Much fuss has been made over the strength of 3D printed parts. These parts are obviously stronger in one direction than another, and post processing can increase that strength. What we’re lacking is real data. Luckily, [Justin Lam] has just the thing for us: he’s tested annealed printed plastics, and the results are encouraging.
The current research of annealing 3D printed parts is a lot like metallurgy. If you put a printed part under low heat — below the plastic’s glass transition temperature — larger crystals of plastic are formed. This research is direct from the Society of Plastics Engineers, and we’re assuming they know more about material science than your average joe. These findings measured the crystallinity of a sample in relation to both heat and time, and the results were promising. Plastic parts annealed at a lower temperature can attain the same crystallinity, and therefore the same strength, if they’re annealed for a longer time. The solution is simple: low and slow is the best way to do this, which sounds a lot like sous vide.
A while back, [Justin] built a sous vide controller for the latest cooking fad. The idea behind a sous vide controller is to heat food in a water bath at a lower temperature, but for a longer time. The result here is the most tender steaks you’ll ever have, and also stronger 3D printed parts. In his test, [Justin] printed several rectangular samples of PLA, set the temperature to 70°C, and walked away for a few hours. The samples annealed in the water bath were either cooled quickly or slowly. The test protocol also included measuring the strength in relation to layer height. The test jig consisted of a bathroom scale, a drill press, and a slot head screwdriver bit.
Although the test protocol is slightly questionable, the results are clear: annealing works, but only if the part is printed at a low layer height. However, parts with larger layer heights had a higher maximum stress. Is this helpful for the home prototyper? That depends. The consensus seems to be that if you’re at the mechanical limits of a 3D printed part, you might want to think about more traditional manufacturing. That’s just common sense, but there’s always room to push the envelope of 3D printing.
The layer-to-layer adhesion is usually the weak link. Hopefully there is some affect that works cross layers. Doing bunches of test samples with layers stacked in the Z-direction and getting an Instron or other pull tester would be a great step. Very important area of research. Keep at it!
interesting, I just did something similar last week with PETG plastic. I took a part, put it in an oven @ 81c for about an hour, and then let cool this did make the part stronger, but also introduced some additional shrink, about 1-2% after cooling. I took a hammer to both parts, the untreated part was able to break after several hits, the annealed part took much more damage and was able to withstand a lot of hard strikes before failure. the untreated part did break at the printed seams, whereas the annealed part broke at stress points of the part, and not alonge print lines. I was planning on another controlled experiment to see if tI get more consistent results, but wanted to share these results to reinforce than annealing on low temp after a print *does* improve the strength of the part after it hardens once again, but take care that there will be some shrinkage.
I don’t have problems with the methodology, but I do wish the sample size was larger.
What she said
Hmm, I am typing this on a peasant grade snaptop I got for less than £30 second hand:
I had the idea to prove people who say, “but you can get a new laptop for $100, why bother getting that old thing for $60???”
So far I am correct in getting a second hand lump-of-metal buisness grade c2d for cheap:
Origionally, the owner brought it and updates it (Windows 8) the updated driver didn’t support mounting the eMMC
upon boot so the snaptop would go to recovery and fail that aswell…
In short, I have a brand new $100-range laptop at about $40-ish.
I got it, and confirmed his claim.
So I got manjaro to work on it, sound et-al. After testing linux/gnu(s) on it and finding the one that mostly works and will allow bootable mmcblk devices, I installed the Manjaro and the RT5640 UCM template.
This usage was about a fortnight.
A few days after I get artifacts onscreen and a halted level crash.
This crash happens a few times so I try to reinstall windows 8 to sell it on, but it crashes again unlike before: definitively hardware. Loosened all the heatplate screws and it seems fixed (Will go sinkless when the failure happens in the future)
So people know, use case:
Goes into a rucksack and gets hauled to and from work, lighter than a real laptop (the ones that last).
Gets warped in thed bag whilst cooling due to weather, riding, what is in the bag at the time and general wear/use.
Lifting the screen warps the base in normal use, I can flop the trhing around by the screen at an est 10* angle from center to the warping edge (10in lcd to give idea of warp radius)
anmd th3 kje6yb oard is this hard to weri8te on, here is the letter [p:
so ik’ll noty p[r9oof resad asny6thing!
