A Little Bit About Me

My name is Brian Thornock and I am a graduate student in physics at Brigham Young University.  I originally come from a small town in Washington State, but have been here in Provo, UT for the past few years.  Though I study physics, my emphasis is in acoustics.  Many people think that because acoustics is fun that it is also easy.  This is not the case.  Anyway, my specialty is room acoustics and I have done several projects in this field.  One of the most intense projects that I have done is that of constructing an anechoic chamber.  Don't know what one is?  Well, keep reading and you may learn more than you ever cared to know.

It may seem like an odd thing to have a page dedicated to anechoic chambers, but after building my own for BYU, I thought it would be useful to provide some practical information that could be used to help others.  When I began trying to design my chamber, I found a surprising lack of information and guidance in this area, so here we go.  I will try to provide as many pictures and diagrams as I can so that should you choose to work on one of these, your experience might be less frustrating and time consuming than mine was.

What is an Anechoic Chamber?

Anechoic chambers are special acoustic test chambers that absorb at least 99% of the energy that hits the rooms walls (In order for a chamber to actually be an anechoic chamber it has to be tested to ensure that at least 99% of the incident energy is absorbed).  You may have seen pictures of one in which there are funny looking wedges covering the walls.  These specially designed wedges are made to optimize the amount of sound they can absorb.  Ideally, an anechoic chamber simulates what is called "free space" or sound transmission in the complete absence of any reflecting surface.  Of course, achieving this is impossible, but a 99% absorption threshold helps ensure that the environment is as close to free space as possible.

History

Anechoic chambers first began to appear in the United States around the time of the second World War when scientists wanted to take measurements in the absence of any extraneous noise to better understand the device-under-test (DUT) (a device under test can be a microphone, loudspeaker, fan or anything else that has anything to do with sound).  One of the first ones was built at Bell Laboratories under Harvey Fletcher who was also the first president and one of the founders of the Acoustical Society of America.  Since then, lots of advances have been made in terms of materials and designs so as to make today's chambers the best to date for any of a wide range of tests.

Purposes

Anechoic chambers are extremely useful places if you do anything in the acoustics or audio industry.  They are used for projects ranging from testing noise floors on small circuits to measuring sound power output by huge earth-moving machines.  I have personally been involved in projects including circuit noise measurements, architectural acoustics measurements, active noise control measurements, directionality and directivity measurements and loudspeaker parameter measurements.  In each case, we wanted to have measurements free from any outside noise but also free from any reflections that might come from walls or floors found in typical rooms.

Design Considerations

This section is really what I consider to be the meat of this page and the main motivation for making it.  I was asked by a professor here at BYU to be in charge of designing, building and testing a small anechoic chamber that would work into ultrasonic frequencies.  I accepted the assignment and began scouring the internet for advice, tips and other information that might prove useful.  Needless to say, I found several pages by companies that had one built for them by someone else, but nothing that really detailed how one is built.  I spent weeks discussing design with my professor and I even called someone at one of the top anechoic chamber construction companies around.  Now that I am almost completely finished with the chamber, I decided that it would be a good thing to share what I have learned with everyone and to put all this information in one place.

The first thing to consider when building a chamber is the room in which it will be installed.  If you are going to be installing one in a new building, then you have much more flexibility and options regarding what kind of chamber you will end up with.  If you can build one from scratch, the first thing to do is make sure that the room is isolated from the rest of the building, including the foundation.  This can be done through various kinds of isolation mounts including specially made rubber mounting mats.  This kind of construction is typically very expensive, but essential if you are planning on constructing a very large chamber.  If your chamber will be very large, then you might want a standalone building.  If your room will be somewhat smaller (perhaps 30'x30' before wedge installation) you may be able to get away with floating the room inside of a room in the building.  This can also be done if you have a large room in an already constructed building.  There are different ways of floating a room, but one of the best ways is to do what recording studios.  When you have the floor that is connected with the rest of the building, put down a layer of concrete about 4 inches thick.  On top of this layer, put some soft fiberglass insulation.  On top of the soft fiberglass, lay down compacted (rigid) fiberglass insulation, anywhere from an inch to two or three.  You can also put some neoprene isolation in there if you want, but it is not vital.  On top of the rigid fiberglass or the neoprene, lay another concrete slab this time 3 to 4 inches thick.  Make sure that the concrete you laid (both layers) does not touch the walls of the room you are working in.  The goal is to eliminate any rigid coupling between the main building and your room.  Failure to do so will result in a vibration transmission path and possibly a flanking path for other acoustic noice, which will degrade chamber performance greatly.  Once the floor is done, build your room walls on top of only this floating floor (you can put some resilient supports between the inner and outer walls to help with stability issues).  Then, cap the room you have just made inside a room with a new ceiling that is also not strongly coupled to the outer room.  Also, you will want to make sure that the door opening into your room is higher than the depth of your acoustic treatment.  How to choose your treatment and depth is discussed below, but in order for your chamber to really be anechoic, you are going to need a cable tension or grate floor that sits plenty high above the actual floor of the room so that you can put acoustic treatement below your measurements as well.

