I’m at a loss here and looking for ideas. Running a 5 Pro with Makita router.
The Problem: I have an item where I cut concentric circles. First pass is with an IDC roughing bit followed by a 0.02 offset with a compression bit. The compression bit removes the roughing marks everywhere except the 5 to 7 o’clock area on the front side, this is consistent around all the circles. I’ve also noticed this on other projects, it removes roughing marks everywhere except the area in the same position. I even changed the offset to 0.03 and it still didn’t even touch the area in question. This is an old file that I’ve run dozens of times without issue.
At first I thought this was a spoilboard or tramming issue. My old spoilboard was chewed up so I replaced it, trammed and surfaced. The X axis is off about 0.005 but the Y is dead on, still produced a nice spoil board surface without any bumps.
I’ve also gave the entire machine a good air dusting, lubricated the guide rails, checked every bolt and screw for tightness and verified there’s no play in the step motors. I also checked my files to make sure nothing was changed by accident and it’s all in place.
I’m not sure where to go from here, but I’ve got a big order coming up in a few weeks and need to figure out what’s going on.
Perhaps the roughing bit is deflecting as it hits different grain? That would explain why it’s at the same ‘clock’ position each time.
If you are willing to experiment, you could try something like this:
Do a circle with the roughing bit, but very shallow - just enough to see the cut, so even 0.5mm. Hopefully this is a good reference cut, with minimal deflection.
Now run the roughing bit as you would normally, at the offset you would normally. Actually, it could be any offset, so if it’s usually very small increase it so the difference is a bit easier to see.
If there’s deflection, you should see it as it changes distance from the reference cut.
I had this thought myself, but it is too consistent and has been appearing on just about every project I’ve done in the last month. Different materials and grain patterns, different cut patterns, it’s always there.
No jiggle in the bits. I have roughly 150 cutting hours on it. I just trammed it again, it’s within 0.002" in all directions and I’ll give it another test tonight. Hopefully this fixes it, but I was getting the same results prior to my recent spoil board update and first tram.
The only other thing I can think of is somehow the machine came out of square, but if that were the case it would be a consistently inconsistent cut and correct its own problems on the second finishing pass.
I’ve been running this setting the last year, these problems just started a few weeks ago. I manually slowed it down today and it still had the same results.
It’s not a file issue, this happens across multiple different files that I run on a regular basis.
That is the problem area though, the finishing compression bit only clears the roughing marks in select areas. It has to be some mechanical issue that’s causing the bit to not make perfect circles as the cut is inconsistent.
The file will show your cutting strategy.
If you’re programming with CC, you’re not cutting circles anyways. It linearizes circles to straight line moves.
So from reading things over, it would seem that perhaps there is some extra play in the movement mechanics possibly just from normal wear or loosened nuts on the X/Y ballscrews?
That would be my thinking as you went from good cleanup passes to missing the marks using the same files over many operations. And you’ve already trammed & flattened as much as possible.
Ultimately, any movement is linearized — the question is when it happens, either deferred to the last possible moment and done by the firmware, or in the CAD/CAM software. Carbide Create is in the latter camp, while some tools can be either/or depending on post-processor configuration (Vectric or Fusion come to mind). The only tool I can think of which was G2/G3 arc-only is PartKam/MakerCAM (which I remember fondly, having used it, and F-engrave and Fontforge for my first project (long story))
If it’s done in the firmware on the machine, the tolerance is ± 1 step. (less than a micron)
Done in CAM or post it’s limited to the resolution of the g-code (3-4 decimal places)
Neither should have an effect on the problem at hand. I wanted to see the file to determine the cutting strategy & further understand how the rough/finish paths were set up.
I replaced the router as a long shot and figured I’d just have another trim router sitting around.
Then I found the problem. I previously noticed the screw rod was uneven between the two sides by about 3mm at the front of the machine. I was able to wiggle the rod front to back and figured this must be the issue. A little bit of looking around and I found this was loose. Removed a few pieces, tightened this back down and the problem seems to have been solved.
In the CNC machining world this would be called excessive backlash and typically the main place you would see it is when you cross quadrants and the ballscrew reverses, obviously this was a mechanical issue but sometimes it may wear. Easiest way to check is with a test indicator moving either axis in both directions seeing if there is play. Not sure if Carbide motion parameters can be adjusted to overcome some of this when wear is the culprit.
(less than a micron)