Introducing: Echelon

[dlneubec]

Introducing: Echelon
Echelon is a full range 3way passive speaker that includes a 12” woofer in a 90 liter sealed cabinet, a 10” planar midrange, implemented open back, and a 1” ceramic dome tweeter in a custom waveguide.


Support

Thanks to fellow DIYer’s who lent a hand to this project. Dave Pellegrene and Dan Poinsett provided waveguides for testing with various tweeters. One of Dave’s was implemented here. Thanks also to Jeff Bagby for directing me to the close-out source for the woofers a couple years back. Ben (Wolf) Shaffer suggested the names, Echelon for the WMT and Echelon Fortissimo for the upcoming WMTM version. Thanks to others for their name suggestions as well. Thanks to all the kind comments in regard to these speakers at the InDIYana event. They are very much appreciated.


Concept

The concept is quite ambitious. The idea is founded in a modular approach, where a fixed woofer/mid module could be paired with multiple tweeter modules to create different WMT’s and/or add a second midrange module for a WMTM. The WMT version would be all passive. The WMTM version, with a 6db increase in midrange sensitivity, would need bi-amping and an active crossover for the woofer, passive for the MTM. A higher sensitivity tweeter module would be needed for the WMTM, most likely a ribbon or planar, perhaps utilizing a waveguide. The tweeter module would be removable and carry the crossover for the speaker (passive or active).


The Drivers: Echelon (WMT version)

Woofer: TC Sounds, TC1000 12”
This is a new driver I’ve had a pair of for awhile, though they have not been available for a number of years. That said, it is likely these could be replaced by the Dayton RSS315 or the TC Sounds Epic12 and others that are currently available. The TC1000 is a high excursion (25mm Xmax) 4 ohm subwoofer. However it measures extremely well, virtually flat from 40hz (sealed) to well beyond 1000khz. The 1m/2.83v sensitivity is indicated at around 87db using a .52ohm series resistance.

The woofer is placed near the floor in this design and on a baffle that is 16” wide at its widest. This offers the opportunity to include little to no baffle step, except what occurs from the dcr of the low pass inductor. The box is about 90 liter sealed. Modeling with Jeff Bagby’s Woofer Box Model and Circuit Designer indicates an F3 of around 28hz. When including room modeling from Diffraction and Boundary Simulator the expected in room performance suggests an F3 of around 18hz. The simulated response has not reach f6 by 10hz.

The woofer can handle over 500 watts of power and still not reach excursion limits. This power handling capability will be important for the WMTM version. In a sealed box, this woofer is capable of keeping up with the fast transient response of the open back planar mid without any detectable differences in the character of the sound through the crossover.


Midrange: BG Neo10 planar
The neo10 is the latest large format planar driver from Bohlender and Graebener. It has been designed to increase power handling, low end performance and sensitivity over previous models, while keeping distortion low and maintaining great transient response. Impedance measures close to flat across the entire spectrum, at about 6.5 ohms.

These are placed in an open back configuration which allows for some adjustment of midrange energy as desired. The open back configuration provides a substantial increase in soundstage width and depth, at no apparent loss of imaging. It also eliminates the coloration/boxy sound that can sometimes be observed in sealed enclosure midranges. The Neo10, perhaps due to its fast transient response and open back implementation, delivers one of the most realistic, live sounds I’ve heard.

Tweeter: Transducer Lab N26C
The N26C is a high purity, molded Al2O3 Alumina ceramic dome tweeter, designed / assembled / tested in the USA, utilizing CNC machined metal parts, a double magnet design, decompression chambers, a fiberglass-under-hung voice coil (1mm linear excursion), silver terminals, with a response range of 950hz - 45khz. They are sold in pairs that are frequency matched to +/- 0.75db.

http://www.transducerlab.com/6501/6522.html

This already low distortion tweeter is implemented in an elliptical waveguide developed by Dave Pellegrene. It is approximately 1.5” deep x 5” tall x 6” wide and provides a low end boost of around 10db. When this boost is equalized out in the crossover, distortion is lowered even further while significantly improving tweeter headroom, allowing substantially increased spl levels without noticeable strain.

