DIY transducers--the final frontier

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
At some urging from others on this forum, I'm starting this thread about DIY transducers, the woefully underexplored final frontier of DIY audio.

A couple weekends ago, I spent a weekend dreaming about transducers (as usual) and thought I'd show you what I came up with, just for fun. It's way off the beaten path, but that's where I like to spend my time.

(See image link below)

What got me started was looking at all the bass drivers I have sitting around in boxes and wondering if I could cannibalize a motor from one as a foundation for something new, maybe a wide-ranger.

These bass drivers have 3" and 4" voice coils, not usually the cat's meow for wide-range use, given the large, resonant dustcap and massive inductance-boosting pole of typical bass drivers. However, I have recently begun a love affair with faraday rings, which have the seeming magical ability to cancel the pole's effect on inductance. I began wondering if a 4" voice coil motor could work for a wide-ranger.

As I see it, there could be several advantages to large-diameter VCs. One could begin with a ready-made bass motor--giant mags driving lots of flux through a thickish top plate. The wide gap spacing of the bass motor would allow room for inner and outer faraday rings without modification. Since the VC is 4" instead of the more usual 1", It would have a 4x advantage in terms of force it could derive from a given field density. Given a thick top plate, say 1/2 to 3/4", and the fact that we'd be winding a short coil of light-gauge wire, an underhung voice coil becomes possible. All else being equal, an underhung coil between faraday rings is the finest recipe I know for extreme linearity.

Furthermore, having played with an online inductance calculator, I believe this short, large-diameter coil should exhibit less self inductance than smaller-diameter, taller coils that are closer to the dimensions of an ideal inductor.

So, to review, we've taken a large bass driver motor and added faraday rings and a light-gauge underhung coil.

Now, the typical way to patch the hole in the cone left by a 4" voice coil is to just slap a dustcap on it. For wide-range, the resonances this would create are unacceptable. So, for this exercise, I decided to try something different. I decided to use a big ol' ugly phase plug--perhaps a wood turning. This phase plug performs a number of valuable services: It replaces the dustcap, progressively horn loads the cone, and serves as a waveguide.

For the cone, I decided on an 8" diameter. With the giant 4" cutout in the center, the effective area of the cone is 16pi-4pi=37 square inches, equivalent to a normal 7" cone. However, instead of the normal cone's 3" of cantilever out from the edge of a 1" former, this cone has a 2" cantilever (from a 4" former). This gives the cone a geometric stiffness profile closer to that of a 4" driver, a significant advantage that will raise the rolloff frequency.

The expanding radius of the phase plug progressively horn loads the inner edge of the cone where the higher frequencies are generated. As frequency climbs and the effective radiating area shrinks back toward the inner edge of the cone, the radius of the phase plug also guides the output away from the center, dispersing the high frequencies and improving the polar response. Of course, the profile of the cone and phase plug could be adjusted for different loading/dispersion characteristics.

To sum up this design exercise, I think this driver could offer the following benefits: use of a production magnetic circuit, high linearity, efficiency, wide-range response, and good high-freq. dispersion.

Allright everybody, let's punch through the glass ceiling and DIY our own drivers!

Bill



An externally hosted image should be here but it was not working when we last tested it.
 
All you need the phase plug to be is stiff and dead. Like maybe a machining made from a dead cat...although that might not be stiff enough.

I just spec wood because (if you have your own lathe) it's a snap to turn whatever shape/profile you want.

Bill

PS. By the way, how do you make the darn pic appear in the dag nabit post?
 
>check out the "Virtuoso" design at http://www.beauhorn.com. They use a
very huge "Phase Plug" made from wood to load a
Lowther driver with a front horn.

Yes, that speaker was one of my inspirations.

>It always seemed to me that electrostatics were fertile ground for DIY experimentation. Agree? Disagree?

Strongly agree! I've seen some interesting examples, though I'm exhibiting my fixation on dynamic drivers.

(Rant follows)

Actually, the whole reason I started this thread is because I believe ALL transducers are fertile ground for DIY experimentation.

For all the hairsplitting DIYers lavish on everything upstream of the transducer, I'm shocked at the dearth of attention transducers get. Aside from the T/S params, which get plenty of attention, the factory transducer offerings available seem to be taken for granted as sealed systems of devine engineering. It's pure audio gnosticism, and I'm trying to punch a hole in it.

Speaking in general and allowing for exceptions, driver cosmetics have far more to do with what reaches production than distortion and dynamics. As publicly-held companies serving predominately ignorant consumers, the big names have certainly learned to avoid the diminished returns of lengthy, meticulous engineering.

