Restoring a Piece of Agricultural History: Recasting Smoke Stack Parts for an Early Rumely 12 HP Steam Tractor

Hello, fellow history enthusiasts and makers! Today, I'm excited to share a detailed walkthrough of a fascinating restoration project I recently completed. As someone passionate about preserving antique machinery, I took on the challenge of recreating the smoke stack castings for an early M Rumely 12 HP steam tractor. These tractors, produced by the M. Rumely Company in the late 19th and early 20th centuries, were workhorses of the American farm, powering everything from plowing to threshing. Unfortunately, original parts like the smoke stack components often succumb to damage, rust, and wear over time.

The specific parts in question are the bottom stack neck interface, which connects the smoke stack to the barrel of the smoke box of the boiler , and the top pieces that allow for the installation of a spark arrestor—a crucial safety feature to prevent embers from igniting dry fields. These cast iron pieces are robust but intricate, featuring flanges, mounting lugs, and precise curvatures to ensure a secure fit.

In this blog post, I'll document the entire process: from 3D scanning the originals to CAD modeling, 3D print prototyping, design finalization, and ultimately sand casting the final products. Let's dive in!

Step 1: Assessing and 3D Scanning the Original Parts

The project started with the original castings, which had seen better days. Pulled from a barn find, these pieces were heavily rusted, pitted, and cracked in places (with blacksmith repair), but they retained enough of their shape to serve as templates. The bottom neck interface is a flared ring that bolts onto the engine, while the top pieces include segmented rings with eyelets for securing the spark arrestor mesh. The bottom casting that connected to the boiler barrel was actually scanned in place on a friend's engine, without the removal of any component. It happened to be broken and fused together in 3 locations, however, an adequate profile was attained so that the flaws could be fixed.

To capture their geometry accurately, I used a high-resolution 3D scanner.

The scanning process involved:

• Cleaning the parts lightly to remove loose rust without altering the surfaces.

• Applying temporary matte spray to reduce reflections from the dark, shiny rust spots.

• Scanning multiple passes from different angles to build a complete point cloud.

The result was a detailed digital mesh of each part, accurate to within 0.02 mm (1/1270"). This step was crucial for reverse engineering, as measuring by hand would have been imprecise given the irregular wear.

Here's a close-up of one of the original top pieces during inspection—it shows the curved flange and mounting lug, complete with historical patina:

Original Casting which has seen better days

Step 2: CAD Modeling the Digital Replicas

With the scan data in hand, I imported the meshes into CAD software (Fusion 360 for its robust surfacing tools). The raw scans were noisy due to rust and imperfections, so the modeling phase involved:

• Cleaning the mesh: Removing outliers, filling small holes, and smoothing non-critical surfaces while preserving key features like bolt holes and lugs.

• Reverse engineering: Converting the mesh to solid bodies by tracing profiles, extruding, and lofting to create parametric models. This allowed for easy adjustments later.

• Adding tolerances: Since these are cast parts, I incorporated draft angles (about 2-3 degrees) for easier mold release and slight oversizing to account for shrinkage in casting (typically 1-2% for iron).

I also modeled the parts as assemblies to simulate how they'd fit with the smoke stack pipe and spark arrestor. The bottom interface was a single-piece ring, while the top consisted of two halves that clamp together.

Check out this rendered view of the cleaned-up CAD model for the top piece—it captures the smooth curves and functional tabs perfectly:

Finalized 3D model of the stack top design, for an early Rumely 12hp steam tractor

Step 3: 3D Print Prototyping for Fit Testing

Before committing to metal, prototyping is key to validate the design. I used an FDM 3D printer with PLA filament for quick, low-cost iterations. The prints were scaled to full size, though I added supports for the overhanging lugs.

Key aspects of this phase:

• Printing multiple versions: The first prototype revealed minor alignment issues with the bolt holes, likely from scan and rendering inaccuracies.

• Fit testing: I mocked up a section of the smoke stack using PVC pipe and bolted the prints in place on a similar engine mockup. This confirmed the interface angles and ensured the spark arrestor mounting points aligned.

• Durability checks: While PLA isn't heat-resistant like cast iron, it allowed me to test mechanical fit and identify stress points.

Prototyping saved time and material—after two iterations, the parts fit snugly without gaps.

Here's a photo of the original bottom neck piece next to its prototype for comparison, highlighting the lugs and rivet details:

Print Prototyping to get the right fit and finish

Step 4: Finalizing the Design

Based on prototype feedback, I refined the CAD models:

• Adjusted lug positions by 2mm for better alignment.

• Thickened certain sections to improve strength without adding unnecessary weight.

The final designs were exported as 3MF files for printing patterns. This step ensured the parts would not only look authentic but perform as originals.

Step 5: Sand Casting the Final Products

With designs locked in, it was time for the real deal: sand casting in ductile iron to match and exceed the originals' durability. I partnered with a local foundry specializing in antique restorations. The process went like this:

• Pattern creation: 3D printed patterns in PLA for higher detail than wood.

• Mold making: The patterns were rammed into green sand (a mixture of sand, clay, and binding oil) to create cope and drag molds. Cores were added for hollow sections.

• Pouring: Molten iron at around 1,400°C was poured into the molds. We cast multiples to account for potential defects.

• Finishing: After cooling, the castings were shaken out, gates removed, and surfaces fettled. A light machining pass ensured bolt holes were precise.

• Inspection: Each part was checked for porosity, dimensions, and fit.

The results? Beautiful, functional castings that weigh about 15-20 lbs each, ready to bolt onto the tractor. They have that classic cast iron texture, and after a coat of high-heat paint, they'll blend seamlessly with the restored engine.

Conclusion: Breathing New Life into Old Iron

This project was a rewarding blend of modern tech and traditional craftsmanship, turning rusted relics into reliable parts for an early Rumely 12 HP steam tractor. Not only does it preserve a slice of agricultural heritage, but it also demonstrates how 3D technologies can make restoration accessible and accurate. If you're tackling a similar project, start with good scans and iterate often—it's the key to success.

If you have questions or your own restoration stories, drop them in the comments below. Until next time, keep those engines steaming!

Posted on August 3, 2025