Build platform woes.

[DBFIU]

So would the syntax for my first attempt look like.

G29 V4 n4 F75

And then get progressively smaller

G29 V4 n4 F50

G29 V4 n4 F40


etc...

Until I get the probe close to the front edge? And do the same with all sides?



I just tried to issue

G29 V4 n4 F75

And it is instantly returning the 0 0 0 0 0 bed matrix with no motion, and if I issue the full G29 V4 n4 F75 B125 L75 R125, it probes the whole bed. I can use the full syntax to tune the bed probing it will just take longer.

[DBFIU]

I just ran this line through pronterface.


G29 V4 n4 F25 B175 L25 R175

It hits 25mm within the ends of the bed. It does a pretty good job of getting the platform covered with samples.

Although I could push it further out, I think it would be getting too close for comfort.

I will be using that line in my Pre Gcode until the new release of marlin. I am ok with that.

Here is my pronterface returns.

SENDING:G28
>>>G29 V4 n4 F25 B175 L25 R175
SENDING:G29 V4 n4 F25 B175 L25 R175
Roxy's Enhanced G29 Auto_Bed_Leveling Code V1.81:
Full support at http://3dprintboard.com
Bed x: 25.00 y: 25.00 z: 3.07
Bed x: 75.00 y: 25.00 z: 2.49
Bed x: 125.00 y: 25.00 z: 1.98
Bed x: 175.00 y: 25.00 z: 1.33
Bed x: 25.00 y: 75.00 z: 3.29
Bed x: 75.00 y: 75.00 z: 2.93
Bed x: 125.00 y: 75.00 z: 2.50
Bed x: 175.00 y: 75.00 z: 1.93
Bed x: 25.00 y: 125.00 z: 3.60
Bed x: 75.00 y: 125.00 z: 3.26
Bed x: 125.00 y: 125.00 z: 2.83
Bed x: 175.00 y: 125.00 z: 2.31
Bed x: 25.00 y: 175.00 z: 3.99
Bed x: 75.00 y: 175.00 z: 3.63
Bed x: 125.00 y: 175.00 z: 3.24
Bed x: 175.00 y: 175.00 z: 2.70
Eqn coefficients: a: -0.01 b: 0.01 d: 2.98
Mean of sampled points: 2.817552


Bed Height Topography:
+1.17384 +0.81609 +0.42592 --0.12006
+0.78617 +0.43964 +0.01706 --0.50399
+0.46831 +0.10806 --0.32199 --0.88792
+0.25266 --0.32947 --0.83931 --1.48501


planeNormal x: 0.01 y: -0.01 z: 1.00


Bed Level Correction Matrix:
0.999956 0.000000 -0.009393
0.000073 0.999970 0.007731
0.009393 -0.007731 0.999926
echo:endstops hit: Z:2.70


The matrix seems to tell me that the bed is very flat, just very tilted, but that is OK with me.

Would like to hear your thoughts Roxy.

[Roxy]

I just ran this line through pronterface.

G29 V4 n4 F25 B175 L25 R175

...
Although I could push it further out, I think it would be getting too close for comfort.

Yes! You need to have both sides when looking for the boundaries. And for that matter, you need both axis too. The reason is if anything is wrong, it won't probe any points.

Usually, most people back off a little bit more and make the probed area a few mm smaller just to make sure the probe is well within the bed area. Missing the bed is horrible because the nozzle is going to get driven straight into the bed.

Here is my pronterface returns.

SENDING:G28
>>>G29 V4 n4 F25 B175 L25 R175
SENDING:G29 V4 n4 F25 B175 L25 R175
Roxy's Enhanced G29 Auto_Bed_Leveling Code V1.81:
Full support at http://3dprintboard.com
Bed x: 175.00 y: 125.00 z: 2.31
Bed x: 25.00 y: 175.00 z: 3.99
Bed x: 75.00 y: 175.00 z: 3.63
...
Bed x: 125.00 y: 175.00 z: 3.24
Bed x: 175.00 y: 175.00 z: 2.70
Eqn coefficients: a: -0.01 b: 0.01 d: 2.98
Mean of sampled points: 2.817552

This all looks good....

