In this blog post I compare well over 50 premium woofers, full range drivers and compression drivers to determine my favorites:
https://ultrahedonism.blogspot.com/2023/06/speaker-design-one.html
https://ultrahedonism.blogspot.com/2023/06/speaker-design-one.html
No Radians?
18Sound 15W500 - nice!
As to the value (mark) there aren’t many sources for GENERAL subjective impressions amongst a substantive grouping of drivers even if the conditions were quite different. It’s basically like saying: these drivers for their typical intended use should be on your “radar” for these particular qualities (and conversely those that maybe should be passed-by).
-in other-words: “maybe investigate these drivers if interested.”
18Sound 15W500 - nice!
As to the value (mark) there aren’t many sources for GENERAL subjective impressions amongst a substantive grouping of drivers even if the conditions were quite different. It’s basically like saying: these drivers for their typical intended use should be on your “radar” for these particular qualities (and conversely those that maybe should be passed-by).
-in other-words: “maybe investigate these drivers if interested.”
Last edited:
Some drivers like the Satori seem excessively damp and unextended, compared with a favorite like the Visaton B200. The former sounded as though it was under a thick wet wool blanket even when considering both drivers within a limited operating band with 24db/oct slopes at both points. Elaborating on that, some drivers present sound in such a way that one is able to observe it through a window. With such drivers one can almost peer through the dust cap as a sort of local region where micro details emerge from. I know it sounds strange but it is true. Drivers that are of the opposite nature tend to be unresolving, shouty, shrill and bad sounding in all respects comparatively.
I'm confident in my ears and intuition to make accurate determinations about what I do and don't like. I trust various other sources of measurements which are readily available on the web in addition to manufacturer specifications. A listening test is more valuable than measurements assuming the listeners ears are good. As it pertains to sound quality and in simple conclusion fast transients supersedes, flat frequency response to a degree, and low distortion also to a degree, in my observation. You don't have to heed my observations or advise and I prefer diversity of opinion though I'm not swayed from mine by a measurement or an alternative opinion.
This was a preliminary exploration. Having singled out the drivers I prefer and established in my mind the direction I want to head with building this and other speakers I will eventually get more granular and hopefully smart with regards to measuring things. I don't intend to answer too much in this thread as it's a burden answering to luddites... though I will update the post and my blog from time to time.
Last edited:
Ah, so you're kind of skipping another dimension of testing.
In a nutshell:
High damping factor / low output impedance is usually considered great for bass, but for mid-range and HF it's kind-of an open question, at least for voice-coil based speakers. Ribbons, planars, ESLs or piezos probably don't apply because they don't have the same distortion mechanism where the inductance is modulated by coil displacement and current. However, for ordinary coils suspended in a magnetic field and surrounded by iron, a relatively large impedance placed in series with the speaker will often help to reduce distortion.
(Padding tweeters down with series resistance versus adjusting the level actively could be an easy test to compare what happens to HD and IMD if you're curious.)
Different means can achieve similar ends though. Tubes or class-A amplifiers often have a high output impedance already without adding any passive filtering components. Whereas with voltage amplifiers, a 1st order passive filter (coil) could help to clean up the mid-range and high frequencies over and above just reducing the output levels with active filters.
That's not to negate your efforts though, as there's probably a strong overlap between drivers that already sound great when "run wild", and ones with the most potential when given individual TLC. Although spending hours tuning passive crossovers by ear and/or mixing and matching amplifiers for best performance will help any driver, there might be a few edge cases where it's factory tuned for a high damping factor and anything other than class-D makes it come unglued.
Subwoofer drivers were tested in an open baffle and or unmounted state. A primary observation is that many if not all drivers under 12" are not adequate for anything other than nearfield listening at moderate volume. For one they do not move enough air and two they can not handle loud playback levels as only a couple drivers, specifically Seas which have high xmax values, and the 15" woofers which are much more robust in every respect with larger voice coils and not necessitating strenuous long excursion to achieve deep bass. The Visaton aluminum cone woofers for example produce a clicking noise at moderate volume levels. In addition to detailedness, and general tonality, it was these sorts of issues that I was stress testing for to determine which woofers were most robust, and also which midwoofers were robust to be able to cross at my desired crossover point. While a bass reflex cabinet may increase the sensitivity of the driver it will not do so to a degree that negates the observed issues at the levels which I would have to drive the driver for midfield playback. Aluminum cones do exhibit lower distortion due to their pistonic nature in that the voicecoil is not modulating the cone. Typically being smaller diameter drivers yet with heavier cones, they sound sluggish and with large excursion the room is overly pressurized in a way that is unnatural sounding and yields lots of room interactions and potential resonances. Issues associated with rear wave reflections increase substantially as cone excursion increases. To reiterate, the 15W500 which is deemed a midwoofer when used for sound reinforcement is ideal for my application which is lower volume midfield listening as the sensitivity in the bass region is still high compared to the various smaller dedicated bass and sub bass woofers I tested. It is more performative in every respect too according to my ears.
