Modifying Speaker Drivers And The
For the purposes of this article, we'll largely be discussing Pioneer B20fu20-51fw as available at Parts Express. This is a darling amongst DIY full-range enthusiasts, as it's very inexpensive ($30 or so usually) and has a very nice sound to it, not harsh or fatiguing, but laid back and enjoyable, with real bass (which is an issue with fullrangers regardless of cost). So, it's not only good and inexpensive, but it's also an ideal platform for modification, being cheap enough to replace if you go too far. This driver is an 8" unit with moderate sensitivity at 90dB w/m. Most 8" fullrangers are more sensitive, in the mid to high 90s. This includes Lowthers, most Fostex 8" fullrange drivers, and many others. The lack of sensitivity in the Pioneer is due to the longer (heavier) coil, which allows the lower Fs (resonant frequency) higher Qts (Fs damping) and higher Xmax (more output before gross distortion). We will also discuss a 3" fullrange, the FF85KeN. This is a fully modified driver, with none of the fiddly, messy processes for the end user listed below. They're all done for you!
Manufacturers Know What They're Doing?
You can also see the use of "Mortite", which is also known as "Rope Caulk". This is a caulking material which remains soft and pliable, and can be removed at any time without any permanent change. It's used in this instance to fill the gap between the magnet and the portion of the frame enclosing the spider. This is done to dampen vibrations and smooth the transition of air around the magnet structure. This can be used on some basket legs in some drivers, when there's appropriate space for it. Do not apply it where it will contact any moving parts (spider, surround, cone). These simple tweaks are the first steps towards significant improvement in many drivers. People will generally notice an improved level of clarity, particularly in the upper midrange through the treble, with less "haze" with the sound. This frequency range also helps with pitch definition in lower frequency instruments.
One can also create a negative effect when tweaking. I tried to implement a shorting ring into the motor, attached to the base of the phase plug. My results are in the chart below. You can notice that the inductance became less symmetrical when the shorting ring was inserted. This means it would actually increase inductance modulation distortion.
Disclaimer aside, if you go slow and take care, you can effect substantial improvements to many drivers. The first, and as old as loudspeakers, is to coat the driver. This is easy to mess up if you're not careful. Many different coatings may be used, from commercially available products like "The wet look" to lacquers (C37) and others. There are also products available which are not intended for this use but are nonetheless appropriate. I've used shellac, diluted woodglue, PVA glue ("ModgePodge" or Puzzlecoat), Damar varnish, and others. Each has its own use. They can largely be divided into "Stiffening" and "Dampening" camps, with varnishes, lacquers, etc providing stiffness, and PVA/glue providing damping.
Stiffening is generally desirable in moderation for drivers who have a gross resonant behavior and need the surface ‘broken up' to avoid this. This means the use of a pattern, rather than a pure coating. Only via experimentation can this effectively be achieved. Stiffening also works in cones you wish to use within pistonic behavior bands exclusively. This moves a paper cone more towards a ‘metal cone' type of behavior. The more rigid a cone is, the more problematic it tends to be when a cone goes into breakup. This gives low in-band distortion but means that you need steep filters and a lot of care to avoid. Damping tends to have a broader band effect. The use of PVA is said to bind the fibers of paper together and prevent "cone self-noise" by its proponents.
You can see an in-process pic of the coating of the B20 driver with PVA glue, diluted 1:1 with water. Parts of the cone have already absorbed the glue/water combo, hence the inconsistent appearance. Consistency in application is important to any full-cone coating method.
Depending upon the type of coating and how much is applied, a specific cone may respond in a variety of ways. Coatings are usually done slowly, meaning thin layers, so one can judge how much is necessary. There are guidelines online for some of the more popular drivers. With the driver shown, 2 thinned coats of PVA (50% water added) were plenty to improve the sound of the driver. The resultant driver had improved clarity and better resolution, but also had less top end. As I usually use supertweeters, this wasn't an issue, but for a purist full range solution, one might have wished to use one, even thinner coat. It's a matter of taste. Go slow, listen, and be careful, or you may wind up overdoing it.
The smaller units are made by Dave Dlugos at planet_10 hi-fi. They are available in a variety of sizes for various drivers, as well as woods and finishes. The larger units shown are homebrew versions of the same, shaped with a lot of sanding and a piece of dowel. Steel screws are used in both sets to allow the magnetism of the magnetic pole piece (where they attach) to hold them in place. Glue could also be used but is not recommended, as anything near the tight magnetic gap of a speaker driver can bind it up and potentially destroy the driver.
A phase plug acts to smooth the acoustic off-axis behavior of the cone, by acting to prevent cancellations from one side to the other. The basic issue is that when one sits off-axis from a speaker driver, the edges of the cone are differing distances from one's ear, so there is a phase differential between the output at different points on the cone, and at some frequency (dependent largely on cone size), they become 180 degrees out-of-phase and cancel. A phase plug helps prevent this to some extent, smoothing off-axis behavior in the highs. Metal phase plugs, as seen in some drivers (Acoustic Elegance, Creative Sound, Peerless, Seas Excel, and others) also help with heat dissipation. Eminence has applied heatsinks/phase plugs to many of its high-power woofers. One other benefit of phase plugs is to eliminate a resonant cavity in the woofer's construction. Similarly to the more commonly seen vented pole piece, the use of a phase plug rather than a dustcap completely eliminates the trapped air between the dustcap and pole. This can be seen on impedance sweeps, but most impedance sweeps are done at lower levels than effectively display the effect.
