DNC Puts The Probe To Work

In this robotic machining cell, a custom DNC link "closes the loop" between the machining center's touch probe and the turning center that feeds parts to it.

Article From: 11/1/1995 Modern Machine Shop, ,

For more than a century, The Falk Corporation, with headquarters in Milwaukee, Wisconsin, and a wholly owned subsidiary of the Sundstrand Corporation, has been known as "a good name in industry." This trademark is a testament to Falk's focus on customer satisfaction and commitment to quality.

The Falk Corporation started in 1892 when Herman Falk first applied a revolutionary cast welding process to the installation of streetcar rails. Today, Falk is a leader in the geared power transmission industry, providing a broad line of standard and special power transmission products. The Auburn, Alabama plant manufactures a variety of couplings, from its Steelflex line to disk-type couplings.

Until early 1995, the Auburn plant operated entirely in a batch manufacturing mode. As part of an upgrade to manufacturing capabilities at this facility, however, the old batch manufacturing method is giving way to cellular manufacturing. The cost/benefit analysis associated with converting the facility from batch to cellular processing showed that the use of automation to produce high volume couplings was justified.

The first automated cell implemented as part of the facility upgrade is called the Turn And Drill Automated Cell, and as the name suggests, the processes performed in the cell are turning and drilling. However, because this cell has a robot, it is generally referred to as the Robotics Cell. The Robotics Cell is designed to manufacture 50 different parts from 2 to 10.5 inches in diameter and weighing up to 80 lbs.

Inside The Cell

The cell consists of a Monarch VMC 150B vertical machining center, which is equipped with probing capability and a 16M control; a Motch 220D-VTC twin spindle vertical turning center equipped with Fanuc 16T controls; and an ABB IRB6000 six-axis robot situated between the machining and turning centers.

The diversity of incoming raw materials and the dramatic geometry changes that occur as part of the machining processes in the cell proved to be extremely challenging obstacles to automation. The cell was designed to run unattended for a period of four hours, with insert indexing and other tool maintenance occurring only at scheduled intervals.

One component to be manufactured on the Robotics Cell starts as a cutoff barstock slug. One of the two robot grippers retrieves a slug from the pallet and places it into a centering stand for part orientation. From the centering stand, the slug is regripped and placed in the first spindle of the turning center for its first operation, turning the barrel diameter.

After the first turning operation, the robot moves the part to the second spindle of the turning center to complete the flange diameter of the hub. The robot then moves the part to the first spindle on the vertical machining center, where the part is placed flange up to have the bolt circle drilled. When the drilling cycle is complete, the robot positions the hub, barrel side up, in the second spindle on the vertical machining center so that a tolerance chamfer can be made on the back end of the hole. These processes are performed in sequence and all machining can occur simultaneously, with the longest of the cycles determining the throughput of the cell.


A primary concern when incorporating automation into manufacturing is always part inspection. This cell presented an exceptional challenge. The wide variety of geometries that the cell is capable of producing required a unique solution to part inspection. In this case, the solution was suggested by one of the cell operators, Chad Ledbetter. He suggested using the probe on the Monarch VMC to inspect the required features (flange and barrel diameters) of the hub and feed that information back to the Motch turning machine. The tool wear offsets could then be calculated for the finish tool that creates the final part diameter.

We decided that perhaps there would be a way to use the VMC's touch probe capability to measure the most critical turned dimensions, and somehow "close the loop" by passing the data back to the twin CNCs on the turning center. Intrigued by the potential of such a system, Mike Thompson, quality control supervisor; Ron Fedder, a project engineer, and I began to look into this possibility.

After speaking with Dan Horn from Monarch Cortland [now Genesis Worldwide], who wrote most of the VMC's probing software, we discovered that the information from the probing cycles could be outputted easily to other devices through the machine's RS-232C serial port. He also recommended that we contact Dan Fritz, from Suburban Machinery, Inc., Software Division (Willoughby, Ohio) to provide custom DNC software for connecting the two machines.

Making The Connection

When we first contacted Mr. Fritz, we explained to him that we wanted to use the touch probe to inspect the barrel's outside diameter on the first operation and then, during the second drilling operation, to inspect the flange's outside diameter. This dimensional information could be fed to our shopfloor 486 personal computer (PC) for performing the necessary calculations. The output from the PC would then be sent to the respective turning center control to make the appropriate offsets to the finishing tool in order to maintain part size in an unmanned mode.

Mr. Fritz replied that this plan was quite workable, but several questions needed to be addressed. First, did the two controls on the turning center currently operate in a DNC mode? Second, could data be transmitted to the control on the appropriate spindle?