In addition, the waveguide provides controlled directivity of tweeter and matched directivity with the midrange at the crossover point. This ensures an improved power response over traditional flush mounted tweeters and a speaker that is less sensitive to room placement.




I will post more on the crossover design and box construction next. When I have more time, I will add crossover detail and sumlation as well as some build pics. It took a lot longer to write this up than expected!

[dlneubec]

Crossover Design

The woofer was surface mounted, the midrange rear mounted and the tweeter waveguide mounted. Combined with the sloping baffle, this allows for the driver voice coils to be equidistant to the listener’s ear, which promotes improved phase tracking between the drivers over a wider range of frequencies and reduces or eliminates the need for asymmetric crossover slopes.

Crossover points are at 325hz and 2100hz. The Woofer to midrange acoustic slopes are approximately 15db/octave, halfway between second and third order. The midranges to tweeter acoustic slopes are approximately 24db/octave. The crossover models at +-1db from 40hz to 14khz, where the tweeter begins to roll off in the upper half-octave (see simulated response graph). The woofer is wired in reverse polarity.

Due to the increased rear radiation of the open back mid, the midrange is very slightly depressed in the crossover design. However, the relative level of the mid is somewhat adjustable to fit an individual’s sound or musical tastes, room response or speaker placement by altering the amount of stuffing in the open back enclosure. A number of different materials where tested in the open back, but the most effective I found was one layer of 2”unbacked fiberglass insulation along the bottom, with poly-fill being employed if further attenuation is desired. I intend to test the recycled denim based insulation material as a replacement for the fiberglass.

Two pair of binding posts are provided on the rear of the cabinet. This will allow for bi-amping of the Woofer and Midrange/Tweeter crossover separately, which will be necessary for the Echelon Fortissimo, where the woofer will be controlled by an active XO, such as the MiniDSP. Bi-amping of the passive XO may also offer some benefits.



Woofer/mid Box Construction

The woofer box is constructed using MDO plywood. This plywood has a thin layer of MDF-like material on the outside layer that does not chip out or splinter when cut and makes it easier to paint or veneer than other plywood’s. It is also much lighter than MDF. The joints around the sides and top are all miter joints, so there is no telegraphing of the joints through the paint. The rear panel is dado’d into the sides and a layer of 10 mil paper back veneer was added over the back to reduce telegraphing.

Bracing includes closely spaced, ½” window type bracing to ensure that unsupported panels are small. The bracing system is tied together vertically and horizontally in the box. If you look closely, you will notice that all the bracing is dado'd to the sides and to each other to increase the glue surface area. One-quarter inch thick rubber tiles have been adhered to the side walls to add some mass and further diminish potential vibrations.

There is a full ½” backing piece behind the laminated baffle that ties the cabinet together with the bracing. The baffle itself is 1-1/2” wide strips of the MDO plywood laminated with the edge grain out. It was then sanded at a local mill to 1-3/8” thick. The drivers are attached directly to the front baffle which is kept floating above the speaker body with ¾”wide x 3/16” thick weather stripping. This is intended to limit vibration transfer from the baffle to the box and also allow for access to the rear mounted drivers.

The sealed portion of the woofer box is heavily stuffed throughout, which helps absorb the sound that radiates from the rear of the woofer cone and flattens the low end response. The woofer was surface mounted, rather than flush or rear mounted for a couple of reasons, as mentioned, one of which was visual. The woofer appears larger and more substantial when front mounted and visually attenuates the width of the baffle.

The many angles to the cabinet sides are for more than aesthetic reasons. They reduce the effect of diffraction and allow the speakers mass to be located near the floor where it is less visually intrusive in the room. Further, none of the cabinet sides are parallel except the top and bottom, which when combined with stuffing reduces or eliminates standing waves inside the cabinet, ensuring that only the drivers are heard, not the box.

