PART THREE OF A SERIES

Methods and Application of In-Die Dimensional Measurement

By Bob Carabbio, Chief Engineer, Signature Technologies, Inc., Dallas, TX

Articles in this series have dealt with the growing necessity of making dimensional measurements while the workpiece is in the midst of the forming process. An extension of making these measurements is the use of dimensional information to reject any workpiece that is outside of established tolerances. In more advanced applications, these data can be used to adjust the tooling to bring the workpiece within acceptable tolerances.

In parts one and two of this series, we looked at how to make measurements of features on the workpiece while it is in the process of being formed in a die. In this concluding segment, we will examine what we can do with those measurements. There are a number of choices when it comes to closing the loop. These depend upon the robustness of the process and the nature of the measured condition.

Stop the Machine

In some cases, we may want to stop the process when things move out of specification. This is the simplest way to close the loop and put the measuring system in control of the process. The monitor screen on the measurement system will indicate which sensor went out of tolerance, and since the machine is stopped, the operator can inspect the tooling and workpieces in the machine to see if something has gone wrong.

If the measurement system has "trending" capability--i.e., it can display the immediate history of the process--a decision can be made as to whether the process actually has drifted gradually out of specification, or whether an "anomalous occurrence" is the reason for the stop, in other words, something suddenly went wrong.

A measurement system can have several levels of fault generation with appropriate actions taken relative to the severity of the indicated problem. One strategy is to set the system up so that a small change can cause a cycle stop, while a larger change performs an immediate stop. Some systems also can supply absolute boundaries to process variables to avoid overloading the machine or tooling.

In any case, it will be up to the operator or tool setter to figure out what happened. Stopping the process is a good way to apply control when the likelihood of tooling damage is high if something goes wrong.

On the other hand, there are cases where we just can't stop the machine. Consider the case of a vacuum former for small creamer cups. The machine cannot be stopped with plastic material in the heating area without causing an even greater problem. In cases like this, a shut-down protocol may have to be developed to cope with recurring workpiece faults. Of course, actions also can be taken to prevent defective parts from getting into the "good parts" bin.

Rejecting the Part

Sometimes, we may want to let the process continue, but segregate workpieces that do not appear to be compliant with specifications. In such a case, the measurement system must have part-tracking ability, so that when a measurement violation occurs, the offending part can be "tracked" out of the machine to the reject station, and thus eliminated from the process stream.

Fig. 19--Reject part tracking.

In cases where progression of the part in the tooling is small, as in the connector industry, the tracking system may need a storage capability from 50 to 75 progressions before the faulted part(s) reach the rejection point, Fig. 19. Obviously, the ability to track multiple parts also is needed.

Ideally, each sensor station should be programmable as to how many cycles should pass before actuating the reject mechanism, and for how many cycles the rejection mechanism should operate--in case a number of parts aren't under strict control. The rejection method could consist of parts sliding down a chute, or being diverted onto a conveyor. It's better to discard three parts to be sure of getting rid of a bad one, than to take a chance of missing it.

To assure machine and tooling safety, the measurement system typically is set up so that a single fault will track out and reject, while multiple faults--on the same station--will stop the process. The measurement system must be sensitive to the likelihood that a faulted part on station number 3, for example, also may fault on station number 4, as well as the following stations. That should be considered "normal" as long as station number 3 has come back into compliance.

If, on the other hand, every part starts faulting on station 3, there's probably something wrong with station 3, and after a determined number of faults, the system will stop the process. The measurement system also should remember that there now are "bad" parts in the strip, which still must be rejected--unless it's told specifically to "forget" them because of rethreading. In cases where the process is running "reel to reel," it may be necessary to limit the number of rejects allowed on each output reel.

In cases where there is a programmable logic control (PLC) system on the machine, some of these auxiliary control functions can be accomplished by providing inputs to the PLC, and adding the appropriate "rungs" to the program ladder. In the case of faster machines, or older PLCs, we must be careful that the PLC is quick enough to provide the needed control.

Automatic Tooling Adjust

The most elegant use of measurement information lies in making the process adaptive. In this case, the measurement information is used to control the process so that parts that are drifting toward an out of specification condition can be adjusted back to nominal values automatically.

In cases where the measurement relates directly to setting of some portion of the die structure, closed loop adjustment is possible. In these applications, the measurement system can control adjustment of the tooling in order to correct a changing process. Before this can be done, of course, the machine and/or tooling must be capable of dynamic adjustment. This probably is the greater part of the application challenge.

A typical application for this type of system is the common practice of dynamic shut height compensation on variable-speed machines used in lamination blanking. Since the penetration of the punch into the die button is very slight, the machine shut height must be set so that the punch separates the slug at low speeds. At high speeds, the punch penetrates too deeply into the die button causing excessive wear from stripping abrasion.

A sensor is applied to the face of the ram, which monitors the lowest position. When the ram starts penetrating deeper at higher speeds, the shut height drive "backs off" the adjustment to compensate. This maintains the punch-to-die button relationship at its optimum value at all speeds. When the machine stops, the system automatically runs the shut height back to its nominal setting for restart.

The same type of dynamic adjustment can be accomplished within the tooling for correction of bend angles, form heights, coining pressures, etc., so long as the die is configured to allow settings to be changed while running. However, first we must confirm the relationship between a measurement and an associated practical requirement.

