The following is a lengthy brainstorm about possible approaches to eliminating the problem of warping. Beware: long read.
Brook
I think you were chasing two different problems.
1. How to fight the warpage that results on large flat layers. The overhang test print presents a strong warping force over a large area from the top area shrinking as it cools. That force can be fought against by cooler layers below that hold it flat due to the stiffness below. Rules that apply here are related to cooling each layer to an optimal temp. All of these help: lower print temperatures, more air volume in fan cooling, optimizing the concentration of the air from the fan(s), thinner layers, slower speeds, wait time between layers.
2. How to fight warping that results on smaller flat areas. The spherical finned print suffers from a stack up of incremental warpage across many smaller layers. Same rules of cooling apply with one significant difference shorter layer print times means higher temps remaining in the layers below. You get multiple layers all retaining heat and cooling at different speeds but working together to warp. There is almost no help from cooler, stiffer layers below. All of the same tips apply. But it’s a purer test since there is no oppositional force of large stiff layers underneath. On this print, the total warping forces are laid bare. The relationship between tweaks in slicing speak clearly. You can’t cheat the results by utilizing the benefit of the counteracting force of stiff layers below.
Put plainly, the plastic has to cool and presents a predictable force. You can always work against that force by cooling each layer to a known minimum value.
I wish slicers would take cooling time into account automatically by mathematically calculating when any just-finished layer will reach a temperature that is cool enough to be stiff enough to proceed. Then all models would perform consistently and you wouldn’t have to tweak slicing settings. Print times would increase though.
An infrared camera that measures the top layer temp and automatically pauses the print until it is cool enough to proceed would be a simpler method since no mathematical model need be present.
Both methods need to know the right temp, which is different across the array of materials with which we print. The programming involved in both approaches is non-trivial of course. Also keep in mind that ambient temperature must be considered.
Current slicers ignore the forces of cooling.
Professional, enclosed machines provide a consistent ambient temperature that equalizes the differences between layer temperatures. A solution that doesn’t require an enclosed chamber seems more robust. Imagine if cheaper machines eliminated warping as well as the more expensive, enclosed chamber machines! Another differentiating factor between expensive and cheap machines would fall. It seems like low hanging fruit and it puzzles me why the reprap movement hasn’t done this. Probably because the cheapest solution seems to always win out even if it requires more knowledge and the added trouble of intuitively knowing how to counteract the built-in forces of warping inherent in certain models and certain orientations of certain models.
One last thought. I suppose that the cooling speed of a print coming out of an enclosed chamber is another factor in the equation. If you have even used an oven to anneal a part for better layer adhesion and strength, you have seen how badly parts can warp if cooled too quickly out of the oven.
There are so many variables in 3d printing. Manufacturers all focus on a different subset of problems. A few problems have barely been tackled yet since other common problems need addressing first. Some not-so-sexy problems remain, but will require costly programming, higher machine cost and a very scientific approach. Few manufacturers have the room in their budget to accommodate, but the payoff would be great.
