How much backlash have you measured? As long as the steps per inch are setup correctly, backlash is usually the reason for bad circles..
Congrats on the first part! Many more to come, I'm sure
I finally managed to complete a part. I'm still chasing a concentricity issue. My circles just aren't quite round yet.
Any tips on sorting it out? The area is in the transition from y to x but only on the north west side of the cut if you're facing the machine.
I'll get a pick of the part put up later tonight. It's not a complicated part but I didn't start with a simple flat part either. It's far from perfect but it's at least a part.
A little video of the part running...along with the previous attempt that I screwed up.
Similar Threads:
Last edited by Kenny Duval; 04-11-2013 at 11:13 PM.
How much backlash have you measured? As long as the steps per inch are setup correctly, backlash is usually the reason for bad circles..
Congrats on the first part! Many more to come, I'm sure
Also, just how far off is it? I just read one of your old posts where you said a circular pocket was out 0.0001" (one ten thousandth..!). If that's the sort of error you're chasing, don't waste your time!
X - .0027
Y - .0023
Z - .0018
I'll go back and double check them again. I tested them all after measuring by zeroing out the each axis and backing them off 3 inches then returning them the same 3 inches to ensure that the dial returned to zero. All of those checked out when I did.
I've got a handful of things drawn out already. They are just waiting on me to learn some more and ensure the machine is good before I try to make them.
Yeah, the backlash numbers are reasonable for most conversions, but slightly out of round circles is the price you'll pay for having backlash, even with backlash compensation turned on.. Most people learn to live with it, as most parts we tend to make don't always have to be accurate to much more than a couple thousandths. For anything round that truly needs to be spot on, I'll finish it on the lathe or use a boring head..
A whole day of checking...rechecking...cutting a 3 inch circle...checking...rechecking...cutting a 3 inch circle.
Best results was
X- 2.994
Y - 2.996
45 deg. - 2.992
The last one I cut after Mach 3 lost my settings was
X 0 2.998
Y - 2.986
45 deg - 2.9986
I measured the other 45 and it was dead on 3.00
I don't know if I am expecting too much out of this little mill or if that's good enough.
I am not sure which mill you have or what screws you are using but if I end up with .014" of error at the end of my conversion I am going to be pretty darn upset. I am going to hopefully make some chips tonight/tomorrow with my G0704 and linearmotions2008 screws and single nuts.
+/- .002" should be expected if everything is working correctly.
DRock,
My guess is your not going to get within .002 with a single nut without some time of backlash comp. I am just starting on my second rf45 build and went with double nuts on all axis, hoping for good results.
Starting to get some time to make stuff again. The results are still improving.
I like my Tormach Super Fly cutter.
The current project is a set of pedals for the Race rig to replace my aging G25 pedals.
The Spring retainers were not done on the mill. I borrowed some time on my good friends lathe to knock those out.
I didn't like the square bracket so off it came a got a little trim around the edges.
I've got to draw out and cam the brackets for the hydraulic brake master cylinder this weekend and them it's on to making the pedal levers and pedals. Once I am done assembling and disassembling all the surfaces will get pass with the fly cutter and then on to powder coating. We set ourselves up to do that as well. No more outsourcing on our projects.
Pedal levers are half done. I have to flip them and mill the other side tomorrow night and put the final bore on the bearing pockets. This was probably the most fun I've had. With the exception of one not quite tight enough clamp, that allowed the first part to shift a little. Everything went according to plan.
Good chips on this one tonight as well.
I guess it all comes down to the tolerances in your work that you are prepared to accept? With your 'best results' you are averaging 2 thou per inch that's 0.2% inaccurate! Pretty good for a non precision machine IMO! Mill cut holes will always show the worst in any 2 or 3 axis cutting machine - if you need ultimate part thou accuracy - use a boring bar!......just my 2 cents worth,
.002" is pretty accurate. Especially for a pedal. That doesn't require any tight tolerances.
The work looks nice.
A lazy man does it twice.
My KC-20 does about the same as yours on circles Kenny. On a good day it might be closer to .001 then .002 but that is about all I can get it to on a circle of that size. I'm going to grab a TTS boring bar from Tormach for when I need perfect holes. Probably a Super Fly too given the surface finish in your pictures- awesome!
I've accepted that that's the case. Good enough for a weekend warrior. I'm not building motor parts or anything that's critical with it. It's mostly learning something to keep my hands busy later when I give up the normal work week. I bought a boring head as well and that works when you need a round hole. The Superfly is something I really like so far. I does a nice job. I'm still working on the details I miss. Like that finish pass on the walls that I forgot to remove the extra material from the floor so there's a thousandth high area of material in the floor of the pocket bordered by the finish pass on the walls. Not going to hurt the functionality of the part but a simple programming mistake that I need to look for next time I simulate before heading to the mill.
**** just got real....
and the full Assembly minus the electronics.
I'm really starting to like this. They aren't perfect and I have made several mistakes and will likely remake the parts that have errors.
Project completed. They are on the race rig and functioning as planned.
This weeks project.
Has a couple issues in the toolpath I missed during the simulation but it fits.
Last edited by Kenny Duval; 12-04-2015 at 02:09 PM.
Those last pics arent showing up for me. Spoke too soon . they are there now. Actually the parts look nice.
Did you ever resolve or improve the circle cutting issue?
Were you using climb or conventional?
How did you lead in the tool?
What are you using for CAM to get the circle or pocket?
Wizards?
Have you started using backlash comp now and if so, what speed is the take up?
How much "look ahead" are you using on the general config page? check the Shuttle Accel setting there for best results using BC. BACKLASH SETTING 50-50-50 RULE
A lazy man does it twice.
Sorry I always forget to make them http instead of https.
I stopped trying. Since the backlash numbers are acceptable I just decided that I would just live with it. I picked up a 2 inch boring head not long after those posts.
That's all climb cut adaptive paths.
Fusion 360.
Most lead ins are 180 degree z radius .025 with the same horizontal radius. Same thing applied to the exits. I went back and beat on the tool paths last night for the finish cut that cause the 1 error in the part. I added another contour to the path that resolved the issue. I could not get the entry and exit to behave the way I wanted using the single previous contour. Once we are done deciding on design adjustments to the part we'll run a second one with the corrected tool paths.
No wizards.
Backlash comp has been on since the original testing. I think Mach is still set up at 200 lines for look ahead.