Hello everyone and good morning I built a plasma cutting table, using Arduino and GRBL as a controller. Until some time ago I had no problem with this configuration, but now that I have completely redone the table the Z axis and the Z probe behave strangely and it seems to me that it could be a problem with the post processor or some setting that has changed during the updates of fusion or the post processor. In short I prepare all the cad and cam and post process until I get the Gcode file, when I start the job by giving all the 0 to all the axes, the Z axis starts to move first with the intent of making the first probe of the material, it reaches to touch the sensor in Z but completely ignores the sensor command (From OpenBuild Control I see that it activates), after a while the Z axis stops, it rubs on the material to be cut to get to the first work point and arrived at the first actual cutting job, starting the Z Probe again from there everything starts to work perfectly, I can't get out of it, does anyone have any ideas on what to check? I will post a video in attachment, I also attach a photo of openbuild where you can see that the path will crawl on the piece to be cut Thanks to everyone for your attention! Video: https://youtube.com/shorts/TpmleqwXt1Y Edit: Post processor version V1.0.41
the machine must be homed before starting a job, before settting the work offsets even. Once the machine knows where home is (at the very least Z at top of travel) you can set the work coordinate offsets (where 0,0,0 is on the material) and then the Gcode will work correctly. I don't see home/limit switches in the video, so if you don't have them you will have to 'fake the home'. How to 'Fake the home'
There is something that escapes me, before it always worked without the need to configure a "FAKE HOME", I gave the XYZ=0 and then I started the job and it did the routine normally, I attach the GCODE that I'm trying to execute, maybe it can help
This just means that you have accidentally done the right thing in turning on/reset the controller with the Z high enough that it acted as a fake home. This time you did not. Your gcode is standard stuff as output by the Openbuilds postprocessor, which contains the following plain English: (the Gcode is just text, and contains many comments to help you undestand it, you should read it, knowledge is power) Code: (This relies on homing, see https://openbuilds.com/search/127200199/?q=G53+fusion ) G53 G0 Z-10 It has relied on homing for several years, so this is not new. Unless you want to be able to repeat jobs based on previous X0 Y0, you only need to ensure that Z is at the top before the controller turns on or is reset. That one simple thing will ensure that initial moves are safe and that the tool will not scrape the surface as shown in your video. Your video shows the initial G53 G0 Z-10 hitting the surface and stalling the Z motor, then the move to the start point, and only then the probe for the surface.
Now I understand where I was wrong, before I didn't have this problem because I had never used probing! I'll try to use the fake home, in the meantime I'll buy the limit switches, I'll let you know if there will be any good news
Also not related. Jog+Setzero or probe setting zero = same thing. Work coordinates. If you don't have homing switches get some. Fake the home belongs in the past decade just makes it harder for folks I wish David would stop recommending it already. Its a couple bucks for switches, not worth the aggrevation of running a machine that doesn't know where it really is in machine coordinates, especially when other coordinates systems all reference mpos. Do it right
Do you have any suggestions for endstops? Should I use shielded cables? I will upgrade the machine with sensors, in the meantime I work with the workaround proposed by David
Yes, Xtension Limit Switch Kit - onboard filtering, no shielding needed when used with a proper controller like a BlackBox Motion Control System X32 (bare PCB arduino will always be a culprit for EMI issues - cable won't really fix it either. A proper controller has active signal conditioning components onboard, optocoupled with noisefiltering)