I appreciate the information about damping factor and amplifier output impedance. I plan to experiment with amplifiers and analog filters as a separate endeavor. I figure starting with the very well measuring Purifi class d amplifier and Topping class ab for the higher sensitivity section would provide a good sense of what premium low distortion amplifiers are capable of. I tested the Topping against other amps I had on hand including a 30w class A 211 tube amp and the Topping was obviously superior and I sold the rest.
I will look into optimizing for reduction of voice coil inductance come the time that I feel the speaker is dialed. I have an ER Audio electrostatic panel on hand to experiment with for the mid section ~100hz to ~2000hz. If it is loud enough and dynamic enough that is likely the direction I will go though I presume it will limited in output level playback.
I appreciate the information about damping factor and amplifier output impedance. I plan to experiment with amplifiers and analog filters as a separate endeavor. I figure starting with the very well measuring Purifi class d amplifier and Topping class ab for the higher sensitivity section would provide a good sense of what premium low distortion amplifiers are capable of. I tested the Topping against other amps I had on hand including a 30w class A 211 tube amp and the Topping was obviously superior and I sold the rest.
I will look into optimizing for reduction of voice coil inductance come the time that I feel the speaker is dialed. I have an ER Audio electrostatic panel on hand to experiment with for the mid section ~100hz to ~2000hz. If it is loud enough and dynamic enough that is likely the direction I will go though I presume it will limited in output level playback.
Last edited:
Or conversely, some drivers are garbage despite being widely accepted as performative and so rather than being a lemming and accepting such marketing as truth, and having the resources to make a determination for myself, I decided to test several and determine which I like...
Removing the dust cap and attaching a cork is an ok idea as it would reduce the mass of the cone further assuming it isn't causing friction during playback. In addition, if horn loading a driver, in the interest of producing a more "laminar" "directive" "non-turbulent" signal, such a phase plug could be optimized like this:
...If anything in addition to providing insight into which drivers I do and don't like; this preliminary exploration informed me of fundamental aspects of speaker design such as the importance of cone diameter, excursion and mass as it relates extension; that 2000hz is treble, 120hz is mid range and not bass, 80hz is mid bass, 60hz is bass, 50hz is upper sub bass; and how to optimize crossover points given this observation of frequency regions in a way that would be initially counter intuitive. Line source speakers with tiny drivers, not to be confused with line array speakers for sound reinforcement, ribbon and electrostatic panels, etc are all interesting and niche, with potential if driver technology is developed radically. Having wrapped my head around things to a greater degree, such designs feel illogical in one way or another compared with a three way speaker featuring a large woofer(s) and mid woofer(s), relative to what hifi stores sell. I pair of er audio electrostatic panels suspended from microphone stands at 2 feet with sealed 8" subwoofers on my coffee table would sound great. I am looking to fill a decent sized living room with sound at which point robust yet low power handling drivers and specifically large ones are the only option.
...Elaborating on the point about the dustcap. Many drivers these days feature a concave dust cap, examples being the SB Satori 7.5" driver which I tested and the 18sound 10W400 which I tested. This renders the driver pistonic in nature. In the case of a traditional cone geometry with a semi flexible cone, the majority of the cone acts as a pseudo suspension where the wave energy is localized to the voice coil (dust cap area) and radiates up the driver cone and is neutralized at the perimeter suspension. Traditionally a cone and perimeter suspension were made from a singular piece of paper. This methodology is best for achieving maximum detail and efficiency. As it correlates to mass, in this instance the voice coil is moving a small localized region to a great degree rather than in the pistonic example the entire cone mass equally and simultaneously. Having to reverse engineer some idiots' desire for 8" audiophile woofers and concave cones because they look slick is an idiotic endeavor. The latter pistonic design is bad for mid woofers above 60hz where distortion is less of an issue and yields terrible sounding resonances and lack of transient retrieval compared with the original methodology. I prefer new technology to dated ways so long as the person calling the shots is not a profit oriented marketing luddite.
You may be right in some ways, but there are other things that, for instance I thought I was right about for years, like the total superiority of active filters, before conceding that passive filters in fact did certain things that actives couldn't.
FWIW, the bending surface of soft cones is very unlikely to be well-matched to the speed of sound in air, leading to odd propagation patterns, such as a virtual sound source that is smeared out in the z-axis in front of or behind the speaker. This effect may be desireable, as it may make echoes (and psychoacoustic localisation effects) less 'sharp'. This also may give a hint as to the source of the "paper cone sound" vs "metal cone cound", going beyond just EQ settings.
