Precision hard turning from Hardinge - extended version

Hard turning doesn't just mean buying a pack of CBN inserts," says Hardinge's Neil Moxey, manager, international technical sales and specialist in precision hard turning/machining. While it is not disputed that many lathes will allow you to turn hard materials with such inserts, the levels of precision – dimension and form – will vary, while consistency may also prove to be an issue. So when Hardinge talks about hard turning/machining, it is not talking about the general area, but rather highest precision machining to tight tolerances, both of dimension and form. And it was at the company's super precision centre in Europe – Krefeld, North Rhine-Westphalia, Germany – that this message was underscored, with the company's US-headquartered expert, Dr Mike Kushnir, principal engineer, laying out the detail just ahead of the biennial EMO show in Hanover, held last month. Dr Kushnir has over 40 years' machine tool experience, with his first 20 spent at the machine tool research institute in Moscow, followed by his Hardinge career, where he heads up machine tool development. Krefeld, under managing director Markus Herdegen, supports all super-precision projects world-wide, with UK-based Mr Moxey (+44 1392 360666) the international roving figurehead for Europe. And it is he that UK seekers of ultimate precision will find themselves talking to, even if first contact is made via UK Hardinge distributor the Engineering Technology Group (01926 818418). MIXED MESSAGE Historically, Hardinge, with its 123-year history, has been associated with highest precision turning. However, it is accepted that this message has become muddied, following, in particular, the acquisition of Bridgeport (machining centres, knee mills) and the mixing of US- and Taiwan-made lathes all under the same brand. To be clear, its super-precision machines are the T42, T51 and T65 SP models (numbers equal bar diameter), plus the Quest GT (gang tooled, 27 mm diameter) SP series and Quest CHNC SP series (27/42 mm bar diameter). The latest generation super-precision (SP) machines and are 2 micron better, in terms of accuracy, compared to the previous family (Quest, RS and Conquest) – so better than 3 micron continuous cutting accuracy versus better than 5 micron. The T42 SP was introduced three years ago, with the others following. These super-precision machines are all made in the USA, at Elmira in the state of New York, but all the company's turning machines have US-built spindles. The company also boasts precision grinding technology, most notably for the UK audience, Jones & Shipman, but also Kellenberger, which includes the brands Tripet, Hauser and Tschudin, while American firm Usach Technologies Inc is now also included. This availability of grinding technology has, in fact, added to its lathe-based super-precision credentials, according to Mr Moxey. "Grinding technology has been incorporated within our super-precision turning solutions increasingly over the past three years, with probably 40% now including grinding; in some cases solutions have featured only grinding, in fact," he offers. Grinding is required where its surface characteristics are preferred; not because of any precision deficiency, it should be emphasised. PRECISION? WHAT'S THAT? So first, what exactly does Hardinge mean by super precision; what levels of accuracy can it machine to? Well, the company aims for the 6 micron to 0.3 micron area. The top end is just above the 5 micron figure achievable with general turning, while the latter overlaps both grinding (with a range of 5 to 0.25 micron) and diamond turning (below 2 micron, to as close to zero as is possible to get). Hardinge is not competing with other turning machines, but, instead, other processes. "We are delivering a process that nobody else can deliver," Dr Kushnir insists, adding that customers are getting: "Hardinge accuracy, productivity and the knowledge that sits behind this process." And it is doing this with machine tools using conventional components, including cast iron, linear guideways, ball bearings, and not via special hydrostatic technology, too, the machine tool expert stresses. Hardinge has achieved this by extending the reach of standard turning technology through good design (see box, below). And what is the benefit for the user? At its simplest, it means a lower investment for a given level of accuracy. A T42 SP with main and sub-spindle, Y-axis and 12 tools is priced at under $300,000, while the next level up, hydrostatic lathe, comes in at over $400,000, with grinding and diamond turning machines above that; the latter put at around $600,000. Of course, this is the machine price; the process development and tooling is in addition, but that is true of all these options. However, a Hardinge super-precision lathe also offers production machining capability versus the more restricted capabilities of grinding or diamond turning, with their more limited tooling and process coverage. A