 12 and 6 or 3 and 9. In my mind at least.
@Ben_Delarre oh that hurts… :-)z
Haha… @Brandon_Satterfield thinking on the motion thought. If you watch the video really slow just before the judder occurs you can see it actually starts as soon as the bit contacts the outside wall. So it’s starting at 3 and moving around to about 12 then again at 8…
Here is the video of the same machine.
https://plus.google.com/u/1/111137178782983474427/posts/LqLeX5kRSnX
hmmm… i was wrong again… 
Twice in a row!
By the way, are you living in the Bermuda’s triangle? This would explain all! 
As per that translation thing, @Ben_Delarre , I think I lost you… I already told about my poor English language…
I live in Los Angeles…nothing works right in Los Angeles. 
what results could be if you try with some tabs ?
let say :
at 3/6/9/12
at 4/8/12
at 2/6/10
big tab, small tab, …
@StephaneBUISSON This actually all started with tabs on a profile cut, i moved to a pocket to avoid any concerns about chip clearance. Same results 
@StephaneBUISSON that was a pretty good suggestion! Ben’s right though we aimed after that originally.
This issue is like one of those strange interview questions that has no answer, but does elicit every single possible response eventually