Standard Test Rig Suits All Light Axles
A new, standardised test rig for measuring axle noise, vibration and harshness (NVH) has been introduced by Burke Porter Machinery. Capital cost of the equipment is said to be significantly lower and lead-time from date of order to delivery and commissioning will be cut by a third. The world launch will be at Automotive Testing Expo 2006, to be held in Detroit from 25th to 27th October, where Burke Porter Machinery will exhibit on stand 7006 a twin-pallet AXT 1800-LS. It is the largest of four models in the company's new standard test rig range and will be demonstrated under power testing a beam axle with tubes and an independent rear for American Axle. There is not a single passenger car or sports utility vehicle manufactured anywhere in the world whose front and rear axles cannot be checked automatically on this rig before it leaves the production line. Axles for some light vans will also fit, although NVH is not so important on commercial vehicles. NVH testing measures acoustic emissions from an axle under drive and coast conditions, accurately predicting its quality when fitted in a vehicle. Such objective testing is becoming essential in the automotive industry, as customers' ever-higher expectations of their vehicles makes it unwise to rely on subjective quality analysis. This is certainly the case in the luxury car sector, as drivers will not tolerate whine from the drive train of an expensive vehicle. OEMs are acutely aware of this and, to minimise expensive recalls that would dent their profitability, place stringent demands on their axle suppliers.
The standard Burke Porter Machinery test rigs are the first to be controlled by a full CNC system - the Sinumeric 840D from Siemens. The AXT 1800-LS (long stroke) has the largest footprint and accommodates any size of light duty axle, including front and rear beam axle assemblies with or without drive shafts (tubes fitted), beam axle centre sections, front and rear independents (side gear or flange output coupling), and outputs coupling with drive shaft hubs or side gear splines.
The short stroke version, AXT 1800-SS, is intended for complete front and rear beam axle assemblies with drive shafts, and the output coupling with drive shaft hubs only. Model AXT 750 tests front and rear independents (side gear or flange output coupling, and beam centre sections without tubes fitted; while the smallest test rig, the AXT 500, is restricted to rear independents with output drive flanges. As the size of the test rig decreases, so does cost and area of factory floor occupied. Burke Porter Machinery believes its new standardised range of rigs to be unique in terms of their universal applicability to any size and type of car axle.
Commented technical director, Bill Coleman, "We used to design every rig around the range of components the customer wanted to test. Obviously we did not have to start the design from scratch each time, as we have years' of experience building end-of-line and audit axle test rigs for many OEMs and Tier 1 suppliers. But there was always such a large bespoke element that we could not start building a rig until the order was placed.
"Now we build all rigs around four sizes of AXT base assembly constructed from standard modules which are common to all models. A number are being built for stock in our Gosport factory, so in future when an order comes in, we will configure the rig to the customer's requirements, which we estimate will slash build time from 36 to 24 weeks. The new approach will also allow us to reduce the price, as series production will result in economies of scale." The driveline assemblies are standard in all rigs and are fitted in advance to the stock machines. The input is up to 6,000 rpm / 250 Nm torque, while the outputs can be controlled to 2,400 rpm / 350 Nm. Within these parameters, a large range of input and output motors can be specified by the customer to suit its requirements, using a common motor frame size.
Market research has shown that the typical displacement of pinion gear to axle centreline is 40 mm, so Burke Porter Machinery has incorporated 60 mm adjustment in the input lift assembly. Control of the offset is by servomotor via a ballscrew under instruction by the Siemens 840D, so adjustment can be done on the fly, allowing the loading sequence to be fine-tuned in two CNC axes. This is crucial for accuracy and repeatability of testing.
The axle is clamped in the same way as it will be mounted in the vehicle so that the geometry is not affected during testing. In the case of a rear axle, the input shaft that simulates the drive from the transmission is engaged and so also are the two output shafts that will, when the axle is in the car, be connected to the back wheels.
Test scenarios can include varying the speed at constant torque, varying the torque from positive to negative at constant speed, and combining the two. In this way, the gear set is monitored for NVH under load in both drive and coast directions and at the optimum load conditions. The speed of the two output shafts is normally synchronised so that only straight line driving is simulated; the differential gears are not usually included in the tests, as noise is more of an issue when cruising rather than when cornering, although differential output speed tests can be accommodated to measure differential 'growl'.
