-
The 11-point checklist
-
1. Match the print core to the profile before anything else
-
2. Run active leveling — and clean the glass first
-
3. Confirm the material profile matches the actual spool
-
4. Verify nozzle diameter in Cura, not just on the machine
-
5. Start from the profile's layer height — don't hand-tune on day one
-
6. Scrub the layer preview before you slice
-
7. Check supports for anything past 45 degrees
-
8. Pick a build plate adhesion type on purpose
-
9. Confirm firmware and Cura versions are in the same room
-
10. Transfer the job one way — and stay consistent
-
11. Watch the first layer — actually watch it
-
1. Match the print core to the profile before anything else
-
What will still trip you up
This checklist is for anyone running an Ultimaker 3 or newer machine through Cura and getting tired of prints that fail for preventable reasons. Eleven checks, most of them under a minute. Two of them are the ones that cost me the most money when I skipped them.
I run additive manufacturing operations for a mid-size industrial design firm. I've been handling Ultimaker machines and print jobs for seven years. I've made and documented nineteen significant mistakes, totaling roughly $5,400 in wasted material, one ruined print core, and three blown client deadlines. Now our pre-print checklist lives on a laminated card next to the S5, mostly because I got tired of explaining the same failure twice.
It took me about four years and 300-odd print jobs to understand that the slicer profile matters more than the printer model (note to self: stop upgrading hardware to fix software problems). That's the whole logic behind this list.
The 11-point checklist
1. Match the print core to the profile before anything else
Ultimaker's swappable print cores (AA 0.4, AA 0.8, CC 0.6, and so on) each carry their own settings. If you load an AA 0.4 profile in Cura but the machine has a CC 0.6 sitting in slot one, the machine will either refuse the job or — worse — run it and produce something you'll throw away. In my first year (2018), I spent six hours chasing what I thought was a clog that turned out to be a profile mismatch. Check the physical core, then check the dropdown in Cura. They have to agree.
2. Run active leveling — and clean the glass first
Ultimaker handles active leveling automatically before each job, but the sensor reads the glass plate, not the air above it. A fingerprint, a glue stick smear, or a stray bit of support material will throw the reading off by just enough to ruin your first layer. I skipped the wipe-down once on a Friday and came back Monday to a nine-hour print that had been air-printing for six. Wipe the plate with isopropyl alcohol and a lint-free cloth. Every time. It takes twenty seconds.
3. Confirm the material profile matches the actual spool
This is the one most people skip, and it's the one that's bitten me hardest. Cura's material profile bundles temperature, flow, and retraction behavior together (the .inst.cfg file carries all three). Generic PLA, Ultimaker PLA, and Tough PLA differ by 10–15°C at the nozzle. Loading generic PLA and printing it at Ultimaker PLA temperatures gave me stringing, weak layer adhesion, and a bracket that snapped under hand pressure. Costs: roughly $40 in material and a two-day reorder. If the spool isn't the one the profile assumes, change the profile. Don't trust the label on the outside of the box.
4. Verify nozzle diameter in Cura, not just on the machine
The machine knows which core is loaded. The slicer doesn't always agree. If you swap to a 0.8 AA core for speed and forget to change line width, you get either heavy under-extrusion or blobs the size of blueberries. Check the nozzle size field in the Cura sidebar before you slice. It's the first thing I look at now, and it's still the thing I catch most often on a team member's screen.
5. Start from the profile's layer height — don't hand-tune on day one
Every stock Cura profile for an Ultimaker has a default layer height that's been tested against that nozzle and material. A 0.2 mm layer on a 0.4 mm nozzle is the reliable default. If you need detail, go to 0.15 or 0.1, not 0.05 because you feel adventurous. I once tried to hit a client's cosmetic spec on a 0.05 mm layer with a generic PLA profile and spent three hours dialing in a print that a 0.1 mm layer would have met in half the time.
6. Scrub the layer preview before you slice
This is the step that separates people who waste filament from people who don't. Drag through the layer view. Look for gaps, floating support, walls that are one extrusion wide, and overhangs with no support beneath them. Two minutes in the preview saves you hours on the plate. Honestly, this is the highest-return two minutes in the entire workflow.
7. Check supports for anything past 45 degrees
Cura's default support placement is decent but not aggressive. Overhangs beyond 45 degrees need support or a different orientation. I've seen good parts ruined because a 50-degree overhang looked fine in the model and lost its shape at layer 120. Scan the preview specifically for angled surfaces and confirm support is generating where you expect it.
8. Pick a build plate adhesion type on purpose
Brim, raft, or none. The Ultimaker glass plate with a glue stick handles most PLA jobs with no adhesion aid at all. But tall, thin parts need a brim, and parts with a narrow footprint need a raft. Don't leave it on the default and hope. Ask yourself what happens at layer 200 when the part is 12 cm tall and only 2 cm of it is touching the plate.
9. Confirm firmware and Cura versions are in the same room
Cura updates every few months. Firmware does too, but on a slower cadence. When they drift apart, menus move, settings appear or vanish, and a job that worked last week throws a warning you've never seen. According to Ultimaker's Cura release notes (support.ultimaker.com, accessed January 2025), version compatibility matters most around print core recognition and material profiles — which is exactly where silent failures hide. Check the release notes before you assume the menu layout is the same as last month.
10. Transfer the job one way — and stay consistent
You can send a job through Digital Factory or over USB. Pick one and use it as your source of truth. If half your team uses USB and half uses Digital Factory, you will eventually print a stale file that someone saved locally two days ago. On a 40-piece bracket order last September, we ran the wrong file on layers 1 through 80 because of exactly this. That mistake cost $890 in redo plus a one-week delay. Now the rule on our floor is: Digital Factory is the only source of truth, USB is for emergency recovery only.
11. Watch the first layer — actually watch it
Stay for the first five minutes. Not from your phone, not from the doorway. Watch the skirt or brim go down, check that the extrusion is smooth, and confirm the first layer is bonding to the plate. If it isn't, stop the job. In month 14 of owning our first S5, I got lazy and walked away — turned out the print core had a partial clog that air-printed layers 1 through 40 before I noticed. A $3,200 job became a $3,200 waste of time, and the client was not amused.
What will still trip you up
This checklist won't save you from every failure, and I want to be honest about the boundaries. I can only speak to our setup: we run mostly PLA and Tough PLA in a climate-controlled shop, on S3 and S5 machines, with a Cura version we update twice a year on purpose. If you're printing ABS in an open workshop, or running a Factor 4 in a humid climate, the calculus changes — temperature management and material storage will matter as much as any step above.
Two more things. First, Cura updates fast, and Ultimaker PLA pricing moves with the resin market. The $60–75 per 750 g spool figure I use for budgeting is accurate as of January 2025 based on reseller quotes; verify current pricing before you commit to a large order. Second, a checklist is a floor, not a ceiling. It catches the mistakes I've already made. Your first three or four failures will be new ones, and you should write those down too.
An informed operator asks better questions and makes faster fixes. That's the whole point of writing this stuff down. If you're going to skip one step on this list, skip step 5 and keep step 3. You'll thank yourself later.