I keep getting a bad first layer with the H2D

Bambu Lab H2 Series
102

Judging from your video, there is a 0.23 lower region in the center of the bed, and the corners seem to be level.
So with tramming you cannot improve a lot (max 0.03mm), and also, this should not be an issue in my opinion.
Unfortunately I can’t judge if the rest of the bed is just as smooth, but it would be weird if that isn’t the case.

If the transitions between all the other locations are smooth, I don’t see why you would get such a bad first layer.

There might have been backlash in the Z-spindles or maybe the Z-guides that don’t run smooth.

I cannot imagine that the firmware is the issue, because my first layers are fabulous.

103

Of possible interest to someone following this thread’s topic:

104

Hi! So you fixed first layer problem only with updating?

105

Updating the firmware, then running the Full suite of calibrations, including the heated bed and anything that wasn’t checked by default. Takes a very long time, but it worked out for me. Hopefully you have the same results.

106

It works for me, the FULL suite of calibrations. Weird but it’s I’ll take it!

107

That is what was thinking

108

I recently started getting first layer issues on the H2S. I’ll report my findings in this thread and (hopefully) the solution.

109

Using a dial gauge and a revised tramming procedure is what finally resolved my first layer issues. I highly recommend it.

110

try turning your bed temperature up 5-10 degrees

111

Alright, here’s my full writeup of the situation:

I’ve been getting first layer problems on my H2S.

I tried numerous fixes recommended from different sources. I replaced the nozzle, I replaced the build plate, I recalibrated the unit, I trammed the bed, and I tightened the ceramic heating element. All of these steps had various effect, but nothing was able to get the bed to a satisfactory first layer, always having large portions of the bed being too low or too high.

At some point, I realized I wasn’t going to make much more progress unless I could get a satisfactory bed mesh. So, I affixed a dial probe to the print head and used some gcode to measure the heatbed countour after a heatsoak at 55C. I then visualized the surface. These are my results:

0 1 2 3 4 5 6 7
0 0.083 0.133 0.185 0.222 0.196 0.152 .143 .131
1 0.171 .089 .079 .089 .066 .038 .083 .196
2 0.254 0.109 0.044 0.046 0.024 0.008 0.081 0.266
3 0.319 0.119 -0.008 -.027 -0.003 0.016 .105 0.298
4 .308 0.106 -0.008 0.010 -0.010 0.014 .093 .228
5 .266 0.083 -0.010 -0.004 0.010 0.032 0.087 0.220
6 0.179 0.067 0.022 0.018 0.052 0.065 0.103 0.149
7 0.056 0.042 0.087 0.079 0.089 0.077 0.064 0.149

Based on the surface, it seems like bed has dip in the center and tacos upward significantly on the sides. This completely contextualizes why I noticed such exaggerated “wobble” (compensation movements) from printer during the first layer test.

This has been a great learning experience for me. The H2S has been my first printer. While it has been a great “set and forget” machine in the past, this issue forced me to do some real troubleshooting and appraise a crucial aspect of my printer.

I was not adept initially, so I cannot say what the bed mesh for my H2S when I first recieved it. Although, based on what I’ve read, this kind of variation is not problematic and 'ought to be compensated for by the bed calibration proceadure.

My plan going forward:

  1. Retram the bed. Previously I trammed the bed at ambient temperature. I am going to repeat the process I used (using the dial probe) after a heat soak at 55C. I’m hoping this will result in a more even tramming that accounts for heat expansion.

  2. I’m going to remeasure the bed mesh.

I hope this results in something a bit more even that the printer can fully compensate for in ~95% of the usable print area. I’m basing my expectations on what Bambu lists on their wiki. If I reach a point with minor defects in only a couple areas, then that’s a win!

I’m still sort of new to this, so I’m not confident that the contour of the heatbed itself is the root of this problem, but I feel it’s major factor. If I cannot reach a satisfactory state, I’m going to reach out to Bambu and see what they can do for me. Maybe I’m barking up the wrong tree, but hopefully it doesn’t come to that.

I like my Bambu printer a lot, but the added experience of this is making me feel like a more compentent printer, troubleshooter, and makes me feel excited about the idea of building a Voron/Ratrig.

UPDATE (9.5.2026):

So, I went through with tramming the bed at 55C, I thought that I managed to get things relatively level (insofar as the four corners).

Rotating counter-clockwise starting from the front right corner, I got these measurements: -0.012, -0.014, -0.018, 0.000.

Getting the corner measurements so close together was a bit painstaking, so I was reluctant to try to get things any closer together than this. I then reran the bed mesh .gcode and this is what I got

0 1 2 3 4 5 6 7
0 0.097 0.121 .165 .173 .157 .135 .124 .072
1 .151 .079 .056 .064 .048 .037 .068 .151
2 .211 .068 .022 .026 .012 .004 .074 .213
3 .262 .086 -.026 -.039 -.020 .020 .111 .266
4 .254 .088 -.018 .004 -.003 .028 .109 .256
5 .217 .068 .004 .010 .028 .006 .129 .228
6 .177 .070 .052 .051 .094 .115 .179 .207
7 .092 .084 .133 .139 .167 .167 .177 .260

It’s almost exactly the same as before, but this time the back right corner looks worse. I’ll need to recheck the corner measurements again, but still, this heatbed is a crater.

I’m running calibration as I write this, then I’ll try another first layer test. Hopefully we get something good. I’m also progressing my ticket with Bambu in the meantime. hopefully this get’s resolved.