By mounting the woofer near the floor, the woofer’s floor bounce null, experienced at the listening position, is pushed much higher in frequency, well above the crossover point and into the stop band of the woofer.




Tweeter Box Construction

The tweeter box is constructed using the same MDO plywood, with the exception of the curved top and side sections. The top was created by laminating 6 pieces of 1/8” Masonite, using strap and traditional clamps to form it over the box. Tightbond III was utilized throughout. The curved side panels are ¼” Masonite glued to the ¾” MDO sides.

The tweeter box is held in place over the woofer/mid box using cap head screws and threaded inserts, with the screws accessible from inside the midranges rear tube. There is a ¼” piece of rubber tile between the woofer/mid box and the tweeter box to help isolate it from vibrations.

The woofer crossover and mid/tweeter crossover are housed on separate boards that are screwed to the top of the woofer/mid box through the rubber tile. There are openings cut in the bottom piece of the tweeter box to allow it to fit over the crossovers. The baffles on both the woofer/mid box and the tweeter box have large facets to further reduce the effect of diffraction.

[dlneubec]

Construction Continued

A number of folks asked about the legs/feet at InDIYana and CJD mentioned them in his post. Here is some info on the legs/feet. I will add a close up photo when I get a chance.

Legs/Feet

There were several questions about the legs/feet (kind of like outriggers, I guess). The legs are 6” x 1.5” x .25” aluminum and are a standard part from 80/20. The part is called 6” pivot arm, part number 4399, page 325 in their catalog. You can go to their site: http://www.8020.net/ and find a local dealer you can order from. You can also download their catalog and price lists from their website. I believe the cost was $5.75 each, or $46 for the total. They come with three 5/16” holes in them, so no need to drill them. In my case, I used threaded inserts in the box base to attach them at two points, leaving the hole in the rounded end for the feet.

The feet were ordered on Ebay from a company called SimplyPhysics. Here is the site: http://www.simphys.com/Pages/isolation%20products.html Mine are the “Medium ToneCones”. They were $6 each, for a total of $48. My one reservation is that the body is made of plastic and you actually create your own threads, they are not pre-threaded. You can’t pivot the speaker on them, the threads are not deep enough to withstand that much pressure. You have to pick the speaker up to move it. They don’t come with studs or bolts, despite what the photos suggest. They are available in various thread sizes, and there are larger and smaller cones available. They have an open slots inside so you can add some anti-vibration material, like bluetack, if you believe that will be beneficial.

What I did was sandwich them with a combination of 1-1/4” x 1/8” thick galvanized washers and doubled rubber washers of the same diameter. There are two rubber washers sandwiched on each side of the leg, followed by the galvanized washer, followed by the foot on one side and a, 1/8" hard black plastic washer, galvanized lock washer and ¼” x 20 x 1-1/4” bolt on the top side.

[dlneubec]

Crossover Results/Design Simulations

Attached are some crossover simulations done in SoundEasy, where all measurement and crossover design were done. Also attached is the box model, courtesy of Jeff Bagby’s fine software tools. I can also post individual measured driver response over the network response, if anyone would like to see that.

Modeled frequency response: based upon in box measurements taken at 15º horizontal off axis, which was the planned design axis. This model doesn't include room/boundary gain at the low end that you see in the box model, which starts at about 300hz and below. The box/room model indicates and increase in bass output starting at about 2db from 300hz down to 70hz and then it increases relative to the sealed nearfield measurment below 70 hz. With that in mind, you can see that the midrange wouls be slightly depressed which compensates for the open back in room contribution. The midrange level was established primarily as part of voicing.


Electrical Transfer Function:


Modeled impedance/impedance phase: based upon in box impedance measurements. Impedance drops below 4ohms briefly at the low end and generally stays between 4 and 9 ohms. Impedance compensation was included in the crossover design to flatten impedance and impedance phase.


Acoustic phase:


Group Delay: This is the result of electrical and acoustic delay. I can also post electrical delay if anyone wants to see it.


Box Model: The box model simulation below includes the effect of room and boundary gain.

