What? More Of The Same?
Nope! After a year and a half with the printer(s) I find myself being asked for advice about buying a printer and what to do when that advice is taken. So, rather than repeating myself in email, I figured I'd post a quick update. If you don't have a 3D printer, I'll understand if you sto... hey - you already stopped? Well, ok.
Microcaliper ($20) - I'm not sure how I missed this the first time. I use this thing a LOT. Not just for prints, but once you can measure something to within 3/100th of a mm, the tool makes its own jobs: how thick is this credit card? Why does this paper feel so light? Is resume paper really thicker? The cheaper models only measure down to 0.2mm, but you want to get finer than that. You can spend more, but it won't really be useful here.
Nozzle/Cleaning Needle kit ($8) - Brass nozzles wear down and, more frequently, get clogged, so you'll want to have some spares on-hand. But before you toss a nozzle you think is clogged, try cleaning it. The needles in this kit are precision-made to make sure they don't enlarge the nozzle's opening. They're also a great way to test your new nozzles to make sure the opening is as big as it should be, which is just what I did after ordering a cheap pack of 20 0.4mm nozzles. In that bag, only 9 were ready to go immediately, 7 needed some assistance from a 0.4mm needle to enlarge the opening and 4 had holes too small to be opened at all. The kit I linked above also includes a variety of nozzle sizes. Why would you want a bigger one? They print faster at the expense of appearance. You can also go smaller to get really fine layers. Just make sure you get nozzles compatible with your printer. This pack will work with many Makerbot, Creality and Prusa, but I make no guarantees.
Splurge Nozzles (variable) - I've been printing for over 18 months and the only reason I've had to discard nozzles is clogging - I've never worn one out, but it is possible if you print a lot. You can spend a bit more to get stainless steel ($7 each ) instead of brass, which not only last longer, but will let you print with more abrasive filament like carbon fiber, fiberglass, etc. Make sure to read the reviews carefully - cheaper ones are usually junk. You could also go all-in and get a ruby nozzle for $100, but if you do, make sure you get a real one. The $60 nozzles appear to be junk too.
What To Print First
(Ok, maybe that should be "what to print second")
Sure, go ahead and print your Benchy. It's a rite of passage. After that, though, you're going to want to learn more about your printer and dial it in to get better, more reliable prints. There's a slew of parameters which can make a huge difference, so the only way to find those is testing: starting temperature, 1st layer temperature, bed temp, 1st layer bed temp, print speed, retraction distance, retraction speed, etc. Here are some of the tests I've found most useful:Retraction test suite : One of the easiest ways to ruin a print is stringing which is just what it sounds like: strings of filament all over the print - caused by molten filament oozing out of the nozzle while the head moves to the next place it wants to print. To prevent that, the printer can retract the filament just before it moves, but how much? How fast? Retracting too much will jam the nozzle. Retracting too fast can break the filament. Retracting too slowly or too little will cause more oozing. This pre-sliced test suite will help you determine speed and distance with a bunch of pre-sliced gcode files. Be aware that these use a nozzle temp of 200C, so if you want it higher you'll need to either edit the gcode (it's easier than you think), adjust on the printer while it's going, or - better still - use Octoprint as described in my last post. That said, here's how I suggest using the test suite for a standard print head (as opposed to all-metal), with a Bowden tube:
Print SpeedTest_speeds-6mm.gcode then pick the speed that looks best for the rest. All other things being equal, faster retraction is generally better because it saves time.
Using the speed you chose above, do the 2-10mm distance test to narrow down the distance range.
Now that you have the range, choose a finer distance test, e.g.: 2-5mm, or 3-8mm. Unless your printer is direct-drive, you probably won't be in the 0-2mm range.
With your chosen distance, redo the speed test at *that* distance.
Use that retraction speed and distance in your slicer. Temperature tower : Another important one, this time to find the best temperature. Higher temps stick to the bed better, but usually cause more stringing and tend to be a bit shinier. Of course you want the print to stick to the bed, so your slicer will let you print the first layer at a higher temperature. I usually go with 10-15C higher. Much like retraction, the optimum temperature will vary by filament brand, though it's pretty consistent within the same brand. Note, that this test is not pre-sliced, so you'll need to edit the gcode to adjust the temperature at each floor after you've sliced it. (Perhaps my next post should be a primmer on editing gcode. Hmm...)XYZ 20mm calibration cube : Helpful to find layer shifting and check accuracy with the micrometer I mentioned above. This is one of the first tests I do when trying a new filament brand or material.Stackable calicat - cuter version of the cube, which can also stack on top of each other. Spiral Thing : Excellent test of temperature and stringing. If you can get this to come out cleanly and without stringing, you know you've nailed it.
All-in-one test : Great way to test a bunch of things at once, especially how well your printer handles overhangs and bridging - basically printing in mid-air with nothing below it. A properly-tuned printer with good print filament cooling can bridge much farther than you'd expect. Make sure to read the instructions because it's going to require some fiddling with the slicer. Mini all-in-one test : If you're just getting started or don't want to deal with support blocking, you can start with this guy.
I use Prusa Slicer
and Ultimaker Cura
. Both support a wide range of printers so there's no need to stick with a slicer with the same name as your printer. Prusa tends to print faster, but Cura does a better job, especially for complex prints and things that need support - their tree supports are MUCH better than anything else short of having two nozzles (which lets you print supports with a filament that dissolves in water.). Both programs work fine with their default settings and are insanely tunable when you're ready to go deeper. They both output gcode, of course, and their save files are compatible, so you can go back and forth with them to see which you prefer.
A Few More Tips
I love Thingiverse and never cease to be amazed by what's posted there. However, sometimes you'll download a Thing and later when you go to print it you'll find that the README.txt file is useless so you don't know what settings the designer suggests. So, as soon as you download the zip file, open it up and look at the README. It should have the URL to the Thing and the full contents of the "Thing Details" for that Thing. If it doesn't - add it. Trust me, you'll be glad you did.
As for organizing your downloads, well, that's a tougher one. I split them into directories: tools, mods (printer modifications, gifts, models, etc. And in MacOS, I use tags on the Thing directories to note whether I've printed that particular item. Using Windows? Well, you could rename the directory instead.
As with any scientific endeavor, you only want to change a single variable at a time when trying to fix or improve your prints. One of the best things I've done is create a spreadsheet for everything I print, and I name the slicer's save file (*.3mf) with the slicer, printer, name of item and line number in that spreadsheet. e.g.: cura-e3p-sunglasscarclip.206.3mf
. Columns in the spreadsheet are: item, printer, slicer, layer height, wall count, infill %, print temp, 1st layer print temp, bed temp, 1st layer bed temp, print speed, 1st layer print speed, mm retraction, adhesion (brim, skirt, raft), slicer time estimate, slicer weight estimate, print spool, file name, changes since the last iteration (if any), start date/time, print time, actual weight, results. Sure, it's a lot to track, but this way I make one change to the print and I can see the results. I'm also using Octoprint
plugins Print Job History
and Filament Manager
, which keep track of a lot of that automatically, as well as taking a picture with the RaspiCamera so I can see it later. In case I haven't mentioned Octoprint enough, let me recommend it once again. Don't want to buy a Raspberry Pi? Run it on an old laptop! I did that for a while and it worked like a charm.