Try to figure out MBL values

Printable View

12-01-2016, 05:55 AM
cabbagecreek

Try to figure out MBL values

I have tried to do Software that improves the manual MBL process.
I had a discussion with Scott Lahteine and he sad that Roxanne may have a clue (Roxy at this board??? )
This is why I try to post my question here.

My intension is to NOT change anything inside Marlin.
I want to begin with the Marlin standard configured and with MBL setup working.

Then I use the G29 to get the report of the FW configured and Manually done MBL. In my case the mesh is 3 x 3 points.
G29 and parse the responce to get number of x & Y ponts and the manually meassured values for each point
The point X1/Y1 has the Z value 6.975. (To keep the description short I concentrate to one single point)

Then I do the G29 S1 = Start probing mesh points.
The nozzle moves to the meshpoint..
M114 gives me the position of the first position ( X35 Y35 = X1 Y1)
Z is irrelevant now.
G29 S2 is Move to next MeshPoint. Redo the same procedure for all 3x3 points.

After this I got 9 points with X/Y where the nozzlepositions for MBL.

I recalculate the X/Ypositions so the PROBE will be at the exact position as the nozzle using ProbeOffsets X/Y

I do G30 on each of the MBL points and gets values. On the X1/Y1 point I get the value 3.01

Now the question is how I get the manual MBL value 6.975 from the probed value 3.01.
The most obvious solution is 10 - probed value = mbl value 10 - 3.01 = 6.99 ( close to manually value of 6.975)

But where do I get the 10 from.. Is this a fixed/hardcoded value inside Marlin or is it possible to configure in configuration.h

When I have the mbl points probe-values recalculated (3.01 become 6.99) I will do
G29 S3 X1 Y1 Z6.99 to adjust the MBL points with the probe-value.

I am sooo close to have a semiautomatic MBL done in software outside Marlin.
I hope that I can get some feedback if this approach is possible and will work for many kind of implementations of Marlin
12-01-2016, 08:04 AM
Roxy

Quote:

Originally Posted by cabbagecreek

I have tried to do Software that improves the manual MBL process.
I had a discussion with Scott Lahteine and he sad that Roxanne may have a clue (Roxy at this board??? )
This is why I try to post my question here.

My intension is to NOT change anything inside Marlin.
I want to begin with the Marlin standard configured and with MBL setup working.

...

I am sooo close to have a semiautomatic MBL done in software outside Marlin.
I hope that I can get some feedback if this approach is possible and will work for many kind of implementations of Marlin

You really should take a look at the Marlin UBL branch. It does everything you say you want, and much more. Besides automatically probing as much of the bed as possible (that the probe can reach), it gives you multiple ways to fill in the remaining mesh points. At that point the typical progression is to use the built in Mesh Validation Pattern to confirm how well you have the mesh defined. And any flaws can be quickly edited and the fine tuned mesh can be saved to EEPROM.

You can get it here:

https://github.com/MarlinFirmware/Marlin/tree/devel-ubl
12-01-2016, 08:31 AM
cabbagecreek

I have done all development against the RC7bugfix. How mutch of the UBL branch will be implemented in next release of Marlin.
12-01-2016, 09:16 AM
Roxy

Quote:

Originally Posted by cabbagecreek

I have done all development against the RC7bugfix. How mutch of the UBL branch will be implemented in next release of Marlin.

0% of it will be in the next release. The plan is to start RC-8 in the next week or so. And if that goes smoothly, to promote it to 'Stable' around Christmas time. I'm not sure that will happen because Scott is so busy. But right now... No new features will be put into RCBugFix / RC-8.

Once that branch gets promoted to Stable, the UBL branch will become the base for the next release cycle.
12-01-2016, 10:05 AM
cabbagecreek

So my idea is something that could work for the RC7bugfix and RC8 Marlin users for at least half a year.

The only thing left to figure out is how the probe values are calculated.. What I want to have confirmed is if my calculation 10 - probevalue = meshpoint is correct and if the 10 is a fixed value.

I have made a tool I call Marlin3DprinterTool that do some visual bed leving charts and also contains a Marlin Editor that add some "arduino-awereness". The MarlinEditor also contains a migration tool that help migrating a working current FW to a new FW (=RC8)
The Automatic MBL is something I have been working on for almost a month.. Its working but I want to e sure of my recalculations of the meshpoints.

I will download the UBL and test it during the weekend... Looked at all the features and like it..
12-01-2016, 10:21 AM
Roxy

Quote:

Originally Posted by cabbagecreek

So my idea is something that could work for the RC7bugfix and RC8 Marlin users for at least half a year.

The only thing left to figure out is how the probe values are calculated.. What I want to have confirmed is if my calculation 10 - probevalue = meshpoint is correct and if the 10 is a fixed value.

I'm not really using RCBugFix. I'm off in my own branch doing work. I know Scott has added the ability to save the mesh to EEPROM and I'm pretty sure the mesh can be automatically generated by using the Z-Probe. Check the latest RCBugFix to see what is there.