However, you and I do not answer to beancounters. We can (and frequently do) voluntarily give up evenings, weekends, and holidays to raise the level of refinement and generally push the envelope of our sound reproduction gear. I'm suggesting that our quasi-religious dedication to tweaking can and should be extended into the realm of the transducer. It's time to split the audio atom. If you doubt the importance of such a crusade, just place the THD numbers of you source/pre-amp/amp beside the THD of your output. The difference is your drivers/cabs.

Are you willing to take the linearity of your drivers' magnetic circuits for granted, or even consider it a luxury? Or are you going to get your hands dirty and make it so?

Have I whipped you all into a seething, chanting fervor yet? :)

Bill
 
<img src="http://www.villagephotos.com/pubimage.asp?id_=215538" width=632 height=901 >

This is a little twist I came up with. How about running the spider inside the VC former? You could damp the resonance of the spider by tuning its air load by stuffing the pole vent that forms sort of a transmission line. Since the spider is isolated in the former, it's noise is less likely to radiate through the cone. Also, the cone's backwave path is opened up. In the drawing, I included some felt on the top plate to further prevent reflections back through the cone.

And here's the side view. Isn't that nice and sanno?
<img src="http://www.villagephotos.com/pubimage.asp?id_=215537" width=632 height=659 >

Bill
 
Disabled Account
Joined 2001
Just wondering how you plan to deal with the resonance caused by the trapped air chamber of the spyder? I see no way to avoid a resonance of this air chamber with the illustration that you provided.

Also, the phase plug you suggest, while drastically creating excellent far off azis HF behaviour, will have the exact opposite effect anywhere near on-axis. Maybe you could consider a smaller phase plug(much smaller).

I am glad to see new ideas in the hobby. Good luck!

-Chris
 
My idea WRT the enclosed spider was to tune the fundamental air chamber resonance 1/2 octave from the fundamental of the spider itself, thereby damping it. There could be other ways to use the pole vent as a transmission line.

As to the phase plug/waveguide, the drawing may not be ideal. The idea is to profile the plug so as to project a ring HF beam that overlaps with itself in the middle enough to create even response there, though the directivity is toward the periphery. I think the overlap could happen pretty nearfield.

Designing transducers is fun! Everyone audio fanatic should do it. If nothing else, it's an education. It also opens your eyes to which drivers out there are quality and which are window dressing.

Bill
 
Disabled Account
Joined 2001
My idea WRT the enclosed spider was to tune the fundamental air chamber resonance 1/2 octave from the fundamental of the spider itself, thereby damping it.

I believe that the fundemental resonance of the spider, while being a minor issue in itself, will not be relevant to the resonance of this chamber you have created. It will not be so simple to damp this. Now if you can find someway to vent this cavity.....perhaps an unusually porous spyder, or special 'cutout' spyder with spaces cut out on it to allow air to pass freely.....I don't know...

The idea is to profile the plug so as to project a ring HF beam that overlaps with itself in the middle enough to create even response there, though the directivity is toward the periphery. I think the overlap could happen pretty nearfield.

This is the purpose of any phase plug. But the dimensions would depend on your intented bandwidth of this driver. Sound waves only bend around an object when the object is less than 1 wavelength of said frequency, otherwise they primarily reflect. The design that you have shown, if the phase plug is anywhere near the scale you have shown..will cause serious on-axis problems, and peaks off axis. Now, reducing the bulge on the phase plug should solve this problem, but then you may wish to shorten the plug also, or you will also be redirecting some lower frequencies which do not suffer from beaming(directional) problems, thus creating further off-axis non-linearities.

Designing transducers is fun! Everyone audio fanatic should do it.

Well, maybe some should do it...LOL. I am a shade pessimistic. I also seriously doubt that without access to finite element analysis software, and the nesecarry hardware and tooling to measure and fabricate these advanced designs...that a superior, or even equal quality dynamic driver can be developed from scratch. However, perhaps you are lucky enough to have those resources at your disposal.

Have a great time building!

-Chris
 
You may by right about the cavities, though there are a number of ways you could vent them to achieve a variety of air loads, it seems to me. Viability of this variation would depend on the tradeoffs involved. A resonating spider parked under a thin cone is a liability of classical design. I don't know whether an innerhung spider is less of one. At worst, consider it a conversation starter!

>Sound waves only bend around an object when the object is less than 1 wavelength of said frequency, otherwise they primarily reflect. The design that you have shown, if the phase plug is anywhere near the scale you have shown..will cause serious on-axis problems, and peaks off axis. Now, reducing the bulge on the phase plug should solve this problem, but then you may wish to shorten the plug also, or you will also be redirecting some lower frequencies which do not suffer from beaming(directional) problems, thus creating further off-axis non-linearities.

Again, you may have a fine point. I'm willing to make the phase plug whatever shape it needs to be to work.

