AIR BEARINGS FOR GRINDING MACHINES
By C. A. SCOLES and H. L. WUNSCH M.Eng., A.M.I.Mech.E.
Reproduced from Machinery Lloyd and Electrical Engineering (Vol. 37, No. 6 - 13th March 1965) With kind permission of the publishers – The Certificated Engineer.
Grinding wheelheads require two main characteristics to be met: the wheels must rotate truly, and the spindles must possess adequate stiffness in both the radial and axial directions. A hydrodynamic journal bearing has the characteristic that its stiffness increases as the oil film thickness are reduced. Because of this, the trend in wheelhead development has been to reduce clearance as much as possible. As a result, unavoidable errors in the forms of the spindle and bearings, such as departures from true roundness, create significant variations in the actual clearance obtained and result in spindle axis displacement during rotation, which must limit the grinding performance of the wheelhead. Final workpiece surface quality is dependent on the wheel head spindle behaviour at the time when the grinding forces are tending to zero, i.e. when the 'sparking-out' phase of the operation is in progress. It is, therefore, vital that the spindle should be stiffly supported by its lubricant film when it is virtually no longer displaced by grinding forces.
With these considerations in mind designers have aimed to produce bearings of minimum clearance at working temperature, or perhaps one ten-thousandth part of an inch or less, and to find thin lubricant mixtures of low viscosity which will limit the temperature rise due to viscous friction so that, at normal running speeds and ambient temperatures, the bearings will just not seize. At any lower running speed, the bearings will be slack and will not produce work of optimum quality. At any appreciably higher temperature, the spindle may seize and ruin the wheel head.
As a result, most precision wheel heads operate at around 150 degrees F, once steady conditions have been achieved, and this means that considerable time must elapse, possibly two hours in extreme cases, before a machine is usable at its best performance, with a 'taut' wheel head and a fully thermally adjusted structure. For this reason, grinding machines are often switched on early in the working day and left running all the time, whether or not they are needed for continual use.
If, however, air bearings are employed, the whole problem of friction and heating is virtually eliminated, since air bearings, thanks to the negligible viscosity of air itself, will run indefinitely at moderately high speeds without any noticeable change in temperature from ambient. The fact that the spindle is supported around its circumference on a film of compressed air results in a truer rotational accuracy than the spindle geometry itself would normally allow, simply because the air film tends to average-out local spindle imperfections. The air film also provides a true elastic cushion, the stiffness of which can be predetermined at the design stage. Thus, for a given grinding force the spindle will deflect initially and then progressively regain its original attitude in the bearings as metal is removed from the workpiece.
In an air bearing, no metal-to-metal contact between the spindle and bearings occurs, irrespective of whether or not the spindle is rotating. As a result, wear is eliminated, and periodic adjustment of clearance is not required.
Material specifications of spindle and bearing bushes are not of prime importance and since the running clearance of a journal air bearing can be quite large, for example, about 0.0015 in on diameter irrespective of basic size, no problems of careful selective assembly arise during manufacture.
Balancing of grinding wheels in situ is also possible since a free journal air bearing acts as a frictionless pivot which is far superior to any knife-edge wheel balancing fixture.
The design of the first air bearing wheelhead, intended for use on a universal grinding machine, was influenced by the need to match its general size to that of the wheelhead it was to replace and by the desire to eliminate a driving belt and separate motor if this were practicable.
The original spindle was belt-driven from a 3 h.p. electric motor, but since a proportion of this power was wasted in bearing friction, it was considered that a smaller power unit would suffice to provide it could be built into the head casting as a stator and rotor assembly; a 1½ h.p. induction motor unit was therefore selected.
The final design is shown in Fig. 1. It will be seen that the spindle, which is of stainless steel, is built up from two pieces to allow its assembly with the rotor unit. The larger of the two air bearing journals are placed at the wheel end and the other is of the maximum diameter that would pass through the stator unit during assembly. The air support in the axial direction is provided by two thrust annuli which are matched for the area, so that the spindle is positioned with equal end clearances. Air is led to the numerous inlet control orifices in the brass bearings through drill ways in the casting, and it will be noted that most exhaust air is directed inwards to pass over and around the stator windings before releasing to the atmosphere. The only air not to do this is that which escapes out along the spindle nose, and this outward flow prevents grinding dust from entering the wheelhead.
The spindle and bushes were made of rustles materials, stainless steel and brass respectively, purely as a precaution against moist air. Electrically, the wheelhead is unaffected by moisture since the stator windings are impregnated with a waterproof material and no commutator or brush gear is involved in this type of motor.
