I think you may be on to something here. When I went back and reread your original post, I noticed that you had experienced some slight burning at these same spots on your X-50 Woodworker. It appears to me that you may now be experiencing “pressure cutting”, where additional material is removed during a dwell due to the “spring force” from machine compliance continuing to press on the cutting tool. An upcut bit generates an upward force on the spindle, and a downward force on the spoilboard. How much they deflect depends, of course, on the machine rigidity.
I know that you don’t have explicit dwells programmed, but low X-Y acceleration rates will have a similar effect. And you did say that you were currently using low acceleration rates.
As for the different results on your new machine vs. your old one, this may indicate that your old machine is more rigid in the Z-direction than your new one.
How are your results when you use higher acceleration rates?
You could also try using a higher-quality bit, to see if the grind geometry of the cutting edge affects your results.
Try a test.
Surface a block at 5 different XY positions one at each corner and one at dead centre of the machine.
See if there is any difference. Could be sagging if it only happens in the centre.
Second try a scrape test on each flute. Using a piece of paper and a flat block lower the spindle till it just touches the paper on one side of one flute- hand rotate the collet to see if the second flute is cutting deeper or shallower
I am just thining out loud based on a quick scan of what others have said and what I see in your pic.
-20 thou seems like a lot of play back and forth
-What method of pocketing are you using?
-Looks like you are using an upcut bit?
-You said you are using a low feed rate, but high RPM?
-Are you using the same exact file, feed rates, spindle speeds as you did on your older machine?
If your couplers are tight on the stepper and the ball screw, and your ball nut is seated properly and everything else is high and tight, then its the bit, the feed and speeds and/or hold down. I would lower RPM of the spindle down to around 12-13k and run your feed at 200-250 ipm and see what happens. The machine is robust enough that you shouldn’t get those kind of marks unless something is off in your set up.
imlooking at these sorts of problems with the gen 2 and not getting much confidence in the new machine but ill keep looking and hope this is a simple thing . lack of rigidity is NOT what I want ?
Yea I’m not sure what to make of it yet, it is a very powerful machine and it feels crazy to think it would be less rigid than my X50. I am in contact with Onefinity who have made this a high priority to figure out and they’re running tests on their own machine.
What other sorts of problems like this have you seen?
That’s what I thought too but with some closer inspection, these dwell marks are actually happening through the entire pass. Especially from left to right.
Results are mostly the same, slow or fast. Slow is a little better, but still pretty much unusable.
I have used the bits and bits downcut bit that was provided by Onefinity with this machine with the exact same results.
What controller did you have on your Woodworker? Were you running the same unaltered gcode program on both machines? Sorry if that was answered already.
-20 thou seems like a lot of play back and forth
-What method of pocketing are you using?
-Looks like you are using an upcut bit?
-You said you are using a low feed rate, but high RPM?
-Are you using the same exact file, feed rates, spindle speeds as you did on your older machine?
-Perhaps, have you measured your play?
-I have tried a few. Originally it was a raster (I think) where it cuts an ever growing offset of the shape until the area is covered, but then I also followed TMToronto’s advice where I babied it with some roughing cuts perpendicular to the grain and then .5mm finishing cuts with the grain with the same result.
-I also tried a downcut on the same program with the same results.
No problem, I don’t think I mentioned. It is the BB, so I had to re-export this program for the RealTimeCNC controller, but these were not massive changes when looking at the code. My Fusion360 setups and programs were still the same so I just exported that with the new post processor. Looking at the code, the path is expressed more simply with G2/G3 because the new PP can handle it, but there is no strange Z behaviour that has been added in.
I use Fusion as well. One thing I was thinking about is how the two controllers process the motion. I believe BB uses S-curve acceleration motion planning, but not sure about Redline - you could reach out and ask.
The two results may not be comparable therefore, and perhaps the smoothing of motion of the BB controller may result in much less ‘jerking’ during abrupt acceleration/direction changes in your toolpaths. The latter may be causing unwanted and unplanned movement of the machine frame, particularly at the spindle nose/tool. Modifying the machining parameters and toolpath strategies may therefore help.
One of the things I have enjoyed most over the short years I have been using my hobby CNC, is understanding its strengths and limitations and learning the skills and techniques to still get it to produce the quality parts I expect from it. This will be different for every machine and user.
I only just mentioned this in another reply but something I noticed is that this pattern isn’t exclusive to turns, but a single axis motion especially left to right on X.
I took the advice you shared from IDC to really try and baby this cut in case I was somehow asking too much of the machine here, and removing the noise of my offset toolpath which demanded multiple axes, i simplified to raster tool paths - first perpendicular to the grain for roughing, then parallel for finishing - and with the X travelling left to right at speed it leaves those swirls along the way. Same goes for the Y axis now that I look at this video again. I will share it here (I have sent this to onefinity).
So almost all variables between tool path approach have been removed as far as I can tell. The problem is happening all the way through a basic cut when it’s already reached full speed, not just during acceleration or direction change.
Is this an issue unique to the Gen 2?
Could it be that because the bearings are C-shaped and not a full circle there is some flex happening?
I’m not an engineer just spitballing
Well I have 2 machines, one is the X50 woodworker, and that’s my only reference but it works perfectly with this kind of cut. I really hope this huge heavy thing is not weaker than my small machine. I am willing to go back to it with the electronics upgrade alone, although repacking and shipping would be such a time sink. I feel like it can’t just be that the machine inherently can’t cut .5mm deep in walnut.
You said you run the same exact file on the X50? Seems like you have a bunch of experience, so I don’t want to go over things you already know. When I pull and push on my spindle threads, I have to put a lot of pressure to get any flex and it’s small. You said you were seeing 20 thou front to back? What happens if you push up on your spindle with a good amount of force? Someone on here should run the exact file on their gen 2. What you are seeing means there is something off that is out of the ordinary. Have you talked to tech support?
Yes support is testing my file on their end trying to recreate it. If anyone has a gen 2 foreman and a 1/4” upcut bit I’d appreciate a facing attempt like this. I can share a file but it’s literally just a facing pass, or even just a single pass from left to right. I suspect it’s not the machine but my machine. I just don’t know what to fix at this point.
One thing, my machine didn’t come with a test cut. Onefinity assures me they test cut everything, did anyone else’s Gen 2 come with a cut? I think the Elite test cut would reveal this it if they actually ran it.
Out of interest, as a fellow Fusion user, what toolpath were you using? I am not aware of Fusion having a raster tool path. I have used many others, including parallel and of coarse surfacing which cut a path similar to what you have shown.
Also, what Feed rates were you when you mention in your original post that you reset acceleration to be very slow after initially being very fast, were you referring to the Redline controller motor settings?
Finally, I live less than an hour from Onefinity and have had the pleasure of visiting them several times over the years. I always have a peak at their production and can say for sure from my observations and questions that every machine that goes out the door is assembled on a test table, run to make sure electronics and hardware motion function, and finally that it cuts the final test piece. It is possible it was not put in the box. This comes up from time to time so I thought I would share my experience.
