Auto Bed Levelling Glitches

[Roxy]

Thanks Roxy, like always I really appreciate your help! The new marlin_main compiled fine against my existing firmware. Manually running G28 like before homes the x and y and then repositions at the middle of the bed to home the z. Subsequent G28's do the same. G29 then probes from the beginning of the bed and then returns the probe once done.

The G29 should leave the probe over the last point measured. If you do the G28 and G29 by hand you will know if that is what happens.

You have:

#define Z_SAFE_HOMING

Enabled, so the G28 will move to where ever you specified (some where near the center of the bed) to make its measurement.

I was a bit excited, and loaded up a test cube and sliced it and clicked print after that G29. low and behold it crashed the y axis, which after crashing proceeded to extrude at the same offset distance. I chalked it up to me being a dummy and not putting the g28 and G29 code into my start gcode.

My existing start code looks like so:
...

For most people, the G28 should be followed immediately by the G29. There may be a reason to do something different. But in general, most people do a G28 and then a G29.

Would I also have to then adjust anything else in slic3r or otherwise then? I was a bit worried with effectively having the print head starting from the middle of the bed, which caused the Y crash, but it could be just cause I never had g28 and g29 in my start code...

I suggest you do what ever startup GCode you want to put into your slicer by hand, one command at a time. Do a G28 and then a G29. And then a M114 to see if it is at the location that look reasonable given where the probe is. If it is... You should be ok telling it to start a print. It is troublesome that it crashed your Y Axis. But if you do these three commands by hand, we will know a lot more about what is going on.

I just wanted to say thank you again too, especially for some of my dumb questions. I've admittedly been rushed with so many projects on the go, that I've probably over looked things I shouldn't have!

This stuff is really hard. And it is a shame to wreck all the hard work it takes to get a printer going just because nobody is there to answer 'simple' questions. Its only simple after you have been through it a few times.

[MiniMadRyan]

Thanks Roxy, to answer your questions, my routine has been connect printer, load repetier and then manually enter g28 and then g29. Like you said, doing so, after the sequence is done, the head remains roughly at the center of the bed. I then loaded a test cube, sliced it and hit print.

I will give it another shot tomorrow, add the g28 and g29 to my start code and see what happens. Interestingly, when the y axis crashed it did so for a few seconds, it was very strange as it was definitely triggering the end stop. Then it just kept printing. The only odd thing was after the crash and killing the print, any subsequent g28 and g29 commands resulted in it homing at 0,0 and not the middle of the bed. It wasn't until I cycled power did it correct back to the middle of the bed.

[MiniMadRyan]

Looking at my start gCode for my print last night, I'm leaning towards the g28 in my start code as part of the culprit. As I had run G28 and a G29 manually before starting the print, I assume then that the G28 that was included in the start code of the print erased any data collected by the g29 manually. Thinking that's what probably led to the crash and issues I experienced. Will definitely be fixing that tonight!

[Roxy]

Looking at my start gCode for my print last night, I'm leaning towards the g28 in my start code as part of the culprit. As I had run G28 and a G29 manually before starting the print, I assume then that the G28 that was included in the start code of the print erased any data collected by the g29 manually. Thinking that's what probably led to the crash and issues I experienced. Will definitely be fixing that tonight!

Yes, normally the G28 resets the bed level correction matrix to unity. But that shouldn't cause the Y-Axis to crash. The change in how far the nozzle travels in the X and Y distance should be very very small. One thing you can do besides the G28, G29 and M114 is to actually issue a command to move. You could say G1 X50 Y50 and see if it goes towards the origin. If that is OK, do a G1 X20 Y20 and see if gets even closer. And you can also just move one axis and say G1 Y5

[MiniMadRyan]

Alright, I'm losing my mind here....this is weird...

I fired up the printer again tonight, having not modified or changed anything from last night. Connected it, and my first command was a G28. The X and Y zero'd out, and then the Z took its probe, so after it was finished, the printer was at 0,0. It did not return to the centre of the bed like before. I then ran a G29, and it ran its probing from 0,0. I ran a few more G28 and they all homed at 0,0 without returning to the centre.

I then unplugged the printer, closed Repetier and started again. the first G28 did the same thing, 0,0. A second G28 then moved the head to the centre, and homes the Z axis there. G29 Probes from roughly the middle like what was intended.