[cjd]

I really like what you've done for the feet. And the baffle.

Wish I could have made it and heard these!

[isaeagle4031]

You missed a real treat cjd. These were amazing! I honestly can not say there was anything about them that I did not like. The bass was powerful without being overbearing.

Beautiful work Dan! You are an inspiration to the rest of us!

[Jeff B.]

Very nice write-up. And exceptional speakers. For those that haven't heard them - they sound as good as they look.

Jeff

[scottsehlin]

I wish I could have heard those too - another reminder that I need to make a point to get to InDIYana one of these years. The cabinet work looks fantastic and I'm sure based on your skills and the driver compliment that they sound great.

I have a pair of TC sounds Epic 10's squirreled away for a future project and have always liked the planar BG mids - they are spooky with their realism and detail at times. I used the TC-1000 10's in a 3-way a number of years ago after hearing them in the Salk HT-3. They are as clean and authoritative as anything I've ever heard from 20 Hz to 300 Hz.

I will definitely have to at some point try out the Transducer labs tweeters - I spent a couple of years developing a ceramic injection molding process for a different application. Given the various choices, I would consider making them out of something other than alumina, though...

[morris]

Wow! That made my heart race! Great work!

[Wolf]

Now that is bracing!!! Thanks for the recognition on the name, and it does suit very nicely.

Looking forward to more pictures of the innards in the top for the xovers.
Later,
Wolf

[Jim Holtz]

Hi Dan,

That is a beautiful speaker! I hope I get to hear it at this falls Iowa DIY event.

Jim

[GranteedEV]

Those look amazing. You must have a brace every 4-6 inches

[Mayhem13]

Amazing job Dan.....WOW!

I want to try the Neo10 in a dipole MTM soooooo badly.

[bolland83]

Wow! Can I get a pair in Cadillac pearl white?

[donradick]

Simply amazing wood work, ambitious audio design, and flawless execution. Sadly, my skills will never be up to the build. Gives me something to aspire to, though.

[mzisserson]

Flawless, as usual! More importanly, inspiring!

[davepellegrene]

Thank for the recognition Dan. Those are absolutely beautiful. These were my first chance to hear one of my guides implimented. I could not tell at all were the tweeter and mid integrated. As I was listening I moved around the room and could not tell any change from one spot to the next in sound. The upper octaves seem to hold steady at all angles. Directly in front of the speaker didn't seem overly bright as I thought a guide may do.
Great job Dan

Dave

[johnnyrichards]

Come on Dan, you gotta do better...

Very well done, very well done indeed. I hope they make the trek to Iowa this fall.

[joeybutts]

uhmmmm.........OMG

Amazing. absolutely stunning. I can only imagine how they sound.

[dlneubec]

Thanks, guys!

Unless life intrudes in some unexpected way I do plan to bring these to MWAF and Iowa this year, for those who are interested. If I get really ambitious (probably not, I'm a bit burned out after this build), I may even have the tweeter and second midrange box to swap out and create the WMTM version by the time Iowa rolls around.

Bracing spacing is in the range of 2" to about 6" or 7", where the side panel miter joint is. The change in direction at the miter joint adds some stiffness, so there was little need for a closer brace. The bracing is 1/2" MDO ply, so even though there is lot of it, the volume it eats up is really not any more than my typical 3/4" shelf bracing would take, spaced further apart.

It was a bit interesting folding that box, wiith the miter joints at all corners, into shape with the tape at the corners holding it until I was able to get some strap clamps in place. Next time I will use a stronger tape than painters tape.

IMO, mostly it takes a lot of patience and time, but anyone here can do this level of woodworking or better. I don't have any high end/special tools, just a Craftsman/Riobi table saw, couple craftsman routers, router table on the table saw, Craftsman drill press and belt sander, a Dewalt orbital sander and a couple portable drills. The highest end tool is probably my Bosch jig saw, which I don't use much on speakers anyway. I don't have a vacccum bag for veneer or a HLVP gun for spraying.