... a bit Suffolk
As it says on our mike position page "Forget all about 1 metre on axis". When development measure a loudspeaker they have an anechoic chamber in the hope that it will give a repeatable free field responses The free field response gives you some indication of our how a speaker will perform in different environments; but when comparative testing you're only concerned how alike are the test unit and the reference. Keeping this in mind the considerations for a production test unit are:
It helps if the environment doesn't put points of steep slope in the response. As a slope becomes steeper the change in y increases for the same change in x. If the environment makes a slope in a unit's response look steeper then differences in that region will be emphasised.
A TEST BOX FOR DRIVERS
To test drivers stand alone you will need a test box. A box volume of 0.5 m3 (500 ltr, 17.5 ft3) will cope with a low resonant frequency 15 in driver but will also be large. Avoid simple relationships between the box dimensions to distribute the standing wave frequencies in the box. (A cube is the worst case as all three dimensions are the same.) Make sure the box ends up at a convenient work height for the operator.
If you really want to go for it design a box with no parallel surfaces. The drawing is not easy and, if you make the top horizontal, it's a real fiddle arranging the box to stand up. If you do build such a box when you stick your head in and say "Boom" there is a satisfyingly unresonant sound. In addition, the box's bizarre appearance impresses visitors making you look really technical.
Test drivers face downward on the box top. (Small systems can be done this way too.) Testing in this position does mean that the coil will be slightly displaced in the gap by the coil and cone's own weight but this is not significant, even with highly compliant high mass drivers.
Mount the mike in the box on desk type mike stand screwed to the bottom or possibly a side. Don't worry about vibration isolation of the mike stand. If vibration of the box is a problem, deal with it by increasing the drive level to the driver, ie increasing the signal to noise ratio. If you use floppy flexible mike mountings, the mike will move about giving inconsistent results. Mark the mike's position in its clip with tape so you can easily relocate it.
Position the mike away from the box sides. For the best signal to noise ratio position the mike close to the test unit but remember if you are going to test systems it needs to be far enough away to integrate between the system's various drivers. If the mike has a wind shield use it. It gives the mike some protection when someone eventually drops something on it and some measure of dust protection whilst having little effect on the mike's frequency response.
Run the mike lead to a connector mounted on the box rather than just passing it through a hole. Then when someone manages to get the external mike lead box tangled up in the fork lift truck the mike will not be wrenched from its clip and the damaged lead is easily changed.
Similarly don't just run the leads from the amp to a pair of croc clips at the business end. Run the amplifier's output to connectors mounted on the box's top. (Perhaps use a small box with two pairs of 4mm sockets wired to the each other and screw it to the test box's top. The leads from the amp plug in one side and the leads the test unit the other.) By having connectors you can easily change leads for products with different fittings, change leads when the connectors wear out with constant use or insert crossovers, capacitors, resistors or what ever your engineer insists are essential to reveal the true wondrousness of the major breakthrough he has designed.
If you want be really polished you may want to put two of these boxes on the box top so that the operator can plug up to whichever is best for them. (Your operator may be left handed!)
Line the box with absorbent. Acoustic foam tiles are best. Tiles don't move about when stuck to the box sides and don't generate dust to gather on the microphone diaphragm, unlike BAF and wool. (If you glue the absorbent, remember to remove the mike first so it isn't attacked by the fumes from glue. Allow time for the fumes to disperse before using the box or they might attack the operator as well.) If you have made a box with non-parallel surfaces give yourself the pleasure of going "Boom" before you put the absorbent in.
Having put all this effort into your box make it as adaptable as possible. Don't just cut a hole in the top and knock in a few nails to make a bench mark but design a "baffle" system. A "baffle" is a ply board with a hole in for the test unit to fire through. Fixed to it is another board cut out to provide a positive benchmark for the test unit. (A benchmark is an arrangement that ensures that the test units are always in the same position for the test.) Clamp the whole lot down to the box top with toggle clamps using battening of some kind to reference it to the box top. Adopt a standard size for the baffles and try to arrange it to cope with large products.
The two things to remember when thinking about all this are ease of use for the operator and compatibility. You don't want to end up with two different boxes with two sets of baffles that won't interchange. Producing a flexible, easy to use, and reliable arrangement takes some thinking about.
With the box finished you plug up a speaker and do a frequency response. You will be surprised at how good a bass extension your drivers have. The reason is that the pressure response inside a closed box mirrors that of the driver in free space and so boosts the low frequency response. It is not a problem but if it worries you the mikes we recommend have bass cut switches, or you could equalise the power amp or mike pre-amp. DON'T change your test box from a closed box type to deal with this problem. You will significantly reduce the signal to noise ratio at the mike and possibly start the box operating as a Helmholtz resonator putting nasty humps in the response.
The reason your engineering department has an anechoic chamber is they want to remove room reflections from the measurement. This is not the reason you have a test booth. You have a test booth to keep noise from reaching the microphone and to create a quiet enough space so the operator can hear the test unit. (As the NaTKiT compares the test unit with the reference it doesn't matter if the response is not anechoic. Excessive room effects will effect the efficiency of the test but it is unlikely.)
If your production is quiet enough then you may not need a booth at all. Periodic noise in the production environment - ventilation fans, electric motors, even ultrasonic welders - tend not to be a problem for the NaTKiT, though it can be a problem for the operator. Impulsive noise - pneumatic tools, particularly staplers - can be a problem for both the operator and NaTKiT.
Two parameters control how much noise gets into a booth. They are how airtight the booth is, and the mass of the walls. The more airtight the booth is the less noise will get in. The greater the mass of the walls the less low frequency noise will get through.
If a production test booth is made airtight then the test speakers cannot move in and out easily. Some companies may find it acceptable for units to pass through a hatch or door that has to open and close but most folks want the stuff to pass through the booth faster than that. Usually the speakers are moving on some track or flow line that passes through the booth. The booth has two openings, one for the untested to enter and one for the tested speaker to leave. A solution in these circumstances is to build a tunnel around the entry and exit lined with absorbent. You have now made a "tortuous" path for high frequency noise entering the booth; ie it will be repeatedly reflected as it passes down the tunnel and consequently repeatedly absorbed by the lining. If you wanted to go for the deluxe version, you could have flexible plastic "streamers" covering the openings.
This tortuous path will have little effect at low frequencies and so constructing high mass walls will be a waste of time. Low frequency noise will get in if there is any opening.
The sound level heard by an operator in a room will have two components. The sound travelling directly from the source to the listener is the direct component. The rest of the sound generated by the source, after it has bounced around the room, makes up the reverberant component heard by the listener. There is nothing you can do to attenuate the direct component except place a panel between the source and listener. The reverberant component as it bounces around the room can be attenuated by lining the booth with sound absorbent. So lining the booth with absorbent will attenuate the reverberant component made by external noise and the component generated by the test unit.
Remember that a closed box lined with acoustic foam is a very oppressive environment for the operator. (They are often referred to by operators as padded cells.) A window makes the booth more pleasant to be in and can make it easier for the operator to communicate with the outside world and vice versa. Covering dark coloured absorbent with acoustically transparent light coloured muslin or something can lighten the place up as well.
In some circumstances, say in very large space with the test point a few yards away from a whining electric motor, the noise will be primarily in the direct field. Under these circumstances you don't really need a booth. A partition (or two) between the noise source and the test point may very well do the job.