Bed Height Topography:
+1.17384 +0.81609 +0.42592 --0.12006
+0.78617 +0.43964 +0.01706 --0.50399
+0.46831 +0.10806 --0.32199 --0.88792
+0.25266 --0.32947 --0.83931 --1.48501

The Auto Bed Leveling code should be able to correct for this. But you have a 2.5mm difference in height from the back left corner to the right front corner. That is a lot! You should push the back left corner down a little bit and pull the right front corner up some. It should print OK the way it is. But you will get better adhesion and a better first layer if you get this more level.

Bed Level Correction Matrix:
0.999956 0.000000 -0.009393
0.000073 0.999970 0.007731
0.009393 -0.007731 0.999926
echo:endstops hit: Z:2.70

The matrix seems to tell me that the bed is very flat, just very tilted, but that is OK with me.

The Bed Level Correction Matrix only tells you about the tilt. But looking at the Topography matrix it does appear the bed is very flat. Looking across the rows and down the columns I don's see any big jumps. If you get the bed more level, you will get a good first layer every time.

[DBFIU]

Thanks for the clarification Roxy.

I just did the best I could to level the bed, since it is on hard fixed supports because I like to run my prints fast with high accelerations. And I generally print very heavy objects 0.5 lbm or more. So I can't always use the spring loaded screws method to do manual leveling, I literally have to dissasemble the entire bed and shim it with a machinists dial gauge that I use.

I just finished doing that and now the probe is showing me that the bed has a high spot in the middle somewhere, this is strange, because before it showed a very flat bed that was just tilted.

There is not stress on the bed, the screws that hold the bed down have zero preload on the bed, it is just floating and not being strained in any way. It is a good setup, hard to explain, but there should be no externally induced bending yet I see a high spot.

>>>G29 V4 n4 F25 B175 L25 R175
SENDING:G29 V4 n4 F25 B175 L25 R175
Roxy's Enhanced G29 Auto_Bed_Leveling Code V1.81:
Full support at http://3dprintboard.com
Bed x: 25.00 y: 25.00 z: 3.79
Bed x: 75.00 y: 25.00 z: 3.79
Bed x: 125.00 y: 25.00 z: 3.71
Bed x: 175.00 y: 25.00 z: 3.44
Bed x: 25.00 y: 75.00 z: 3.82
Bed x: 75.00 y: 75.00 z: 3.91
Bed x: 125.00 y: 75.00 z: 3.87
Bed x: 175.00 y: 75.00 z: 3.63
Bed x: 25.00 y: 125.00 z: 3.77
Bed x: 75.00 y: 125.00 z: 3.93
Bed x: 125.00 y: 125.00 z: 3.88
Bed x: 175.00 y: 125.00 z: 3.71
Bed x: 25.00 y: 175.00 z: 3.83
Bed x: 75.00 y: 175.00 z: 3.88
Bed x: 125.00 y: 175.00 z: 3.88
Bed x: 175.00 y: 175.00 z: 3.71
Eqn coefficients: a: -0.00 b: 0.00 d: 3.81
Mean of sampled points: 3.784238


Bed Height Topography:
+0.04386 +0.09622 +0.09372 --0.07705
--0.01098 +0.14857 +0.09996 --0.07830
+0.04012 +0.12613 +0.08375 --0.15558
+0.00397 +0.00522 --0.07456 --0.34506


planeNormal x: 0.00 y: -0.00 z: 1.00


Bed Level Correction Matrix:
0.999999 0.000000 -0.001186
0.000001 1.000000 0.000883
0.001186 -0.000883 0.999999
echo:endstops hit: Z:3.71

Let me know your thoughts,

Thanks

[Roxy]

It is easier to understand your high spots with more points. Probably a n5 would give you more information. And with the bed more level it is easier too. (But you do have it fairly level now.)

[DBFIU]

Have about a dozen prints on the new firmware, so far the first layers are coming out spot on. I brought my first layer height down to 75% and width to 135% and the brims and rafts look beautiful.