Hardinge lathe has a tool turret with up to 24 tools (BMT55 tool system, 12 live, 12 static) or 32 (BMT45, 16 and 16), and can have a sub-spindle and Y-axis; it can turn, mill and grind, and produce complicated forms, completing a part on one machine. "Hydrostatic and diamond lathes typically have no more than four tools. Grinders, one or two wheels. Hardinge super-precision lathes offer high accuracy and production machining," Dr Kushnir explains. "High accuracy machines typically achieve their capability by having a simple layout that gives a short force loop [which gives high stiffness]," he adds. The demand for Hardinge's super precision capability has been increasing over the past five years, says Mr Herdegen. For aerospace and defence, steering systems, fuel injection systems, navigations systems and fasteners are areas he pinpoints. In the medical field, implants, prostheses, surgical instruments and medical equipment components are highlighted. In automotive, the areas of application take in fuel injection systems, ABS/ASR systems, power steering, gears, gear shafts, bearings and more, the managing director adds. Fluid technology sees hydraulic pistons, hydraulic housings and pumps tackled, while the mould and die sector is another customer. At the moment, Hardinge Super Precision Europe is processing about 50 machines each year. And he emphasises that selling catalogue machines is not the company's business. "We don't sell any machine without proving the process to the customer; we don't want any surprises," offers Mr Herdegen. "We are a bit like a laboratory for our customers; customers have a problem and we try to take care of it." UNDER WRAPS A difficulty for Krefeld in exampling what it does is that much of its work is covered by non-disclosure agreements (NDAs), which means that full details of projects are not available in the main, while taking pictures within the Krefeld facility is a plain no-no, for the most part. But Dr Kushnir was able to give some instances of what the company is able to achieve. A first one is hard through-boring of a 66-64 Rc flange/boss forged component. The bore is 20.96 mm diameter nominal and 28.70 mm in length. The critical requirement was cylindricity of 1.5 micron. The customer initially approached Hardinge for a machine for semi-finishing prior to grinding, but was convinced that super-precision turning could provide a better solution. Precision hard turning delivered a result for cylindricity of 1.19 micron for the first run-off part; 1.21 micron for the second; with all others up to part 10 coming below 1 micron. And looking at the cost benefit, a comparison of the equipment plus operator in each case is given. This was put at €140,000 plus €46,000/annum for the general-purpose lathe and operator; plus €360,000 plus €64,000/year for the grinder and its operator. Using Hardinge super-precision technology, the process route is: semi-finish, using the sub-spindle, then finish turn, using the main spindle, on a T42 SP sub-spindle machine. The cost of the machine is €266,930, with the operator cost put at €62,000/annum. A crude cost saving of €279,070. An important point is that the operator cost for the super-precision machine is the same as that for a grinding machine. "He must have the mindset of a grinding specialist," Dr Kushnir underlines. And he also highlights that with a variation in hardness (64-66 Rc), the ability to maintain the requirements under varying conditions demands a machine with high stiffness. Important in this particular application is the ability to separate the machine from sources of heat and vibration (chillers, power cabinet), which Hardinge machines allow for, as standard. The customer returned after one year for a second machine, incidentally. In another example, the ability of Hardinge super- precision machines' axes' positions to be software compensated by increments of 0.1 to 0. 2 micron was pointed out. The company was tasked to demonstrate its ability to produce a large radius with form tolerance of 0.5 micron within an 80 mm length of a bore through a part of 120 mm thickness. To achieve this, software compensation of X-axis positioning at the levels stated above was undertaken at length increments of 2 mm in Z. "This is only possible if your machine can move this very small amount," Dr Kushnir emphasises. REACTION IS EVERYTHING It is quite possible for many machines to have such axis compensation entered, of course, but the key is the minimum increment to which a machine will react. Hardinge claims particular credentials, as this example underlines, while it can show tests that highlight that another machine (boxway guides) will not react at all to even 10 successive 1 micron increments. But that is stick-slip, a well-known issue. Finally, concerning form accuracy of single-point turning versus its own Kel-Varia grinding technology, when producing large convex radii (5,080 mm) of within 1.5 micron form accuracy on a diameter, hard turning produced a smoother