Correct, it makes no sense to look at percentages when there are highly nonlinear materials effects that occur within specific narrow temperature ranges.
So which percentage is more useful, Windyās at 7% (0C to 100C) or the physics professor at 1.9% (-273C to 100C)?
Taken in isolation, without any way to know critical thresholds, I would argue that neither is useful. If someone worked out the critical thresholds through experimentation, Iād probably go with the Kelvin number, if only because no thinking required. With the other approach, it might be fine, but Iād have to think harder to be sure. So, based on expediency and conservation of brainpower, Iād go Kelvin. Thatās not a proof that Celsius would be wrong though. If you like puzzles, and have nothing else to do, see if you can prove it. Iām guessing you probably can.
[Edit: I think we can all agree that Kelvin would be objective truth. As a possible quick answer, Iād try to see whether I could find two pairs of celsius values with the same percentage difference in Celsius but a different percentage difference in Kelvin. If I could, then that would be proof by contradiction that the Celsius method isnāt as accurate. ]
1 Like
Nope. For the weather, you want Celsius, which is tied to the freezing and boiling point of a particular, if important substance - water. For properties of materials, you want the (real) absolute temperature
Well, thatās some fine gibberish. Celsius uses the same ādegreeā as Kelvin, just offset by 273.3 degrees for convenience in day to day usage.
Yes, but the original question (if you scroll back) had to do with which unit you did the percentage differences in.
Iāll defer and let the people who actually care about this argue it out.
This is the wonder of the internet - that people think their opinions - on matters of which theyāre (obviously) wholly ignorant - are worthy of discussion. I did the percentage difference in Kelvin, which is the relevant scale for discussing the physical properties of materials. Iām an engineer with 45 years of working and teaching experience. Itās not my āopinionā. I certainly ācareā about the answer, but because itās correct, not because itās mine. Any competent person would give the same result.
1 Like
The percentage is a fact, the relevance is your opinion.
Itās my opinion that the percentage difference is irrelevant when important physical transformations take place within very narrow temperature ranges.
butā¦ without reducing it to an idiotic % difference, how else can we over-simplify the issue which may not even be an issue???
Personally, I think the Celcius āextra idioticā % difference really draws attention to the serious concern for those who may not even have the printer yet and are basing all of this on day-one youtube ādrivelā.
I feel left out, practically attacked, because I have the printer and it just isnāt giving me any issues with big parts yet, unless I purposely use the worst material. Iām really glad this no-name glitter PETG is finally giving me adhesion failure over a 450mm long test part 
trying to cut through all of the chatter in this thread. Are there plans on trying to correct this through the firmware. Mine is way off by it is fairly uniformed by off by about 15 degrees. as a temp fix i have edited the fdm_filament_pla file in configuration to up the temp. so i dont have to change it or have a thousand filament profiles. however when printing from handy it does not carry the information over even when i click use my profilesā¦ so i fix would be niceā¦
I think weāve been trolled over this percentage question, and the tip-off was all the heavy bluster that accompanied it, together with a lack of mathematical proof.
So, here, Iāve got a counterproof, based on math, which says that itās the absolute difference, not percentage difference, thatās the better predictor of warpage. Itās not my proof, but instead chatgptās. Therefore, after presenting it, Iām totally washing my hands of this. Iāll give points if someone finds a flaw in its argument, but zero points or even negative points if itās just a general whitewash smear. Hereās my query to chatgpt, and itās response. If somebody wants to take issue with it, go ahead, but start your own thread about all your credentials and everything thatās wrong with the argument, and how stupid everyone but you is and blah, blah, blah. If you want to, go ahead and post one link to your thread from here, and then anyone who is interested can follow you there and you all can hash it out to whatever level you wish there. However, thatās the last I want to hear about it on this thread, because as OP Iām just going to flat-out say that abusive rhetoric doesnāt belong in my thread. I am 100% sure any of the moderators will agree, so consider this a warning.
My Prompt to chatgpt: which is the better predictor of warpage? Percentage difference in temperature, based upon kelvin measurements, or just simple difference in temperature?
ChatGPT said:
The better predictor of thermal warpage is the absolute temperature difference in Kelvin, not the percentage difference. Hereās why:
Key Principle: Warping is driven by absolute temperature differences, not percentage differences.
1. Thermal expansion is linear with absolute temperature change (ĪT):
- The formula for thermal expansion:ĪL=Ī±ā
L0ā
ĪT\Delta L = \alpha \cdot L_0 \cdot \Delta TĪL=Ī±ā
L0āā
ĪTwhere:
- ĪL\Delta LĪL = change in length
- Ī±\alphaĪ± = coefficient of thermal expansion
- L0L_0L0ā = original length
- ĪT\Delta TĪT = absolute temperature difference (in Ā°C or K)
This means 10Ā°C = 10 K of expansion effect, whether the object is at 20Ā°C or 200Ā°C. No percentage involved.
