Manufacturing surgery - medical implant grinding and surface finishing best practise

German surface finish specialist Rösler (0151 482 0444) said that it would reveal 21 things about medical implant manufacturing at the event, where it was supported by grinding machine expert Haas Schleifmaschinen (Dorman Machine Tools, 024 7622 6611) and inspection technologist Taylor Hobson (0116 276 3771). Finishing supplies/tools companies Pometon (01952 299777) and Moleroda Finishing Systems (01722 711988) were also in attendance on the day. Image: Roesler's medical implant event drew over 60 delegates Cornelius Wecht, sales engineer at Haas Schleifmaschinen, kicked off the presentations; in the order of production, grinding would be the first process, after all. Having a three-machine range, two of these are suited to medical implant duties – Multigrind CA and CB, both 5-axis models. Tools for medical implant use can be manufactured on the third model, the AF, which is similarly a 5-axis model. Parts that have been made on the CA and CB machines include femur (top part of a knee implant), tibia (bottom part of a knee implant) and modular hip implants. On the AF model, hip broaches and reamers are parts that can be made. With the majority of 'the big four' implant manufacturers (Stryker, Zimmer, Du Puy and Biomet) said to be using Haas grinding machines, up to 80% of all knee implants made world-wide are produced on the company's technology, it is claimed. There are around 50 machines in the UK and Ireland. Machines feature composite material beds and no cantilever structures, making them very stable. Toolchangers change wheels, cutting tools, coolant nozzles, probes and linishing belts, while part loading/unloading can also feature. The CB model is, uniquely it is claimed, able to grind two femur implants at the same time, a process that sees up to seven discrete operations undertaken: grind articulating surface; mill periphery; mill box; grind intersection cam condyles (projections on the lower extremity of the femur); mill cam; blend box by milling; and belt linish articulating surface. Grinding wheel dressing in parallel with grinding itself is also a claimed unique capability, as is the ability to use a linishing belt system in a grinding machine as part of the process. Image: Two femurs at one time for Haas Schleifmaschinen Depending on complexity and size, this can be undertaken in 12 to 20 mins per part and form tolerance achieved is better than 20 micron, while surface finish is better than 0.5 micron Ra and which minimises finishing processes (see later). Versus a milling process alternative, tooling costs are reduced by 80%, Mr Wecht says, while the process chain is also simplified to a three-stage 'cast, grind, polish' route. Changeover between jobs is under 30 mins, too. A major development revealed was the forthcoming launch of a new programming system. Currently, toolpaths are generated externally, typically using Siemens NX (formerly Unigraphics, UG), although belt linishing is still a tricky area, Mr Wecht offered. Post-processed programs – taking CLS data (co-ordinate data) and converting (via Haas CLS converter) it to Haas '.prn' format – are then fed into the Haas machine control, where the co-ordinates are linked to geometry, probing/measuring, technology data and wheel/tool data, with a G-code editor available to make modifications at the end. There may be several iterations of this process before a successful program is generated, Mr Wecht says. The expertise required to generate a program for a femur implant, for example, is high. It is often the case that Haas specialists undertake this on behalf of customers. But the company's new control/software combination, Horizon, aims to take all this programming (based on imported 3D models of parts, tools, wheels etc) into the machine control itself, making it easier for end users to generate programs. With this, Mr Wecht suggests it will take an end user one day to program a part, making in-house programming viable. Moreover, the original 3D models can also be edited and exported back to the CAD system, so that the part design model can be updated. Following Mr Wecht, Rösler's national sales manager, Colin Spellacy, explains the drag finishing process and Rösler's particular take on it. An automated process, it sees fixture parts rotated and moved through media held in a bowl. More fully, parts are located on the end of a leg that orbits a centre of rotation (the Earth, if you like) that itself orbits the centre of the machine (the sun, if you like). So, to continue the astronomical analogy, parts rotate as the moon to the Earth and the sun. Depending on the condition of the part, drag finishing is either a two- or three-stage process. If surface finish is greater than around 1.2 micron Ra, a three-stage process is required, typically taking up to 180 mins. A two-stage process would typically be 80 mins smoothing (plastic media), plus 40 mins polishing (graded crushed walnut shell, impregnated with a polishing paste). The finish after polishing, incidentally, is 0.02 micron Ra. Three-stage sees grinding (ceramic media to remove excess roughness), smoothing, then dry polishing. In drag finishing, finish follows the form of the part, producing consistent surface finishes and maintaining geometry, avoiding any flattening/variation that might occur via manual or robotic linishing belts (linishing belt wear in robotic operations poses problems), while drag finishing also allows different parts to be processed simultaneously (from 12 to 30 pieces), including different sizes of the same part. SHARP EDGES ISSUE Sharp edges are broken on the back of components in drag finishing, while linishing leaves sharp edges that demand further work. And worn linishing belts can expose particles that scratch the part surface. Furthermore, studies have shown that implants finished in this way, by