contour, closely aligning to the theoretical profile, while grinding delivered a profile result that oscillated far more and at more frequent intervals along the theoretical profile, although was still within the required tolerance. It should be added that a competitor's turning machine's results did not deliver the radius faithfully at all, while oscillations along its path where up to in excess of 2 micron versus the Hardinge T42 SP machine's tenths of a micron figure. So, if turning just doesn't cut it and you are considering hydrostatic lathes, grinding or diamond turning; before such a change, talk to Hardinge, is the message. Box item 1 Dimensional tolerance and form tolerance According to Hardinge's Dr Kushnir, who is referencing Mahr data, at the 15 micron tolerance level (semi-finish turning), form tolerance would account for 15% of the total dimensional tolerance (2.25 micron); halve that, for finish turning, and form tolerance is 20% of the total budget (1.5 micron); at the 2 micron level, grinding, form is 30% of total tolerance (0.6 micron); and at 1 micron, honing, form tolerance is 40% (0.4 micron). Form tolerance becomes a larger part of total error, as overall dimensional tolerances tighten, he observes, while adding that it is easier to achieve high form tolerance with a single-point turning tool than with a grinding wheel, for instance. For example, a 25.4 mm sphere in 62 HRc material with a +/- 1.5 micron form tolerance can be achieved easily and quickly by turning, but Dr Kuchnir says colleagues at Kellenberger said they would not attempt this; it would take far too long. Box item 2 Bearing races Bearing races are still mostly produced via grinding. The requirements are profile accuracy of about 2.5 micron; roundness of 3 micron and surface finish of Ra 0.15 micron. Hardinge has been developing a process for bearing races and has achieved these or better (roundness was 0.5 micron), but the question is one relating to fatigue life of a turned surface, and that requires lengthy testing to be undertaken by the manufacturers to usurp the established grinding practices, the related performance data for which are probably 60 years old, offers Dr Kushnir, highlighting that few companies do tests today. That said, he believes this will be a large future application area for Hardinge super precision lathes, but there are probably only a handful of companies across Europe and the US currently applying the process to bearing races today. Yet the companies "keep knocking on our door," says Mr Moxey, adding: "I don't think it will be long before it becomes an accepted process. History states that bearing races are ground, so they continue to be ground. We had similar problems before, years ago. The grinding fraternity said that hard turning gave rise to white layer, but it has been proved not to be a problem." White layer was accused of decreasing the strength or fatigue life of a part, but in a detailed article in the Industrial Diamond Review, by M.A. Fleming, C. Sweeney, T. J. Valentine and R. Simpkin, published in 1998 and titled 'PCBN hard turning and workpiece surface integrity', the conclusion is: "This investigation [on PCBN hard turning of steel parts] has confirmed that hard turning does not' have a detrimental effect on component fatigue life." Box item 3 Machine design for super precision Through years of experience and measurement, Dr Kushnir has formulated static stiffness values in X-axis/bar diameter that indicates whether a machine will be good (above the line) for super precision hard turning or not (below the line). But in additional to static stiffness, dynamic stiffness is also required and equally important in hard turning. Hardinge has developed a simple chatter test to indicate this. Interestingly, the T42 and T51 machines outperform both its own hydrostatic guideway QC51 Hydro machine and two competitors' box guideway machines. This, points out the machine tool expert, confounds those that say damping with linear guides is not as good as for boxways or hydrostatics. It depends on how linear guides are applied. "This all takes time and money, but we have done this over years, because know we will need it for the next step in development." On basic construction, the machine bed is a slant bed set at 45° and also employs a Y-axis at 90° to the X-axis, not a compound slide wedge-type that requires 2-axis interpolation to generate Y-axis movement (X plus the compound slide on the X-axis slide). And Harcrete is placed in strategic locations in the base, to provide good damping and stiffness. Dr Kushnir explains that Hardinge knows for each machine element how much of the stiffness budget can be taken by the various elements of the complete assembly. So, for the spindle, it's 10%; the base, 10%; Z-axis assembly, 20%; X-axis assembly, 30%; turret, top-plate and coupling, 15%; and for the Y-axis slide, it's 15%. This supports a stiff final design and allows for incremental use of finite element analysis of modules, prior to having