Another thing to consider when building a chamber is the operating frequency range.  This, along with the size of the room to begin with, will in large part determine the amount of usable work space inside the finished chamber.  When building an anechoic chamber, the acoustic treatment on the walls, ceiling and under the floor needs to be at least 1/4 wavelength of the lowest frequency that you are interested in.  Say for example you wanted your chamber to work down to 100 Hz.  Since the speed of sound is roughly 340 m/s, then one wavelength of 100 Hz sound will be approximately 3.4 m (~11' 2").  Then 1/4 of this wavelength is 0.85 m (~2' 9").  Remember that you have to put this on every wall in the room, so if you start with a room that is 20' wide, you will have an inside dimension of 14.5'.  But also keep in mind that the usable work space inside a chamber is all the space that is located at least another 1/4 wavelength of your lowest frequency from the tips of your treatment.  This would mean that your actual work space in a 20' wide room would be about 9' across.

The next thing to consider is what kind of material will you be using for your treatment.  You can use prefabricated wedges or panels from companies that specialize in acoustic test facilities or you can make your own or even contract out someone else to make your treatment (from now on I will typically refer to the acoustic treatment as wedges).  If you are operating on a tight budget or building one for personal use, fiberglass insulation works great.  You can put it inside a chicken wire frame in the shape of the wedges and it will work pretty well over most frequencies.  In my case, we used open cell foam because it can be cut into wedge shapes without any kind of wire or perforated metal covering, allowing us to reach into even higher frequencies for testing ultrasonic sound sources.

In conjunction with your treatment material, the shape is extremely important.  You can't expect to just throw big blocks of foam on the wall and it will work.  There is a lot of science at work here.  Leo Beranek has written what has been called THE guidelines to follow when creating anechoic treatments for rooms or tubes.  He specifies the angles and optimal air gaps for one, two and three tipped wedges.  Since he has written them and I don't have them on hand, I will refer you to his books Acoustics and Acoustical Measurements which are available through ASA.  This is where you will need to bust out some equipment and savvy acoustics skills to do some work.  The first thing you need to do is get samples of your material.  If you can get them full size and fit them in an impedance tube, that is best.  If not, you can get proportionally sized samples and put them into an impedance tube and proportionally scale the frequencies you test it with.

An impedance tube is a rigid walled tube that has a source at one end (usually a loudspeaker mounted in a box so that there is an airtight fit between speaker and tube) and either a rigid or anechoic termination at the other end.  In this case, you need both a rigid termination and an anechoic termination (but not at the same time, of course).  Make sure that when you are taking measurements that you don't excite the tube above the first cutoff frequency.  What this will do is introduce cross-modes into your tube and throw your measurements way off.  If you use a square or rectangular cross section tube, measure the largest of the two dimensions (either width or height) and divide the speed of sound by twice that length.  This will be the frequency below which you need to make sure your measurements are made so that the plane wave assumption holds true.  There should be at least two holes in your tube for you to insert microphones into.  By microphone, I don't mean anything you can get at Radio Shack, I am talking precision microphones, like ones available from Bruel and Kjaer, G.R.A.S. or PCB.  You will also need a quality fft analyzer with at least two channels.  The holes that you put your microphones into need to have a spacing that is known exactly so that you can take accurate measurements.  The spacing between microphones also plays a role in what frequency range you can measure over.  If your intermicrophone spacing is d, then the usable frequency range over which your two microphone technique will work is given by the equation .05*f<d<.4*f.  A way to increase the frequency range over which you can make good measurements can be increased by adding a third hole in such a way that the distance from the first hole to the third hole is not simply double the distance between the first and second holes.  This allows you three different configurations that you can use to measure the largest frequency range possible, but be sure that you do not drive your tube above the first cutoff frequency!  Also remember to plug whichever hole you are not using during measurements with a rubber stopper or a nicely machined metal one to prevent any sound leakage to the outside.  Also make sure that your tube is plenty long (probably at least 10 feet for most measurements).  Now that you have your tube set up, you have to relatively calibrate your microphones so that you can actually use the date that you collect later.  This is done by putting the anechoic termination on the end of the tube and inserting your microphones into two of the holes (the placement of the two holes doesn't matter here).  Make sure to designate one microphone as microphone one and always know which one it is (if you are using a microphone capsule with preamp, like a PCB mic, make sure that it is always the same capsule with the same preamp).  Take a transfer function measurement between the two microphones while driving the tube with random noise that has a spectrum that goes from 20 Hz (lower if your chamber is intended to extend below 20 Hz) to just under the cutoff frequency of the first mode.  The transfer function is a common measurement and any decent analyzer ought to be able to do it easily.  Make sure your microphones have the power they need to operate.  After taking the transfer function measurement with at least 15 stable averages, switch the location of the microphones (just swap holes).  Take the transfer function measurement again.  Make sure to save the two files separately and in the frequency domain.  These will be vital for later.

