The Problem With Strength

Why PLA is the 'strongest' 3d printing material and how to improve it

As a 3d printing hobbyist with opinions about the strength of different types of materials I’ve sometimes had discussions with people with Real Engineering Degrees who feel the need to mansplain to me trivia about the different types of strength and how I’m using the terms all wrong. It might sound weird that I’m being dismissive of people who have an actual degree and are using real scientific jargon, but this is a case of knowing a lot of trivia while missing the underlying point.

Within engineering there’s a range of measurement types between ‘fundamental physics’ and ‘empirical rules of thumb’, with weight on one extreme and coefficient of friction on the other. It isn’t that weight has no artifacts: Objects are affected by moving them close to other objects, and when in orbit the centrifugal effect changes it a lot, but it’s so close to the fundamental physics concept of mass that we’ve relied on it for making very prices scales since antiquity. Coefficient of friction on the other hand is a vague concept about ‘how well these two things mash into each other’, strongly affected by how long they’ve been meshed together, how much force has been pushed into them orthogonally, whether they’ve been moved already, the phase of the moon, and the purity of the soul of the person conducting the experiment. When building things which rely on coefficient of friction we run the numbers, add an order of magnitude to it, and then test the real thing to find out when it actually breaks.

Measures of hardness are much closer to coefficient of friction than mass. Take the Mohs scale, which is literally a metric of who beats who in a scratching contest. It’s more meaningful than IQ scores, but directly comparable to Chess ratings, which are at almost exactly the same point on the metric accuracy scale (but to their credit both are closer to real than the the metric accuracy scale itself).

For measuring strength there’s all manner of terms used, but they’re all different benchmarks based off what a particular measurement happens to say. The Shore measurement scale literally specifies the size and shape of objects to attempt to jam into the material and then measure penetration at different pressures. The different shapes have a fair amount of correlation but often deviate based on their size and spikyness. There’s no platonic ideal being measured here, it’s just an empirical value.

When you bend a ‘strong’ object it tends to snap back to where it was when you’re done but there are a lot of things which could happen during bending. Maybe it got little microfractures from the bending which build up at some schedule if you bend it repeatedly. Maybe it underwent some amount of plastic deformation. If it did maybe it lost some of its strength, or maybe it will stay a bit bent permanently. Maybe it undergoes plastic deformation slowly, and will snap back varying amounts based on how long you keep it bent, possibly on a schedule which doesn’t look very linear. The details are so varied and hard to measure that most of the measures of strength simply ignore the amount of material failure which happened during the test. This is an expedient but dubious approach. John Henry only technically defeated the steam engine. He died with a hammer in his hand. By any reasonable standard the steam engine won.

A material can be ‘stiff’, meaning it’s hard to bend in the first place, and it can be ‘tough’, meaning it doesn’t undergo much damage when it’s bent. PLA is still but not tough and tends to fail catastrophically. TPU is tough but not stiff. Nylon is both stiff and tough, but not as stiff as PLA or as tough as TPU. In general PLA+ is PLA with something added which makes it less stiff but more tough so it doesn’t undergo catastrophic failure. The downside is that it then undergoes material failure much sooner. This makes it do better on strength tests while making it a worse material for making real things out of. For some niche safety related applications it’s important that things fail visibly but not catastrophically, but for the vast majority of practical applications you don’t care how gracefully things fail, you care about them not failing in the first place. For that you need stiffness, and as boring of a result as it is PLA wins the stiffness competition against the other common and even not so common 3d printing materials. If your parts are failing you should design them more robustly not try to switch to some unobtanium printing material.

With that very disappointing result out of the way, the question is, what if you want to find some material which really is stronger than PLA? Having a 3d printer which could work with solder would be awesome, but there aren’t any of those on the market right now and I don’t know what it would take to make such a thing. Short of that the best material is… PLA. Even within PLA there are different levels of quality based on how long the chains are at the molecular level and how knotted up in each other they are, but as you may have gathered from the tirade about PLA+ about the PLA vendors are less than up front about the quality of their material and it isn’t possibly to simply buy higher quality PLA right now. Within what’s available now you can use PLA a bit better. If you print in a warm chamber and only slowly cool it down once printing is done you’ll get some annealing in and have a stronger final product which gets soft at a higher temperature. Ideally you’d repeatedly reheat the entire chamber to the temperature you’d properly anneal at and cool it back down again after every layer, which would result in an amazing quality product but take forever.