hello ... I'm new here and the question has probably been asked before, but I can not find the answer. so here comes my problem. I have built a CNC mill, which actually works pretty well according to my expectation, There is always room for improvement ... but it's the first ... ihha it works. anyway. I have a software problem.... I think.. When I want to cut large circles / round items, they become very uneven. as if the circle consists of a lot of straight lines. I know that a circle in cnc is just lines from point to point ... but do ther need to be 3 ~ 4 mm between each point. I'm looking for a way to get a higher point resolution. I use fusion 360 + ugs platform.
Most of my circles are small for bolts and such. I know in Sketchup, when I make big circles, I can increase the number of points so it becomes smoother. Have you looked to see if that is an option? I don't have access to Fusion 360 at the moment so I can't check, but I believe it does.
Hello Thanks for your quick response ... I had the same thought. But I can not find this feature in fusion 360. I can not help thinking that it has something with arduino grbl firmwear or the settings in the ugs platform??
Make sure you are using a postprocessor that supports G2/G3 arcs. Some of the postprocessors forces circles as lots of G1 moves (check your gcode output in a text viewer - if theres no G2/G3 commands, then you need to use a better postprocessor : (You dont mention which post processor you are using?) GitHub - Strooom/GRBL-Post-Processor: Start for example is quite popular on here (not sure if it does G2/G3 - i dont use Fusion myself (I use DXF -> github.com/LaserWeb/LaserWeb4-Binaries -> Machine workflow - LW4 has built in CAM thats easy to understand and also host interface for the machine)
To update my own doubt above, GRBL-Post-Processor/OpenbuildsGRBL.cps at master · Strooom/GRBL-Post-Processor · GitHub shows that Strooom's is using G2/G3 IJK format (ie this is PROPER ARCS! - not segmented G1 moves) - so give it a try
Fusion 360's included Grbl post-processor also outputs G2/G3 arcs - I just checked a recent file of my own. I've had no trouble with circles using Fusion 360 (largest so far though is 24mm diameter). Under post-processor properties for GRBL, there are settings for maximum circular radius and minimum chord length. Mine are set to 1000mm and 0.01mm respectively (but I think those were the default settings).
Hello friends... Sorry, I first respond now, I've been away. + I have done all sorts of experiments... Now I have taken a picture of the problem, it is a diameter of 70mm. The 2 pictures are the same one indicates with arrows what I want to solve... I wonder if it's the step size of my stepper motors... I have a travel of 5mm 1 round on the x and y axis....
I don't know how Fusion360 handles circles but I know that we had this problem with SketchUp too. In SketchUp a circle is drawn in lots of segments. In order to get a perfect circle in the cutting the g-code output from SketchUcam is handling the segments as arcs and the results are a perfect circle. Look into it in Fusion, maybe there is a setting for it in the g-code settings.
Hi! The suggestion has been before, just I can not figure out where to change the parameters in fusion360, it's probably a bit over my head,
What has been done to ensure this is not machine related? Backlash, steps, etc. could all be at play here. I would make a 2x2 square, and see how close the cutout is to that.
a little follow-up ... in the match, I have switched to mach3 ... still the same problem ... but in mach3 post processor I can choose " Radius ARCs yes or no " description: if no so the output is IKJ based if yes then it is based on the radius... ?? Does it make any sense ... the choice is now "no" so ikj based... I'm trying to change to Yes and see what happens,... I'm just wondering that even if I go with an arduino GRBL or a MACH3, I have the same problem ... but not the same options
TO cncmd: Is tried and it's spot on. Sharp corners and everything perfect ... But I doubt that it's hardware ... so I've ordered a gearbox for the motors 10 -> 1
TO cncmd: Is tried and it's spot on. Sharp corners and everything perfect ... But I doubt that it's hardware ... so I've ordered a gearbox for the of motors 10 -> 1 If it is hardware any suggestions
Under the Cam -> Actions tab, when you select the Post Process option, do you have "Grbl" selected as your Post processor (as shown in the image)? If not, try that.
gearbox will introduce backlash, and really slow things down! you said you have a 5mm lead screw? what microstep are you running? hard to tell the issue with no scale on the pics. Also try in a soft material like foam, could be bit deflection from hard material.
Hello Thanks for your support! First: Scotty Orr comes with a suggestion ... I do not know which vertion of Fusion 360 you are using, but that's not how my menu looks But I think it's the same... and for gery caruso: it's a ring with a outer diameter of 70mm and a inner diameter of 50mm so the ring is 10mm thick ... hope you can use it.+ The resolution itself is 800 on the driver. The next step is 12800 see the picture
Ok I have read and seen something about the phenomenon of backlash ... and apparently a bigger problem when making round holes ... I will go to buy a measuring device ... I come bake with some results ... when I have the device to make the measurement
Hi Kim, you mean 8000 (1/40th microstep)? And your the screw is 5mm per turn lead? that is very high resolution. Actually too high in most opinions, as holding torque for a given step diminishes rapidly with increase microstep. Really you should do great with 1600 (1/8th microstep) with a 5mm lead screw this would get you 1600/5 0.0031mm per step travel. At the most I would recommend using 1/16th or 3200 steps per turn. this would get you 0.00156mm per step!
Ok i see what you write The change will clearly make a huge difference, what effect does it make, I can not conclude it's going to be tested ... I live in Denmark, the time is 22.43 so it's late .. I will try the first thing tomorrow morning. You get an update. Many thanks for the help
Hello, although it's late, I could always make the adjustment ... and surprise .. I had a digital callipers ... x.xxx mm and with a little tweaking I could actually measure some backlash... actually a lot ... I think ... 0.015mm regardless of that then I got the measurement into the backlash adjustment in mach3 and now I'm sure ... I have not dared to turn on the miller itself because it makes a lot of noise so it has to wait ... I'm somewhat excited on the result Regards Kim jessen
Hello again... Well, there was a little clarification about the problem... The actual freezer is actually accurate enough... The problem is that the Z axis is not stable enough ... baa .. there will be a new gantry.... at some point ... when time comes to it... Right now I live with the problem Otherwise, many thanks for input Regards Kim Jessen
