I am having HORRIBLE problems getting my M2 calibrated closer than 7mm which is wholly unacceptable at this point. My main problem has been that it is next to impossible for me to do the one thing that WebControl does: allow you to move the left/right motors independently so you can be assured that both of them are at 12 o’clock. I can’t find anything documenting how that might be accomplished in the console or else I’d just do that myself.
So the question remains, when will WebControl be able to control an M2? Holey-calibration and such would be much appreciated but merely getting the ability above would help me be 1000% certain my chains were all the same length.
M2 will not work with webcontrol. The DUE controller doesn’t talk to webcontrol. If you had a mega controller with your M2, you could use webcotrol, but that is not an option for sale. I’m not going to make webcontrol work with the M2 because I’m not convinced the M2 is actually any better than the Mega nor that I have the time or honestly the skillset to do so currently. I do have both controllers, so I can speak from some limited exposure to them. Plus a newer version of makerverse is rumored to be in the works, so I’d hang tight and see what comes out.
With the M2, you have to find dead center, calibrate, move it back to dead center, then mark it and then mark your chains and sprockets. I’m not clear on why or how it works like that, but it does. Webcontrol and groundcontrol count links to ensure your length. Makerverse lets you set it. The math is all in the open source software anybody can fix it, but it is fairly involved.
Makerverse is much more sensitive to a beam parallel to the workspace and evenly spaced proud from the frame. I found with makerverse that I couldn’t get very good calibration until my measurements were nuts on. measuring to some imaginary plane where the motor sprockets touch the chain and measuring between them, then measuring to the plane of the center of the motor shaft down to the workpiece to the nearest mm took me some time to get right. Perhaps you might try remeasuring? Someone recently said their calibration was sub 2 mm and the error was higher on a measured part, so I also question the true uncertainty in the system. I have not seen data anyone has shared of a systematic study to verify accuracy, but that doesn’t mean it has not been done.
I know little of the M2 set up but couldn’t a solution be to acquire a Keystudio Megga and a a regular maslow shield? i mean, it sucks to have to throw money at a problem but geez… that calibration error is horrible.
you totally could. The shield will run you about $40 typically unless you can find someone with a spare or maybe you could have one made with the plans that are out there. There are options. The due is supposed to be a smoother cut with acceleration planning and true GRBL compatibility that never made it into the Mega. It is a 32 bit processor compared the the Mega’s 8.
The issue with calibration is the centering. If it is off then the sprockets may not be lined up correctly unless they are spaced precisely the correct distance apart. Measuring from the outside of the board changes the reference point to make it more dependent on frame build quality than on the absolute chain position. I’m not saying it is wrong to do it that way, it just changes the reference points that are most important.
I actually went out and bought a tape measure that does metric just for making sure my measurements were spot-on, and that’s how I got the numbers as “low” as 7mm. What I think I’m gonna do is follow the general procedure for how GC/WC handle making sure the chain is centered where it starts with the first link of the chain on the sprocket(with each in the 12 o’clock), and then you jog the Y axis so you know for certain that both chains are 100% equal. It seems that’s the hardest part of the process with Makerverse because it makes an assumption about the initial length of the chains I simply can’t make apparently. I had actually thought of doing this in the beginning but dismissed it as useful since I was having trouble with getting my motors to move correctly(because I hadn’t “reset the chains”).
I did have a look at the Due firmware and it’s possible I could hack a command into it that would rotate the left/right motors independently.
I’ll consider getting a Mega but I’m going to try to fix the problems with the M2 in hopes I can help others down the road.
UPDATE
Ok, I remeasured Every. Single. Thing…TWICE. I ended up being off by about 2 cm on my motor height. I suspect I was using numbers from before I had to make an adjustment to the frame. Then I discovered that one of my motors was about 5mm lower than the other. I made that adjustment and I think I’m at 1.1mm accuracy. Now I can start doing some of my more simple cuts.
Does anyone know why it is that if you jog the sled over to what is actually home and then you zero out all the axes, the machine thinks you’re off by about 1 cm? I swear, I’ll zero it and the machine seems to report that I’m actually at -8mm. I’ll actually have to move it over -2 cm(in both directions usually) to get it to actually register where Home is. I’ve never had that kind of issue from my other CNC or any of the other software I use. I don’t know if this is a Makerverse-only thing or if it’s something with the M2 or what… If nobody knows I may post this to the troubleshooting section to see if anyone has any insight.
I haven’t seen that, but it doesn’t seem right. Maybe that is where the error is coming from. Maybe some pictures would help us to understand exactly what is going on?
Well, at least the software thinks I’m off by almost a cm. It literally won’t set itself at actual zero and instead will say it’s at around -4mm…usually up to -8 mm at times. This literally has nothing to do with my frame, etc. though because this is entirely in the software and/or machine. I know because if I leave it in the same place and tell it to zero the machine axes, it will start moving the location by 4-8 mm at a time, so if I hit zero 3 times, it’ll be -12mm off in the X and/or Y axes even with the machine having moved not the least little bit. I mean that last part literally. It won’t move but the machine will register as though it did. But if I can trick it to get it to accept where home actually is, physically and in the software, it’ll move right back there with zero errors every single time.
I’ve already got it set to zero and it isn’t showing any errors from this point so I’m not super keen on trying to break a system that is now working(even if by some weird voodoo) considering how long it took me to get to this point. If I have issues in the future, I’ll be sure to post pics, etc.