I experienced the opposite with Accuton speakers, which were very impressive, but also very sensitive to diffraction and undamped rooms. If I had known some of these things earlier, I would have built horns around them. But my first customer (dad) didn't want horns, so there. Pros and cons both ways.
FWIW, the bending surface of soft cones is very unlikely to be well-matched to the speed of sound in air, leading to odd propagation patterns, such as a virtual sound source that is smeared out in the z-axis in front of or behind the speaker. This effect may be desireable, as it may make echoes (and psychoacoustic localisation effects) less 'sharp'. This also may give a hint as to the source of the "paper cone sound" vs "metal cone cound", going beyond just EQ settings.
I experienced the opposite with Accuton speakers, which were very impressive, but also very sensitive to diffraction and undamped rooms. If I had known some of these things earlier, I would have built horns around them. But my first customer (dad) didn't want horns, so there. Pros and cons both ways.
I'm skeptical of Accuton drivers. They produce very low distortion within a certain range, and then are very distorted outside of that range. They seem to be very insensitive and overly damped. I understand the diaphragm has ringing issues and long term durability issues. They seem too exotic and fragile. Based on measurements and listening test feedback I don't foresee my curiosity getting the best of me to order a pair. I recall reading of a couple instances where they loose out to various Seas drivers. Experimenting with exciters on a low mass low modulus stretched canvas or similar, for example would likely yield better midrange reproduction than most conventional cone drivers. While not as tech oriented as an Accuton driver, this sort of approach is more interesting to me and likely more performative. Such a thing is just hard to market and sell when the average customer is accustomed to seeing tech oriented circular woofers of a certain form factor.
Last edited:
My ones have been going strong for almost 20 years. Apart from once blowing up a pair of unprotected tweeters due to amplifier thumps, only the rubber surrounds of the midwoofers need replacing, and those have seen a lot of sunlight.
They have a characteristic harshness, indicative of the same kind of inductance modulation that I've heard in other speakers. But because the XO is active, they've always been run straight through (except for a capacitor for the tweeter).
It's an old build and used daily, so I'm not touching it, but experience gained in the meantime tells me the likely biggest improvements:
1) change the active-only XO to active + passive, with a passive notch filter open-circuiting the speaker at the break-up frequency.
2) Upgrade the amplifier. I personally like the challenge of developing my own, but there are some very promising designs if you look around the forum.
3) Waveguides with a smooth overall box shape for minimal diffraction. Maybe not saccharine like a chocolate seashell, but ideally I think something inspired by a Le Cleac'h-like wave guide should continue all the way to the back of the box. Or an in-wall baffle, or large pyramid stands that reduce floor bounce (by which time you might as well add a large woofer underneath).
I'd have to pry off the metal grilles to get closer to the cone edge, so there's a risk that hacking away at them might break them, so that's a risk.
Most recently, I'm mulling a project with the CHR-90 or CHP-90. My CHN-50's got creased and some gentle rework with cotton picks revealed that they may be closer to aluminium foil than aluminium cans. But the M.A. drivers offer an interesting technical choice with "mono" vs classic spider suspension, so I'm torn between the 'ms' range which is a bit heavier, or classic spiders that siphon off a bit of energy and potentially generate reflections. A high quality oval cone would be very interesting for me if they ever decide to make one.
They have a characteristic harshness, indicative of the same kind of inductance modulation that I've heard in other speakers. But because the XO is active, they've always been run straight through (except for a capacitor for the tweeter).
It's an old build and used daily, so I'm not touching it, but experience gained in the meantime tells me the likely biggest improvements:
1) change the active-only XO to active + passive, with a passive notch filter open-circuiting the speaker at the break-up frequency.
2) Upgrade the amplifier. I personally like the challenge of developing my own, but there are some very promising designs if you look around the forum.
3) Waveguides with a smooth overall box shape for minimal diffraction. Maybe not saccharine like a chocolate seashell, but ideally I think something inspired by a Le Cleac'h-like wave guide should continue all the way to the back of the box. Or an in-wall baffle, or large pyramid stands that reduce floor bounce (by which time you might as well add a large woofer underneath).
I'd have to pry off the metal grilles to get closer to the cone edge, so there's a risk that hacking away at them might break them, so that's a risk.
Most recently, I'm mulling a project with the CHR-90 or CHP-90. My CHN-50's got creased and some gentle rework with cotton picks revealed that they may be closer to aluminium foil than aluminium cans. But the M.A. drivers offer an interesting technical choice with "mono" vs classic spider suspension, so I'm torn between the 'ms' range which is a bit heavier, or classic spiders that siphon off a bit of energy and potentially generate reflections. A high quality oval cone would be very interesting for me if they ever decide to make one.