Suburban Machinery Software Division 37777 Harlow Dr., Willoughby, OH 44094 (440) 951-8974 Fax: (216) 420-9208

Home

Contact

Software

Putting The Probe To Work

An article By Kim Smith,

Quality Control Engineer - Falk Corporation

For more than a century, The Falk Corporation, headquartered in Milwaukee, Wisconsin and part of the Sundstrand Division, has always been known as a "Good Name in the Industry". Good reputations don’t come easily. Falk Corporation started in 1892 when Herman Falk first applied a revolutionary cast-welding process to the installation of streetcar rails, and continues today with the broadest line of standard and special power transmission products in the world. At our Auburn, Alabama facility, The Falk Corporation manufactures s variety of couplings for our best-selling Steelflex line as well as fluid disc couplings, always priding itself in the quality of our manufactured products.

The Falk corporation recently decided to move into the robotics age for manufacturing its highest volume product lines. With the help of our corporate headquarters in Milwaukee, our Auburn engineering department embarked on this task. One of the most important questions Auburn had to address was how to produce quality products with its unmanned robotic cell. The robotic cell Auburn had constructed contained an outstanding vertical machining center made by Monarch Corporation (Cortland, NY), which provided great probing capability. It also included a 2-spindle turning center from Motch Corporation (Cleveland, Ohio), and an ABB robot.

One component to be manufactured on the flexible cell starts out as a slug, sawed from barstock, which would be taken by the robot’s gripper from a pallet and set on a recentering stand. This slug would be regripped, then loaded into one spindle of the Motch turning center for its first turning operation. When the first turning operation was completed, the robot would again grip the slug, and move it to the second spindle of the Motch for a second turning operation on the opposite side. The robot would then regrip the part, and move it to the Monarch machining center for its first operation of drilling. When drilling was completed, the part was then turned over and again loaded on the Monarch for a second operation to chamfer the backside of the part.

With a suggestion made by one of the robot cell operators, we decided that perhaps there would be a way to use the Monarch’s excellent touch-probing capability to measure the most critical turned dimensions, and somehow "close the loop" by passing these data back to the Motch’s twin CNC controls. With the help and assistance of Ron Fedder, an engineer involved with implementing the robotic cell at the Auburn facility, Mike Thompson and I began to look into this possibility. After speaking with Dan Horn, an applications engineer at Monarch Cortland who wrote most of the Monach’s probing software, we discovered that information from the Monarch’s probing cycles could easily be output to other devices through the Monarch’s RS232 serial port. He also recommended that we contact Dan Fritz, from Suburban Machinery, Software Division, in Willoughby Ohio, who could possibly provide custom DNC software to connect the Monarch machining center to the Motch’s twin CNC controls.

When we first contacted Dan Fritz, we explained to him that we wanted to use the Monarch probing cycles during the first drilling operation to measure the barrel’s outside diameter, and during the second drilling operation to measure a flange outside diameter, then automatically transmit that information to a 486 66 mhz computer, which would send these data to the Motch’s CNC controls. Dan explained that it could easily be done if the Motch’s CNCs were running in a DNC mode. This presented us with a problem, because the CNCs on the Motch were not equipped from the factory with the software needed to run in a DNC mode. Even though the twin Fanuc CNCs on the Motch turning center had RS232 serial ports, they could only be used for loading files in an "edit" mode, and could not be used for direct DNC operation. Dan suggested that we contact the Motch engineering department because the first step of our project was to add this DNC capability to the turning center’s PC (Programmable Controller) software.

We next contacted Bill Minerich at Motch Corporation, who had the necessary source code and knowledge of the Motch’s programmable controller software to make the changes we needed. By putting Bill in contact with Dan Fritz, we were well on our way to completing the first step of our project: Making our Motch turning center accept information with the DNC link that Suburban Machinery was prepared to supply. In a few weeks, Bill sent us a new set of EPROM chips containing the revised Motch Programmable Controller software.

When the EPROMS in the Motch turning center were replaced, we could then "drip-feed" the twin CNCs with part program data from our 486 computer. This was necessary because the custom DNC link that Dan Fritz was developing for us was designed to work as follows: Both "sides" of the Motch’s CNC control (actually, 2 CNCs in one box) would be receiving data from an RS232 port on our shop-floor 486 PC. A third serial port on the PC would be connected to the Monarch’s CNC control, and would be used to "listen" for information from the Monarch probing system. Each time one of the Motch’s CNCs started a part cycle, the DNC software on the 486 PC would send a "G10" command to update the tool wear offsets in the lathe’s memory, and would then send the correct part program file in direct DNC mode.

The DNC software from Suburban Machinery had several tasks to perform simultaneously. First, it had to "drip-feed" two lathe controls independently of each other. It also had to receive commands from the Monarch, calculate new tool offsets, and send them to the correct side of the Motch turning center at the beginning of each part cycle. Also, the software had to detect several possible error conditions which would result in an orderly shut-down of the cell’s operation. Finally, since the 486 PC was drawing NC program files through a network, it had to download a "core" NC file from a network drive onto its local "C" drive at the beginning of every part cycle. This way, we could update the Motch’s programs from any PC connected to the network without disturbing the cell’s operation.

The next step in our project was to devise some way for the Monarch machining center to send probing data to the DNC system. We decided that whenever the Monarch measured a part, it would transmit a letter ("A" or "B") followed by a number. The letter would identify the measurement as being for Motch side "A" or Motch side "B", while the number would simply be the measured part diameter. Since the Monarch probing cycles were really macros programs stored in its Fanuc 15M CNC control, we were able to add special "DPRINT" commands to the probing cycles supplied by Monarch. This Fanuc macro command enabled us to send the measured diameter of the part (along with an "A" or "B") to the DNC system each time a probing cycle was completed. It was then the responsibly of the DNC software to receive this information, compare the measured dimension to the "target" value and create a new tool offset command for one of the Motch’s CNC controls. If a measured value required too much of an offset change, or if the tool offset value went beyond a pre-set limit, an alarm would be generated on the 486 PC, and one side of the Motch would not begin the next part cycle. If an alarm occurred, however, the DNC system must permit the programs running on the other side of the Motch to finish, and must not adversely affect the Monarch or the ABB robot.

After discussing our software requirements with Dan Fritz, he assured us that the DNC software could, in fact, perform all the functions we required. He also cautioned us that any software they supplied would probably not work 100% correctly on the first try, and that some quick and easy method must be devised for software revisions and "de-bugging". We decided on the following procedure: Suburban Machinery would send us the first draft of the software, along with the source code, the compiler, and a special "batch" file for compiling the DNC software. We would then test the software and report any bugs back to Suburban. If a change needed to be made, we could simply call Dan, who would suggest some source code changes. We would then edit the DNC system source code file, re-compile, and try the system again. This procedure gave us a greater understanding of how the software worked, and eliminated any costly travel between Suburban Machinery and The Falk Corporation. Also, by giving us access to the source code, we could be sure that we could make any future changes without involving Dan Fritz or Suburban Machinery.

Although there were many technical problems to overcome with this project, the results of our efforts have been quite satisfying. What is even more important to report is this: We at the Falk Corporation were able to successfully coordinate the efforts of several talented engineers from several different companies to accomplish a complex task. Each one contributed greatly to the project. Each engineer’s contribution hinged on the contribution of someone else. All information was exchanged by phone, FAX or MODEM, and none of us have ever met in person to discuss this project.