By Larry Stockline
Senior Applications Engineer
MTS Systems Corp.
EDITED BY John R. Gyorki
Springs with steep force-deflection curves are difficult to install automatically and accurately in mechanical assemblies using presses that depend on displacement and position feedback loops to set the force. The problem arises from the variability or tolerance of the absolute spring constant from one spring to another. For example, a spring with a nominal rate of 1,000 lb/in. could have as little as a 0.01-in. deflection variation from unit to unit.
This represents an error of 10 lb under a retention force of 1,000 lb, and for some applications, 10 lb is grossly out of tolerance.
Conventional wisdom says the typical approach to solving the problem is to tighten the manufacturing tolerance of the spring so that the final product falls within a narrower window established by the Quality Assurance department. This calls for higher quality springs, and of course, higher cost. The approach also increases the cost of the assembly press system.
But Promess' President Larry Stockline looks at the problem from a different point of view. His customer handed him a specification that the final product must meet. It's a functional specification, so setting the spring to produce a specific retention force from unit to unit — regardless of the length of the spring — is the problem to solve, not a displacement which depends on spring constant.
The goal, then, is to develop a motion-control system that can respond to a nominal displacement command to establish an initial spring force, then set the final specified force under a new, fine-tuned displacement. The system measures and records the final displacement required at the specified force setting, and the spring moves on to a second assembly station where it is inserted into the final assembly and capped in place. The cap can now be inserted under the final measured displacement required for that particular spring, so the precise force is accurately set in the finished product.
After having evaluated numerous control systems without success, Promess teamed with the Automation Div. of MTS Systems Corp., Eden Prairie, Minn., and developed a technique that simultaneously measures displacement and force to ensure a constant spring retention force, regardless of the length of the spring. This system allows for an even wider spring-constant tolerance than required in previous tests which, in turn, reduces the cost of the spring by 50%.
Although it's not particularly difficult to manually set a specific retention spring force using a pressure or force gage and dial indicator, doing it automatically to meet production rates is another story. The problem with automating the assembly machine hinges around the servosystem's ability to deliver step-by-step precision and maintain tight repeatability for consistent results. In fact, precision indexing requirements for some applications need accuracy measured in microns, and this is difficult, if not impossible, to set manually.
Most off-the-shelf servosystems can't do the job. They typically use tachometers in a velocity feedback loop and cannot stabilize sufficiently well under the extremely small incremental indexes commanded by the controllers. The instability stems from slow loop update rates in the amplifiers and drives, as well as small physical perturbations coupled back into the control system from electrical noise riding on relatively high values of motor current. Other factors also contribute to the instability, such as minute temperature changes that tend to alter the spring rate, dynamic changes under force that produce a bouncing or hunting response, and simply no response to a command because the drive's deadband is larger than the signal that commands a small, incremental step.
These were the obstacles facing Promess before it started testing the MTS MaxPlus FLX drives and servomotors. Larry Stockline recounts, "When we began to do development work with MaxPlus drives, the incremental steps at the output were so small we couldn't possibly see them. From past experience, we didn't realize that drives could step one micron and stabilize. The computer indicated we had done so, but we didn't have an instrument with high enough resolution to measure it. At first, we didn't believe the results because we had never seen anything so stable." One tap on the keyboard produced one micron of displacement, but it took 10 taps to move 10 microns before the motion could be measured.
"One significant difference in the MTS System," says Dierk Benecke, MTS Automation senior applications engineer, "is a quadrature, incremental encoder that computes a 1/T algorithm for velocity and position. These digital encoders are capable of finer resolution than analog tachometers with outputs that must be integrated to establish position." And tachometers are more susceptible to signal drift which adds to a shift in tolerance. For another, the MTS MaxPlus FLX drive is sinusoidally driven. This approach eliminates torque ripple from stepping motion and motor tooth detenting. The stand-alone, fully digital MTS amplifiers deliver a 25- sec current-loop update rate and maintain low electrical current noise, says Benecke, along with the flexibility to increase machine output and productivity.
