Heat Soak? Help me tune this bad boy. (H2D)

Bambu Lab H2 Series
1

After going through roughly 4kg of filament trying to optimize my prints, I’m starting to think I may have been chasing the wrong problem entirely. My current theory is that thermal expansion is changing the bed geometry enough after full heat saturation to cause the nozzle to scrape the tops of taller flat surfaces — and I’m curious if anyone else has experienced this.

The problem: I have a cylindrical model, 130mm diameter and 45mm tall. I’ve done countless flow and PA calibrations, dried filaments repeatedly, and tried every optimization I can find — with inconsistent results and a lot of conflicting information along the way. Interestingly, when I slice off the bottom and print only the top 5mm of the model, the surface looks great. Those same settings on the full 45mm model show what looks like nozzle drag on the outer edges, as if the nozzle is sitting too close to the surface at height.

My heat soak theory: I’m wondering if calibrations done on a relatively cool bed don’t reflect the actual geometry once the bed reaches full thermal saturation during a longer print. Metal expands when heated — even a slight cup or bow changes the relative nozzle-to-surface distance at height, independent of the model itself. Short prints may finish before the bed fully stabilizes, masking the issue.

Questions for the community:

  • The H2D has three bed leveling options — the standard calibration menu, high temp calibration, and the per-print checkbox in Bambu Studio. Which one is actually used during a print? Does it transition between them, and if I run a high temp calibration does it get used for the full print, or just when the printer thinks the bed is fully heated?
  • If flow and PA are calibrated on a cooler bed, will those settings still be accurate once the printer is fully heat soaked? And after 40mm of layers, will any small error compound enough to be visible?
  • Has anyone with an H2D actually reached a point where they feel dialed in and get consistent results on both large and small models?

I’m open to being completely wrong on the heat soak theory — just trying to find a direction that actually sticks.

NOTE: I ran this through AI, because i was rambling like crazy in my initial draft and it was hard to follow.

EDIT: In my subsequent reading, I’ve learned that this variation in bed height at different temperatures is common enough. I’m curious how people are solving this issue in the easiest way.

2

I had made some measurements on the calibration results and the high temp calibration kicks in as soon as the bed is at temperature.

You printed PA and you get nozzle scraping when you print your taller model rather than the thin one. It seems to me this is evidence of warping. Is it pure PA or PA-CF or GF? Its is not always obvious to see warping on PA especially glass or carbon filled, have you tried putting a ruler on the bottom surface, is it flat or does it lift around the perimeter?

3

Sorry, PA in this instance is referring to Pressure Advance or K value.

My prints have been in PLA Matte (Bambu Brand)

4

:slightly_smiling_face: :+1: Oops, sorry, OK so I guess no warping :wink:

5

Try running your next print on “silent” mode just to gain the data point - it will slow you down by about half and can make for a good qualitative comparison to see if you indeed do need to slow down for real.

6

Try running your next print on “silent” mode just to gain the data point - it will slow you down by about half and can make for a good qualitative comparison to see if you indeed do need to slow down for real.

I have run some prints on silent, and manually adjusted some acceleration and speeds manually. I did not have improved results, though my testing was not extensive for this.

7

Oops, sorry, OK so I guess no warping

Some warping of the plate, even if slight. But no warping of the filament, or at least that’s not what I was discussing in my original post.

8

Don’t underestimate warping of PLA Matte.

You might want to watch this video with some very good tips.

1 Like
9

And wood PLA also shrinks significantly, in my case always after removal from the print bed (cooled down).

10

That could also be an explanation.

11

Good watch, thank you. I have used an engineering level straight edge ruler on these parts in the past and had not observed any level of warping. However, there was a lot of good information in this post I can consider. Many of the features of the part I modeled do already have these deterrents baked into the print, but now I can be more aware of how they work and monitor the results a bit better.