Marlin firmware fork for MakerFarm i3v w/ auto bed level & Roxy's enhancements

[clough42]

Doing a Topography report would be helpful. Something like G29 V4 P5 T That will give you a map of the height of the bed at different points.


I think that is tell you that the extruder is not moving perfectly flat or linearly.

Got a topography report. If I'm reading this correctly, I have a bow in both directions, which is probably a glass or mounting issue, right?

Also, since 0,0 is in the back, right corner, are the matrices rotated 180 degrees?

Send: M80
Recv: ok
[...]
Send: G28
Recv: ok
[...]
Send: G29 V4 P5 T
Recv: G29 Auto Bed Leveling
Recv: Bed X: 30.000 Y: 30.000 Z: 3.448
Recv: Bed X: 90.000 Y: 30.000 Z: 3.686
Recv: Bed X: 150.000 Y: 30.000 Z: 3.851
Recv: Bed X: 210.000 Y: 30.000 Z: 3.952
Recv: Bed X: 270.000 Y: 30.000 Z: 3.997
Recv: Bed X: 270.000 Y: 90.000 Z: 3.996
Recv: Bed X: 210.000 Y: 90.000 Z: 3.974
Recv: Bed X: 150.000 Y: 90.000 Z: 3.879
Recv: Bed X: 90.000 Y: 90.000 Z: 3.742
Recv: Bed X: 30.000 Y: 90.000 Z: 3.492
Recv: Bed X: 30.000 Y: 150.000 Z: 3.569
Recv: Bed X: 90.000 Y: 150.000 Z: 3.816
Recv: Bed X: 150.000 Y: 150.000 Z: 3.927
Recv: Bed X: 210.000 Y: 150.000 Z: 4.007
Recv: Bed X: 270.000 Y: 150.000 Z: 4.044
Recv: Bed X: 270.000 Y: 210.000 Z: 4.000
Recv: Bed X: 210.000 Y: 210.000 Z: 4.013
Recv: Bed X: 150.000 Y: 210.000 Z: 3.947
Recv: Bed X: 90.000 Y: 210.000 Z: 3.834
Recv: Bed X: 30.000 Y: 210.000 Z: 3.634
Recv: Bed X: 30.000 Y: 270.000 Z: 3.601
Recv: Bed X: 90.000 Y: 270.000 Z: 3.780
Recv: Bed X: 150.000 Y: 270.000 Z: 3.900
Recv: Bed X: 210.000 Y: 270.000 Z: 3.970
Recv: Bed X: 270.000 Y: 270.000 Z: 3.981
Recv: Eqn coefficients: a: 0.00186917 b: 0.00031375 d: 3.51421260
Recv: Mean of sampled points: 3.84165048
Recv: planeNormal x: -0.001869 y: -0.000314 z: 1.000000
Recv:
Recv: Bed Height Topography:
Recv: +-----------+
Recv: |...Back....|
Recv: |Left..Right|
Recv: |...Front...|
Recv: +-----------+
Recv: -0.24115 -0.06140 +0.05835 +0.12860 +0.13910
Recv: -0.20715 -0.00765 +0.10510 +0.17160 +0.15835
Recv: -0.27265 -0.02565 +0.08585 +0.16560 +0.20235
Recv: -0.34990 -0.09965 +0.03785 +0.13285 +0.15435
Recv: -0.39390 -0.15590 +0.00960 +0.11010 +0.15535
Recv:
Recv:
Recv: Corrected Bed Height vs. Bed Topology:
Recv: +0.07745 +0.14505 +0.15265 +0.11075 +0.00910
Recv: +0.13028 +0.21763 +0.21823 +0.17258 +0.04718
Recv: +0.08360 +0.21845 +0.21780 +0.18540 +0.11000
Recv: +0.02517 +0.16328 +0.18862 +0.17147 +0.08083
Recv: +0.00000 +0.12585 +0.17920 +0.16755 +0.10065
Recv:
Recv:
Recv:
Recv: Bed Level Correction Matrix:
Recv: +0.999998 +0.000000 +0.001869
Recv: -0.000001 +1.000000 +0.000314
Recv: -0.001869 -0.000314 +0.999998
Recv: ok

[clough42]

I just reran the report with the bed freshly heated and again after soaking. It warps another .05mm when heating and then settles somewhere in between. Time to start swapping glass and adjusting insulation, I suppose.

[Roxy]

Recv: -0.24115 -0.06140 +0.05835 +0.12860 +0.13910
Recv: -0.20715 -0.00765 +0.10510 +0.17160 +0.15835
Recv: -0.27265 -0.02565 +0.08585 +0.16560 +0.20235
Recv: -0.34990 -0.09965 +0.03785 +0.13285 +0.15435
Recv: -0.39390 -0.15590 +0.00960 +0.11010 +0.15535

That's interesting... I didn't know there was a Corrected Bed Height vs. Bed Topology:
section to the report. That is new. I'll have to look at the code to see what that is doing. But if the report is orientated correctly for your bed it is saying the front left is almost .4mm below the mean and the front right is .15mm above the mean. Going diagonally, you have .5mm of difference from front left to back right. And .4mm difference from back left to front right. That is a lot. The Auto Bed Leveling should be able to adjust for that, but you will get better prints if you physically adjust out some of that error.

