Vibration is that repeating switch from kinetic to potential energy resulting from pushing, stretching, or suspension. Because it’s so often observed as waves through matter, it’s also the most convenient means of communication around. But what do vibrations have to say? It’s easy enough to understand another person speaking, because we’re wired to process words. For vibrationally complex working industrial machines, it’s impossible to interpret signals with perfect accuracy. They way systems vibrate depends on complex mass, damping, and energy propagation properties. But with a few reasonable assumptions, sophisticated materials effectively damp, and new software and measurement equipment allow analysis of critical vibrations.
A real world, empirical approach is increasingly effective insurance against machinery wear and failure caused by vibrations. In fact, more plants are making vibration analysis standard protocol so that machinery runs safely and efficiently between maintenance times. For a better picture of plant machinery health, vibration, speed, and other condition indicators are increasingly integrated into a common database. In addition to including this integration capability in their diagnostic system, Rockwell Automation, Milwaukee, Wis., has recently improved their information system to sift out error data points. This equates to fewer false alarms and more accurate maintenance calls. The Microsoft Windowsbased software system is also compatible with portable data collectors, which have becoming increasingly popular since their introduction in the 1980s. As an economical monitoring alternative, National Electrical Carbon Products, Inc., Greenville, S.C., has developed one such portable FFT vibration analyzer that displays frequency and time wave form information on the unit’s screen as a technician makes measurements. The unit includes spectral analysis software to help grade overall machine condition picked up by the unit’s dual accelerometers.
Results are compared to automatically generated spectral alarm bands based on a reference guide called The Proven Method. Created by Technical Associates, a vibration analysis teaching and consulting firm founded in Charlotte, N.C., The Proven Method is a 70-page compilation of hundreds of machine types and their corresponding application formulas. The formulas define how to apply six different band alarms across one frequency range; each band is focused on a certain machine problem (such as alignment, imbalance, or looseness). Jay Gensheimer at National explains, “Data for very specific types of machines provides an advantage when compared to the ISO standards which are acceptable, but very general, and watered down. It’s sort of like driving across country with a GPS system in your car versus without one. You gain a lot more confidence about what you’re doing right off the bat.”
Teaming up with International Source Index, Inc., Williamsville, N.Y., National also loads premium units with data software that includes fault frequencies for more than 120,000 bearings. Besides indicating when something is damaged, the software can also listen to an unidentified bearing’s frequencies, and return a list of possible manufacturers.
Another route is to continually monitor a system for vibration changes. Because noise usually has many sources and is masked by structure-borne or amplified vibrations, isolating and investigating it is difficult. Reproducing noise in a laboratory is even more challenging. To bypass these difficulties, Intonix Corp., Roseville, Minn., has developed an acoustic machinery monitor that activates alarms when constantly monitored sounds begin to deviate from their norm. Using 96 frequency bands, it learns what a system or process sounds like, then continually compares records to present signals. Used on motors, fans, and gearboxes, the monitor provides another economical alternative to direct vibration monitoring. Although its output results are qualitative, not quantitative, the mountable spectrum analyzer does help keep personnel out of potentially dangerous locations. Senior product engineer James Radmonski explains, “The sounds coming from equipment and processes in a plant are not just noise. They contain information — information that can be used to prevent breakdowns and mishaps that damage expensive equipment, cause costly loss of production, start fires, and even injure or kill workers.”
Besides transmitting waves through the system, vibrating machine surfaces radiate pressure waves into adjoining air as sound. When these noises distract or even injure personnel, and when monitoring, tracking, and analyzing a system’s vibrations is of little diagnostic value, the best action is to damp it. In air, all sound frequencies travel easily, at a set speed. But inserting a soft material such as foam rubber or carpeting helps destroy sound’s pressure fluctuations. Wave energy is spent deforming the materials, and converted into thermal energy.
Continue on page 2
Common in industry, foam is usually produced in large loaves, then sliced into useable sheets, with varying degrees of accuracy. E-A-R Specialty Composites, Indianapolis, has developed a urethane acoustical foam made in sheets, for effective damping of frequencies 1,000 Hz and above. Used in vehicle cabs, lab equipment, and on mechanical devices housed in enclosures, the foam has a skin-like surface on both sides to protect it from contaminants and decrease noise permeability. Urethane rises between and bakes on to carrier paper for a smooth, denser surface; then, the paper is peeled off.
When possible, another way of damping empty, echoing cavities is to inject foam in them. By mixing quick-acting polymer catalysts during filling, an instant, lightweight, full acoustic seal forms. However, injecting does present challenges. Foam mixing releases a carcinogenic fume called methylene diphenyl isocyanate, or MDI, which must be within OSHA limits during and after filling.
