Ferrari 250 GTO

The process of creating this hammer-formed body panel began with a digital model of the Ferrari 250 GTO. We discussed the layout of the buck, and worked out how to divide the 3D model in a smart way with Fusion 360 Slicer. We knew we needed to retain as much detail as possible, so ensured that the main crevices and edges showed through on the buck. This gave us a closer reference for fitting. It was important for the whole team to be present in the first few stages, as each decision we made would affect all of us in some way.

We settled on 6mm MDF for the buck ‘slices’ because it is sturdier than 3mm MDF, and plywood was unavailable. The laser-cut file had to be corrected after the initial cutting, as we found that some parts fell off due to an error in the pattern generation. The final version of the file had a few issues, but laser-cut correctly and allowed us to see an adequate level of detail.

We covered the buck in tape to allow for the contours of the car to be more accurately perceived, and to prevent some of the outer slices from falling out due to gravity. Taping the buck also allowed us to better see where the major edges and valleys were, so we could draw boundaries between parts. Once this was done, I chose to hammer-form the rear fender, as it had a roughly bowl-shaped form that appears almost as its own object. Its transition to other panels of the car were on significant edges, meaning I had clear boundaries to work within.

Looking at the digital model, Ivan and I initially thought that creating a blister would be the easiest way to form the rear fenders, but Daniel explained that doing so would complicate things. As such, we used a similar method to the bowl¸ and made extensive use of the sandbags to achieve the general shape.

I worked with Ivan to trace the rear fenders by wrapping the paper tightly around the buck and tracing the outline of the rear fender with a permanent marker. We tried to keep our patterns as close in size and shape as possible, but eventually decided to use the same pattern, just reversed. This ensured that our panels had the same foundations, which could reduce the number of issues later.

I shaped the upper curve of the panel first, to give me an anchor for the rest of the shape. I used the teardrop mallet and cylindrical mallet with a rounded edge. There was a particularly deep area that I was wary not to overstretch.

As we tried to fit our panels onto the car, the overhang of metal towards the underside prevented us from fitting the forms to the buck.  I found a block of thick, laminated MDF slices in the scrap box, and hammered four nails into it as a makeshift plinth for the buck. The nails fit between the slices of MDF, and ‘locked’ the buck in place quite effectively.

It was much easier to work on like this.

As the hammer-forming continued, we repeatedly placed our panels onto the buck to check our progress and see how they fit together. I checked this quite often to avoid over-stretching the material. We used images, Joel's 3D printed models, and the digital model of the Ferrari as references too. I specifically tried to follow the curve from a top view, and in doing so, lost track of the flatter upper surface.

Tail cut-out. Leaving out the tail made the rear fenders look really neat.

Test fitting.

The turning point was when I cut out the wheel well at Ivan’s suggestion, which made it much easier to shape the piece. This essentially freed the upper surface from the side by breaking it up into smaller pieces. I continued to draw directions for flattening and stretching the panel, especially in repairing the upper surface. This continued for a while, and the upper surface became flatter and flatter, and fit the buck more closely.

There was a unique problem in that it was easy to write on the outside of the piece, but you have to hammer from the inside.

I shrank the back edge of the piece with the shrinker/stretcher and sandpaper between the jaws and the workpiece to avoid leaving such big marks. This successfully made a bowl shape, but only at the very end of the piece. As such, I flattened the edge back down and decided to use the nylon mallets to shape it instead.

Sandpaper helped to prevent big marks from appearing.

...but marks still appeared, as the jaws ripped through the sandpaper.

I used a folded sandbag overhanging the edge of the bench to push the side surface down to a flatter level, as the area above the wheel well bent outwards quite easily due to its thinness. I changed my method slightly at this point by ‘clamping’ the panel onto the buck with friction, and lightly hammering onto it. This gave me a better idea of how I was going, and I depended less on images and the 3D model for reference.

The panel was much more curved than the buck at this point.

I used a sheet-metal vice grip to curl the uppermost edge of the panel so that it sat nicely on the buck, This also gave it an ‘anchor’ underneath Logan’s roof panel.

The sheet-metal vice grips were great for this purpose.

Shortly after this, I annealed my panel in the workshop oven at 385 degrees Celsius, as quite a bit of work hardening had occurred, and sharp corners of the piece began to split. This softened the panel by a lot, and I resumed shaping. It was much easier to shape it now. When I began to hammer the panel directly on the buck again, I noticed that certain points that contacted the MDF slices on the inside of the panel were black with adhesive from the masking tape that covered the buck. This meant the workpiece was deforming between these points, so it took on a very slight polygonal shape. I took note of where these points were, and shaped the upper surface back to an even curve.

The circled areas were rubbing against the buck quite forcefully.

Note the worn-out masking tape towards the top of the curve.

We removed most of the masking tape, as it had become worn and dirty. The outer MDF slices of the buck immediately fell out, so I decided to nail them to the perpendicular slices. The nails I used were too long, so I clipped them with some end-cutting nippers. This made it a little harder to begin nailing the MDF, but it reduced the risk of splitting. Despite this, the MDF began to split in a few places. The nails could also slip out if enough force was applied to the slice. With care, I managed to affix the outer slices to the rest of the buck with only minor splitting.

When I had shaped the panel to a level I was happy with, I sanded it with a handheld electric sander and 120-grit to reveal the lumpy areas. I used water to remove the aluminium dust as I sanded it, like how you would use wet-and-dry sandpaper. This made a huge mess, as the 120-grit sandpaper was not suited to wet use. The front edge disintegrated after a while, so I took it off the sander and cut it into shape again.

I carefully smoothed out the lumpy areas of the panel with a flat nylon mallet and a curved piece of plywood found in the scrap bin. I modified the plywood form to allow it to sit better in the vice by slicing part of it in half with the band saw, and shaped the outside edge to a curve. This allowed me to position the work piece in more useful ways. I also had to touch-up the main curve of the work piece with a line-shaped nylon mallet and a sandbag. This long process of sanding and reshaping was valuable, as it made the surface pretty even.

Initial sanding revealed the uneven patches.

Touching up the main curve.

After shaping the plywood edge on the disc sander.

After cutting on the band saw.

The 120-grit had done a nice job of smoothing out the surface already, so I hand-sanded the panel with 400, 800, and 1500, using a circular motion whenever possible. I used wet-and-dry sandpaper with water, and they smoothed out the surface quite well. I then polished it with Brasso, which gave the panel a nice sheen. There are some crevices which the sandpaper could not reach, such as the upper lip that meets with the roof panel. However, this would be hidden on the assembled model.

800-grit with water.

Half of the piece is polished.

I’m satisfied with the final piece, as it fits very tightly onto the buck and has a nice, satiny finish. It does have some surface inconsistencies, though, which may have been fixed if I spent longer smoothing out the lumps before sanding. Overall, this was an interesting undertaking that taught me a lot about controlling the shape of a hammer-formed piece of aluminium to fit a buck.