Reactive coupling improves test accuracy Worldwide patents are pending on a new system invented by Burke Porter Machinery for the consistent coupling of the driveline to the axle under test. This is the factor that most influences test rig repeatability, ie the variability of measurements obtained when repeatedly measuring NVH values for the same axle. The difficulty with coupling an axle in a production test rig is that the action must be effected quickly to minimise cycle time, yet the linkage has to be rigid to avoid play and unwanted vibration during the tests. Repeatability is strongly influenced by how well the coupling is connected. Positive pressure by the input adapter assembly on the axle companion flange would solve the problem, but this has to be avoided, as any axial force would influence the test results.
So clearance is always left between the face of the input adapter and the axle flange. However, the exact position of the latter is uncertain due to the variability of tolerance build-up during manufacture. Until now, it was necessary to take the worst-case scenario, ie the longest axle flange, and program the test rig slides to position the adapter as close as possible, without touching it. The problem was that axles with shorter flanges ended up further away from the adapter, causing a larger engagement gap that made the test results less accurate.
To overcome this difficulty, Burke Porter Machinery conducted a series of tests at the same oil temperature to find out what the clearance should be ideally. The test cycle involved acceleration, coasting and deceleration so that both faces of the meshing gears were tested and the dynamic torque measured, which translates into a decibel figure representing the noise generated. Having identified the optimum engagement clearance, the company devised a method for positioning the input adapter accurately to that distance from the flange of any axle as it comes off the assembly line, ie the machine reacts to the real position of the flange and not to the predetermined slide position of the machine. This is the crux of the Burke Porter Machinery reactive coupling patent.
Further tests showed the system to be remarkably effective. With reactive coupling turned on, repeatability of measurements was improved by 50 per cent on the drive side and by 70 per cent on coast.
Other steps taken to improve the repeatability of results have centred on the design of the machine structure and of the driveline. A SolidWorks CAD model of the machine was sent by Burke Porter Machinery to Anthony Best Dynamics in Bradford-on-Avon, Wiltshire. ABD used its specialist software to model mathematically the amplitude and frequency of resonation around the structure during axle testing under various simulated operating conditions. Such dynamic mapping is able to suggest design modifications that minimise resonance at operating speeds where it is most likely to affect the accuracy of the NVH measurements.
Burke Porter Machinery has been co-operating with ABD for several years and is now on its fifth generation of test stand, every design iteration having been an improvement on its predecessor. Each time, the predicted distribution of resonance and the expected positions of dead zones around the structure were correlated against measurements subsequently made on the machine itself, enabling the design to be progressively refined. ABD has also co-operated with Burke Porter Machinery in the development of on-shaft torque sensors and torsional accelerometers to measure accurately the dynamic torque signal, or ripple, to within 0.01 Nm, giving an accurate measurement of gear mesh noise. Concluded Bill Coleman, "There is a tendency for axle assembly line suppliers to focus on the manufacture of the product and think about end-of-line testing at a later stage in the project.
"So we are often asked to supply our production test rigs in a very short time frame so that the axle manufacturer can correlate test results with those from the audit rig in their R&D department, or indeed the vehicle itself, which is essential before the axles can be brought to market. Our new, standard test rig design will greatly assist in meeting these tight deadlines." Burke Porter Machinery is one of the world's leading producers of automotive test stands, having custom-designed, built and installed more than 1,000 since 1960 for most of the major names in car and truck production. Although the company specialises in test equipment for automotive power train assemblies, it has also manufactured test cells for steering components, cylinder heads, hydraulic valves, alternators, starter motors, clutches and both electric and hydraulic motors.
A standard CNC axle test rig manufactured by Burke Porter Machinery.
An output driveline with motor, flexible coupling, flywheel, slipping clutch and intermediate bearing block.
With reactive coupling deactivated, the gap between the axle and input shaft is 0.5 mm. The dynamic output torque (Y axis) plotted against static input torque for the same axle tested 6 times shows a 1.8 dB variation in the results leading to uncertainty as to how noisy the gears are.
With reactive coupling activated, the gap between the axle and input shaft is maintained at 0.1 mm and the spread of results are much smaller at .9 dB allowing gear noise to be predicted much more accurately.