Hi Guys, Need a little help as I can admit I am a little stumped and one of our customers is having some issues we would like to get solved as soon as possible! They are running a standard OX with GRBL (one of the early 0.9 releases from memory) and a protoneer CNC shield on an arduino uno. I have given what help comes to mind already but would love a few more minds on the issue as I may have missed something very simple and keen to see him up and running again! The original issues: The homing function sometimes works and sometimes not? The homing cycle has begun to behave erratically after functioning fine all day, it moves a bit, sometimes not and sometimes it doesn't bounce of the switches The machine keeps stopping mid project, it did not do this to start with? The machine and work coordinates do not seem to be tracking the project At first it sounded like electrical noise issues as the customer recently installed a 300w quiet cut spindle (are they prone to putting out bad noise?) but after the below troubleshooting via email and a phone conversation I am not so sure about my assumption. How much cutting have you done before this issue popped up? No exact amount given but sounds like the rig was up and cutting before this issue popped up. Have you changed any hardware/settings since this issue arose (adding new spindle, wiring or machine settings)? No new hardware and only settings changed after issue arose. Is the machine just stopping or is it showing some sort of warning/alarm? No warnings or alarms but also noted that the issue was happening in similar locations on the table, so we checked all mechanical movement is free and as smooth as expected. Happening in the same place on the table to me suggests mechanical issues but as it was smooth could it be a pause/stop/hold call from within the Gcode or possible comms issue causing the arduino to run out of things to do maybe? Have you checked all cables are seated well and that you have no loose connections? All looking good What sort of cables are you using running out to your homing switches and have you grounded the shields at one end only? Shielded cables but not sunk to a star ground, so I have asked the customer to have a good read up on CNC grounding, start grounding and ground loop issues to help give some clear background and ideas as to how to hook them all up. I have suggested a single star grounding point within the enclosure and bond to it. Protective earth (from the wall supply) The earth lug from each of the SMPS All cable shields (but only from one end only) An earthing strap that runs to each part of the CNC to assure all exposed metal is bonded back to the protective earth pin by no more than 1 ohm (as per our local regulations). I have seen people bond 0v from the different power supplies also to the star ground point but always been a bit confused about that myself as in NZ I have electricians both agree with the practice and disagree. Can anyone tell me why you would and would not bond your 0v of your different power supplies to the star ground point? I have also asked for the results of M119 as to see if the switches are in the correct state as well. I had wondered if the machine was losing steps due to cutting/acceleration forces being greater than the stepper motor power but there is a strong fan to keep things cool and the machine stops and holds with no errors on screen, so not really sounding like a lost step/overheating situation either. Can anyone else suggest any other troubleshooting tests that could help track down this issue? Thanks in advance Krishna www.CNCkits.co.nz
after a bit more digging might just be hitting a soft limit... Seems that sometimes it is just something simple!
Fitting an ac noise filter on the mains supply to spindle as close to the spindle as you can get it I.e an inch or two, is always a good idea @ just a few quid each.
