Hello so toady i was doing a nesting work.496x650mm into smaller parts from 180 to 220mm. Than i noticed that some parts are 2-3mm smaller!!So i did a quick testing and i noticed when the cutting size gets bigger(vector) the actual finished parts gets smaller and smaller.I made a program for a 58x58,92x92 and 142x142 square.So the finished measurments: X58,03Y58,08////X91,42Y91,67////X141,24Y141,16 Also tried adjusting the depth of cut and feeds and speeds.It stayed the same. Iam using a black box controller.With a workbee cnc.
1) CONTROL > Wizards and tools > Calibration to check machine calibration 2) Make sure its not mechanical slippage (Shaft couplers tight, etc) 3) Make sure acceleration/max rate is not set high enough to stall (Grbl v1.1 Configuration)
The lead screw was loose on the y axis so i fixed it.After that i tested it and the results was the same before. Wat about the acceleration/max rate.Can you explain how to set it up?Thanks.
The Grbl Wiki explains it in good detail, but basically try half the current values and see if it makes a difference
Fixing loose only make it mechanically sound. You still need to calibrate for accurate distance movements.
I tried out some thing but nothing changed.I bumped up the current on the black box,tried out some acceleration settings.I went from 150 to 300 on all axes.Now looks like the machine is moving quicker/faster?! The 50x50 test cut is allright +-0,1. But the 120x120 test cut is off 0,8-1,3mm.
150 to 300 is double, not half Instead of random things, 1) Half Acceleration and 2) Half Max rate. Leave the current setting alone, it wasn't mentioned Then 3) do the Calibration... In that order, to the tee, and only that.
Yes, i already did that based on a square test 50x50.It was allright +-01,-0,2.But when i try to cut for example 100x100 square or 120x120 its always of 0,8-1,2mm.
You cannot calibrate accurately by cutting - that introduces other variables that affect the calculation. You also cannot calibrate accurately over such a short distance. Measure the actual distance moved over the longest distance that you can measure accurately. If you cannot use the wizard in Openbuilds Control (if your controller does not run grbl or grblHal) the PDF attached explains the principles of calibration. Alex.
Like you mentioned,i did the calibration in the wizard using a vbit and a tape measure(x500y700).After checking the travel it was spot on.But when i tried tu cut a test square 50x50 and 100x100 it was off.The cutting setting was 15mm/s and 2doc 1000rpm 3,175mm upcut bit. The 50x50 square came out x50.4 y50.2 and the 100x100 one x100.1 y99.9 The collet chuck is clean.The tool is in good condition.In the proces of cuting i didnt saw any flexing or heavy vibrations on the z axes.All the bolts are tight.Everything is greased up.
Then if 1) and 3) has been done, it still leaves you with more of (2) to look for Assuming of course calibration was done correctly and over the longest possible distance
Also, is your endmill truly 3.175 mm in diameter, or is it smaller. Often times they are smaller than specs so you end up with a larger exterior cut and pockets that are smaller than expected. This is particularly true of cheap Chinese made endmills. You need to accurately measure your endmill with a caliper and use that as the diameter in your CAM software to get the most accurate cuts. Also you get more accurate cuts if you do a roughing pass and leave a small amount of material to remove with a finishing pass. This is particularly true with smaller endmills because they flex more.
I was testing all day long,still 0,6-0,9 off.Now iam sure it must be a mechanical error.But i dont know what can it be haha.
I think i found out something,now all the test cuts came out with the same wrong size same on the x/y from small part to bigger. Maybe the machine is not square.or something is bended.I need to check that out.
Measure the tool diameter with a micrometer (1)and enter that into the CAM so that the Gcode is generated with the correct offsets. 0.6mm can easily be the difference between a 6mm and 6.3mm bit (.25"). Then there is a tool runout, if you turn the tool 90 degrees in the collet (keeping the collet where it is now), what is the result? (1) someone mentioned using a vernier caliper for this, but those are not accurate enough even if they do display 4 digits of imaginary precision (-:
the exact dimensions of the tool are important enough that some very expensive equipment is made to measure them, no you don't need a Spironi, you do need a micrometer (-:
Thanks for the tip! I measured my endmil 3.175 it was 3.170 with this tool the 50x50 test cut came out 0,8-1mm off almost same error with the 100x100.After that i measured my 6mm endmill it was 5.99 the 50x50 test cut came out like 0,2 off and the 100x100 was almost perfect +/- 0,1.The machine calibration is spot on(The steps).Than i did a test project using the 3.175endmill end the 6mm one.The pocket with the 6mm came out with the error of 0,1mm in woodworking this is basically nohing.And the countour cut with the 3.175 mm endmil (170x220) came out with 2-2,5mm error.Tomorrow i will test out more cuting pocketing using the 6mm endmill. On my home cnc i just did a handfull of project mostly decorative stuff,so most of the time the precision was not really necessary.I also just noticed something when i did the pocketing toolpath with my 6m endmill when the endmill goes from the right direction there was no chattering.But when the endmil travels the left side there was a noticebla chatter.Than i incrased the rpm durring the work from 13000rpm to 2000.The chattering was off both sides.It was like 15 or 20mm/s feed.So i just wonder whats the case with this because i saw a lot of folks on yt using a workbee cnc milling with lower rpm (10000-13000rpm) with no problem?!And yeah i dont remember their feed or doc,so maybe thats the case with this.So i noticed maybe the high rpm and diferent endmill will do the job with no to minimum cuting size error.Also i have to check if the machine is square.Can that cause the errors in my cuts?I hope i will found the solution after all.And thanks EVERYONE for the tips!
Tramming can effect squareness, which can affect measurements. See Tramming: A generic guideline For example, X and Y not square to each other Also check for unwanted play / looseness in your machine, the chattering combined with dimensional issues could indicate something moving that shouldn't. Grip the endmill, give it a wiggle, if you feel movement, follow the feeling of the play to find the source.
This is really helpful!Thanks.First thing i did today is that make sure that my table is flat as it could be.It was WAY off.Noticed that when i did the spoilboard surfaceing.Tomorrow i will make sure that the machine is square and so on. I didnt checked these things when i was buildin the machine,cause iam a beginner in this area.But this is a great place to learn from others.So again,thanks people!
As well as checking the diagonals of the frame to check for square, and checking that your X axis is square to the Y axis, tape some thin thread across the diagonals of the frame - they should JUST touch each other in the centre. If they don't touch, or one pushes the other out of line the frame is twisted and you need to adjust it by putting shims under the lowest corner. Alex.
I did some work on my machine also on my table.I rebuilded te machine completly and found some errors.Now they are gone.The machine is square now.The parts are in right angle,the bolts are tight etc.The tramming was done.The wasteboard has never been this flat...I did the calibration using the wizdard tool,tape measure and a 60deg. V bi.It was done right.But hen i try some test cuts it was off like before.Like Peter suggested,maybe this error has something to do with the acceleration or something wit the stepper motors.Maybe they behaving different under work/tension.
