Mike Kelly's Solenoid Auto-Level for Stock R1

Whats your Z offset set to?

That's bizarre on the Y. You might try manually tramming the bed a bit and make sure all the magnets are in place.

For tramming people have good luck just sticking the holes from a hole punch in them. Each of them is .1mm

 

The z offset is -0.6. I've got an early model that hasn't got the magnets, it's pretty much a beta but I've changed the draw sliders to linear bearings. I'm using a PrintinZ plate so I'll stuff things in there and it'll probably work, problem is that will lift it off so the glass so the heat transfer won't be as good :/. And if i do that I may as well not have the probe on there?

It's strange how it reads the bed position fine but then doesn't use the information when printing, is there any code that's meant to go after the G29 to save the measurements before printing?

 

G29 is the code for saving it...

 

Thought that may have been to just measure it. Oh well, I'll probably just not bother with the probe in that case as it doesn't seem to work for me :-(

Thanks for all the help though

 

It was obvious on my R1plus that the results from G29 were automatically having an effect, as the head would move from one X extreme to the other I'd observe the Z axis tweaking slightly as well. So it seems to know how to make this adjustment on the fly just by running G29.

For me though this happened because my carriage was a bit skewed left-to-right. I'd have to track down the video link but Novice Expert had a nice video talking about leveling that mentioned checking the G29 output and turning your Z axis rods to make the leveling more level. So I got the side-to-side readings coming back almost identical (took many back and forth very minor tweaks on the rod turning), and once I did this I could sweep the carriage from end to end and the Z rods didn't turn at all. It may not make a big difference in the end, if the leveling software does its job then it should just work, turning the rods basically makes the leveling process unnecessary I think. But I guess the less work the firmware has to do the better the chance for success.

Now front to back is a whole different story for me, but I think it's the whole way the R1plus does this. By lowering the carriage until it hits the bed, the entire weight of the carriage is now resting on the bed. On mine this causes a visible deflection in the middle 3 points. So the leveling comes back with results around 0, -0.2 and 0. But when it's printing it won't be putting all of its weight on the bed, so these middle points seem rather pointless. Seems to me that it should just probe the 4 corners, the other 5 points don't add much benefit. I also don't get the impression this algorithm really does anything about bed warpage, it just handles leveling but maybe I'm wrong. I certainly don't see the Z axis trying to compensate for this sag when I sweep the bed to its extremes.