When Automation Has A Ring To It

How do you automate a CNC lathe? If the job can be done from bar, put a bar feeder behind the headstock and a parts catcher or a robot in the work area. If it is a chucking job, then apply a robot. It seems, however, that once the system is set up and gets going, then the problems start.

I visited a company up in the middle north of Yorkshire that makes millions of selector forks and synchromesh rings for Europe's automotive transmissions makers. The company forges and machines some 8 million rings per year for a number of OEMs, including Renault and Ford. Deliveries are controlled, sequenced, just-in-time batches; there is no room for rejects.

On my trip, I learned that the simple shape of a synchromesh ring was anything but simple as far as automation is concerned.

The Yorkshire company precisely saws individual rings from seamless brass tube. After forging and heat treatment, the forgings exhibit hardness in the range of 240-250 Brinell. They are hard and springy and rapidly form a built-up edge on turning tools if tool geometry is not right. It is not easy to turn these springy rings to within a couple of ten-thousandths of an inch in a batch size of 5,000 to 8,000.

The company used to use front-loading, twin-spindle, CNC chuckers, which were acceptable until the 1990s. But as gearbox designers sought quieter, quicker meshing, synchromesh ring tolerances tightened.

Typically, a ring can have an OD of 80 mm (3.15 inches), an average wall thickness changing in section from 8 to 3 mm (0.32 to 0.12 inch) and a depth of up to 10 mm (0.4 inch). It has teeth and needs to be faced, chamfered, bored and internally grooved. The OEM wants out-of-roundness held within 12 microns (0.0005 inch). Individual detail relationships may have to be held within 5 microns (0.0002 inch).

The springiness of the synchromesh ring material, and also the revised forged engagement teeth layout, require that facing and OD turning be performed with the ring located on the teeth and be clamped internally. Then chamfering, boring and internal threading can be carried out with the ring clamped externally.

On the chuckers, a rotating clamping mechanism on a sliding tailstock had been necessary to clamp the front face of each ring during OD turning and chamfering. The actual orientation and location of the precision forged rings in each chuck were critical. The limits for clamping pressures and spindle speed had to be determined before ‘lobing' occurred in the rings.

Trials with a twin-gripper gantry robot-loaded Mori Seiki CL 20 single-spindle CNC chucker showed a cycle time gain. Impressed, the company began a replacement program. As things moved on, the relatively complex sliding tailstock arrangement was replaced by a German Röhm double-acting, pullback chuck, which further shortened cycle times.

Then came the idea that the gantry robot could work better if two Mori Seiki CL203 CNC chuckers were placed in line, back-to-back. So now, a Röhm double-action chuck grips the ring externally, then internally. After machining, one gantry robot places the ring in a turnover device, before the second picks it up and loads it into a collet chuck in the second CL 203 for back end facing and chamfering.

Now the whole two-stage operation takes 23 seconds, floor-to-floor, giving some 120 rings/h compared with some 60/h on the original front-loading, twin-spindle, CNC chuckers or 80/h on two chuckers working as stand-alone automated cells.

The rings are machined under strict CPC. Achieving near-zero rejects in a batch of 8,000 speaks well of the company's methodology and machine tools.

Would automation stop there? Is that the peak achievement with current technology? I mentioned the type of job to pick-up spindle vertical turning machine manufacturer Emag. It was considered that a Röhm chuck could be accommodated and the job done without the complexity of using a robotic gantry loader.