How to shake the building
Dr. Don Stierman and students
Date of activity: March 22, 2005
Date posted: July 1, 2008

Materials: "Ranger" SS-1 seismometer and PS-2 smoked paper seismograph (manufactured by Kinemetrics).
1 tall building (but not too tall - 8 to 18 stories is usually about right, although shorter bell towers work well).
Seismology instructor and/or students (sense of rhythm is more important than body mass).

Theory: when a periodic force is applied to an oscillator that matches its natural frequency, resonance occurs.  Vibrations increase in amplitude (size).  When the periodic application of force ceases, vibrations decay to background levels.

Many of you may have experienced a sense of swaying on the top floor of a tall building, particularly in high winds or as waves from an earthquake shake the building.  This swaying is regular, like the rocking of a boat.  The frequency of oscillations for a building depends on the building height, it's cross-sectional plan, and the materials from which it is constructed.  I suspect that computer simulation of how a tall building will respond in event of a strong earthquake, based on analysis of the building's structural plans, is a complex and costly procedure.  A rule of thumb is, the natural horizontal period (period is the inverse of frequency) is about 1 second per 10 stories.  This rule of thumb suggests that a 10-story building will swing one complete cycle once a second, while the Empire State Building (102 stories) will sway about once every 10 seconds.  These are approximations - buildings of equal height but constructed in a different manner will probably not exhibit equal free periods.

Procedure:  Select a tall building and obtain permission to conduct this experiment.  We used Parks Tower, a 16-story dormitory on the main campus of the University of Toledo.


Parks Tower as views from the ground, northwest corner.

 

 

 

1. Set up the seismograph on the top floor, in the horizontal mode with the axis parallel to the direction you plan to shake..

 

 

 

The cable connects the SS-1 to the PS-2 like a wire connects a microphone to amplifiers and speakers.  Follow manufacturer's instructions on unlocking and centering the mass.

I reviewed a report in which an engineer conducted a ground noise test at a petrochemical site.  I could feel the ground vibrating, hut the report stated that ambient noise was quite low.  The figure showing his tap test on the seismometer revealed his error: he had not unlocked the mass, so the only vibration was from one of the internal springs.  I discovered how to make this error - and learned its signature wiggle - while still a student.  I suspect the engineer's bill for his 'test' was several thousands of dollars.

2. Turn on the PS-2, run a maximum drum speed.  Use the high-cut filter to eliminate vibrations other than the slow rocking of the building.  Sources of high-frequency noise include 60-cycle AC electrical power, motors for ventilation systems and elevators, and vibrations from dorm room appliances and speakers.   Note the gain setting that permits you to clearly observe the rocking (video) but no more.  The PS-2 has amplifiers capable of clipping the waveforms if the gain is set too high.

You may hear us talking about decibels.  As a rule of thumb (seismologists have lots of thumbs), a 6 db increase in gain setting doubles the trace amplitude (size of the wiggle) on the seismograph.  The lowest gain setting on the PS-2 is 0 db, but there is a preamplifier that kicks up the signal even when the panel dial reads '0'.

3. Set the PS-2 so that the needle swings parallel to the direction you are able to 'rock'.  As you begin to shift your weight in beat with the needle, the amplitude should begin to increase (video).

At 12 db, I was pegging the needle (driving it to its maximum).  We cut the gain by 12 db (a factor of 4) and tried again.  Even at the lowest PS-2 gain, resonance was observed.  Our first try did not produce good results because we were a bit out of sync.  The second try was more successful (video).

Shaking buildings is hard work.  I'm running out of gas (video) as we finally get our act together.

Note that the SS-1 and PS-2 are calibrated.  Using the generator constant of the SS-1 (about 337 volts/m/s) and the millivolts/millimeter table in the PS-2 manual, we can calculate just how much the building moves naturally and how much it moves when driven.  I'm not going to reveal here just how small these motions were (needless to say, students sleeping in nearby rooms did not notice the building shaking).

Trish and Levi were UT seniors when we conducted this lab.  Trish has since earned her M.S. in Geology (applying geophysical measurements to sediment characterization) and in now gainfully employed with an environmental consulting firm in Virginia.  Levi was, the last I heard, working out of Woods Hole, installing and maintaining scientific instruments just outside the surf along our East Coast.  Somewhere in my files I have a photograph of him wearing his wet suit.

We attempted to shake the 22-story One Government Center in Toledo but did not get good results.  The SS-1 has a natural period of 1 second and is rather insensitive to long-period vibrations.  One Government Center rocks, but too slowly for the SS-1 to accurately measure.

In another life, my students shook the bell tower at The University of California (Riverside) vigorous enough that the shaking was felt by people on the top platform.  It is my understanding that the Hoover Tower on the Sanford University campus has also been shaken, and not just during earthquakes.

Hit Counter