The making of: The Two Towers (a 25 driver Full Range line array)

Still puzzling about the cross talk cancelation I'm using. And in my head comparing it to the shuffler I tried before it. It's not that I don't like what I got now, just wondering if there could be more. Right now I use a few (P)EQ cuts to even out the sound (!). Somehow I think I should be able to come up with some better arrangement that doesn't need any cuts at all.
There's no doubt in my mind this cross talk problem within Stereo is real. And my current solution, however simple it is, does a better job than not having it. But it is a puzzle that might have an even better answer. What can I say, I'm still too curious. I have some ideas I'd like to try, I'll work on those in theory first with REW simulations. If I get something out of it I'll share.
Have I ever been this vague before? :)

(!) = right at the frequencies that get boosted by the cross talk cancelation. Like around 1850 Hz and 5500 Hz. I bet I can change that with a phase deviation like the shuffler did, but this time using a phase shift between the direct sound and the cancelation. There, that's less mysterious.

I did do some REW simulations but did not find a "better" solution. I did try a couple of things. With RePhase and REW I could quickly simulate it and compare to my current solution. In my head it worked way better :spin:.

So no cigar this time around.
 
On to the next experiment. I figured why not try something that would be like the phase shuffler, but different.

The phase shuffler did do some extraordinary things for my perception. But it was very hard to get my desired tonality. So why not use the cross talk in a similar way. I have played with cross talk by cancelling the opposed sound 0.28 ms after the direct sound, and I have placed it in front of the direct sound. There are perceptual changes between the two. But the sides were remarkably close to each other. The biggest difference was noticeable in the (phantom) center.
So this time around I'm playing with cancelling one side before the direct sound, and the other channel I'm cancelling after the direct sound. Very interesting results so far. But only time will tell if I really see (or actually hear) it as an improvement.

But it remains fun to experiment. To learn more about "how we hear". I was enjoying this test, but have limited time. I wanted to play a lot more songs, the 3 songs I tried were really grabbing me. Ah well.... in about 2 weeks I'll have more room to play again. Meanwhile I do get a Forum visitor. Exiting and sort of a sanity check. I can only hope he will like what I'm doing (lol). I can present him a few different options DSP wise and I do hope he will write about it afterwards. (you know who you are ;))

OK, off to listen to one more song!
 

Boden

Member
2010-03-02 9:29 am
Last friday I paid a visit to Wesayso, not only to see and listen to the Two Towers, but also to familiarize myself with the REW, Jriver, DRC and Convolver FIR linearization procedure. Wesayso was most patient and helpful in his explanation.
I have listened for about 30 minutes to the Two Towers, with some material I brought myself. I can say this is one of the most impressive loudspreakers I have listened to in a long time. Apart from detail and depth, for me one of the most striking and unexpected feature was the deep and completely effortless bass, which is amazing given the size of the drivers. But then again, cone surface/displacement matters. As the late John Wright of IMF/TDL advertised: “we rather move a lot of air little than little air a lot”. And so it is with the Towers
Although the listening session was very brief, and I could not compare different approaches of digital equalizing, it was clear form the first notes something truly excptional has been achieved by Wesayso. The measurements of the Towers do not lie.
Very very few system I have heard in the last 30 years come even close. Sometimes reproduction of the Towers is almost intimidating. Some 34 years ago I had a magnetostat line source system with 3 stacked Streathern units per side, which units were also used in ithe Infinity QRS at the time. The was certainly resemblance between these two -seemingly very different- systems. I never managed to properly integrate bass with my Streatherns, so finally I gave up. Wesayso has succeeded here: this is were i.m.h.o. the Two Towers excel. But only praising bass does not do justice to the rest of the reproduction including extra-ordinary transparance.
Some may like to know whether this a “they are here” or a “you are there” system. I always struggle with this description, but I tend to say this a “you are there” system.
For me, this system is an outstanding example of what can be achieved by very clear thinking and application of full blast DSP instead of the typical Hi End obsession with exotic drivers and electronics, but missing the basic requirements a la Toole and Geddes for a worldclass in-room loudspeaker system by far.
What a job!

Eelco
 

Boden

Member
2010-03-02 9:29 am
Hello Dublin78,

Given the limitations of my short listening experience, I found the Towers much much more forgiving than my 1982 Streatherns , which imaged breathtakingly, but only in the sweep spot with head-in-vise.
The Towers are very non-sweet spotted, but cover a large listening window.
Hope this helps,

Eelco
 
Thanks for this write-up, Eelco, much appreciated!

I'll be honest, in a way it was kind of scary for me, having someone come over and "judge" my project. Especially using your own recordings, not me using a select few songs hand picked by me ;).

For me, an outside opinion like yours counts way more than if I state my own, as I created this system. Making me biased as a source of "information" on how good this system really works or performs.