To safeguard against damage resulting from air supply failure during use, the electrical supply to the wheel head stator windings includes the following safety features. Firstly, the motor cannot be started unless the bearings are pressurised. An adjustable pressure switch set to operate electrical contacts at an air pressure of 50 lb/in2 provides this requirement. Secondly, if during operation the air supply should fail, the same pressure switch will break the 3-phase supply and then switch in immediately a d.c. supply to two of the three-phase windings of the stator, so applying eddy-current braking to halt the spindle quickly. The same sequence of events is used when the normal 'stop' button is pressed since merely switching off the power supply to the wheel head would leave the spindle 'freewheeling'.
The air supply to the wheel head is passed through a ceramic type of air filter to remove foreign matter. The bearings are not adversely affected by reasonably small amounts of moisture or oil vapour in the airline and since these contaminants, once in the airline, are usually difficult to filter out, this is useful toleration. It should be mentioned at this point that a major criticism against the use of air bearings under factory conditions is that most compressed air is dirty, wet, oily and of very variable pressure. This criticism, unfortunately, is true in many cases. Air is the Cinderella service in many factories, where its sole purpose is for blowing swarf out of components after machining, and, as a result, its quality is unimportant. However, all manufacturers of air compressors can supply efficient aftercoolers and filtering systems and providing the pipes in the factory are correctly installed and fitted with drain points, there is no need for this unsatisfactory state of affairs to be perpetuated. Also, carbon absorber-type filters, which will remove oil and water vapour, are now commercially available and can be used to improve the quality of existing air supplies. As a short-term measure, of course, a local portable compressor can be employed to supply dry, clean air.
Upon completion, the wheel head was subjected to a series of tests and measurements to verify its design characteristics. The free air flow was first measured at the normal working pressure of 60 lb/in2 and found to be 2.5 ft3, a value within the predicted range. This is equivalent to 0.7 h.p. in terms of compressor capacity, and offsets to some extent the saving in the power of the driving motor.
Next, the static load capacity and stiffness of the bearing were measured in both the radial and axial directions. These measurements showed that the spindle would support, at the wheel position, in the radial (vertical) direction, a load of 85 lb additional to its weight, and a load of 150 lb in either axial direction. The derived stiffness values were 130 000 lb/in and 240 000 lb/in respectively. These values related to an air supply pressure of 60 lb/in2; at higher pressures, proportionately greater load and stiffness values would apply.
The radial and axial loads on the spindle arising from normal precision grinding are not likely to exceed a few pounds, but it will be seen that the wheel head has a reserve of strength to cope with accidents, such as inadvertent heavy feeds.
Since the grinding process is essentially a dynamic one, the behaviour of the bearings under forced vibration was also studied. These tests showed that resonance frequencies of 250 and 220 cis occurred in the radial and axial directions respectively, both values being well above the normal wheel speed frequency of 25 cis. However, since the possibility of resonance excitation depends also on the grinding operation itself and on the vibrational behaviour of the machine including its other motors, pumps, etc., the most straightforward method of finding out if these resonance frequencies or their harmonics were likely to be excited during use was to carry out a comprehensive programme of grinding tests on a wide range of workpieces.
Under power, the spindle was found to run very silently, reaching full speed almost instantaneously and no vibration could be detected on the casing of the wheelhead. The eddy-current braking system was found to operate satisfactorily and was capable of bringing the spindle to a halt, whenever necessary, within three seconds.
Long-period running tests were undertaken to determine the thermal behaviour of the wheelhead both when running light and when under dynamometer loads of 0.3 and 0.6 h.p., which were considered to represent typical grinding conditions.
Whilst the air bearings themselves would not generate a measurable amount of heat in this application, the stator windings were a direct heat source, and these tests were therefore carried out to determine the effect of the exhaust air-cooling arrangement in limiting the temperature rise within the wheel head. Measurements showed that in five hours the stator temperature had risen by 14°C and the casing temperature generally by only 8°C. The effect of dynamometer load was insignificant in these tests, and after a full day's running the wheelhead casing felt barely warm to the hand.
Before commencing grinding tests on the wheelhead, the effectiveness of the air bearing spindle as a frictionless pivot for the in-situ balancing of grinding wheels was investigated. This showed a clear improvement over the conventional knife-edge method.
The need for a dummy spindle was obviated and the sensitivity of unbalance detection was found to be extremely high.
The initial workshop tests consisted of normal grinding on a variety of jobbing components to enable several different operators to become as familiar with the use of the wheelhead and to form general impressions of its performance.
