By Dr. Andrea Allen
In my previous blog, I shared the excitement of printing my first shark fin and bringing the digital scan to life. This post dives deeper into that process. I compare TRAFFIC’s ultra-realistic nylon replicas with the Creality K Plus filament printer I use, and highlight how AI helped guide my learning.
Getting Started
Before beginning, I immersed myself in learning the basics of 3D printing—and I want to assure anyone new to this technology that it’s absolutely achievable. With the help of AI tools, Creality K Plus guides, and instructional videos, I mastered steps such as orienting files on the build plate, setting supports, slicing models, generating G-code, and safely removing finished prints. Though initially daunting, persistence and proper resources made success possible.
Comparing Printing Technologies: Filament vs. SLS
TRAFFIC, a leading NGO combating illegal shark finning and wildlife trade, developed ultra-realistic 3D shark fin replicas using Selective Laser Sintering (SLS) technology. This technology uses a laser to fuse nylon powder layer by layer into a solid, durable model without the need for support structures. The resulting texture closely mimics the sandpaper-like surface of real dried shark fins and improves paint adhesion, allowing for accurate species identification—an essential tool for customs officials worldwide. SLS printers are highly advanced and expensive, costing upwards of $20,000, which limits accessibility for smaller organizations or individuals. Pictures of their technology and prints can be found at their website, linked at the beginning of this paragraph.
Comparing Printers
| Aspect | Creality K Plus (FFF with PLA) | TRAFFIC’s SLS Nylon Printing |
|---|---|---|
| Printing Process | Melts and extrudes filament layer-by-layer | Uses laser to sinter nylon powder in layers |
| Surface Finish | Smooth but plastic-feeling with visible layers | Slightly rough, realistic texture |
| Cost | Affordable ($1,500 printer, $20 filament spool) | Expensive (SLS printers cost $20,000+) |
| Detail & Strength | Good detail, moderate durability | High detail and durability, very realistic |
| Support Needs | Requires supports for overhangs | No supports needed |
| Material | PLA filament (plant-based, eco-friendly) | Nylon powder, recyclable and highly resistant |
Comparatively, my Creality K Plus is a consumer-grade fused filament fabrication (FFF) printer costing about $1,500, using PLA (polylactic acid) filament spools priced around $20 each. Filament is a plastic “ink,” melted and extruded layer-by-layer to build models. PLA is plant-based, eco-friendly, and excellent for detailed educational models, though the prints feel smoother and more plasticky compared to SLS. Despite these differences, filament printing remains a valuable, affordable method to produce wildlife models for training and conservation education.
The cost difference is stark. While SLS may be preferred for ultra-realistic, tough replicas, filament 3D printing can make these models widely accessible for educators and conservationists on a budget. After mastering the mechanical aspects of printing, I turned to understanding the digital side—how 3D model files translate into physical prints.
Understanding File Formats and AI Assistance
Before I started printing, I needed to understand the files I’d be working with and how to prepare them for my printer. TRAFFIC’s 3D shark fin files come in STL format, a common type of digital 3D model that represents an object’s shape with tiny triangles—like a digital sculpture. This widely compatible format meant I didn’t have to convert files for my Creality K Plus program, which is where I prepare the prints.
However, simply opening an STL file doesn’t mean it’s ready to print. Often, the model needs to be rotated or repositioned on the virtual build plate—the flat surface where objects print—to fit properly within the printer’s limits. For example, “flattening” parts of the model against the build plate can reduce the amount of tree supports needed and save filament. Proper orientation not only optimizes material use but also improves print stability and reduces errors.
To help me through this entire process, I relied heavily on AI. I described my printer model and the challenges I faced, such as rotating or moving shark fins on the virtual build plate to ensure they fit properly. AI guided me on which functions to use to achieve these adjustments. It also helped me select “tree” supports to save material while stabilizing difficult overhangs—the parts that extend beyond previous layers—and assisted in slicing the model into layers that generate the G-code, which is the printer’s detailed, step-by-step instruction set. Even with careful preparation and AI support, trial and error remained an important part of the learning curve.
Overcoming Challenges
During my printing journey, I encountered a few challenges that are worth sharing for anyone starting out. One issue I ran into early on was orientation. I oriented a couple of prints upside down. This resulted in the fin name being printed on the bottom, with a rough and uneven surface, rather than on top with a smooth finish (seen in the picture below).
Why the rough surface on the bottom? The underside inherits the texture of the support structures and build plate. These contact points leave circular or layered patterns, which give a rougher finish. The top layers, however, print at slower speeds and often include an ironing pass for extra smoothness.
Another issue I encountered while preparing the STL files for G-Code is that several fins in the TRAFFIC set are simply too large for my Creality K2 Plus’ 350×350 mm build area. One is the Great Hammerhead dorsal fin. To fit the plate, I tried rotating it vertically and including substantial tree-like supports. The slideshow below shows the G-code files I prepared for this particular fin. The support trees are visible in green beneath the fin.
Unfortunately, the setup was unstable and fell over mid-print. When I discovered this, the printer was generating stringy threads back and forth instead of a solid model. I did not capture a picture or video of this, but the picture below shows what the incomplete print looks like.
After this failed attempt, I moved on to completing other fin prints that more easily fit to the plate. This past week, however, I revisited fins I previously skipped due to their large size. Rather than not print them at all—as they are intended as gifts for my partners at the Wildlife Forensic Academy and the set would be otherwise incomplete—I decided to scale them down.
The photo below shows the Porbeagle pectoral fin model, for example, overlapping and extending beyond the build plate. The parts of the green fin that appear grayish or shaded are the places where the print is outside of the allowable print zone. The red message box is an error warning.
To scale down the Porbeagle pectoral fin (original dimensions: 263.25 x 385.41 x 58.02 mm), I relied on AI. I asked it to help me determine practical scaled dimensions that accounted for my printer’s build volume constraints. The software proportionally adjusted the fin size to 341.91 x 300.30 x 37.54 mm.
The photos below show the resized model and how shape and details are preserved even with size reduction. The rescaled dimensions reflect typical biological variation and retain educational value by focusing on shape distinctions rather than strict size exactness.
This approach allows me to print all fins in the training set. The finished rescaled Porbeagle pectoral fin is shown below.
Lessons Learned
This project is as much about learning the process as it is about producing the fins themselves. Each challenge—from upside-down prints to scaling oversized models—offers insight into the patience, precision, and creative problem-solving that 3D printing demands.
One major lesson is the importance of orientation and stabilization. Even small adjustments in how a model sits on the build plate can determine whether a print succeeds or collapses mid-process. The Great Hammerhead dorsal fin taught me that supporting structures must not only be strong but well-balanced, especially with tall or irregular shapes.
I also learned the value of AI for troubleshooting and planning prints. By providing exact model dimensions and printer specifications, AI helped me rescale fins, choose more efficient supports, and reduce wasted material—all while maintaining realistic proportions.
Another practical takeaway concerns scaling for purpose. The primary goal of developing these prints is to create effective teaching tools. As long as the general shape and color variations of the fins are preserved, slight variations in size are, arguably, inconsequential.
Finally, every misprint provides data. Mistakes reveal where a setup, for instance, can be improved. These moments become valuable teaching opportunities about the capabilities—and limits—of 3D printers.
In the end, 3D printing is a technical and artistic exercise—a blend of engineering, experimentation, and persistence. What began as a test of equipment became an exploration of process, showing how innovation often grows through iteration.
Leave a Reply