An essential part of the assembly holding the rocket motor in place when the Bloodhound supersonic car travels at over 1,000 mph during its bid to break the World Land Speed Record in 2016 has been made using Edgecam CNC software.

Andrew Wright, production engineer at the Sheffield-based Nuclear Advanced Manufacturing Research Centre, which made the rear sub-frame for the car, said: “The accuracy of Edgecam’s tool-paths was vital in allowing us to achieve the extremely tight tolerances required for this large and complex assembly, which sits inside the exterior titanium skin.”

The Bloodhound Project is led by Richard Noble, who opened last week’s MACH show and took the record in 1983 with Thrust 2. As well as aiming to break the existing record and top 1,000mph, Bloodhound is intended to excite young people about manufacturing and engineering. The car is a mix of automotive and aircraft technology, powered by the engine used in the Eurofighter Typhoon aircraft, along with a hybrid rocket; the jet engine will take the car to 300mph, after which the rocket will boost it to 1,000mph. A third engine — a 750hp 2.4-litre Cosworth CA2010 Formula One V8 petrol engine — is used as an auxiliary power unit and to drive the oxidiser pump for the rocket.

The body and chassis are relying on a range of advanced design and manufacturing techniques, including a specific production engineering solution based on Edgecam software (www.edgecam.com) that prevented distortion of the rear sub-frame side wall structural panels. The 1.6 x 1m panels, which were produced by the Nuclear AMRC on its Starrag Heckert HEC 1800 horizontal boring machine, have to mate up with other parts in the rear assembly that are vital in keeping the rocket pointing perfectly backwards and providing downward thrust when RAF fighter pilot (and current world land speed record holder) Andrew Green drives into the history books in South Africa.

Maintaining flatness

While the typical machining tolerance for milling was ±0.1mm, some of the wall-thickness tolerances were ±0.05mm and those for hole diameters were ±0.025mm. Mr Wright adds: “The original billet of aerospace grade 7075 aluminium was 80mm thick; the finished component is 20mm thick — and some wall thicknesses are just 6mm. Removing such a large amount of material while maintaining the flatness and shape of a component this size was quite a challenge.

“My main concern when I started programming was that the part would distort and we would struggle to maintain the required wall thicknesses. If some of the walls became too thin, the component might not have been strong enough. However, Edgecam’s roughing strategy and profiling cycles worked perfectly.

“We used a three-side machining strategy. This entailed roughing one side, then rotating the component and roughing the opposite side; we then released it and re-clamped it to finish machining that side. We then turned it back round again to finish the side that we had started cutting originally. That way, we minimised distortion and any chance of the walls ending up too thin.

“Using a 3-D design package, I built the machine set-up, which Edgecam allowed me to import directly. This included the part model originally supplied by Bloodhound in NX format, the modified stock model and all clamps and fixture elements. Edgecam’s ability to read a wide range of model formats and to handle assemblies was invaluable, particularly as multiple set-ups of the part were required.”

Although some 3+2 operations were used, much of the work was three-axis milling, contour and profile milling, with fourth-axis rotation to reach additional features. “Edgecam was absolutely perfect for that; we set multiple datums and then indexed between those datums — all within the same set-up. I built as much of the manufacturing process as possible into the Edgecam part files before taking it to the final simulation.”

Edgecam tailored the tool-paths exactly to the features they needed to machine, particularly when it came to leaving extra material for clamping. With scant excess material on the length or width of the billet, clamp areas were required; these were removed later. “With the difficult shape of the component and small amount of stock material, Edgecam’s ability to tightly control the link moves and feed-in/feed-out moves between sections enabled us to produce smooth and safe transitions between machined features and areas.”

For the main machining, Mr Wright kept the number of cutting tools to a minimum. “As we were looking for a secure and accurate process, rather than a high-productivity operation, I only used three cutting tools to do most of the milling — a solid-carbide end mill, a solid-carbide ball-nose end mill to finish the profiles, and a chamfer mill to deburr as much of the component as possible in the machine.” Mr Wright said Edgecam’s Code Wizard was another big plus.

“As we were using a brand-new machine, we only had a simulation post processor for its predecessor — the Starrag 1600 — but Edgecam made it very easy to create, in effect, a ‘daughter’ post-processor for the new 1800. This gave me absolute confidence in the tool-paths that we transferred to the shopfloor; and I didn’t have to do that much on machine prove-out, knowing that the X-Y G-code between the two simulations wasn’t affected.”