This hippie voodoo process does seem to do what it's said to, creating a more open and intelligible presentation to my ears. There are some measurable effects, mostly in time-domain behavior. Cumulative spectral decay plots show a more ordered decline of energy. My own experimentation with this (on several sets of drivers) shows it to be a worthwhile process, creating better intelligibility and helping the speaker disappear. The less refined a cone design, the more benefit there seems to be.
Dave Dlugos at planet_10 hi-fi has been modifying speaker drivers for many a year, and the result of those years is a keen understanding and skillset for modifying speaker drivers. I requested a sample of the FF85KeN to try in my Open Baffles. These drivers incorporate many of the processes described above, and you can clearly see the efforts taken in applying EnABL. There are varying thicknesses of the PVA cone coating and EnABL gloss coat, determined experimentally for each driver. The effort put into coating and EnABL'ing the Pioneer driver is duplicated, but with a more experienced touch. Effectively, Dave's drivers are very high quality units, and well worth the premium over untreated, stock drivers. The Planet 10 drivers are $170 a pair, at the time of this article, compared to a price at Madisound of $71.80 for a pair of untreated drivers.
My results in the open baffle speakers were excellent, with the insertion of the FF85KeN bringing a significant improvement over the already impressive stock FF85k. The first thing I noticed was a significant increase in spatial resolution vs. the stock driver. Images were much more stable and the location of the baffle itself was further masked. The transparency and intelligibility became much greater, and more like a high quality electrostatic, planar, or ribbon. The ff85k, stock, is a very impressive driver, with extended treble, and tonal purity, as well as not making offensive noises when asked to play loud. Being a 3" driver with a powerful, advanced motor, cone/dustcap/suspension design, it is superb. The strengths are enhanced in every way by the full Planet10 modification process. You can DIY it if you want to put the effort in, that's why I wrote this article. Dave would be the first to offer advice on doing so. I'll let him explain his processes in more depth within his manufacturer's comments.
When a batch of drivers is received, every driver is unpackaged, inspected and put onto a break-in fixture for at least a week. Upon completion each driver's impedance & T/S parameters are measured. Any defective or out of spec drivers are set aside (few & far between, but not unknown). If drivers are available in different color schemes or some are special ordered, smaller sets of drivers are streamed into smaller batches based on T/S match.
Whizzer cone drivers then get their dustcaps removed (and measured again). Paper drivers get a pre-treatment to seal the cone. A more elaborate pre-treatment, specific to FE126/127, is done to ameriolate some significant peaks at approximately 7 kHz. If drivers are being streamed into smaller batches they are again measured. Spots are then applied. 229 for the FF85, FE126/127, EL70, 265 for the CHR70, 324 for the whizzer coned FE166/7/206/7. Two or three coats of conformal coating applied on top of all this (fewer on whizzers & dust-caps).
A PVA teat is applied to the point of the dustcap, this covered with a touch of conformal coating. Some drivers get an addition set of mid-cone stealth rings (72 more spots -- FE126/7, EL70). Whizzer cone drivers get phase plugs. These are EnABLed (229 spots each), get a conformal coating, PVA teat, and that coated with conformal coating. Each step requires 6 to 24 hrs minimum between steps. Drivers then have their final impedance, T/S measurements and are matched into pairs. Any outliers are set aside.
Larger Fostex drivers (FE126/7/166/7/206/7) have the back of their bezels filled with a marine latex compound which stiffens the bezel., and damps the entire basket. It takes 3 or 4 applications to fill. They are then sanded & sealed.
All drivers have the backs -- particularly at the basket/magnet junction -- is faired out for less turbulent air flow, with a compound that also supplies some damping. Drivers can be special ordered with this step down using the same material as is used on the bezels. It is quite labor intensive and time consuming and has exponential costs for the gains.
We continually work to improve the process and a certain number of drivers are set aside for R&D.
Is All A Bunch Of Hooey!
The top graph is stock while the bottom one is with modifications as above. As you can clearly see, there is a major improvement in the smoothness of this trace. This means less stored energy, and smoother overall driver behavior. These were done on a single unit, before and after modification, in the same test cabinet.
Thanks To Dave Dlugos And Bud Purvine
www.planet10-hifi.com -- A commercial site, modified drivers and some cabinet plans.
www.frugal-horn.com -- A non-commercial site, with plans for a variety of drivers, mostly large tower "BVR" designs
www.t-linespeakers.org -- The original transmission line hobbyist site.The Patented EnABL process is owned by Bud Purvine, and he has given free usage to hobbyists for personal use, but if you plan to paint and sell drivers, you need a license. Dave at Planet 10 and a few others are licensees. If you're in Australia, and some other locations, there are other vendors.