I appreciate you taking the time (and gamble) to come over for a visit. And of course for this write-up. You're making me blush :).

Ronald
 
My Son is back to school, girlfriend back to work... more time on my hands again.

I plan to revisit parts of the phase shuffler. From time to time I like to revisit a theme I did before, but armed with the experiences in between. The Haas kicker was a good example. I've managed to make some improvements during the second run I did with that subject.

The phase shuffler got a long first run from me because if did some things no other experiment could match. Yes it got me stuck eventually, a tonality problem I couldn't solve at the time. A fresh look might bring some new views to this subject, and I could try to mix in some of my later experiments with cross talk cancelation.

I'm not sure yet on how I want to conduct these tests, I've got some more thinking to do, some simulations with JRiver and REW, we'll take it from there.

If I find something even remotely interesting I promise I'll share it.
The cross talk project worked quite well, but it didn't solve everything I wanted yet, that's where the phase shuffler differed from everything I tried so far. But as we all know, different isn't always better.... :dunno:

It deserves another shot, I might learn a thing or two....
 
Well, it's not like I'm not taking the time to enjoy the listening :D....

But there's always room to learn a bit more. For today I just tried some songs convolved with the phase shuffler and listen to it with a few variants of the cross talk. Just to see if I could spot differences in staging and imaging. Tonality is usually a bit harder to really notice, across the stage left to right and the phantom center. That takes a bit more songs for me to judge.

I had fun listening, part of the "why" I still do different tests ;). I'll probably throw on a movie tonight. All the hard work has been done, it's easy enough to try a few changes.

I still need to bring my mic to work and do some near field tests on the monitor speakers I have there. Make a workshop out of it for my students...
 
Hi guys,

I've followed some of this lengthy thread and Ronald's incredible speaker project. I am contemplating a more conventional two-way but ultimately plan on a three-way DSP electronically crossed over system.

I am very interested in doing translam boxes and would love to get some feedback. I LOVE Ronald's enclosures, but mine will be somewhat shorter! I am an experienced woodworker and rabid audiophile, but new to speaker building. Here are a few questions that would be terrifically helpful:

- How do you go about making router templates? Should they be aluminum or wood and how do you fabricate them?
- Is there a router bit you recommend (I am planning on using high grade birch plywood)?
- What bonding techniques are recommended?
- Glue brand/type?
- Dowels or steel bolts through the layers?
- Recommended wall thicknesses for maximum rigidity?
- Baffle
-This is where I am most...baffled! Should I mount the drivers directly in the translam layers (terrified of screwing that up!) or should I do a baffle of MDF, Baltic birch, or perhaps a heavy Corian like composite (I've used this type of material before and I think it could be a great baffle solution)
- What is the best way to mount the drivers into the baffle?
- Recommendations on cutting perfect driver cutouts
- Should the interior by ridged and offset like in Ronald's speakers?

Thanks in advance for any help with these questions. I'll probably have more!
 
I'll let the OP answer most of these when he has the time, because he dealt with all that.

Some I can answer:
-yes, dowels, metal rods and alignment cuts to mount and glue the layers... remove any metal rods or bolts before finishing up.
- DO NOT mount the drivers on the translam layers. Allow for mounting a baffle on the front, and yes, Corian should make for a very nice baffle. Mounting the baffle with a layer of dampening butyl between the enclosure and the baffle.
- any router and measuring tool will allow you to cut perfect circles and recess for the drivers.
- Ronald used the ridges and offsets to minimise backwave reflections inside the enclosure... always a good thing, I believe.
 
Hi guys,

I've followed some of this lengthy thread and Ronald's incredible speaker project. I am contemplating a more conventional two-way but ultimately plan on a three-way DSP electronically crossed over system.

I am very interested in doing translam boxes and would love to get some feedback. I LOVE Ronald's enclosures, but mine will be somewhat shorter! I am an experienced woodworker and rabid audiophile, but new to speaker building. Here are a few questions that would be terrifically helpful:

- How do you go about making router templates? Should they be aluminum or wood and how do you fabricate them?

As I used to be a mechanical engineer specialised in CAD for me it was obvious to draw the whole thing in 3D CAD. But a template could also be drawn in 2D. I chose aluminium as the material because I needed to do a lot of cutting. I figured the aluminium would last me longer than any wood template. I used a water jet cutting service to cut me the templates. All they need is a 2D drawing of the template, that way I had a lot of freedom in the design of the outer and inner shapes.
chambershadow.png


- Is there a router bit you recommend (I am planning on using high grade birch plywood)?

I used B-B grade, void free Birch ply. The router bit I used was: Makita D-09472 HM 9,5mm S6. I must have ordered about 10-15 of them total. Not cheap but worth it to me. Just to ensure I always had a sharp one available. I also used those to cut my aluminium baffles. The Router itself turned out to be important too, the first cheap router I bought lasted me a week before the bearings gave up. The second one (4 times the price) is still working excellent after all those months of abuse.