the whole machine design available for FEM analysis, which is really verification at that stage. Interestingly, the machine tool expert underlines that 80% of the performance of a machine is derived from the first 20% of fundamental design choices, ahead of availability of FEM data. Putting anything right later is expensive and time-consuming, he underlines. And over many years, tests have shown that this approach can deliver a final result that is within +/- 10% of theoretical figures, Dr Kushnir says. Other key design details include scraped slideways (not just ground or milled) for the joints between base casting and headstock unit mating surfaces (expensive and time-consuming), with ground surfaces (not plain milled) for the linear guideway connection surfaces. In previous times, when this was not done, results were "disastrous", Dr Kushnir underlines, and, in relation to this, he says that joints can be responsible for 30-50% of a machine tool's compliance (movement under load). The heart of any machine is its spindle. Many designs use roller bearings in the front row and then angular contact bearings behind. Good for stiffness, but not accuracy. Hardinge employs angular contact bearings in total, three rows at the front and two at the rear, which is superior, but more expensive and more sensitive to any assembly variation, so requiring more time and skill. The results speak for themselves. At 2,396 rpm, total error in the former is 0.91 micron, while for Hardinge it is less than one third of that, at 0.28 micron. In addition, as rpm goes up, the errors worsen, due to residual unbalance. "Most of our competitors believe their spindle will be balanced by design. We don't think so," Dr Kushnir says. Spindle balancing sees two-plane balancing of the spindle shaft and two-plane balancing of the assembled spindle and closer. The established spindle nose collet-ready system is also employed, of course, placing the part closer to the front spindle bearings than in other designs. So, taking a T51 SP, at 1,000 rpm both spindles display virtually nothing; at 2,000 rpm, the box guideway competitor is now displaying over 1.5 micron eccentricity versus Hardinge's 0.25 micron; at 3,000 rpm, the competitor's machine is up to 3.5 micron, with Hardinge under 0.5 micron. At 4,000 rpm, Hardinge has fallen to around 0.4 micron and the competitor's has also improved, displaying about 1.2 micron. And for roundness, Hardinge again beats the competition, although the differences are smaller than for concentricity. The worst case is at 4,000 rpm, where Hardinge manages 0.5 micron, but the competitor's figure is approaching 0.8 micron. And the fact that plate vibration in the X-direction, caused by the spindle unbalance, is greater in the box guideway machine than in Hardinge's linear guideway T51 SP again underlines the impact of spindle design – it can clearly mitigate the benefits of other machine features. So, we have a rigid, stiff assembly, but the ability of the machine to react to small programmed steps to faithfully follow super precision forms or compensate for mechanical errors is key in application. Hardinge can demonstrate that, for example, its T51 SP can respond to 0.2 micron increments faithfully. It can even respond to steps of 50 nanometres faithfully; and it will even react to 15 nanometre increments, although the fidelity is not so good at that level. Okay, so what about temperature influences in production? Thermal compensation is not a favoured route for Hardinge; thermal management is. By understanding what happens to the machine under various duty cycles by measuring 36 data points, thermal growth can be predicted and its effects managed, by coolant and spindle chiller adjustment. So, for any installation, variation due to thermal influences can be kept within 3 micron on diameter, it is claimed via thermal management. What this all points to is that there can be no weak link in the machine tool design, when talking about super precision machining. And also that you can put together general components and get a good machine; it is how they are put together, of course. "We haven't done anything fancy; we make no compromises and we apply good design principles. But it is expensive," Dr Kushnir offers. And all this design stands behind each machine's cutting performance certificate showing continuous cutting accuracy of better 3 micron; part roundness of 0.25 micron; overall axis repeatability, 0.76 micron; and surface finish of 0.15 micron. This is a standard super precision machine's capability. But Mr Moxey says that there are examples where continuous machining accuracy of less than 1 micron is being achieved in the field. Indeed, independent measurement by America's Lawrence Livermore National Laboratory of a 50 mm brass sphere machined on a Hardinge super precision lathe demonstrated roundness and profile accuracy of within 1 micron. First published in Machinery, October 2013 Edited on 3 January 2014 to remove use of SP²; replaced with SP