Here is the configuration for my MP3DP. It’s based off of 2.0-bugfix, from 9/17/2018 (commit c5e5cc5e9f20e82cbd15b47a85bf08bbca542961, if you are particular). I view this printer as a sort of test platform/configuration fun house, so please don’t anyone take it as-is :).
Configuration-1.h (71.8 KB)
Configuration_adv.h (70.6 KB)
You can also get the matrix without running a G29. Use “M420 T4” to set the output format to “compact” followed by “M420 V” to print the stored mesh data.
Some progress here:
Autolevel is working (Zmin)
Home working (Xmin, Ymin, Zmax)
Whoa! That is a beast of a machine, and seeing it autoleveling makes it look so elegant.
So, have you done any cutting with it? Can you tell if zero in the middle is closer to zero at the edges? I’m also curious to see the numbers for your mesh. I am curious just how much flexing it’s doing over that span, and what other shapes there might be.
You made my day. Nice work!
Nice work, can’t wait to see how the cuts come out!
Hi. these are the values after the auto leveling process. I have a problem with the table that make the flex on pipes a minor problem.
Wow. That is great. That is an impressive amount of compensation.
Thanks for trying that. I will have to add this to my project list for sure.
I can’t find in the documentation how to read the leveling data. the cnc size is X 2800, and Y 2000.
home is in the lower left corner, X increase to right, and Y to up. When I start the autoleveling process, the first point is taken at x2770, Y0, the first row ends at x0y0.
Is the attached image correct?
Is there any software that graphs this data in a more visual way?
Octoprint has a bed leveling visualization plugin. I tried it on a pi zero though and it didn’t handle it.
I would put the data into excel, sheets.google.com or plot.ly and use that to plot a surface.
I think so. That’s what mine is, but it starts probing at 0,0 for me. I ended up doing the leveling, but I put a spacer on the second probe location. Then I could see which point that was, because it was a lot higher than the last bed level grid.
You probably also want a different number of X and Y points, right? It makes more sense to have them about equal in distance. That would help you make sure you have X and Y right.
Before going into the autolevel route, did you make absolutely sure that your problem comes from tube flex and not the table itself? On my machine, I have a marble surface and even this is not flat, I get more or less one millimeter between the highest and lowest spots.
Given the huge size of your planks and given the fact that it is not marble but wood, I’m pretty sure you will end up with issues because of this. In my opinion flex should be very minimal with a 2 meters tube.
If you are absolutely certain than this is a flex issue and not a table issue, then thicker tubes will help, that’s for sure. You will always have flex, no matter what, but it can be reduced dramatically. Tubes thickness is not the only thing to take into consideration, material is important too, there will be a big difference between, for instance, standard steel and chromoly steel, so you can also check that.
On my machine, I’ve used very thick chromoly tubes, especially made for guide rails on heavy duty industrial machines, the tubes were 2 meters long when I received them and there were no noticeable flex, I compared them with regular 25mm stainless tubes with thin wall and there was really quite a big gap in terms of rigidity, standard tubes looked like chewing gum in comparison.
But again, I think the main issue is the table itself, you might want to find a way to make it flat. Autolevel is definitely great, but given the size of your machine it won’t be very convenient because the autolevel routine might take even more time than the job itself…
Dui, didn’t you just figure out that you can save it in memory and only remeasure when there’s a problem? That’s what I assumed he was doing (or going to be doing).
you are right, the table is the main problem, the flex too, but is not that important
yes, the gird is saved in the eeprom
I did some research about this a while back (also not a MechE), and found this thread: https://cr4.globalspec.com/thread/66957/Ideas-for-Reducing-Sag-on-25-Tube which covers a lot of different strategies for stiffening long round tubing.
This reply in particular details what I believe to be a workable approach to stiffening the gantry tubes (if that end up actually being needed, an open question given the performance of the mesh leveling): https://cr4.globalspec.com/comment/707325/Re-Ideas-need-for-reducing-sag
If the bending loads are not high (so compression failure in the top region is not imminent) and a cable can be located inside close to the edge the pipe (ends attached to end plates etc.), the effect will be to make the beam bend around a axis along the bottom edge of the beam and not its centre line as before. This is called the neutral axis where the sum of compression forces and tension forces is equal.Bending stiffness is a function of beam depth squared and so in the perfect world the result would be to increase stiffness by a factor of 4.
Frosty that only works if the cable is independent of the tubes, but our tubes move. So the cable mount would have to move.
Independent it what way? The cable sits inside the tube and is attached to only to end plates which bear on only the tube ends. There are no external attachments.
That would exert pressure on tubes. On top of that the cable method will only off sag support not support in any other direction. The most efficient way to make stronger tubes is larger diameter. This comes up every week or so, and is even stickied in other forums. The tubes work as is for the intended size. Larger tubes, mean slower accelerations, more expense, etc.
