Ultrasound can certainly be a stand-alone program.  Today’s instruments give the end user the ability to perform the duties described above and more.  For instance, some of the manufacturers include waveform analysis software and have features such as an RPM sensor, infrared temperature sensor and onboard recording chips to store and play wave files for diagnostics.



Would you believe that during a recent maintenance conference, 60-65 percent of the maintenance professionals polled did not use ultrasound?  Most of the same group that was polled do have vibration, oil analysis and infrared programs, but no ultrasound.  I find this incredible.  Can you imagine all that low-hanging fruit?  Why are there so many  plants and businesses that rely on vibration or infrared technology but do not have ultrasound?  Is it that ultrasonic technology is not considered technical or objective enough?  That’s only a guess.  But whatever the reason, many companies are simply losing out on a quick and easy way to save good money for a relatively small investment.  And with a little more investment in training, they could be realizing tremendous returns.

How much training is needed to use an ultrasound instrument?  It simply depends on what you want to accomplish with your ultrasound program.  Do you want a world-class predictive maintenance program or simply a program to locate compressed air leaks or to listen to a bearing or two?  Certification is a noble pursuit (particularly if you are compensated financially) because it expands the user’s knowledge and capabilities. But, otherwise a 2–3 day ultrasound workshop is all that is needed to start saving a company money.

You may be reading this article and thinking that all you need is a little familiarization with the technology.  So, I have written a few briefs in this article to help you locate the leaks, listen to a motor bearing, diagnose a steam trap, and possibly hear corona tracking and arcing.

Air Leaks — Over the years I have had technicians say that their work environment is too loud for ultrasound to be effective.  Nonsense!  I have been in some of the loudest environments, everything from being next to a 650MW turbine set (low and high pressure) running in a power plant looking for leaks to being under a Boeing 737 jet aircraft trying to locate control air leaks (25-30psig) while the aircraft’s auxiliary power unit (APU) is turning within three feet of me.

Ultrasound is truly remarkable for finding air or gas leaks.  Yet, many people still think of it as snake oil.  I can’t tell you how many technicians would rather take a bottle of soap and water into a plant, hitting every single pipe joint or potential leak site looking for bubbles rather than take an ultrasonic instrument into the plant and find the leaks.  These same individuals will have to use a man-lift or climb up a ladder to get to a leak with the soap and water technique.  This is not only dangerous but a waste of time.  If you have an ultrasound receiver and the right accessories, you don’t need to climb or reach for the ceiling.

In a previous article, I used an illustration of a technician who needed to find a leak that was creating a terrible hissing in his plant.   No one had been able to locate this leak.  The fact is low-frequency sound (which you and I hear) is high energy, but is omnidirectional.  Who knows what direction it’s coming from?  This part of the plant had a small office and above that office next to an air compressor was a mezzanine in which several 3/8” and 1/4” nylon tubes were routed for control air.  When measured with an “A” weighted db sensor attached to the SDT170, this noise measured 110 db’s!  Wow…this sound was deafening. (What if this sound had been related to a potential production stopper?)  After using ultrasound he was able to locate the leak hidden high above the mezzanine.  A knife and a repair union was all that was needed to remedy the problem.



That leads me to another point.  When using ultrasound for surveying air leaks, let your eyes have time to coordinate where you are pointing the instrument.  Simply walking into a plant and swinging the instrument back and forth is not the most effective way to survey your plant, and many air leaks will simply be missed.



I prefer my “X” pattern survey (see Figure 4).  That is to survey in front of you with a slow and deliberate “X” pattern.  Whether you are using a parabolic dish or a close or long-range horn attachment, aim the ultrasonic receiver high and to the right of your filed of view, then slowly and deliberately bring the receiver to the left of your body and downward to the far left, making one leg of the X.  Then bring the receiver up and to the far left of your field of view and bring the receiver down and across your body to the lower right, completing the “X” pattern.  I have performed countless demonstrations and air leak audits and found this method less tiring.  My arms do not feel so sore after a survey and I could probably survey for several hours instead of just a couple of hours.



When it comes down to measuring leaks, you can either use one of the many manufacturers’ charts for converting decibels to cubic feet per minute, or, my favorite is to simply note the leak as either a small, medium, or large leak on the survey sheet.

Since many of the ultrasound instruments are not repeatable among instruments of the same model, you might want to reevaluate the idea of putting a cost on the readings you get.  Some of the earlier ultrasound instruments manufactured could differ from as little as two to as much as forty decibels between units of the same model, from the same manufacturer.

I should point out that there are many variables that can apply to every leak.  For instance, the material, the size and the configuration of the leak site (well-rounded or a jagged opening) can determine how quickly the leaking air can exit the leak site.

Today with “Cap and Trade,” and other environmental and energy reduction programs, plants have to be accountable for their reports of savings through audits or studies by an independent lab.  I think that the “decibel to cfm charts” can lead to plants claiming inaccurate energy savings.  Fair and accurate reporting of savings is needed, with clear documentation of those savings.