Why percentage difference is misleading:
- Letās say your print bed is 100Ā°C (373.15 K) in one spot and 110Ā°C (383.15 K) in another:
- Absolute ĪT = 10 K ā relevant for expansion
- Percentage difference = ~2.7% ā but that doesnāt reflect physical stress directly
- Now suppose you had 20Ā°C vs. 30Ā°C:
- Also ĪT = 10 K, but the percentage difference is 50%.
- Still, the warping potential is exactly the same if material and ĪT are the same.
Summary:
| Metric | Usefulness for Warpage | Reason |
|---|---|---|
| Absolute difference (ĪT in K or Ā°C) |
Accurate predictor |
Directly drives thermal expansion |
 Percentage difference |
Misleading |
Depends on reference point, not physical expansion |
So:
Would you like help applying this to a real-world printer bed measurement?
4o
You said:
double check
ChatGPT said:
Double-checked and confirmed ā in thermodynamics, material science, and mechanical engineering, the absolute temperature difference (ĪT), measured in Kelvin or Celsius, is the correct and physically meaningful predictor of thermal expansion and thus warpage.
Sources and Scientific Basis:
Thermal Expansion Equation (Standard Engineering Textbooks):
ĪL=Ī±ā L0ā ĪT\Delta L = \alpha \cdot L_0 \cdot \Delta TĪL=Ī±ā L0āā ĪT
- ĪT is the absolute temperature difference ā not a percentage.
- Used across physics, engineering, and FEA (Finite Element Analysis) simulations for stress prediction.
Reference: āEngineering Mechanics of Materials,ā B.B. Muvdi & J.W. McNabb ā standard in structural analysis and thermomechanical design.
ASME and ASTM Standards:
- ASTM E831 (for thermal expansion measurement) and similar ASME guidelines rely on absolute ĪT.
- Neither mentions percentage temperature differences as relevant for mechanical deformation.
Even in complex simulations (ANSYS, COMSOL, etc.), the input is always ĪT, not %.
Why percentage is wrong:
- Temperature is not a ratio scale ā 0Ā°C doesnāt mean āno temperatureā, so % comparisons are meaningless.
- Example:
- 20Ā°C to 30Ā°C = 10Ā°C rise = 50% if you foolishly treat 20Ā°C as ā100%,ā which is scientifically invalid.
- 270K to 280K = 10K rise = ~3.7% ā but same warping risk.
Final Answer:
Use the absolute difference in temperature (ĪT in K or Ā°C) to predict or model warping.
Do NOT use percentage difference ā itās not valid for thermal stress or expansion predictions.
Would you like a visual diagram or thermal expansion calculator for a specific material (like polycarbonate, ABS, etc.)?
1 Like
The assumption of linearity is not always true, for example when a phase change is involved.
2 Likes
Just to update those who are curious about the developments, I printed the test model that the support sent me, and obviously it came out perfect.
However, I didnāt stop there, and I did other tests being very careful not to change anything except the bed temperature, but Iām really demoralized by the quality with which the prints come out, in addition to the ever-present warping I also noticed obvious problems in multicolor, with artifacts visible right in the layers where the H2D swaps the nozzles. Iām attaching some photos that speak for themselves, both for one and the other defect.
I sent everything to the support 10 minutes ago, Iām waiting for a response.
I want to point out that all the filaments were dried with ams 2 pro, the black, green, white and light blue are bambulab filaments, the red is azurefilm.
5 Likes
@mteomarch Did you see any improvement after adjusting your filament profiles, as we had discussed earlier in the thread? I was hopeful we had found your root cause, or at least part of it. If I were in your shoes Iād pull out all the stops to get to an absolutely plain vanilla configuration as a baseline and then work on troubleshooting that, if itās still needed. Otherwise youāre introducing more than one change at a time, and that quickly leads into uncharted waters, especially on a machine that according to you has never worked right in the first place.
Thanks for the advice, thatās exactly what I did. Factory reset and first complete calibration. Printed the test that the support sent me following their instructions, that is, position a bambulab spool and start the print without touching anything. It came out perfect.
Next step was to load the two red and blue filaments in the two ams, modified only and exclusively in the filament profile the temperature of the bed. Sent the print from BambuStudio and the result was, a warping less than before, but that line visible right in the layers where it swaps the hotends. Then I printed the large two-color basket, also in this case a filament for ams, profile with only modified temperature. result a very light warping on one corner but a print absolutely not up to the machine! I donāt know what to do other than wait for more info from the support. I only hope that these days when Iāll be away from home for Easter my H2D will be convinced that in its new home we love it!
@mteomarch May interest you:
