Separate names with a comma.
Discussion in 'Troubleshooting' started by Avrution, Jan 12, 2018.
Indeed, no idea how it gets worked out other than stock or mesh firmware
It essentially works like the stock firmware. Ended up back on Z1.0 like before. Odd thing though is 20mm cubes are now coming out 1mm short, though 10mm ones are fine. Still tweaking.
M565 does not work with Marlin 1.1.7, it was deprecated long before that.
Using M851. Most likely going to have to start a new thread because I just can't make sense of this thing at the moment.
Moved the trouble shooting portion here for @Avrution
So where it stands right now is if I don't use any autoleveling (No G29) and just print with my normal Z offset then my prints come out fine (20mm test cube). If I send the G29 before the print then I tend to end up with my print height being 1mm short (Z is 1).
Part of the problem is repeatability. I can print with the same code 3 times and have 3 different results. I have printed from .8 to 1.2 (offset) and they can all give the same results.
With MESH (AFAIK) you do it and then you are done. No further use of a G29.
The MESH is stored in the flash EPROM and used. Not the best source since I am still using oldSkool ABL on all of my printers (Marlin 1.0)
Correct. The "G29 S4 Z-0.350; offset for mesh positive is further from bed" isn't setting an autolevel routine its setting the "Z Offset" for the Mesh that is stored in EEPROM. If the G29 is messing with you and not giving repeatable results I'd assume you'd have a hardware issue (something sticking/binding) or a very odd bug. @WheresWaldo may be able to comment as I'm still in last years Mesh leveling firmware (because its not broke). bit at a loss on that one. Are you 100% certain....like 100%...maybe even more...that the z level switches are not sticking or loose on your machine? That is the only way I can come up with any kind of bed mesh that if you run the same gcode multiple times without changing anything the Z height changes.
Z switches not being repeatable is a very distinct possibility.
It is also likely why Robo no longer uses that on the R2/C2
Its the only thing that makes any sense from the information I'm wrapping my head around lol. Granted...it is Monday . @mark tomlinson I'd suspect your right, all the issues people have with z switches and seeing the "reliability" of Jims IR likely helped push the way
I am having a ton of problems with my z switches, that was part of my issue last week with the bed constantly moving up on the g28 command. I got them aligned again, but it has happened a few more times since. That is also when I noticed that the screw on the left Z stop is stripped, so I won't be doing much with that - unsure the condition of the right one.
With no adjustments my bed is almost 100% flat. I have moved around to different points and adjust the Z by .01mm with the display and at most the extremes are off by .02. My first time with UBL worked great, then after the PEI I changed the settings for the number of points that defaulted at 7 and moved it to 12, thinking it would be more accurate - I'm guessing that could be part of my problem during that go round.
If the firmware has enabled repeatability (M48), you could just test the repeatability of the probe. You would likely find that using the switches, the repeatability is very poor.
I see #define Z_MIN_PROBE_REPEATABILITY_TEST //--ROBO-BH
What will the M48 command do/show?
It repeats probing the bed and then it will show you the minimum, maximum, mean, range, standard deviation
Mean: 0.015125 Min: 0.013 Max: 0.017 Range: 0.004
Standard Deviation: 0.001278
X:114.00 Y:127.00 Z:3.02 E:0.00 Count X:9120 Y:10160 Z:241
Also, my G28 always ends on or very close to Z:3 - what part is it says is 3mm away? It certainly isn't the tip from the PEI
G28 lifts by 3 mm after probing, its built into the firmware. The numbers look pretty good with M48.
Here is the issue with a flexible build plate like used on the R1 and why you have to create a mesh in the first place. First, the switch triggers after the probe hits the bed and actually continues to move lower until the axis brackets lift off the switches. Then second, the bed flexes downward from the weight of the extruder. It flexes differently in different section of the bed.
The reason I forced the homing close to a corner of the bed is that the bed flexes a bit less neat the supported corners. Another reason to choose MESH or UBL is that you can adjust the actual probe points so that it doesn't have to fight the bed flex.
Just one note, do not use M851 and G29 S4 at the same time, you do not want to double up on the adjustment. With UBL I don't use either M851 nor G29 S4, I use home offset of M206 Z instead. You can then start with the inverse of the Z value you get with M206.
After homing issue a G1 Z0, then raise or lower the extruder until it barely touches the bed. Then issue an M206 command. At that point you will know exactly how far off the actual 0 point is from the 0 point that the printer thinks it is..
Not sure what the G29 S4 command does, but I have only used G29 L1. I have been using that along with the M851 though. I just went back and redid UBL with the defaults and of course when I add my M851 back in there I am dragging into the PEI.
These are the first three lines I have been using
G28 ;move X/Y/Z to min endstops
M851 Z1 ;set Z-Offset in mm (may need adjusting for your specific printer)
G29 L1; leveling loaded
So G28 - G1 Z0 - lift Z axis by .01mm at at time. Sheet of paper can be slid in and out with resistance at .2mm.
At .1mm I need much more resistance. Technically at 0 if I try I can push the paper under the extruder. It looks like it is on the bed though.
I also moved my G28 to the center as opposed to the corner - maybe that was a bad idea? I figured since my prints started there it would be the best place to home.
Printing does not rely on triggering the Z axis switches, homing does. that is why the offset exists the center flexes downward and the Z axis has to lift the extruder of the axis brackets, so Z0 with no adjustment is below the top surface of the bed.
In your example above, if you are at M851 Z1 and you still need to lift .2 mm to slide the paper under the nozzle then the printer still believes Z0 is 0.2 mm - thickness of paper below the top of the glass bed. So maybe M851 Z needs to be 1.15 or 1.16 to be accurate?
Right now when doing the G1 Z0 I have no offset set into the machine. This is just turning the machine on, running G28 and then G1 Z0.
Or does the offset not come into play outside of prints?