Just jumping in the thread to say that I’ve seen similar issues. I haven’t quite eliminated it, but Jim’s suggestion of reducing the acceleration helped the best.
I also made my table more rigid with better supports. If you have the aluminum table, add sheets of MDF bolted to the sides to reduce flexing. I’d bet that if you put a dial indicator on the top of the machine along the y-axis, you will see quite a bit of movement when the machine switches 180 directions – particularly with the default acceleration. This causes chatter.
I don’t know if it is the best solution. It’s definitely one of the first things I’d consider if I had the aluminum legs. I think one of the reasons people use it is to also add weight. I’ve seen some people just add a ton of t-nuts into the extrusions and a ton of bolts.
Just about any plywood would be stronger and stiffer at the same thickness. Strategically gluing on straight pieces of hardwood would reinforce a plywood skin akin to a structural beam.
Massive improvement after slowing acceleration on all axes to 15, and rewriting the toolpath to use spiral as much as possible. Not sure which one of these fixes helped more, but I’m going to do some more testing.
Latest fun issue is that the toolpath froze about 5 minutes away from finishing this four hour cut. Machine and mach4 were responsive, spindle kept running as well. The movement just stopped. When I used start from line without re-zeroing off the piece it gouged the part, ruining the bowl and meaning it lost its xyz position. Replacing the laptop tomorrow to see if that could have been the cause.
I’m averse to using hardwood to reinforce the table, as it both expands and contracts in response to ambient humidity, and is substantive enough for that change to impose on connected structure. Moreover, that dimensional change is not uniform, but grain-relative.
Yes, this is invaluable. If you have the time, please keep it coming.
It could still work well. Yes, you have to account for wood movement with hardwood, but wood sheet goods have wood fibers, too. Plywood and MDF change dimension with RH up to 0.5%, while solid wood changes less along its length, up to 0.1%. If it’s a concern, use standard techniques: break panels into sections and cut slots in the appropriate directions at panel-to-frame connections. If you use solid wood, keep the pieces narrow in the plane of the panel, as width varies up to 5-10% with RH.
Almost all of us have a one-piece MDF spoilboard bolted to our frame. It adds stiffness, and I’ve not heard of it causing issues, probably for a variety of reasons. For example, T-nuts slide with wood movement, or people already have slotted connections for other reasons, like locating fasteners. I suppose side panels are less proven and could cause more of an issue. Maybe in addition to the above, first add 8020 beams across the bottom.
While not speaking to MDF, quality engineered plywood may enjoy greater dimensional stability than many common hardwoods (white oak, some maple, for example); when laying wood floors, a larger expansion gap is indicated for the natural product. A legacy of joinery designed to account for the mobility of hardwood in construction and furnishings attests to the possibility of surmounting these challenges, but I would still not invite the need. (There are some hardwoods enjoying remarkable dimensional stability- I’d be happy to encourage someone else to experiment with teak plywood and report their findings!) I reckon the saving grace of MDF is the uniformity of expansion and relative softness.
It depends on how you use it. If its longitudinal, then wood is about as good as aluminum because the thermal expansion is low compared to aluminum, and the longitudinal changes for humidity aren’t bad. Cross grain though, ya, thats not great, but plywood isn’t nearly as bad.
I sheeted the sides on mine with 3/4 mdf and used aspen 4x4s for my legs and to stiffen the panels. It’s a pretty stable wood as far as movement goes. I’m sure it changes a little but I haven’t notices any issues with what I do with my machine. Ill post some pictures here next time I have it cleaned up a bit.
A big thank you to everyone who helped me out on this thread! Machine is working great again with the acceleration set to 15 on all motors, and toolpaths written to be as smooth as possible. After some more testing there is still some slight gouging on fast corners, however it’s very clean compared to before the changes.
Following up on the machine pausing issue I was experiencing - I surge protected the laptop and cnc electronics and everything is working properly again.
As far as increasing rigidity, I have some extra diagonal struts and y-axis-direction extrusions that I will be installing once this current job is complete, and I will likely add some extra plywood bracing as well.
Has anyone gone nuclear and added sand/epoxy mix into their extrusions for vibration reduction?
I wouldn’t do that, especially on any moving parts. Making them heavier won’t help. For the non-moving parts, if you added sheet goods to the sides, or do corner to corner cross bracing, thats about as good as it gets.
The rest of the undesired movement (the relative movement between your workpeice and the bit) is due to flexing because of the stress on the system due to acceleration. Less weight, more strength, and controlling the acceleration profiles are really what helps that.
Some new interesting developments. The machine pausing issue persists, and I will be trying to catch it on a log in mach4 to see if anything shows up. I am working with avid support on this, but I figured I would continue to post about the ongoing issues here in case anyone else experiences something like this.
This occurred at every similar move around the protrusions I was tracing using an adaptive clearing toolpath, and it seems that the first one caused the pause. The stutter continued, but the pause did not reoccur. The initial failure occurred around line 13082, but it seems to jump a few lines when I hit feed hold so I am not sure exactly. Very conservative 120ipm, .2" stepover, .375" doc, 1/2" upcut bit in mdf.
The machine was extremely laggy while jogging for my tool change, and gave a “waiting for in progress flag to go low” message while I was jogging it.
Apart from the jogging, it sounds like a buffering/data starving issue. Do you have smoothing on, and what is your tolerance? If you follow the recommendation in the video below, set the tolerance to 0.001" with smoothing turned on at 0.001".
Great call, that is definitely something to try. I agree that it looks like a throughput issue. The gouging was also occurring in areas with complex g-code during transitions.
I usually have it turned on, but I did not for this toolpath in particular. I have the tolerance set to .0004", the default value in fusion. I will test out .001" with smoothing today!
I did try a second laptop, same issue occurred immediately so I switched back.
Edit: Just applied those edits in fusion - the toolpath is 10x smaller. I really hope that is the issue, this is the most promising fix yet.
