It’s not always possible or convenient (or quick, or cheap) to go through a robust process of testing and listening to find the driver you want.
Sometimes you just have to make an educated guess based on the manufacturer’s datasheet. So, I’ve come up with a guide, listing some of the parameters that I consider to be the most important. If you already know what you’re looking for, this may not be for you, but if you
First, try to be clear on what purpose the speakers should serve. Anecdotal experience could be very helpful, such as having an idea of some of the basic parameters like cone size, displacement, as well as the size of the listening space.
Let’s say you’re building a 2-way system and you want a woofer that will have good bass extension without any additional help from subwoofers. But it also needs to perform well, up to mid-range frequencies, where a tweeter seamlessly takes over.
Right away we can see that there may be some competing requirements, and sorting through the capabilities of various speakers without knowing what you want can get very messy. So, in the beginning I would get clear on what purpose the speaker system is going to serve.
What are the speakers for?
What sized woofer would be appropriate for your needs?
| Type | Suggestions |
| Desktop speakers, Bluetooth-style portable systems | Full range drivers with 3″ – 5″ nominal size (approximately 25cm² – 85cm² cone area). I’d advise against a multiway system, unless you are certain that you want the additional complexity for this project size. |
| Book-shelf, stand mount | 5″-8″ mid-woofers in 5L-20L Golden Ratio boxes or other interesting shapes. |
| Floorstanding, big | 10″-12″, even 15″ may be achievable with only 2 ways. |
For a smooth mid-range
Here are some of the key factors that I look for:
| Suggestion | Reasoning | Caveats |
| Low ratio of voice coil height to magnetic gap height; go for the lowest Xmax that you really need. | Most speakers are an “over-hung” type, which means that the ends of the voice coil protrude slightly past the region where the magnetic field is strongest. This is a compromise that fixes some problems but creates others. On one hand, substantial overhang allows for a generous amount of coil displacement (Xmax), which may help to keep bass tones linear. On the other hand, tall voice coils are more prone to inductance modulation, which can make mid-range frequencies distorted due to IMD. | ‘Overdrive’: a characteristic form of distortion occurs when the bass signal exceeds the capabilities of the speaker. |
| Low MMS or MMD | The cone mass (with or without the air load) should be as low as practical. Some argue that it doesn’t matter because you can just throw more power at the speaker to get the same SPL. But high power also promotes distortion, whether in the amplifier or in the speaker motor. Often the ONLY reason a speaker cone is made extra heavy is so that it’s tuned for deep bass, and to reduce flexing imposed by a stiff air cushion in very small boxes. A heavy cone also necessitates more self-damping to reduce resonances. However, mechanical damping is often non-linear with hysteresis-like effects that add low-level distortion. On the other hand, a lightweight cone has less momentum when it moves, allowing the surrounding air to act as a more effective damping mechanism. | Higher bass resonance tuning, affecting bass efficiency at the lower frequency limits. Large box size; cone flex if the box is too small; disconcerting break-up sounds if the speaker is overdriven. |
| Non-conductive voice coil former (bobbin) material | This is an anecdote without hard evidence, but I’ve noticed that better sounding speakers tend to use non-conductive bobbins. I have a hunch that metallic bobbins may be a source of distortion: as the conductor moves through the magnetic field, eddy currents are generated and a braking force is generated, but it’s likely to be non-linear and a source of intermodulation distortion. | Probably lower power handling: a metallic bobbin prevents hot spots at either end of the coil, despite also producing its own heat. |
| High VAS vs cone area | The speaker generally has to include a mechanical system that keeps the coil suspended in the magnetic gap. The spider and surround form a system of springs that return the cone to its middle resting position. This is yet another potential source of distortion, and it can be reduced by selecting speakers that have softer suspension: high VAS (equivalent volume of air suspension for a given cone area). | Loss of mechanical tolerances and sagging if the suspension is too soft. |
| Low inductance value; shorting rings | Inductance is a necessary and unavoidable feature of a dynamic speaker, but it can be a source of distortion due to its variability, as described above. A very low value indicates that the motor design likely uses “shorting rings”. Conductive magnet poles near the coil may act as a short-circuited secondary transformer winding, which reduces inductance (albeit inefficiently because of the air gap), and in a similar vein, shorting rings complement the geometry of the magnet to reduce modulation as the coil moves in the magnetic gap. | I’m a little wary that it’s possible to overshoot a sweet spot, as there seems to be a balancing act between low harmonic distortion vs electrical braking of the voice coil, which could cause mechanical reflections and resonances. |
| Low power rating | This becomes almost inevitable and a package deal, considering the other factors. | It can’t be too low for your needs. |
| Frequency response smoothness and sensitivity | As a last resort, check the frequency plot for any glaring anomalies. The exact testing conditions are rarely described in detail and are by no means standard, so I would take it as merely a rough guide for equalization and crossover filters. | I wouldn’t get too wowed by a frequency response that looks impressively smooth and flat. It says nothing about HD or IMD. In addition, it may be difficult to assess the significance of low-level resonances. High-Q resonances may have a significant detrimental effect on sound, yet be practically invisible on a frequency plot due to their narrow bandwidth. Conversely, wider peaks may be more benign despite their appearance. |
YMMV
This, I think, should provide a solid starting point for choosing drivers with a good mid-range sound, and also for finding mid-woofers with a good balance of bass extension and smooth tone. You still need to do your own research on simulators and check things like cone displacement vs frequency.
Taking the above guidelines to an extreme, you may find yourself looking at guitar speakers (Eminence, Celestion, Fane, etc.) but I’m not sure if that’s advisable. It’s difficult to get good information about these sort of speakers, and people who use them are often only able to describe the sound in non-technical terms, so you’re left guessing as to whether the “rich mid-range harmonics” refer to actual distortion, or a natural, undistorted sound. (Buy the speaker first and find out later, right?!)
However, maybe I’m overly cautious and missing out on some amazingly clear sound that requires nothing more than a bit of EQ, and to make sure the bass never exceeds some hard limit where overdrive kicks in.
Most recently, I’ve been looking at base models of “pro-audio” speaker brands. The ‘budget’ models tend to have the lowest Xmax and power ratings, but are likely to offer the smoothest mid-range response. On the other hand, an extended bass response is costly, not just in terms of price, but the maximum usable frequency is often divided by 2 or 4.