At this point if you manually adjust your bed to bring all those numbers closer to zero, you will be able to see if there is a warp or a bulge in the bed. What will happen in the case of a warp or bulge is you will see a row or column start off +, go to - and then flip back to +. (or start -, goto +, and then back to -) If you see that on any column or row, you have a non-flat bed and there is some kind of warp. If you see that happen in both a row and a column, it will be likely the intersection of the two is the location of the bulge. But that will only show up on the Topological Map if you have it adjusted closer to level.

[clough42]

Hi Clough,
Do I take this post to mean that you've been running the most recent version of Marlin with changes to support your geometry and not my old fork? If so, that's great news.

Yes. At the moment, I still have a fork of your fork published and recommended for my customers, but I have been running the RC on one of my production printers for a while now, and it works great.

[clough42]

Recv: -0.24115 -0.06140 +0.05835 +0.12860 +0.13910
Recv: -0.20715 -0.00765 +0.10510 +0.17160 +0.15835
Recv: -0.27265 -0.02565 +0.08585 +0.16560 +0.20235
Recv: -0.34990 -0.09965 +0.03785 +0.13285 +0.15435
Recv: -0.39390 -0.15590 +0.00960 +0.11010 +0.15535

That's interesting... I didn't know there was a Corrected Bed Height vs. Bed Topology:
section to the report. That is new. I'll have to look at the code to see what that is doing. But if the report is orientated correctly for your bed it is saying the front left is almost .4mm below the mean and the front right is .15mm above the mean. Going diagonally, you have .5mm of difference from front left to back right. And .4mm difference from back left to front right. That is a lot. The Auto Bed Leveling should be able to adjust for that, but you will get better prints if you physically adjust out some of that error.

At this point if you manually adjust your bed to bring all those numbers closer to zero, you will be able to see if there is a warp or a bulge in the bed. What will happen in the case of a warp or bulge is you will see a row or column start off +, go to - and then flip back to +. (or start -, goto +, and then back to -) If you see that on any column or row, you have a non-flat bed and there is some kind of warp. If you see that happen in both a row and a column, it will be likely the intersection of the two is the location of the bulge. But that will only show up on the Topological Map if you have it adjusted closer to level.

Just realized I never came back with an update. I eventually got the bed shimmed to within .06mm of flat and it's good enough if I split the difference on the Z height.

After talking with a mechanical engineer at work, it seems that this is typical. Flat glass really isn't very rigid. It tends to take a bow, usually across the diagonal. Once it takes this bow, it becomes very rigid and stable, but it's very tricky to keep it entirely flat.

[dacb]

Just realized I never came back with an update. I eventually got the bed shimmed to within .06mm of flat and it's good enough if I split the difference on the Z height.

After talking with a mechanical engineer at work, it seems that this is typical. Flat glass really isn't very rigid. It tends to take a bow, usually across the diagonal. Once it takes this bow, it becomes very rigid and stable, but it's very tricky to keep it entirely flat.

This is almost certainly why borosilicate glass is preferred for the application. I'd be curious to hear from someone who has used it.

[clough42]

This is almost certainly why borosilicate glass is preferred for the application. I'd be curious to hear from someone who has used it.

You're hearing from one now.

This is what I'm using:

http://rover.ebay.com/rover/1/711-53...mtid=824&kw=lg

I have a high-power silicone bed heater, and I was exploding ordinary window glass. It turns out I was going a little too hot (140C) due to a thermistor configuration problem, but the heater was getting there so fast that the glass couldn't take the thermal stress and it was shattering. The borosilicate glass takes it just fine. I have since fixed my thermistor issue, but I still use the borosilicate glass.

[dacb]

You're hearing from one now.

This is what I'm using:

http://rover.ebay.com/rover/1/711-53...mtid=824&kw=lg

I have a high-power silicone bed heater, and I was exploding ordinary window glass. It turns out I was going a little too hot (140C) due to a thermistor configuration problem, but the heater was getting there so fast that the glass couldn't take the thermal stress and it was shattering. The borosilicate glass takes it just fine. I have since fixed my thermistor issue, but I still use the borosilicate glass.

Terrific information. So even with the borosilicate you see deformation of the glass plane? Interesting.

[Roxy]

That doesn't sound right. Whether the GCode comes from the host control program or from the SD Memory card, the commands get handled the same way.

I think there is something else going on here.

[Roxy]

As a side note to all the MakerFarm people running the BeckDac fork at: https://github.com/beckdac/Marlin

Here are my thoughts: This branch was very helpful and was extremely stable for the last 18 months. People with MakerFarm machines could just load it and use it. And for sure... It would just plain work. It was very stable! (Good job BeckDac!!!!)

Probably what makes sense now is if anybody with a MakerFarm machine has moved over to RC-7 of Marlin, we should set up a directory with your Configuration.h files and make it easy for people to get their machines running RC-7 too. If you have Configuration.h, Configuration_adv.h and a Pins.h file that are working with MakerFarm hardware, please let me know. I'll make a directory in the examples folder and people will be able to duplicate your setup very easily.