Very Frustrated. First layer, every layer, problems

I am going to have to check the standard firmware again, but I believe that Robo still uses Linear auto-leveling, so what happens with Linear is this, they take 9 measurements then average them all so Marlin can create a single plane that fits most of the points, some will be below the plan some above the plane. Linear auto-level does a poor job of following the contours of the bed. Linear planes are only sloped in one direction. A step up is Bilinear auto-leveling, I had asked that they at least do this, but reverted the change without ever releasing A FIRMWARE WITH Bilinear enabled. Bilinear takes the same 9 measurements, but this time they create a plane than will encompass those points in two directions. Better but not perfect either. Unified bed leveling takes each group of four points and creates a plane that encompasses just those four points, then it moves to the next four and so on, you end up with planes stitched together that looks more like a topographical map. UBL also extrapolates the plane beyond the outside of the probe points instead of assuming that the bed is perfectly flat around the outside like Linear and Bilinear does.

That is why I asked them to consider the change, but they never did. They played with it some, then went on to some other firmware modifications with little to no value added for R2/C2 owners. Since then their GITHUB repository has been virtually silent on any changes to the firmware, almost as if they really want to put the R2/C2 behind them (this is just a personal observation and I have no evidence verifying this opinion).

 

Either is possible, but if it is the gantry then unless it is somehow specifically mangled in both X and Y in a fixed location I don't think you would get that sort of pattern. I will ponder this further though

It is usually the rather simplistic autoleveling method they chose combined with subtle warping in the bed.

If possible you should attempt as mesh level and try mesh leveling (or at least one of the better ones) : http://marlinfw.org/docs/gcode/G029-mbl.html

 

To increase the number of probe points would require downloading the firmware from Robo's GITHUB repository (https://github.com/Robo3D/Marlin) and change the lines in Configuration.h that specify the number of X and Y points, recompile then upload to your R2.

The lines will look something like this:

Code:

  #if RBV(R2)
    // Set the number of grid points per dimension.
    #define GRID_MAX_POINTS_X 3
    #define GRID_MAX_POINTS_Y GRID_MAX_POINTS_X

You would need to change the line that shows X points set to 3, to whatever number you want within reason, I believe the maximum for this leveling scheme is 7. Also note that you only need to set X as Y is then set equal to X.

Since I was working from memory earlier I misquoted Robo's Marlin leveling choice, at some point they did switch to bilinear which is still only marginally better than linear which I had said it was set to.

I honestly cannot tell you which branch is current since I am not using Robo's firmware. I believe it is the "RC" branch, but cannot say that with 100% certainty.

 

I am running the Community firmware with UBL enabled. there are a few people running it. Please note that some of the wizards do not work with the community firmware. If you run any of the offset wizards you will screw up the EEPROM settings. Also using this firmware is not for people that don't understand how UBL works. There are also changes that must be made to the OctoPrint scripts or else you will get bad results and endless frustration.

It would be a lot better if Robo just rewrote some of their LCD code (RoboOS) and enable UBL on their side. I have asked but they seem to be uninterested, instead focusing on worthless (my opinion) additions to Marlin code that add no value to us as endusers (again my opinion).

Disclaimer: I do not work for Robo, myself and the other admins are all Robo users and volunteers on this site. That is why I feel justified in giving my opinion outright.

 

You could always try something like putting a thin sheet of plate glass over the bed (cheap and quite likely to be more flat). Get it custom cut to your size at a local hardware store. I have done this on a few R1 style beds, almost always fixes the high/low spots for a few bucks.

 

@Gary Boyce Just for future reference, borosilicate glass is okay but not really a requirement for 3D printer beds. The thermal extremes used in 3D printing do not stress the bed enough to require its use. If a printer already has borosilicate glass then that is fine. If you are just putting something over the top of an aluminum bed, plate glass is cheaper and works just as well.

 

Yep, good old (cheap) plate glass like for a picture frame is the best choice

The kind you can get (did I mention cheap?) at a local hardware store.

 

It is probably more likely to be completely flat in a thin sheet too

 

Adjust the Z offset if the IR sensor sees through the glass.

Only needs to be a 1/8" sheet or so, that is easy to cover with the Z offset.

Or put small sections of tape at the spots it is sensing (under the glass naturally edit: actually over would work too if you make them small or use a black permanent marker)