- The model came out in fine shape.
- Supports held through the print.
← Back to projects
SLA Resin Printing · CHITUBOX · UV Cure
Printing in resin, one failed orientation at a time
My second project in the San3a Tech technical circles: learning SLA resin printing end to end. This page follows the process from a first print with a missing head, through a three-orientation experiment, to washing, curing, and the lesson that surprised me most: with resin, it can be hard to see a print fail until the job is over.
How SLA printing worksBackground from Menna's onboarding session
I started with an onboarding session delivered by Menna, who walked me through the machine and workflow before I used it. Unlike an FDM printer, which melts and lays down plastic, an SLA printer cures liquid resin with UV light. The build plate lowers into the resin tank, a light source cures one thin layer, and the plate lifts so the next layer can form. The model is built upside down, hanging from the plate.
Each new layer is pulled away from the tank and held by the layers already printed. That makes orientation and supports structural decisions, not cosmetic details: they determine whether a feature forms and whether the print stays attached to the build plate.
UV cureLayers hardened by lightInvertedGrown from the plateSupportsStructural, not cosmeticOrientationChanges what survives
Build plate descendingThe bed lowers into the resin tank to begin a layer.
Entering the resinSide view of the plate meeting the vat.
Seated in the tankThe build plate is ready for the layer to cure.
Attempt 1 — the headless printA short first trial
With Ahmed's support, I fabricated the first iteration from an STL of a human body. The result was not satisfying: the model printed without its head. I think that part may have failed around the supports and fallen into the resin, although I could not confirm exactly what happened. I closed this iteration as a small learning trial rather than continuing to troubleshoot it.
- Started with an existing human-body STL.
- Prepared and ran it as my first hands-on print.
- Found that the body had printed but the head was missing.
- Ended the trial and reconsidered support and orientation for the next run.
Headless first printThe body formed; the head did not make it onto the final print.
The orientation experimentThree copies, three setups
For the next run, I turned the problem into an experiment. I set up three copies of the same model, each in a different orientation, so I could compare what survived rather than guess:
- Auto orientation — the angle selected automatically by the slicer.
- On its side — the body laid along its length.
- Upright / vertical — standing directly on the plate.
Menna helped me slice the file in CHITUBOX and gave me the sliced job ready to run. Comparing the orientations side by side made the trade-offs visible before the print started.
3 copiesOne model, three setupsAutoSlicer-chosen angleOn-sideLaid along its lengthUprightStanding vertical
Orientation view 1The three model setups in the slicer.
Orientation view 2Comparing supports and orientation from another angle.
Orientation view 3Final look at the layout before export.
Running the printBed prep, tank fill, overnight job
With the sliced file ready, I prepared and started the printer:
- Cleaned the build plate with alcohol to prepare it for the print.
- Filled the resin tank and started the job.
- Noticed the slicer estimated about 4 hours, while the printer displayed about 6 hours.
- Left it running overnight and returned the next day to check the results.
~4 hoursSlicer estimate~6 hoursPrinter displayOvernightLeft to finishAlcoholPlate cleaned first
ResultsTwo survived, one fell into the tank
- The model also printed well.
- A second orientation that survived the run.
- The print detached and fell into the resin.
- I found only part of its leg in the tank.
Two of the three orientations survived. The upright copy reinforced the lesson from my first attempt: orientation and supports can make the difference between a finished part and one that detaches. It also revealed how difficult it can be to spot a failure while an SLA job is running.
Final result 1One of the finished resin prints.
Final result 2Another view of the finished prints.
Final result 3A further view of the finished prints.
Post-processing & cleanupWash, cure, and leave the machine spotless
The work was not finished when printing stopped. The surviving parts needed washing and curing, and I had to reset the station carefully:
- Removed the supports from the surviving parts.
- Washed them in alcohol to clear uncured resin, then let them dry.
- UV-cured for about 11 minutes total: first 4 minutes, checked the parts, then added 7 more minutes.
- Poured the resin back into its bottles. The first bottle overflowed slightly, but the loss was small.
- Cleaned the whole machine with alcohol, checked every surface, and returned the tools to their place.
Supports offBefore washingAlcohol washThen dried~11 min UV4 min + 7 minResin returnedTank emptied & cleaned
What I learnedTakeaways for the technical-circle review
- Supports and orientation matter. The missing head and the lost vertical copy both pointed me back to support placement and orientation.
- Test orientations in parallel. Printing three poses at once turned a guess into direct evidence of what survived.
- Upright parts can be risky. The vertical copy detached and fell into the resin tank.
- SLA progress is hard to see. Unlike an FDM printer, it is difficult to judge progress by eye; a failed or dropped small part can be hard to notice in time to stop or restart.
- Time estimates can differ. The slicer estimated four hours, while the printer showed six. Plan around the machine's displayed estimate.
- Post-processing is part of the print. Washing, staged UV curing, careful resin pour-back, and a full machine clean are the real finish line.