The success of the MTS system let Promess develop their new Electro-Mechanical Assembly Press, called EMAP. With the MTS servosystem at the heart of the EMAP press system, it can function at high speed, under high loads, with high accuracy. The product line uses MTS MaxPlus FLX drives and servomotors ranging from 10 to 1,000 lb-in. torque and delivers precise linear force from 5 to 120 kN. The result is an assembly system in which products consistently achieve functional specifications. Within user-defined limits, it is theoretically possible to "clone" good assemblies and reject bad ones with near100% reliability.
Fine-Tuning The Meaning Of Quality
Although many companies have programs aimed at improving product quality by orders of magnitude, the underlying goal by all is to drive component parts tolerance to zero, regardless of where that part resides. It may be a final assembly, or a part of another, and its tolerance dictates whether it fits or not. But more important is the intent of tightening the tolerance. And that is to make the completed assembly meet the final product specification — the functioning specification — and survive adverse environments or just live a long time under intended use.
But for virtually all of the last century, inspection-based quality control usually consisted of checking part dimensions after the manufacturing process was complete. This means if a process defect results in bad parts, it is not discovered until after that fact. This can be a costly mistake.
In addition, product quality is often not a function of individual components meeting physical specifications. Systems that are assembled improperly will not function as intended, even when every component is made to specifications. That's why it could be a smart move to take another view of quality and what it means in assembly operations.
Take assembly presses for example. EMAP has built-in force and distance sensors coupled to servomotors containing encoders that drive a ball-screw ram. They also can process a variety of external sensors and gauging signals. This brings the same closed-loop precision and repeatability to press-assembly operations that CNC machine tools bring to metalcutting operations.
All data gathered from assembly operations are displayed with force over position curves. By setting upper and lower tolerance limits for this signature curve, good assemblies are certified in process without subsequent inspection. The precise shape of the signature also provides information about individual assembled parts that can generate recall information for warranty claims and input to control the strategies for other processes. For example, events such as no part present, wrong part, part upside down, and wrong-size part produce a distinct signature change.
But an EMAP system is not just an alternative method to consider for the mechanical assembly of two parts, it is a whole new way of thinking about assembly operations. It integrates assembly, quality assurance, and process monitoring and control into one highly flexible system, and the implications are far reaching.
For example, EMAP is an ideal solution for flexible assembly operations on families of parts or even a series of entirely different parts. Like a CNC machine tool, an EMAP can store hundreds of part programs and execute them on command.
Assuming common tooling, a family of parts can be assembled. Even where tooling must be changed, it can be done rapidly. Because the EMAP stores the assembly signature as well as the assembly program, multiple assemblies can be produced with complete quality assurance, eliminating the need for multiple inspection-gaging systems downstream.
EMAP also operates in applications where one component must be inserted into another and then either retracted or stroked to test its operation. Many fluid assemblies fit this category, as do products like medical syringes and antilock brake assemblies. The system's ability to monitor force over distance can be used to check the fit of the two components. It can also be equipped with external sensors to measure and gauge other properties such as part or fixture deflection, temperature, flatness, dimensions, and so forth. One possibility might be to cycle a valve detail while flow is measured, as part of the final assembly calibration. Hydraulic-flow-control valves can be calibrated in this manner as can gasoline and diesel fuel injectors using a nonflammable fluid.
EMAP is also used to assemble rubber bushings and other flexing materials into metal parts. Here a push-measure-push strategy can be used to acquire the final required tolerance, even when the materials are flexing during assembly and a high rate of response is needed. For example, consider installing two bushings, inserted simultaneously in automobile suspension control arms. The control arm, as well as the rubber bushings, move as they are being inserted. This is easily compensated with the EMAP and can reduce scrap rates by 80%. Because quality is gauged and monitored in process, in real time, not after a part is produced, such low scrap rates should come as no surprise.