>I am a shade pessimistic. I also seriously doubt that without access to finite element analysis software, and the nesecarry hardware and tooling to measure and fabricate these advanced designs...that a superior, or even equal quality dynamic driver can be developed from scratch. However, perhaps you are lucky enough to have those resources at your disposal.

Well, they were developing pretty fine drivers even before T/S parameters came into vogue! An intuitive sense of what's going on, grounded in good ol' cut-n-try, can take you a long way, IMHO. Though they do have FEA on their side (grumble grumble), most manufacturers are hobbled by a little critter called profit margin. This often means cutting corners on hidden details that only affect often-ignored metrics like distortion and harmonic components.

Take, for example, the issue of faraday rings. It has been long known that faraday rings, especially full-length ones, greatly enhance magnetic circuit linearity by eliminating the negative effects of eddy currents and induction in the pole piece, and to a lesser extent, the top plate. Trouble is, full-length faraday rings require additional spacing in the magnetic gap which results in lower flux density. To hit your target BL, now you've gotta spec significantly larger magnets. Some of the best drivers have faraday rings on their poles, but usually not full-length. And I don't know of any that go the extra mile and faraday shield their top plates.

This is just one example where manufacturers' compromise tolerance is significantly different from that of most Hi-Fi nuts who are generally willing to make epic journeys deep into the realm of diminished returns to find the next gram of performance.

The problem I sense is that there is general ignorance about the root causes of speaker distortion. Most subscribe to one of two attitudes: either they are blissfully unaware of driver distortion, trusting fully in the art (marketing?) of the manufacturers, or they consider the downfalls of certain drivers inherent to the design (breeds comments like, "Electrostatics are inherently more detailed than dynamic drivers.") Neither attitude puts pressure on manufacturers to refine their products, much less innovate. The kind of envelope stretching I see in DIY amplifiers is what I miss in transducers, which are arguably a much more critical link in the reproduction chain.

I'm ranting again, sorry.

Hey Chris, thanks, by the way, for the quality feedback! When you want to toss around a concept, it's nice to have someone besides yourself to play catch with.

Bill
 
Disabled Account
Joined 2001
I agree that many companies do negatively effect the performance of their drivers in order to save a buck or two. But their are many fine drivers available that don't seem to sacrifice any performance, such as scan-speak woofers/mids and SEAS excel lines, for example that have extremely linear motors, solid construction and high quality cones and suspensions.

As for old dynamic drivers being in the same ballpark as todays? Uhm, I would have to disagree. Only a distinct handful of antique drivers are of any value, IMO. They were likely a result of 'luck'. LOL. And I do not think any 'antique' dynamic drivers come too close matching any prime examples of the dynamic drivers of today. Though, I'm sure some people may feel otherwise and disagree with me.


I have one question. :)

How do you intend to achieve acceptable sensitivity levels from this cone driver? It is obvious that the moving mass is nothing less than huge, compared to surface area of the diaphragm. Even if the phase plug is installed to the pole piece, then the mass is still nothing less than huge, or at least it appears to be.

Again, if these pictures are not to any scale whatsoever, and only very rough ideas, then i do understand.
 
Bill,

topic phase plug:
Did you consider what Lowther did with their holy (holey?) doorknob phase plug?

I made a listening comparison on my PM6A-equipped Lowther Acousta short before i sold it.
A really nice improvement, compared to the classic, paraboloid phase plug. Particularly the driver's beaming was reduced and spatial radiation was dramatically improved. The phase plug's horn loading of the VC edge and whizzer cone area od the diaphragm is softened by the phase plug's holes and the holes also may let pass high frequencies better on-axis, maybe also by forming secondary sound sources, edge reflections, dunno.
 
Chris,

Of course, the phase plug is not attached to the diaphragm!

As to sensitivity, that big motor is packing a pretty hefty magnet. Although my underhung topology and full-length faradays use up some of its grunt, I'm guessing it could have pretty decent sensitivity (though, of course, I don't know for sure). Though the VC former is a good deal larger than most, it's being driven by a whole lot of BL.

I suppose I'll just have to slap one together to find out. Where to find the time...

Dice,

Yeah, I've puzzled over that pierced phase plug. I'd love to see a white paper thoroughly explaining its function.

Bill
 
Hello Diy fiends!
This is my first post here.
I too have been bitten by the transducer bug.
But on a smaller scale....for headphones.

I have thus far made my first working prototype pair of phones using transducers that work on the planar magnetic principle.

I will refrain from expanding at the mo as this is a first post and
feel it would be rude to barge in on this thread.
It was this thread that caught my eye though and encouraged
registration.


Setmenu

Team flatstuff:)
 
Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.