It would be pretty embarrassing for a technician to document 1200 cfm’s of leaks in his plant on a system that can only supply 1100 cfm of air.  I have heard about reports like this over the past 15 years.  Ultrasound can contribute to significant savings, but some numbers I’ve seen seem just too large to be true.

Mechanical Survey — When performing mechanical surveys, you should remember that sound travels faster in solids.  This is because the particles (atoms or molecules) in a solid are touching each other and rather fixed together.  Since the particles are bonded together, a sound wave moving one particle can immediately transfer the motion to another particle touching it.  However, temperature, material, and frequency can affect the outcome.  Generally speaking, the speed at which sound travels is fastest through solids, then liquids and, finally, gases.  Sound waves travel faster (velocity) through harder materials of a lower density.  For instance, titanium has about half the density of steel and aluminum is about one-third (Table 1).  So, all variables being equal, sound travels slightly faster (velocity) through aluminium than it does through titanium or or steel. 



So keep this in mind when using the contact probe on motors with different materials that house the bearing and motor.  Use of the contact probe to diagnose bearings is preferred. Most companies have a contact probe with a rod or stinger to touch the surface of the motor.  Using the rod or stinger, simply  touch the bearing housing.  If there is a zerk fitting atop the bearing housing, I like to seat the rod end or point in the threads at the base of the zerk fitting.  If there is not a zerk fitting, but there is a plug where the zerk would have been, use this location (see Figure 5).  Whichever location is used, care must be taken on subsequent readings to come back to the exact location.



If you were to take a reading at a point where the metal was thicker between the housing and the bearing versus the distance from the previous read, you may end up with a reading of less decibels than previously read.

Magnetic sensors are great too.  If you have one that has the power to hold to the surface (vertically or horizontally) no matter what, than you have the correct strength magnet.

I am not a big fan of listening to a bearing purely on the sounds I hear, unless I am using an analog ultrasonic receiver.  Today’s instruments are mostly digital instruments.  For me, using an analog instrument may be the difference between hearing a ball out-of-round or not hearing it, or hearing the sound of a defect on a race or not hearing it.

Many of today’s instruments offer waveform analysis to help you diagnose a condition or defect. Still, unless you’re qualified in vibration analysis, many of us cannot define what we are seeing.

In general, first establish a “baseline.”  This may be accomplished by two methods: comparison and historical.  Comparisons can be done by comparing one motor to another motor of the same rpm, loads, etc.  It must be performed using the same location, i.e. the drive end of one motor to another motor’s drive end, nondrive end to another motors nondrive end, etc.  If you have several motors that are alike, take readings and compare all the readings of the drive ends.  The bearing with the lowest decibel reading becomes the baseline for the others (do the same for the non-drive end bearings).  However, if a bearing is removed and replaced it is important to give the motor some “run-in” time before taking a new reading to be used as a baseline.

I prefer creating a baseline and then watching the bearing over a period of time.  When the bearing decibels rise in the 8–10 db range over the baseline, I would lubricate the bearing, either  acoustically using an ultrasound instrument, or, manually using a grease gun and following local standards or procedures called for by the plant.



When bearings reach a preset (predecided) decibel level that warrants removal and replacement, I might then (if applicable) invite a vibration analyst to analyze the bearing.  Sometimes, integrating other technologies like vibration, oil analysis, infrared imaging and ultrasound may allow the bearing to continue running for several more weeks or months. 

What decibel range would be considered a failure?  Who really knows?  A failure rate or decibel that represents failure can be made up from many different factors.  A very small bearing or a very large bearing may have the same decibel range for failure. 

So, how do you determine a failure rate?  By first taking several readings and comparing them to acquire a baseline.  After a baseline decibel is established, you can then set a decibel level as a “failure.”  One manufacturer says that when a bearing exceeds the baseline by 12 decibels, it’s in an “incipient stage” of failure.  Incipient simply means the beginning or to become apparent.  So, once it enters this stage you are to make a determination as to what would constitute a catastrophic failure or a reason to remove and replace.  Which manufacturer’s instrument you are using will determine what decibel range you consider failure and/or a catastrophic failure.

NOTE:  Always consult with the manufacturer of your instrument directly about what decibel range above baseline may be considered a catastrophic range.

I have searched many owners/user manuals available online for the top instruments and cannot find a definitive answer in any of them.  They mention pre-failure and incipient failure, but nothing saying what decibel range indicates a failure and when a bearing should be removed and replaced.

If you start a trend of reading your bearings and you see the decibels steadily rise, then you can assume a problem is occurring.  For example, let’s say you have a two-year trend of a bearing that was reported for several months with readings between 25–30 decibels, yet slowly over three to four more months the level continues to climb steadily.

