so, my physics professor approached me the other day with an idea to integrate 3d printers into the thermodynamics portion of the class. the idea is to come up with some good thermo problems to solve involving the printer that would be doable for a 2nd or 3rd year engineering major. so give me that thing you allways wanted hard data and calculation on, but never had the time to run the calculations or collect the data.

for your information the universiy that I go to has access to two printers currently, a Kittaz that I bought and built, and an old prusa i3 (with some modifications). there will be about 25 minions to work on your problems and your problem will most likely be rephrased as it will go through the prof. we do not currently have access to a thermal camera, BUT given a good enough reason I bet I could convince the guys over a FLIR (who work in the same complex as the University I go to, OIT ForTheWin) to loan us one.

so submit away! the crazier the better, but normal busy work is very good too.

I’ve recently had some odd issues with PID - my best guess is that it has to do with the E3d Chimera I’m now using.

Essentially, before the chimera, PID tuning was simple and generally worked with no issues - no temperature overshoot, maybe a little slow as it ramped up but otherwise fine. The Chimera, however, has two heater blocks connected to the same heatsink, with a fan on the heatsink. This, apparently, is causing havoc on the PID tuning - I just can’t get it to be happy. I had to settle with massive overshoot (like +10 degrees) and a bit more oscillation than I’d like, because the other option is it levelling out 3 degrees below the set temp - meaning with a ‘wait for temp’ command it never actually reaches it and just sets, heating, forever. Clogged a nozzle beyond repair that way yesterday.

There’s a lot of variables that seem to affect the PID algorithm in Marlin - PID_MAX (which is from 0-255 in the firmware, but appears to be from 0-127 in reality) which is the PWM duty cycle maximum value, PID_FUNCTIONAL_RANGE which is the range of temperatures that the PID algorithm will be active in, centered around your set temp, and PID_INTEGRAL_DRIVE_MAX, which I’m not mathy enough to really understand, as well as your usual P, I, D values.

I spent about three hours yesterday trying to get the settings tuned nicely, didn’t really have much luck.

What I’d like is some sort of double-loop PID, based on the set temps of BOTH extruders. So, for instance, if Heater 1 is set to 245°C, and Heater 2 is set to 0°C, it uses one set of PID values. If they’re both at the same set temp, it uses another one - and blends in between if they’re at different values.

Barring that, I think there’s some good room to grow in the temp tuning realm on Marlin. I don’t know enough about the math to make a good judgement on it, though.

Edit: Here’s a link to the Chimera page:

I’ve been meaning to run a few experiments. A FLIR would be great for analysis.

#1: At what temp does a 3mm plate of ABS become rigid enough to resist deforming when removed from plate? 3mm because, that’s what seems to produce a fairly rigid ABS plate. Therefore to reduce warping from contraction of the cooling upper layers, set the heated bed to the temp determined after 3-5mm.

#2: The rate at which ABS naturally cools. Or more basic, determine the distance from the top at which filament becomes less than the temp determined in problem #1.

#3: Determine the height at which the bed heated layers drop below the temperature in Problem #1. For example, a 50x50mm box with 3mm walls, 100C for the bed and no fan for cooling. With the print continuing at 230C, at some height there will be an equalization of temps (natural cooling of ABS, and diminishing effect of heated bed) in the “center section” of the wall.

#4: Now enclose the 50x50mm box with 50% infill and solid/diaphragm layers every 10mm. What’s the difference in cooling? The diaphragm should trap heat lower the rate of cooling.

#5 (last one): An experiment considering aerodynamics… What’s the temp difference between the different sides of the box above? I’m guessing the faces of the Y-axis will be cooler than the X-axis due to the bed movement. If you rotate the box, 45 degrees is there a difference in face temps? Or would they all be equal because all four sides receive equal cooling from the bed movement.

Okay… those results would be interesting… possibly create a good formula for creating “non-warped” prints. Of course, the ultimate answer would be an enclosure to maintain a consistent ambient print temp. :slight_smile:

I’d like to see the thermal profile across the print bed and how it’s affected by enclosing the printer.
Also the thermal profile of a tall and/or wide part as it’s being printed and how this is affected by the bed temperature, print speed, etc.
All the above with different filaments.

Finally a thermal analysis of what causes ABS to warp and what factors affect it such as bed temp, extruder temp, printer enclosure, print speed, etc.


  1. At what thickness does an Aluminum plate become a comparable or superior print bed to Borosilicate Glass under the criteria of Rigiditiy of the surface at temperature (initial, and over prolonged and repeated use), Lower mass, and thermal conductivity. For example, is a 3mm thick aluminum plate with a PEI surface superior to a standard Borosilicate glass print bed? If not, is there a thickness of aluminum that is?

  2. How do various thicknesses of PEI plastic sheet influence thermal adhesion properties of ABS plastic when overlayed over a given substrate?

  3. Is a thick (10-12mm) piece of PEI plastic strong enough at temperature to act as a print bed with a bed heater attached directly, and will a sufficiently thick PEI plate still conduct enough heat for good adhesion

  4. Lasers. Lots of lasers. I dunno, what happens to a 3d print when you shine all of the lasers at it?

ok, I am submitting all of these ideas to my professor this week and will hear back from her soon (a week or two).