Leveling Bed - Mapping Build Surface With Marlin

Is it possible to create a topographic map of the glass build plate surface utilizing multiple locations, for example the 4 corners, the center of the bed and maybe a few other points utilizing Marlin so the first layer always goes down the same layer height everywhere?

Due to the large number of variables like material type, bed temperature, how much the x-axis dips down in the center of the bed, etc it is difficult to always get a good first layer the first time every time so I am looking to eliminate one of the major problematic variables, namely unequal height of the hot end from the bed all across the bed, by creating a topographic map that the hot end can follow to create a uniform first layer.

Any suggestions on how to do this? A probe? The TAZ Mini bed leveling technique? Can Marlin even create a topographical map from input data?

You can use a probe to do this. There are a couple designs on thingiverse that will work with the Taz. Here are a couple of them:

The enhanced version of G29 will print a topographic map of your bed but it’s main use is to see the low and high spots your bed has. Then you make manual adjustments to get your bed as level as possible and if it still isn’t level the auto bed leveling code will fix it. I just level my bed as best as I can then I use auto bed leveling. My prints are way better than they were without bed leveling.

I’ll post pictures of my Taz using the Mini bed leveling technique now that I have my parts cut and assembled.

Techsavvy34, thanks for the info because I really want to get my TAZ 4 dialed in excellently.

I’m curious to see how well the Mini bed leveling technique compares to the other methods you mentioned.

Off the top of my head, I like your idea of a proximity sensor (your second link) taking measurements across the bed because the center of my bed is 0.15 mm higher than its corners which is a fairly high percentage (42.9%) of the nozzle diameter and it creates a drastic 71.4% error in my 0.21 mm first layer height (0.15/0.21 = 71.4%). This makes it nearly impossible to get a good first layer on large prints that cover a significant portion of the bed.

One of two things happens to mess up my prints:

  1. I raise the hot end so my nozzle doesn’t drag in the center of the bed, but then my first layer doesn’t stick in the corners.
  2. I lower the hot end so the first layer sticks in the corners, but then it drags or creates ripples in the center due to too much pressure against the nozzle.

I cannot manually dial in a good happy medium nozzle height with such a large discrepancy between the center of the bed and the corners so I’m hoping one of the bed leveling techniques solves my problem. Any guidance you can offer is appreciated.

I look forward to seeing your results.

Many people, including myself, have thought the center of the bed was high. In all the cases I have seen explored here on the forum, the problem has actually been that the x rods flex a bit when the extruder assembly moves to the center of the x axis. You can check this by rotating the heated bed 90 degrees and checking you measurements again. You will likely find them to be pretty much the same.

My extruder tip runs about 0.010 inches closer to the bed in the center than at the sides. It has done it since day one. The printer needs stiffer x rods to correct the problem.

I realize this doesn’t solve your problem but it might help you to understand what is going on with the printer.


Thanks for your help. Certainly the TAZ Mini Auto Bed Level is very eloquent (https://forum.lulzbot.com/t/update-video-of-taz-auto-bed-leveling-mod-based-on-taz-mini/1063/1), but given my situation I’ll take your suggestion and lean towards a probe.

  1. In your opinion will your proximity sensor (http://www.thingiverse.com/thing:617615) auto bed level give similar accuracy to the TAZ Mini in the four corners?
  2. Of the three probes you mentioned which one do you think is easiest to build and integrate into the TAZ 4?
  3. It looks like your proximity sensor can stay rigidly fixed in one spot which seems like it would be an advantage, but if it doesn’t actually touch the bed I’m wondering how accurate it can be?
  4. Lastly, I have 0.010 PEI on my glass. Do you know if capacitive or inductive proximity sensor would be preferred?

For just starting out I would use the fan servo or 1013’s design. The proximity sensor is a bit more complicated. You would need a capacitive sensor because inductive only senses metal. You should also have a good understanding of the Marlin firmware because the proximity sensor is a little harder to configure. I’m not sure how accurate the cheap proximity sensors on Amazon are because I used one that cost over $100 and was made for factory automation. From what I hear endstops are pretty accurate.