The first question presented a problem because the CNCs on the turning center were not equipped from the factory with the software needed to run in a DNC mode. Although the twin Fanuc controls on the turning center had RS-232C serial ports, they could only be used for uploading CNC programs in an "edit" or memory mode. It would be necessary to convert the turning center from CNC to DNC mode.

Bill Minerich from the engineering department at the Motch Corporation, who had the necessary source code and knowledge of the turning center's programmable controller software to make the changes to DNC mode, got in touch with Suburban Machinery. The two organizations worked together to be sure that the turning center could accept information with the DNC link that Suburban Machinery was prepared to supply. In a few weeks, we received a new set of EPROM chips containing the revised Motch programmable controller software.

After the EPROM chips in the turning center were replaced, we could then "drip-feed" the twin controls with part program data from our PC. Drip feeding was necessary because the custom DNC link was designed to have both controls receive data from the RS-232C serial ports on our shopfloor PC. A third serial port on the PC was connected to the Monarch CNC so that the PC could "listen" for the information from the machining center's probing system.

DNC In Action

Each time one of the Motch CNCs starts a part cycle, the DNC software on the PC sends a "G10" command to update the tool wear offsets in the turning center's memory, and then downloads the correct part program file in DNC mode.

The custom DNC software has several tasks to perform simultaneously. First, it has to "drip-feed" the two controls on the turning center independently. It also has to receive commands from the Monarch, calculate new tool offsets, and then send them to the correct side of the turning center at the beginning of each part cycle. Finally, because the PC was drawing NC program files through a network, it had to download a "core" of NC files from a network drive onto its local C drive, at the beginning of every part cycle. This method allows us to update the turning programs from any PC connected to the network without disturbing the cell's operation.

Along with the DNC software, we created a PC database to assist in calculating the new tool offsets. This database includes a table of upper and lower tolerances for each part dimension, the mean of the part dimension, the amount of tolerance allowed, and the tool number to offset in the turning center CNCs. Unless the proper tool number associated with NC program is found in the database, the DNC system will not respond. This ensures that the data is "drip fed" to the correct side of the Motch turning center at the beginning of each part cycle.

One other feature of this DNC software is very important. If a part is out of tolerance, the system is able to identify several possible error conditions. This information allows an orderly shutdown of the cell's operation for correction.

Integrating The Probe

The next step in our project was to devise some way for the 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 the mean of the dimensional tolerances allowed. The letter would then identify the measurement as being for side A or B of the turning center.

Because the machining center probing cycles are macro routines stored in the Fanuc CNC control, we were able to add special "DPRINT" commands supplied by Monarch to the probing cycles in our part program. The DPRINT command enables us to send the measured diameter of the part, along with a letter A or B, to the DNC system each time a probing cycle was completed. It is the responsibility of the DNC software to receive this information, compare the measured dimensions to the "target" value and create a tool wear offset command for one of the turning center CNCs.

If a measured value required too much of an offset change, or if the tool wear offset value went beyond a preset limit, an alarm message is generated on the PC and the designated side of the turning center is prohibited from beginning the next part cycle. However, the DNC system permits programs from the other side of the turning center to keep running so that operation of the cell may continue.


After discussing our software requirements with Mr. Fritz, he assured us the DNC software could, in fact, perform all functions we required but cautioned us that any software supplied would probably not work 100 percent correctly on the first try, and that some quick and easy method must be devised for software revisions and debugging. We decided that 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 Machinery.

If a revision were needed, we could simply call the company for suggested changes to the source code file. We could then recompile the software and try the system again. This procedure gave us a greater understanding of how the software worked, and it minimized travel expenses in the development process. Also, by having access to the source code, we could be sure that we could make any future changes without necessarily involving the software developer.

Currently, we are running statistical analysis to ensure that quality parts are being produced under the DNC system and the PC database. By manually inspecting finished parts, we can use X-bar, run charts and histograms to determine our process capability. Our target capability index is 1.5 Cpk or greater. Over a period of time, manual inspection will be greatly curtailed to reflect the high level of confidence in the capability of our process.

Although many technical problems had to be overcome with this project, we were able to accomplish our objective of creating an unmanned robotics cell that produces quality parts. The results of our efforts have been very satisfying. Perhaps more important, The Falk Corporation was able to coordinate the efforts of several talented people from several different companies to accomplish a complex task. Each person contributed greatly to the project and the cooperation of everyone involved was essential to success of the project. It is also a sign of the times that virtually all information was exchanged by phone, fax or modem, and most of us have never met in person to discuss this project.

The author wishes to acknowledge with gratitude the substantial contributions to this article made by her colleague, Mike Thompson, and Dan Fritz of Suburban Machinery.

Comments are reviewed by moderators before they appear to ensure they meet Modern Machine Shop’s submission guidelines.
blog comments powered by Disqus
The Field Guide for Automotive Management
Channel Partners
  • Techspex