You might say a 30–35 decibel increase over the original baseline would warrant removing the bearing at the next convenient opportunity.  A “catastrophic failure” may be 45–50 decibels over the baseline, which means remove and replace now rather than later.  Due to its early warning (there is actually ultrasonic energy before there is vibration and heat), ultrasound can be very subjective.

Personally, I believe the number of decibels is just a reference point anyway.  Watching the decibel level climb over a period of time means the bearing has a problem.  It’s either degrading or there is a lack of lubrication.  Simply graph the readings over a period of time and, as the decibels reach a previously decided level or range, you can move the inspections closer together.  For instance, instead of every 90 days, now you might inspect every 60 or 30 days.  Depending upon the criticality of the bearing or motor you may take a reading weekly or even daily.

Many of today’s manufacturers have some very good trending software as well as waveform analysis software.  But, even the basic pen and paper will allow you to monitor motor bearings.  Once you start, who knows?  Maybe you’ll be buying a bearing trend phone application for your Blackberry or IPhone.

Electrical Inspections — These applications are still among the most underutilized of all the capabilities of ultrasound.  I think the main reason is because most technicians are unfamiliar with the application, and don’t understand what they are supposed to hear.

Unlike listening to bearings and watching decibels climb, when scanning electrical switchgear one must listen to the sounds and try to identify what he/she is hearing.  Is it arcing, tracking and/or corona (either destructive or nuisance)?  Fortunately, there is software out there with which a technician can record the sounds and play them back through FFTs or in Time Waveform (see Figure 7).



Sounds to remember:

  •  Arcing, an abrupt start and stop popping sound.
  •  Tracking, this sounds like baby arcing,climbing then
      discharging and starts all over again.
  •  Corona or nuisance corona has an egg frying or bacon
      sizzling sound.  It may not require attention now, but
      left unchecked could find a path to ground.
  •  Corona or destructive corona exists to find ground. Corona
      should be removed and a remedy to prevent corona
      should be put into place.

Ultrasound should be used to scan any and all electrical panels or doors for arcing, tracking or corona discharge prior to opening.  If you don’t learn anything more from this article regarding electrical inspections than the prior sentence, you have still gained a lot.

If you are an infrared user and do not use ultrasound in tandem with your electrical inspections, or your third-party inspector does not use ultrasound, you are missing out!  Accidents, downtime, and possibly saving a life can be attributed to those that use ultrasound during electrical inspections.  Unlike infrared thermography and corona camera inspections, ultrasound does not require line-of-sight, so it is effective when some other technologies might not be.

Steam Traps — These are defined as either working or not working.  Ultrasound, along with infrared, is a surefire method of determining the condition of a steam trap.  The three more common traps are the Dish Trap, Inverted Bucket Trap and Thermostatic Trap.

All traps make a particular sound of distress. Motor boating or quick popping sounds could describe the disk trap when the disc is badly worn.  If the trap has failed you will not hear it open and close.

The inverted bucket trap has a bucket and linkage.  Condensate and air enter under the bell or inverted bucket and air flows through the discharge orifice.  After all the condensate and air are removed, steam reaches the trap and floats the bucket, closing the valve. When condensate and air enter the trap, the bucket loses its buoyancy and drops.  This opens the valve and the condensate and air are discharged until the steam again floats the bucket, closing the valve.  Normal failure may be either open or closed.  The thermostatic traps are designed with a float inside that is raised by condensate.  When released, steam can be heard escaping through the vents with ultrasound.

Simply listening to any steam trap is not enough.  Knowledge of the cycle, temperature and load is very important to know.  Start listening now, as the more you listen to, the better you will become at diagnosis.

So, no matter what your skill level is, pick up your ultrasound instrument and use it.  Using this instrument can easily improve your troubleshooting skills, and who knows, it may locate that anomaly early enough to avoid some real headaches or downtime.  If you do not have an instrument, contact the different manufacturers and ask for demonstrations.  Do your research, buy an ultrasound instrument and start using it.  Believe me, the savings will add up in no time at all.

Unless noted otherwise, all photos courtesy of Ultra-Sound Technologies, Inc.

Jim Hall is the president of Ultra-Sound Technologies (UST), a vendor-neutral company providing on-site predictive maintenance consultation and training.  UST provides an Associate Level, Level I & II Airborne Ultrasound Certification.  Jim is also a regular provider of on-line presentations at Reliabilityweb.com and is a contributing editor for Uptime Magazine.  Jim has provided airborne ultrasound training for several Fortune 500 Companies in electrical generation, pulp & paper, petrochemical and transportation (marine, automotive, aerospace).  A 17-year civil service veteran, Jim served as an aerospace engineering technician for Naval Aviation Engineering Service Unit (NAESU) and with the Naval Aviation Depot, Jacksonville Florida (NADEP). Jim is also president of All Leak Detection, LLC, a leak detection company providing air leak audits above and below ground leaks.  Jim can be reached at (770) 517-8747 or  jim.hall@ultra-soundtech.com