Some foams on the market do have lower MDI emissions; however, their chemistry requires a ratio of the isocyanate reactant to the polyether polyol base of 24 to 1. Dr. Sven Meyer-Ahrens of the Bayer Automotive Business Group, Pittsburgh, Penn., explains, “The higher ratio, along with the high temperature required to lower the viscosity into a range that allows the use of low-pressure equipment, can lead to production difficulties.” In other words, mixing equal amounts of two ingredients is much easier than mixing wildly different proportions. Recognizing this, Bayer has recently developed a foam that requires equal amounts of isocyanate and polyether polyol. Called Bayfill, its lower MDI emissions meet OSHA standards, making it safe for workers. With its repeatable mixing ratio, it also satisfies OEM standards. Gary Karas of Orbseal LLC, Plymouth, Mich., the company distributing Bayer’s new foam, says he believes in a few years denser versions might further reduce vibrations by adding structural stiffness to the structures they fill.
Baffling the source
Instead of containing the noises of moving components, damping greases actually reduce them. These specialty lubricants decrease wear, contamination, and moisture like traditional greases, but their primary function is to control motion and noise. Focusing threads of binoculars, zoom lenses, and other optical instruments are improved by the velvety motion and silent operation afforded by damping grease. The lubricant also keeps lenses from coasting. However, high-shear situations quickly degrade the grease’s damping characteristics.
Seeing the potential uses in document printers, automotive tie rods, and other applications, engineers at Nye Lubricants, Inc., Fairhaven, Mass. sought a more robust version. By mixing high-molecular weight-base oils of their traditional greases with PTFE, they developed grease that damped motion, even after 153 hours of continuous high shear.
Though just released, a lighter weight of this grease has already been specified for automotive plastic switches. The off-white substance comes in a variety of viscosities; this makes it appropriate for softening the sound and feel of everything from computer keyboards to automotive suspension systems.
Noise isn’t always audible, or even mechanical. With the increased use of controllers and microprocessors, industrial environments are saturated with electrical signals that can cause electromagnetic interference, or EMI. Perfectly good output, by the time it reaches receiving electronics, can be corrupted to uselessness. Les Wolff of BEI Technologies’ Industrial Encoder Division, Goleta, Calif., explains that in encoder signals, EMI translates into detrimental noise, miscounts, and positional errors. A proper setup, he says, includes a compliment channel line — a twin — for every data line, to iterate signals. By twisting the two together, any EMI effects on one are cancelled out by the other; remaining noise problems are uncovered if the two channels’ signals don’t match at output.
Another problem is electrostatic interference. Woven metal shielding helps eliminate this problem by isolating the cables wrapped inside. Wirebraided, wire-spiraled, and tapewrapped shields made of copper, aluminum, or a combination exist. A drain wire (with lower impedance to carry more charge away) contacts the shield, and is eventually terminated with a connector pin. In many cables, an aluminum tape contacts the drain wire, and a copper braid is wrapped over that. If no aluminum tape is used, end users piggyback the shield, and then terminate. To cover all bases, LAPP USA, Florham Park, N.J., has developed a connector that combats EMI/RFIC problems. The contact spring in the SKINTOP MS-SC grounding connector establishes a continuous electrical path from the shield of the cable to equipment ground for a secure shield connection. The contact spring offers the added benefit of accepting cable diameters with wide dimensional tolerances. In a few tool-free steps, line strain is relieved, moisture and dust are sealed out with IP 68/NEMA 6 stringency, and the cable is grounded.
Continue on page 3
What’s that ziz?
Chicago Rawhide, SKF Group, Chicago, Ill.
When it’s hushed, smaller noises seem to come out of the woodwork. Drippy faucets boom, wind howls, and in quiet new cars, formerly unnoticed power steering units suddenly squeal. Because this increasingly common noise often seems to originate from seals, engineers at SKF’s Seal Division seeking criteria to predict rubber-on-solid sliding noises decided to investigate.
Typical of vibrations, the steering gear noise proved elusive; it only occurred in some cars and was extremely unstable, often coming and going from one day to the next. So the engineers removed noisy units from cars and set them up in a quiet laboratory. With a test rig to perform the difficult task of precisely modeling steering cyclic speed, pressure, oil temperature, and rack loads, torque, and noise measurements eliminated all suspect sources but the rack seals and piston seal. From there, a matrix experiment eliminated the rack seals. Laboratory tests confirmed the noise signature to be similar to that in field tests. Apparently, noise from the steering pump and other mechanisms excited the seals’ natural frequencies.
These kinds of friction-induced noises are the result of stick-slip instability of shaft-on-seal rubbing. The friction force depends on relative speed of the two components, decreasing with acceleration.
Usually quite helpful for analysis of harmonic oscillations, a noise’s frequency spectrum appears on a graph as a stand-alone peak. Even if it is masked by other noises, the background noise can be cancelled from the spectrum by noting where it is present and absent. However, noises like the steering seal zizzing are nonharmonic vibrations, appearing as sharp peaks in the time domain. This makes the time domain representation quite useful, because matching signatures verifies that samples come from either one vehicle, or a vehicle and a good model. However, those peak-like effects decompose into their separate frequencies in the frequency domain, making the representation practically useless. Those corresponding to the structural resonance frequencies would be amplified, and the others would eventually become dampened. Consequently, noise spectrum analysis could only give information on resonant frequencies and other harmonic excitations.