A Case in Point

The can industry found that, within limits, the depth or condition of the score was less important to the proper initiation of a score fracture (opening the can), than the amount of work hardening introduced by coining the initial break area. No amount of measurement and adjustment of score depth would correct a poor opening condition, which actually was caused by insufficient coining pressure three stations earlier.

However, when we're dealing with bend angles, the solution generally is fairly simple. We simply must adjust clearance on the wiping edge for more or less clearance.

Fig. 20--Sliding wedge adjustment.

In the application shown in Fig. 20, a "sliding wedge" adjustment method is used to elevate the lower die portion of the station that produces the final form on the curls of a female blade-type connector. The brunt of the force is taken on the wide surfaces of the sliding wedges. The stepping motor, pictured in Fig. 21, only has to move the lower wedge by turning the adjusting screw during portions of the press rotation when no force is applied. If the curl starts coming in too close (measured by the system shown in Part Two, Fig. 15), the stepping motor can withdraw the wedge gradually, while the measuring station keeps track of the dimension of the curl. This principle is amenable to the upper punch, which can be spring loaded against a sliding wedge system and adjusted the same way.

Fig. 21--Linear stepping motor servo adjuster.

Many times, in both warm and cold forming, tooling is set up so that individual stations can be adjusted using a wedge-type adjustment mechanism that can be retrofitted with a dynamic adjusting system. Other times in a similar fashion, arrangement of parts in the tooling is such that a direct screw-type adjustment is more practical.

Fig. 22--Original station configuration.

As shown in Fig. 22, the original configuration of the station was amenable to both force measurement and adjustment. The realization of this is illustrated in Fig. 23. Since the stepping motor cogs, the adjustment screw is held in position when not being adjusted. Because of the small motion steps of the motor, a very fine adjustment is possible. Of course, if more powerful adjustment is required, different types of drive mechanisms can be used. A large AC motor, for example, can be bumped in small increments by pulsing its contactor. This may not be considered a precise method, but it can be a valid economic alternative to installing servos.

Fig. 23--Direct screw adjustment.

The Test Bench

The use of a sensor test bench is a common-sense approach to sensing and measurement, which has been introduced to the metalforming industry mainly by George Keremedjiev in his seminar presentations and monthly columns in MetalForming magazine. Simply put, before attempting any kind of measurement in production, we should simulate it under known circumstances. We also must have the ability to move elements of the system in a precise and repeatable fashion. This requires that a test bench be based upon a stable platform, Fig. 24. It should be equipped with a multi-axis position table assembly and associated bracketry and fixtures that can be used to mount measurement devices easily, but solidly, so they can be applied to the workpiece in a known manner.

Fig. 24--"Quickie" laser test jig for Faston terminal measurement.

In many cases the user will have to custom fabricate bits and pieces of the fixturing. Access to a drill press, grinder, lathe, milling machine, etc. will be helpful. Often an in-house tool shop will provide needed support in this area. If, for some reason, this isn't possible, other arrangements will have to be made. In extreme cases, a small lathe, grinder and drill press in the test lab usually will suffice. Test fixturing need not be good-looking, but it must be acceptably solid to achieve physical repeatability.

There also must be electrical test equipment including a good quality digital multimeter (two if differential measurements are planned), and a power supply with several adjustable outputs capable of delivering up to two amps at voltages up to 24 volts. This will be used for operating elements of the measuring system.

It goes without saying that someone with a good understanding of the processes must be available consistently to use the test facility. In choosing this individual, it probably will prove easier to train somebody who has good mechanical feel but no electrical background to use the equipment, than it would be to train an electronic technician with no mechanical sense to define the measurements. A metalforming operation that has a skilled diemaker that works with electronics as a hobby truly is fortunate.

The idea in using the test bench is to mock-up the measurement we want to make and see if we can perform it repeatably and get the proper results. We also can apply contamination (such as lubricant), and experiment with part positioning and the like, to get a feel for the robustness of the process.

Conclusions

While in-process 100-percent measurement of critical parameters is relatively new as a metalforming practice, most sensing equipment has existed and been applied in other industries. It really is a mature technology. It's newness in some metalforming areas stems from the fact that many potential users haven't bothered with it until now.

The German "ghost-shift" operation refers to a practice where, at the end of the day, they load stock onto the machines, hit the "start" button, shut out the lights and go home. This started back in the 1970s, yet, it still is almost unheard of in North America. The ability to run untended depends upon having and using reliable controls, which can ensure that we are making the right thing, and can have confidence that either we will stop it before we waste resources by making junk, or automatically correct the process if it strays.

The justification of any type of automation lies in its ability to increase the production of salable goods and reduce the labor content thereof. Making a valid measurement is crucial, but we must be able to realize significant benefit from having made the measurement for it to make economic sense.

We also must be satisfied to attack problems that are within reason. Starting with the "worst nightmare" potentially may get us the biggest and fastest return, but it also may prove to be out of reach for the people commissioned to do it. There is a learning curve, and it's best to address things that are more intuitive so that we can achieve a small but measurable success quickly and build from there. As confidence and expertise grow, so does our ability to benefit from the technology. MF

To obtain a reprint of this series, write number 400 on your reader service card.