I’m not sure but I think it was the really old Batman and Robin series with Adam West where the show would end with the dynamic duo in sort of real trouble.  Some place along the line they’d pick up the story with a line something like “when last we checked in with…” well that sort of applies to my efforts to collect new data on Cleveland Vibrator’s line of pneumatic piston vibrators, both impact and air cushioned. Well…when last we check in… Cleveland Vibrator was making efforts to better document the performance of the piston vibrator line and we’d just recently purchased a CoCo-80 Dynamic Signal Analyzer as part of that effort.  I’d Blogged about my efforts and initial results a while back.

The initial plan was to capture the acceleration of a heavy plate to which we bolt an impact vibrator or air cushioned piston vibrator and isolate the plate with very soft airmounts.  Then using the equation F=ma, we’d calculate the force output of the vibrator, it all seemed reasonable.  At the time I wrote about data collected on the VM-25, miniature piston vibrator.  I was able to capture the acceleration of the heavy plate as it was produced by this tiny vibrator.  This gave me a calculated force output for the VM-25 of something in the neighborhood of 60 pounds of force, a force output that really just doesn’t make sense given the size of the vibrator and the piston.  The problem only got more perplexing as I collected data on other units.  Data collected on the 1125 VMR, rectangular based impact piston vibrator, gave me acceleration values in the 60-70 g’s range.  As I increased the sampling rate of the CoCo-80, I was recording higher levels of acceleration as a result of the steel on steel impact as the vibrator’s piston strikes its base.  The CoCo-80 was catching the very brief impact and displaying that acceleration.  Unfortunately, if we were to take the 60 g’s of acceleration and calculate the force output, we’d be looking at a force output in the area of 1800 pounds of force… highly unlikely.  So as I’d mentioned in the previous blog, it was back to the drawing board, approach number one wasn’t providing the sort of information that was meaningful.

Taking a step back and looking at what tools and equations we have to work with that might give us more meaningful numbers, better tools for our customers to work with.  During the design of vibrating equipment we will typically calculate the amplitude of vibration of that unit.  It’s helpful to know what the expected peak to peak displacement of the unit will be based on the mass of the unit and the input force of the rotary electric vibrators.  We use the equation; Amplitude = (70452 x F)/(W_V x RPM²) where W_v is the total vibrating weight of the equipment.  So for example, on a recently designed Electromechanical Vibratory Screener (EMS), the estimated vibrating weight from the computer 3d model was 2875 pounds.  The unit is designed with two rotary electric vibrators, RE-34-6, producing 7480 pounds of force each.  Using the amplitude equation we would expect to see a peak to peak displacement for this particular design to be 0.255”.  Useful information when selecting isolators and verifying the performance of the unit.  Potentially helpful when thinking about testing pneumatic piston vibrators, we’ve seen this equation hold up when we do our final test for our equipment.

So stepping back to our original problem of trying to quantify the performance of our line of air piston vibrators, it’s helpful to think about what an end user is trying to do with the vibrator.  When applied to a hopper as a flow aid, the end user wants it to work, get the material flowing.  Part of getting the flow going is movement in the hopper wall produced by the vibrator, what “work” is being done.  So it made sense to me to look at what “work” was being done by an 1125 VMR while mounted on my test plate.  The 60 g’s of acceleration that happens almost instantaneously as the piston strikes the base and what that would calculate out as a force really isn’t helpful in solving problems.  Yes, I can capture that data now but so what?  However, the displacement of the plate and the “work” being done by the vibrator to move that plate probably is meaningful.   So let’s go back to the amplitude equation and solve for Force by measuring the displacement or “work” done to the plate.

Rearranging the equation we get; Force = (Amplitude x W_V x RPM²)/70452, so now all we need to do is measure the displacement of the plate and solve for the force output of the vibrator.  Bam!  No Sweat… can do.  Just a few changes in the parameters of the CoCo-80 and instead of looking at the acceleration of the plate we can now look at the displacement.  Some things seem so simple and then it hits you.  While the CoCo-80 can in fact display the displacement of the plate, this is done via a double integration of the acceleration data.  So if I’m seeing a 60 g impact acceleration due to the spike of impact, the calculated displacement is huge.  It’s been a few months back but I think the calculated displacement of the plate when vibrated by the 1125 VMR was something like 0.31”; this simply didn’t pass the eye test.  While operating the unit it was clear by inspection that the displacement of the test plate was well below the value calculated and displayed by the CoCo-80.  So the double integration from acceleration to displacement was working off that high impact acceleration value and giving us an unrealistic displacement.  So it would seem that it’s back to the drawing board, again.

Fundamentally I think the approach was good, collect data on the displacement and calculate the force using an equation that we’ve used before and are comfortable with the results.  What was required was a way to directly measure the displacement of the plate.  After some searching of the internet I found a laser sensor that touted its ability to quickly and accurately measure distances and displacement.  I was able to bring in the local sales engineer, we discussed the application and he demonstrated a laser sensor.  We have purchased the sensor and now bring that analog signal into the CoCo-80 as well as the data for the accelerometer, channel 1 – acceleration, and channel 2 – displacement.

Now armed with a dynamic signal analyzer,  two different types of sensors, an upgraded flow meter and plenty of industrial pneumatic vibrators, the horizon if full of future tests.  My first tests have been on our 1125 VMR, had it in place previously so it made sense to start there.  I’ve in fact tested three different units, collecting performance data on all three.  As one would expect as we increase the inlet air pressure the force output of the unit goes up.  Does it match the performance data passed down to us by The Elders, in a word, No.

So what do you do with that information?  Right now I’m still evaluating and considering next steps.  Some here have suggested mounting the piston vibrator on a heavy test block as used with turbine or ball vibrators and running again.  This type of mounting is beneficial to the performance of pneumatically powered rotating vibrator such as ball or turbines.  Heavier more rigid mounts will result in higher operating speeds and with the higher speeds; the calculated force output will be higher.  This method of determining turbine and ball vibrator force output seems to be the industry standard but the minute you mount the vibrator on something less rigid, a hopper for instance, the rotational speed drops and with it the force output.  The idea behind the suggestion was to see if the vibrator would run at a higher frequency on the heavy block.  Not sure that what we see with rotating vibrators would hold true with a piston vibrator.  Higher frequency, higher force output.  I’ll probably do that testing soon just to see if there’s a difference in frequency, probably learn something one way or the other.

Now the question is what to do with the new data as we work through the product line?  Hmmm, that’s the big one for sure.  As I mentioned in my first Blog on “Vibrator testing and data collection,” the industrial vibrator industry most certainly seem to function in the world of “close enough for horseshoes and hand grenades.”  Most folks want the “right” vibrator for their application, I don’t know that it matters if that selection process is based on hard empirical data or over 90 years of real world experience, or a combination of both, it’s just got to work.  So if my new force output is twice as much or half as much as what we’ve worked with for years, it’s highly unlikely that it will impact what vibrator our sales team recommends.  If we know that a 1200 VMS will work on a hopper with particular characteristics, that’s not going to change.  We won’t down size the recommended vibrator because we now have a better understanding of the force output of the unit.  The 90 years of experience is probably going to trump that new data.  That’s fine, at the end of the day we just want to get the best industrial vibrator into the hands of our customers, period.  The decision making process behind that recommendation probably isn’t important to 98.4% of our customers, just the solution.