Now you are ready to actually measure the frequency dependent absorption coefficient of your material.  Place your microphones in the tube with mic 1 located closer to the source and mic 2 at least one duct diameter away from the tip of your sample that is placed in the tube.  Put the rigid termination on the end of the tube and then take the transfer function measurement between the two microphones while driving the source with random noise (still limited to the frequency range used above).  If you are using more than just one microphone spacing, rearrange the microphones and take the measurement again, making sure to keep track of which measurement is which.  I recommend doing this not only for each sample of material, but for different air gaps between the treatment and the rigid termination.  The air gap plays a vital role in making a room anechoic.  For example, I had 9 different wedges (one, two and three tipped wedges in three different materials) and tested each with 3 different air gaps (1", 2" and 3") for a total of 27 measurements with just a single intermicrophone spacing.  Yes, it can be very time consuming, but if you don't get this right, then the whole rest of the chamber won't amount to anything more than a monument of shame.  Once you have these transfer function measurements, you need to use a program like Matlab to process the data (you might be able to use Excel, I have no idea if it can handle this though).  I will see if I can get my code posted on here for everyone to use (though I claim no responsibility for any damage or anything like that that may result from its use) since there are some ugly equations that have to be used to extract both the reflection and absorption coefficients from the data.

Now you have your measured absorption coefficients as a function of frequency.  Make sure you have plots of these so that you can determine which sample and air gap combination works best.  Remember that in order for your chamber to be anechoic, it has to have an absorption coefficient of at least 0.99.  I usually plot this as a straight line versus frequency so that I can see wherever my measured coefficient might dip below.  Anywhere that it dips below after initially going above the line is called a pass band.  Most likely, all of your samples will have some kind of pass band, but the trick to picking which sample is best is by limiting the width of the pass band, the minimum value in the pass band, and the location of the pass band in the frequency range you measured.  Don't worry too much about higher frequencies, because as you go up in frequency (especially above your tube's cutoff frequency) the samples will tend to absorb plenty of high frequency energy.  This measurement is primarily for the lower end, or low frequency cutoff of your samples and chamber.  If you can make your pass band appear as low in the frequency range as possible, that is best.  Also, try not to let your minimum value in your pass band be below about .97 or .98.  The width of the pass band is really up to you and what you find acceptable.  Once you have determined which sample works best, it's time to design the room itself.

In order to design the room itself, you will need a CAD program.  AutoCAD works great, but if you are doing this project yourself and don't have access software like AutoCAD, you can find some free programs.  I like a free one called A9CAD (you can download it here, though again I claim no responsibility for anything it may do to your system, works on mine just fine.  I hope this is legal and I don't get sued).  The first thing you want to do is create a floor plan of your naked room and create scale drawings of your wedges.  I did all my drawings in AutoCAD and did them in decimal feet (I recommend using engineering units).  Then you want to determine what kind of layout the wedges will have in your room.  Remember to alternate the position of wedges by 90 degrees from one to the next, including side to side and up and down.  If this doesn't make sense, take a peek at some of the pictures below.  Also, something you must remember is to put fills in all the corners (wherever two planes meet, so in the corner between walls, corner between wall and floor, wall and ceiling, etc.).  Fills are typically just large rectangular blocks of your material and typically aren't seen very much at all when construction is finished.  The pictures below ought to give you an idea of how this is done.  Make sure that your layout takes into account your cable tension floor, doorway, light fixtures and electrical outlets and anything else that may be in your chamber such as cable feedthroughs.  You will want to make layouts for your ceiling, walls, floor, door and a bird's eye view of the room without the roof on.  Make sure that all your wedges, fills, and floor fills (discussed later) are taken into account here.  A good strategy is to number all your fills and floor fills to help the three dimensional visualization process.  Make sure that your wedges alternate from one wall to next and from the walls to ceiling and floor.

