Automated Gaging

George Schuetz, Mahr Federal Inc.

 Every shop should have this gaging "problem."  A Tier 2 automotive supplier was required by his OEM customer to perform 100 percent inspection on parts for dimensional tolerances and several other characteristics.  The parts were aluminum forgings for air conditioning compressor pistons, and the inspection requirements included dimensional tolerances for bend and twist, and checks for fill at four positions, presence of two radii, and flash removal at two positions.  The problem was the size of the contract: the supplier was obligated to deliver, and thus to inspect, nearly 100,000 parts per day.

 Some gages lend themselves to faster throughput than others.  Generally, the more specific the gage's purpose, the more quickly it can be operated.  For example, fixed-size bore gages (i.e. plug gages) are quicker than adjustable (rocking) bore gages.  If the task is more complex—for example, inspecting multiple diameters on a part such as a crankshaft— then a purpose-built fixture gage will allow faster throughput than a selection of snap gages, or the use of surface plate methods.  But there comes a point when even the most narrowly focused, manually operated gage must give way to automation.  It's either that, or hire a whole roomful of inspectors to keep up with production.

 Automated gaging is usually cost-justified in applications where a part must be inspected every 45 seconds or less.  This figure includes not just gage operation, but the entire gaging cycle.  While the part measurement itself may take only three seconds, the complete cycle includes at minimum: placing the part in the gage; operating and reading the gage; and removing the part from the gage.  Other required actions may include: recording the measurement; sorting parts into appropriate categories by size; and removing rejects from the lot. 

 Many additional variables influence the speed at which a part can be measured, and hence influence any gaging setup, whether manually operated or automatic.  These include: the number of features to inspect; the need for a dimensional reading versus simply go/no-go results; how measurement data will be used (e.g., for export to SPC, or for direct process feedback); the level of accuracy required; and whether gaging occurs in-process or post-process.            

With so many variables in play, it is hardly surprising that automatic gages can rarely, if ever, be bought "off the shelf."   In the case of the Tier 2 supplier, we custom-engineered a fully automatic gage, capable of inspecting one part every 3.5 seconds (i.e., 1,030 parts per hour, or 24,720 parts per day).  Four identical units were built and installed, providing total throughput of 98,880 parts per day.

 Reliable parts handling was obviously one of the most important engineering considerations.  Accordingly, the gage was designed to make use of the most reliable parts-handling mechanism ever developed: gravity.  Parts feed in at the top of the machine, sliding down a 45° chute to a dead stop, where a proximity switch senses the part and triggers a locking mechanism.  An air-driven cylinder then raises the holding fixture, and a nest of electronic gage heads descends until it contacts the part.  The gaging device traverses the part, checking for true position, material fill, flash, and the presence of radii.  Bend and twist are checked as independent features by comparing position and diameter measurements at opposite ends of the piston.

 When the inspection is complete, the holding fixture descends and the part is released to drop down the exit chute.  Out-of-spec readings trigger an escapement, which diverts bad parts into a reject bin, while good parts pass straight through to the next production process.

 Gage head signal conditioning, data processing, and data storage are controlled by a gaging computer, from which measurements are downloaded daily.  Programmable logic controllers (PLCs) control all of the gage's logic functions.  The systems have proven to be extremely reliable, operating around the clock for months between downtime for preventive maintenance.

 To equal the throughput of the automated gage, the Tier 2 supplier would have to keep 13 human operators working around the clock, at a minimum throughput of one part every 45 seconds per person.  Even at minimum wage, and assuming no coffee breaks or sick days, it wouldn't take long before those human operators, each with a manually operated gage and a master, started to look pretty expensive.

 Few machine shops face throughput requirements even close to this, but any shop involved in a large production run can usually benefit from some form of specialized gaging, to make inspection easier and faster.   And for larger shops where throughput requirements are very high, and the production run will last for a year or more, customized, automated gaging may be the only practical approach to inspection.