So last night I reported that it was always returning to the middle of the bed, well...I guess it's not...it still always homes at 0,0 on the first G28, and only does it return to centre on subsequent G28's.


Edit.....CSI: Bed Levelling Edition is getting stranger...here's my output from Repetier....let's have a look:

In the screenshot, you can see i've connected the printer, and it's online at 7:52:45

Screen Shot 2014-11-06 at 7.54.07 PM.jpg

At 7:52:52 I issue an G28 command. The G28 Homes to 0,0 and takes the Z probe from that location. It does not return to centre at any point.
I issue a M114 at 7:53:14, and the results are: X:93 Y:131 Z:10.00

Another G28 is run after this at 7:53:27, this time, the X,Y go to 0,0 and then return to the centre of the bed where it takes the Z probe. The G28 completes and the extruder is in the middle of the bed.
My last M114 is taken at 7:54:01 with the results of X:93 Y:131 Z:10.00

Summary?....two identical commands, two completely different results on the physical printer, and yet the output are both identical.....WEIRD!


For fun, I then ran a G29 followed by a M114, the results being X:118 Y: 156 Z:4.95, This correlates with the position of the extruder after it's done it's final probe
I ran one last G28 and M114 after that G29, the results then being the same X:93 Y:131 Z:10

Screen Shot 2014-11-06 at 8.00.28 PM.jpg

So, I think I can understand why my Y crashed last night. When the print ran, it ran a G28 as usual, the resulting physical position of the extruder was 0,0. I can only assume that the values returned were in fact X93 and Y131, which, with the bed physically being at 0, would have caused the grinding crash when it attempted to move the bed.

Last Edit.....I'm an moron...I assumed that a single G28 command was the same as manually homing each axis, either through code or Repetier's on screen controls. Apparently I was oh so wrong....I can't believe I over looked that!.....well it no longer crashes, but when it prints, the head is still far too high above the print bed.....time to play some more!

[Roxy]

Last Edit.....I'm an moron...I assumed that a single G28 command was the same as manually homing each axis, either through code or Repetier's on screen controls. Apparently I was oh so wrong....I can't believe I over looked that!.....well it no longer crashes, but when it prints, the head is still far too high above the print bed.....time to play some more!

Ah... I think it should be the same. I couldn't read the attached screen shots because they were so small. So I bumped up the size.

But if it is working except for the height of the nozzle, that is easy to dial in and get correct.

At 7:52:52 I issue an G28 command. The G28 Homes to 0,0 and takes the Z probe from that location. It does not return to centre at any point.
I issue a M114 at 7:53:14, and the results are: X:93 Y:131 Z:10.00

So you issue the G28, it moves to (0,0), measures a point and then without moving any where else, a M114 claims it is at (93,131,10) ?
It kind of looks like it isn't moving to the right place for the G28 ????

[AbuMaia]

If he has safe Z homing enabled, and G28 does not move to the middle (or to his specified coordinates) then something is wrong. It should move prior to Z homing on every G28, not just a second command.

[Roxy]

If he has safe Z homing enabled, and G28 does not move to the middle (or to his specified coordinates) then something is wrong. It should move prior to Z homing on every G28, not just a second command.

Agreed... I'm hoping he can confirm that. If so, I'll start looking at the code he is running and see if I can figure out a way that is happening.

[MiniMadRyan]

I'm just working on some welding projects in between the snow here so I will post a video later today, but here's what I've observed

Power the printer and connect to repetier. At this point the head is in a random position on the bed.
I type in a single g28 and it goes to 0,0 and homes thru read at 0,0 and then retracts the probe and stops. It does not return to Centre of the bed

Any additional g28 will then return to the Centre of the bed.

What I observed last night was that after I connected, if I used repetiers controls to home the x and y axis first and then run my initial g28, it would then return to the Centre.

I assumed that a g28 command was the same as homing each of thru axis manually, when in reality I still needed to home the x and y axis before issuing a g28...if that makes sense!

[Roxy]

We might have to add some debug code so we know what it is thinking during the G28 process. It might take a few iterations of loading firmware but I suspect this isnt too difficult of a problem to isolate.