the mass media process, also demonstrate harder wearing surfaces, Mr Spellacy offers. In addition to these benefits, drag finishing also delivers a cost reduction. No programming is required (robots), there's no need to change linishing belts (manual or robot) and throughput is faster (20-30 mins per process – grind, smooth, polish) versus both manual and robotic approaches, plus up to 30 pieces are finished simultaneously versus a robot's 1-piece flow approach. Versus manual linishing, drag finishing eliminates some 60 to 80% of that manual work, Rösler claims; although Mr Spellacy adds that there will always be a need for some prep work. Drag finishing reduces the number of process stages, reduces inspection time and delivers a fully automatic process. Summing up, bringing it down to the numbers, drag finishing versus manual or robot linishing offers identifiable cost benefits, quite apart from anything else (see box, below). Taylor Hobson's business development manager, Jon Gardiner, moved the day forward onto the measurement of geometry, form and surface finish, highlighting recent technology developments in this area. The three specific areas that drew his focus were roundness measurement, with a resolution of 0.8 nm; non-contact interferometer instruments, with a resolution of 0.1 angstrom (one angstrom is 10-10 m); and surface instruments, with 0.2 nm resolution. The tasks that these three tackle are: cone and surface analysis; wear scar and surface analysis; and spherical form and surface analysis, respectively. NEW PROCESSES DETAILED Taking these in order again, Mr Gardiner drew attention to a new process for measuring cone tapers (critical in implant manufacture) on Talyrond instruments; highlighted that, with non-contact interferometry, it is possible to clearly distinguish between the surfaces produced via the three stages of Haas machined part, smoothed part and polished component; and underlined that the spherical and form analysis technology (PGI Dimension) was a recent development that had originally been developed for optics measurement and which supported far more accurate sphere measurement than is possible with CMMs, measuring down to 200 nm form accuracy (but to 0.2 nm resolution). The new process for measuring tapers sees a ball or diamond stylus capture cone and spherical form, taper angle and surface finish, all in 3D – the measured data can be joined to make a complete tapered cylinder or sphere, ready for volume analysis to support wear investigation, for example. In doing this, measurements are collected against noise that is no more than 30 nm – "We are getting very, very high accuracy form and surface finish readings," he confirms. Moving to non-contact interferometry (coherent correlation interferometry [CCI], to be precise), the CCI machine measures 3D surfaces, step height and roughness over small areas – 7 mm square or smaller – with an X-Y table supporting the capability to gather data from larger areas. Resolution is 0.1 angstrom, as already stated, which does shock some people, Mr Gardiner offers, as what look like smooth surfaces take on the form of mountain ranges in a 3D plot – "The reality is, we are looking at things you couldn't see before," he adds. Such 3D measurements parameters as Sa (average deviation of surface), Sz (maximum height of the surface), Sv (valley void volume) and Sp (maximum peak height) are derived, along with the hybrid Sdq parameter (the rms slope of the surface), an important parameter, in implant terms, that relates to friction. As with the Talyrond-based 3D measurement, the 3D surface in this case can again be variously analysed, for wear, for example. Turning to the PGI Dimension, developed for the optics industry for measuring spheres and aspheres, this is a very low noise system that combines elements of roundness measurement devices with those from surface measurement, explains Mr Gardiner. That combination delivers very high precision 3D form measurement (normally a 2D measurement) and is able to gather form in a circular, as opposed to transverse, direction, too. There are two ways to gather data using this machine. Draw the stylus over a sphere to just over halfway point; rotate the sphere 180°and repeat; join the two lines. This can be done at a number of positions to gather a 3D map. But a recent development sees the part rotated as the stylus travels over the sphere. Not as accurate as the former, but faster and still able to return form error better than 300 nm on a calibrated ball, while surface finish down to 10 nm can be measured. All in all, the event, which attracted more than 60 delegates, had much detailed information to stimulate the production engineering juices. Box item Drag finishing cost benefits Manual linishing and polishing takes between 30 to 60 mins for each femur implant. Over an 8-hour shift, the net result is a minimum production of eight components, with a maximum of 16 produced per person. So, the cost to manually linish and polish a femoral part is put at €15-20. Based on a batch of 24 knee implants, at an average 9 mins per part, the robotic linishing route would take 216 minutes versus drag finishing's 150 (two stages), which means the latter has a 66 minute saving. Moving to the polishing stage and the robot takes an average of 6 mins/part against drag finishing's 44 minutes, giving the latter a 100 min advantage. Total advantage is 166 mins per 24 (almost 7 mins per part). The costs of the drag finishing process for 24 parts are <€4 in consumables; in a two-step process, the cycle time per part is 5 mins, and in a three-stage process, 7.5 mins. So Rösler claims that, compared to manual linishing and polishing, drag finishing is 68% less expensive, while throughput is increased by 84%. Versus robotic linishing and polishing, costs are lowered by 41%, while throughput increases by 63%. These figures do not include capital investment.