TL;DR:
Got cheap brand new laptop of the $100 category, it broke, sorta fixed, I AM right: get REAL hardware
I’ve experienced plastics that run hot tend to become more brittle, harder and more rigid, would performing treatment like this strengthen the case?
I feel like this is a copy pasta joke.. The only thing relevant to the article is the last sentence.
[Unferium] probably is going to sous vide a netbook.
They blather on about getting cheap hardware as if it’s something amazing in every post. Given I’ve refused more free hardware than they’ve ever seen, I assume their links are affiliate, or self promotion.
I got the snaptop cheaper than normal to prove a point, the yellow text isn’t links, but Too Long Didn’t Read [TL;DR] so you can skip the rant and get to the point easily.
.
.
.
On earlier posts (i.e. Libreboot post on HaD), I’ve seen people rant about decent 2nd hand laptops going cheap and they’ll say their $100 plastic laptop is better (1Ghz vs 2.5Ghz) and stronger (Plastic vs metal) and lighter (Ok they win this one point).
Another side note: I intended to reuse the hardware in another project when the hinges snap within the expected a-few-month lifetime.
then:
I finally get to the point of asking for confirmation: would heating the plastics strengthen the casing, thus making the thing slightly more reliable?
You would have to know for sure what kind of plastic the case is, there should be markings stamped on the ‘unseen side’ if it. A search reveals that most plastics including ABS can be annealed, but if it shrinks or warps then that would be a problem.
I’d worry that it would become more brittle, that a cheap laptop case would protect the device better than running the halves through the dishwasher.
No, it would probably not. Those laptop parts are injection moulded not produced through addative manufacturing. The plastic already cooled relatively slowly compared to a FDM parts and was probably hotter to begin with
My daily driver is a $40 Panasonic Toughbook CF-T8, has Win7 and I added a SSD so it can be tossed around. The battery life is great. I don’t know how its capabilities compare to your cheapo slaptop with win8, but I mostly putty into routers with it.
If you want a cheap laptop that can take the abuse, buy one meant to take abuse.
Unless you 3d printed the case, annealing will not make it stronger.
Sure just leave it in some water at 70°c for 2 weeks what could go wrong
I am confused. Wikipedia says the glass transition temperature of PLA is 60-65°C. I have had PLA parts warp when left on the parcel shelf of a car.
Excellent job, [Justin Lam]! I hope to see more experimentation like this, and nice website design.
from the article : “The inherent property of these parts is that they’re built layer upon layer, with different areas being rapidly heated and cooled at different rates. This causes internal stresses and they end up acting like perforated lines that are prone to snapping apart. The Solution: Heat Treatment”
I just watched one of Shapeways production video’s, and they mention that cooling the parts after printing takes about a day. I assume they print with pre-heated powder, or else there would not be much to cool down. So keeping everything just below melting point before and during the printing, and then do a very slow cool down may be an alternative to heat treatment. (Shapeways may do heat treatment too, maybe also to dry the parts after dying, but they have not mentioned it.)
No mention of significant dimensional changes (shrinkage) which I’ve seen mentioned in all YouTube videos where this was experimented with.
My takeaway from this is that my next printer will need to have a closed build chamber with external motors and electronics (CoreXY, bowden extruder) then that buildchamber will be heated to around 70°C and parts printed in hot atmosphere. After the print job is done, temperature inside the chamber will be held at temperature for some time before controlled cooling starts. This should give the strongest possible printed parts. The only thing that could get tricky is cooling the heatbreak of the hotend (watercooling or external air supply?) and how to realize a part cooling fan in the hot atmosphere to make shure the solidification is quick enough to get clean prints.
This is contrary to the practice of using a cooling fan to give a rapid transition of the plastic to a solid state.
A more repeatable method of testing the failure point of the samples is to simply use a bucket as a weight and slowly fill it with sand or water until it exerts enough pressure to cause a failure in the plastic (using whatever jig you like to apply the force) and then you can measure the weight or volume of the material in the bucket and get a very repeatable value for comparisons
.
Yeah Justin!!!!!
Sure, works great for little rectangular test pieces. Try putting a real part through the test. Any non symmetrical parts will warp like crazy as the stresses on the different shapes and thicknesses of the part are relieved differently and with differing timing. Strong is of no value if that parts don’t work.
Iterative design.