Lunenberg Piston
(ongoing)
Each year seniors in MIT’s Pappalardo apprentice program collaborate together on a large machining project. Apprentices build on the skills they honed while fabricating their individual Stirling engines to each make one part of a larger machine. In spring 2019 the project was building a 1918 Lunenburg Foundry single cylinder marine engine from original drawings. My contribution to this effort was the fabrication of the piston, which involved modeling, sand casting, and finish-machining the part.
Original 1918 assembly drawing of the ‘Atlantic’ series single cylinder marine engines produced by the Lunenburg Foundry in Nova Scotia.
Original detail drawing of the Atlantic piston. The curved fin on the piston crown helps separate the entering air-fuel mixture from the exiting exhaust fumes.
CAD model derived from original drawings.
Sand casting pattern created by removing features on the piston and adding draft. A flange on the bottom makes space for supporting the core (shown to the right).
Sand casting core for creating internal features on the piston. Cores stay in the mold cavity during casting and are made from resin-bonded sand so they can be broken out from the cast part later.
Pattern and core superposed to show the volume that will be filled in the casting process.
Pattern shown in its casting orientation. This model is color-coded according to standard foundry practice with the core print in yellow, unfinished surfaces in black, surfaces to be machined in red, and risers, gates, runners, and sprue in green.
CAM simulation of scalloping operation on pattern with a 1/4" ball end mill.
Scalloping operation performed on a TRAK DPM at 4000 RPM and 150 in/min. Video shown at actual speed.
3-axis milling of core box. Both the pattern and core box were machined from RenShape polyurethane foam.
Core box machining simulation.
One half of the pattern after an initial adaptive clearing operation.
Pattern half after a finishing scallop operation and a contour pass around the perimeter of the part to clean up the bottom edge.
Completed pattern along with wooden runners. The central runner extends past the gate-runners and tapers down to capture any slag or dross during pouring.
Completed core box halves. Two dowel pins help the core box come together after the halves are packed.
Core box halves packed with resin-bonded sand. Since the core will be suspended upside-down in the mold by its flange, resin-bonded sand provides added strength over traditional 'green' casting sand.
After both halves of the core have cured, they are brought together and glued with core paste (an acetone-based adhesive specific to sand casting).
Pattern and central runner mounted to a board that is temporarily fixed to the bottom of the flask (wooden box that will hold the sand mold). Both pieces are dusted with talcum powder to help release them from the sand after packing.
Flask beggining to be packed with green sand. The green sand mixture used was ordinary play sand mulled with Bentonite clay and water.
Once the drag (bottom half of the mold) is packed with sand, it is flipped over and the gate-runners are added. This surface is then dusted with talcum powder, another flask is mounted on top, and the cope (top half of mold) can be packed.
After packing the cope, the cope and drag can be seperated again and the pattern removed. The pattern was designed to come apart in wedges to make removal easier.
View of mold cavity and central runner cavity in the drag portion of the mold.
Pouring at the MIT foundry. Nearly 50 lbs of gray iron were poured into the mold.
Removing the part from the mold before it has sufficiently cooled can compromise the crystal structure of the cast iron as well as produce internal stresses. This part was allowed to cool off for a full day before being broken out of the mold to mitigate this risk.
Cast piston with gates and risers trimmed off.
Setup for rough machining of the piston's OD. The boss on the piston crown was drilled to accept a live center.
Rough machining of the piston's OD. It was a relief to see very little porosity or inclusions under the rough cast surface of the piston.
Piston after initial machining of the OD, but still oversize of its finished dimension. This machined OD served as a reference to clean up the bottom surface of the piston skirt so that it could provide a flat and square datum for future operations.
After turning the Piston OD to its finished dimension, the ring grooves are cut with a carbide parting tool. The magnets stuck on piston are an old machinist's trick for reducing chatter by adding some extra damping to the part.
Checking ring groove depth using a test indicator. The rings have to have enough radial clearance between the piston it so that gases can flow behind the rings and press them against the cylinder wall.
As of spring 2020, I'm still popping in to the MIT machine shop from time to time to make progress on this project. The two main milestones left are cleaning up the piston crown (removing the live center boss and machining the crown flat) and boring the wrist pin hole. Both of these operations require some clever fixturing and machining techniques and I'm excited to share some updates on those processes.
