Speaker Design By RJB
Baffle Layout and Cabinet ConstructionIn this section I will discuss how to determine the baffle dimensions and driver locations for the best performance. I'll also discuss cabinet construction techniques including cabinet material, wall thickness, bracing, stuffing and lining. I will also share some of the methods that I use to construct the cabinet and the tools that are required. This section will complete the topic of enclosure design and construction and the next section will begin to talk about crossovers.
Baffle LayoutI will be assuming rectangular baffles in this discussion because they are the easiest to build and I have the most experience with them but I will also discuss some of the benefits of using irregular shaped baffles. I will discuss baffle dimensions and driver locations as well as the topics that are related to these things including baffle step loss and baffle diffraction. I'll also talk about the benefits of using roundovers on the baffle and how flush mounting versus surface mounting affects the response. I'll finally discuss grill frames including construction strategies and effects on performance with references to some of my designs for measurements.
Baffle DimensionsThere is a strong trend towards narrow baffles in the speaker community that it has become quite common for baffles to be just slightly wider than the largest driver on the baffle. It is said that a narrow baffle provides better imaging than a wide baffle but I haven't experimented or verified this in my experience with speaker design. I prefer narrow baffles because they look better and with most speakers it is nice to compensate by making the cabinet deeper which ensures that the drivers are further away from the back wall which is a good thing in terms of imaging. Narrow designs are also probably more popular because they have a higher SAF (spouse acceptance factor). I'll usually use a relatively narrow cabinet but if I need to squeeze out an extra few liters of volume without changing the height or depth I'll sometimes make the baffle wider. The difference in baffle width will change the diffraction effects and baffle step response. A wider baffle will have diffraction anomalies that are lower in frequency than that of a narrow baffle.
Baffle Diffraction and Baffle Step LossBaffle diffraction and baffle step loss are both phenomenon that are based on the wavelength of sound and its relation to the baffle dimensions. A speaker that is located on an infinite baffle is said to radiate in 2pi space since the baffle is a reflective surface that focuses all of the sound waves in one direction. With a narrow baffle the driver will continue to radiate in 2pi space until you reach a low enough frequency where the wavelength of the sound is about the same as the distance to the edge of the baffle. When the wavelength gets longer than this distance the waves begin wrapping around the baffle and radiating into 4pi space. When you get low enough in frequency there will be a half power reduction do to the full transition to full 4pi space because half of the sound is radiating around the baffle which equates to a drop of 6 dB. Near the transition point which is determined by the baffle width there will be a slight peak in the response and then the response will drop until it stabilizes at a point 6 dB lower. The wider the baffle, the lower in frequency the baffle step loss transition will occur. The peak in the response is due to baffle diffraction which I will discuss next. I have a few links to sites that go into more depth on the topic of baffle step loss and baffle step compensation here...
True Audio TechTopics: Diffraction Loss
ESP- Baffle Step Compensation
Baffle diffraction is a phenomenon that causes irregularities in the frequency response of a speaker when sound waves reflect off of the edges of a cabinet. The frequencies where baffle diffraction irregularities occur is based on the distance between the driver and the edge of the reflecting surface. Baffle diffraction occurs at the edges of the cabinet and can also occur at the edge of a driver flange if it isn't flush mounted. The closer the driver to the diffracting edge, the higher in frequency the worst diffraction will occur. Diffraction responses are caused by the constructive and deconstructive combination of direct radiated sound and sound reflected by the edges of the baffle. There is a great article that discusses this topic in more detail at the following link...
Understanding Cabinet Edge Diffraction
You can reduce the effects of baffle diffraction by using roundovers on the cabinet edges. The larger the roundover, the greater the effectiveness. I've only used 1/2" roundovers on my projects because I don't have a larger bit or a router that could handle it. There might be some gain in using a 1/2" roundover but I think that the overall improvement is minimal. It really takes a roundover of 1" or greater to get the best results. Chamfers offer similar diffraction improvement. If you want to see the difference experiment with the Baffle Diffraction Simulator (BDS) spreadsheet and compare the responses with different types and widths of edge treatment.
Driver LocationLocating the drivers is usually much more critical for tweeters because the typical distance between the tweeter and the edge of the cabinet is equivalent to a wavelength that is within the tweeter's passband. Woofer and midrange location on the baffle is a little bit less critical and should be chosen according to other factors including placing it close enough to the next higher frequency driver such that it is within one wavelength of the crossover frequency which I will elaborate on a bit later. Also multiple drivers (an MTM design for example) should be placed as close together as possible because the comb filtering between the two drivers' responses will be less severe. I'll discuss methods of locating drivers in a three way design to minimize floor bounce cancellations. When locating drivers in a three way design it is often desireable to locate the midrange high enough on the baffle so that the floor bounce cancellation frequency is lower than the midrange high pass filter crossover point (essentially putting the dip below the driver's operating range and eliminating it from the total response. You'll want to locate the woofer in a three way lower on the baffle which will put its floor bounce cancellation frequency higher. The closer the driver is to the reflecting surface (the floor in this case) the higher in frequency the floor bounce cancellation will occur due to the shorter wavelengths involved. Floor bounce cancellations occur when the wavelength of the direct sound between the speaker and the listening position is out of phase with the wavelength of sound that has reflected off of the floor and travelled a longer distance which differs by half a wavelength. The only problem with optimizing a design to have minimal floor bounce interactions is you will be ignoring the ceiling bounce cancellations which might even be more severe relative to the floor bounce issues since the ceiling isn't carpeted and the reflections will be stronger. Ceiling bounce cancellations should be low enough with the midrange driver to be well below the passband of the driver. However, ceiling bounce cancellations are going to interfere with the woofer's response since it will usually be located pretty low making the wavelengths long between the driver and the ceiling resulting in lower frequency dips. This frequency may even be low enough to be in the region where the room is effected by room modes especially due to the ceiling height since that is usually the shortest dimension in a room and may combine with the peaks or dips caused by room mode resonances. I would have to simulate a design to determine the overall effect but I think you get the basic idea behind locating driver for minimal floor and ceiling interactions.