The next big thing is the cable tension floor.  This is a special floor made by weaving steel cables in a basket weave pattern to create a floor that will support lots of weight without being acoustically obtrusive.  This is probably the single most important element when it comes to safety in your chamber.  Remember that the bottom of your door should be raised from the floor at least the depth of your acoustic treatment and preferably one and half to two times that depth.  The first thing to consider is how much weight you want to be able to put on the floor.  Since the chamber I built was small, I chose a max load of about 1500 lb.  For a larger chamber you may want it significantly more so that you can support large groups of people or lots of equipment.  In order to make this floor, you will have to mount I-beams all the way around the room so that the top of the I-beam sits flush with the bottom of the door opening.  Make sure they are good and level.  You also want to make sure that the I-beam is strong enough to handle not only the tension from the cable, but also the intense horizontal loads that people and equipment on the floor with produce.  For my case, I used 6x25 I-beam (I-beam identification is the width by weight per foot, so 6x25 beam is 6 inches wide and weighs 25 lb. per foot).  The spacing between cables is up to you.  Fewer cables means less acoustic scattering, but also means lower max loads.  Typically floors are woven with cables spaced every 2 inches.  The type of cable is also up to you, but typically 3/32" stainless steel cable is used.  The best way I found to make the floor was to have the steel supply company cut my I-beams to length.  Then when the beams were just sitting the chamber floor, I marked where the mounting holes would go and also where the holes for the cables would go.  Make sure the make the holes for the cable as close to the center webbing as possible to reduce the shear stress that the cables will put on the beam.  This maximizes the usefulness of the inertial moment of your beam to prevent bending or anything else disastrous like that.  After you have drilled all the holes (and there could be several hundred for a larger chamber), then mount them on the wall (you can use either concrete anchors or if you are building one from scratch, put in some uni-strut and that will make life easy).  Connecting the cables to the I-beam can be a little tricky.  You want to make sure to attach the cables to the beam via springs in some manner (see picture below) so that there is a little give to floor.  This will help reduce any really high stresses that may occur from people jumping or falling loads, etc.  Actually weaving the floor will need to wait for now, because you have to put in all your treatment below the floor first.

Actually installing the wedges takes a whole load of planning, which should have been done with your CAD program.  You will need to come up with some sort of way to mount the wedges to the wall.  What I did was mount EMT electrical conduit using vibration isolating pipe clamps that were attached to the uni-strut using coupling nuts at 1 foot intervals (since the footprint of my wedges was 1 foot square).  You can use whatever method works best for your particular situation.  Make sure that when you mount them, you don't put down the floor fills or the lowest row of wedges first, since you may need to make adjustments to the cable floor later.  You will want to make sure that the support system is properly spaced and setup (some wedges may not be exactly the dimensions you thought) by experimental verification.  The corner fills should go in first, then all the floor wedges and then all the wedges from the second or third row up.  Then install the ceiling and make sure the cable floor is all adjusted.  Then put in the floor fills and the last of the wedges.

The final step in constructing a chamber is the testing.  There is a professional standard (I'll try to find out which it is) that governs how the chamber should be qualified, or tested and verified.  Basically, the testing is to verify that the sound in the chamber decays as 1/r where r is the distance from the source.  This is what happens in a free field and is what will happen when the chamber is truly anechoic.

My Chamber

Here is a rundown of how I designed and constructed my chamber.  I will try to get this all up to date as soon as possible.  To start, I had a room that was already located in the university's science center and was already isolated (at least it was until some contractors decided to pour cement into the isolation cavity between one of the walls and the building).  The room was 9'9" wide, 11' long and 11'2" tall.  The doorway was located in the north wall and the bottom of the door opening was 32" up from the ground.  Due to the small dimensions of the chamber, we decided to attempt an operating range of 150 Hz up to ultrasonic.  This meant that we needed a treatment depth of 22.5".  We decided that if we had wedges 20.5" deep with a 2" air gap we would be in business.  We used 6x25 I-beam and 3/32" stainless steel cable.  To mount the cable to the beam, we used 3/8" - 6" long black oxidized steel eyebolts with 4" of thread and 3" long springs with a spring constant of 544 lb/in.  We used swaging sleeves and thimbles to attach the cable to the eyebolts.  A picture is shown here.

Before weaving the floor, we mounted all of our wedge support system to the floor, walls and ceiling.  Below are pictures showing how we went about putting things in.  They should be somewhat self-explanatory after the discussion given above.  Hopefully this page has been of some use.  If there are any mistakes, I will try to update them and make this a more complete reference.

                                                          Mounting of the I-beam to the walls

                                                Mount the beam flush with the bottom of the door

                                                  Mount the beam and weld the joints together

The wedge support structure mounted on the wall

Install wedges under the floor and install bottom floor fills

Install corner fills and then wedges from row three and up

What a finished wall looks like

Arrange wedges around feedthroughs