Their reports were unanimous in its favour. All found that the head would grind as well as or better than the original wheelhead, that 'sparking-out' time was reduced and that close tolerances could be worked to immediately the head was switched on. No difference in workpiece quality, i.e. dimensional accuracy and surface texture, was detectable between the use of the wheel head within minutes of switching on or after several hours of running under power. A variety of splined and serrated components was included in the test programme, and in no case was the spindle found to vibrate during grinding. It was found that excessive feeds could be applied without adversely affecting the wheelhead in any way.
Using this wheel and workpiece feeds of 0.008 in could be applied if a slow traverse was used, but the actual metal removal was about half of this nominal amount due mostly to machine deflection. Under this excessive feed, an air gauge showed that the wheel head spindle deflected by only 0.00016 in within its bearing radial clearance of 0.0008 in. Wattmeter tests indicated a power requirement of about 2 kW but slowing down of the wheel was not obvious despite this overloading.
The performance of the wheel head at the other extreme was then examined, i.e. its ability to produce a workpiece of optimum surface quality. Three cylindrical workpieces of hardened steel, 6 in long and 1¼ in diameter, were ground, each with a wheel of a different grade and using, first, the conventional wheel head, so that direct comparison of results could be obtained using the same workhead and centres but different wheel heads. As nearly as possible, the same grinding and wheel dressing conditions were employed.
The results obtained showed a slight superiority for the air bearing unit and the operator reported that much less sparking-out time was needed to obtain an optimum finish with the air bearing wheelhead than with the conventional one.
The specimens were not examined for errors of roundness since such errors result from both workhead and wheelhead behaviour.
Intervals of time between necessary wheel dressings were also found to be increased by a factor of about four with the air bearing unit, resulting in increased wheel life and productivity.
Because of the success of this wheelhead it was decided to use air bearing in a wheelhead for a small tool room surface grinder.
In this case, existing belt drive facilities were, for simplicity, retained and the bearings are shown in Fig. 2 were designed and manufactured. Once again, assembly requirements with the existing machine-controlled to some extent the overall dimensions of the new unit. Two new considerations arose from this exercise; firstly, the need to control very accurately the air pressure to the bearing, to avoid any rise and fall of the wheel due to air pressure variations and, secondly, the need to avoid excessive tension in the driving belt. The first was dealt with satisfactorily by using a pilot operated pressure regulator commercially available, and the second by calculating the necessary belt tension and obtaining it by monitoring the pulley displacement through a dial indicator until the correct tension was applied. This could be done because the air bearing stiffness was known and the bearing could therefore be used, in effect as a load dynamometer.
The wheel head was operated at 2800 rev/min from a 1 hp. motor.
Grinding tests proved most satisfactory: the wheelhead ran extremely smoothly and showed no increase in temperature from ambient irrespective of the length of time it was used. The surface texture of workpieces was excellent, despite significantly shortened sparking-out times, and so-called mirror finishes of 0.5 micro inch CLA were readily obtainable when using, for example, an Aloxite 60 I vitrified wheel.
Once again, a fail-safe electrical circuit was used to prevent the motor being started before the wheel head bearings were pressurised, but in this case, the motor bearing provided sufficient friction to slow down the wheel head in a reasonable time once the power was switched off, hence no eddy-current braking was necessary. This wheelhead has now been in constant daily use for some three years and has given unfailing service throughout this period.
A third wheel head in the series was designed as a replacement for a ball-bearing internal grinding head. Again, this was a belt-driven unit running at a nominal speed of 10 000 rev/min.
Despite the small space available for the air bearings in this head, it was nevertheless possible to design them to have a radial stiffness some five times greater than that of the extension arbor.
Excellent grinding results were obtained, and workpiece surface texture values of 1.5 micro inches CLA could readily be achieved using an A.A.60.J.5 grinding wheel.
Further developments, meanwhile, have been under consideration; the use of PTFE bearing bushes, in place of brass bushes, to prevent bearing damage in the event of extremely sudden air failure, has been studied with satisfactory results, and the design of grinding machines equipped with air bearing table slides and/or workheads is envisaged.
Precision micro-feeding is also a feasible proposition with air bearings since asymmetrical adjustment of orifice pressures can be used to displace the spindle axis, or the worktable height, by micro inch amounts under the control of a manual pressure regulator valve.
Thus, extremely fine metal removal can be achieved without any need to in-feed the wheelhead during final sizing operations.
Finally, air bearing wheelheads should be of great use in grinding research, since the wheel speed can be varied without affecting the bearing stiffness, and spindle deflection can be measured by air-gauging techniques, provide data on grinding forces simply and straightforwardly.