- What bonding techniques are recommended?

Another time consuming job, the glue stage. Here's what I did,
Glue 2 layers together and make them fit the baffle:
glue1.jpg

(here you see the countersunk hole, the layers are glued together at the non-countersunk side)
glue2.jpg

(the thread cutting tool is used from time to time to rid the bolts from residue glue)
glue3.jpg


After preparing 2 of these I glue them to the stack:
stage-1.jpg

Glue the bottom one first, slide down the middle one, again glue and clamp:
stage-2.jpg

(check level while fastening)

- Glue brand/type?

I used a water resistant wood glue, PVAC based: Bison KM45 Professional glue

- Dowels or steel bolts through the layers?

I assume you saw what happened to my stack?

crack-1.jpg


You don't want that, obviously. My stack was huge of coarse, but wood will always expand and contract, even after being sealed inside and out. I can tell you what I would do, actually I already did. A next time I would repeat what I did to fix my stack, I'd still use metal rods, slice the enclosure up regularly (fill with a damping layer) and cover the whole stack with epoxy and fibreglass matt.
There are other ways of coarse. I know we have people on here making horns and use stacked Birch for them, bolting each layer to the previous one. As I have no experience with that, I cannot say what it will do over time.

- Recommended wall thicknesses for maximum rigidity?

My minimum wall thickness is 18 mm, the wavy shape makes it irregular and the thicker parts are ~35 mm. Of coarse I do have a lot of bracing. For a 2 way I'd look into the bracing as suggested by Dr. Geddes. We touched on that subject in this thread: http://www.diyaudio.com/forums/multi-way/295467-box-damping-bracing-question-maybe-idea.html

More specifically this post sums up those ideas: http://www.diyaudio.com/forums/multi-way/295467-box-damping-bracing-question-maybe-idea-5.html#post4803827

- Baffle
-This is where I am most...baffled! Should I mount the drivers directly in the translam layers (terrified of screwing that up!) or should I do a baffle of MDF, Baltic birch, or perhaps a heavy Corian like composite (I've used this type of material before and I think it could be a great baffle solution)
- What is the best way to mount the drivers into the baffle?
- Recommendations on cutting perfect driver cutouts
- Should the interior by ridged and offset like in Ronald's speakers?

I had a reason for every step I took, and have no regrets. I liked aluminium for the baffle as a material for it's rigidity. I wanted to make a durable product. I took the steps needed to damp the aluminium and believe damping it is crucial. Corian like materials should work too.
After I started my project and cut the first few pieces I stumbled across the Magico Mini speakers. I swear I had not seen them before making my own decisions, but took note on what they did and how they did it after finding them. They have some cool features in the Magico Mini II enclosure. I like how they connect the front and back with tensioned rods. That way they can have an uninterrupted stack of ply and still mount the baffle in a sleek way without hindering the expansion and contraction of the wood. I doubt it would work with a big stack like mine though.

attachment.php


attachment.php


Magico went to an all aluminium enclosure after this one, but optically for a mini monitor (of huge proportions) I still think this one wins the price. But if the Sonics rule.... :rolleyes:

I am more than pleased with my enclosures, they seem to be the only part in my living room that's completely inert when the music is blasting. You can lay your ear on them and not pick up any sound or movement while the music is blasting full force. I used at least 5 different damping techniques to get there though. I figured every detail counts.

Perceval already went over the reasoning behind the wavy shaped interior. One is to disrupt the back wave (though one has to consider wave lengths to determine how successful that can be) and the other reason to vary the wall thickness was to prevent it from being one big resonating surface.

Thanks in advance for any help with these questions. I'll probably have more!

No problem, I used it as an excuse to (re)post some pretty pictures :D
 

Attachments

  • baltic_birch.jpg
    baltic_birch.jpg
    76.9 KB · Views: 315
  • 2XMini2.jpg
    2XMini2.jpg
    145.2 KB · Views: 305
Last edited:
Thank you, that is a fantastic reply! I appreciate the detail and will follow your links.

The Magico Mini was actually part of my inspiration. I heard it once and it was terrific, I liked it better than the current aluminum body Magicos. The baffle mounting is indeed interesting but may be a bit ambitious for my first project!

Thanks again, I'll post my progress once I get started.
 
It should be (the reply, lol) it withheld me from blasting a few songs :D... Now I got to run to work! Do post that progress! Plan your project carefully, it will pay off. And do not forget to chamfer behind the driver, making it breathe on the back side....
uitsparing-a.jpg


chamferback.jpg


I was able to do that with a template made out of left over aluminium baffle material and a 45 degree router bit.
 
Good to know about the Carbide tipped router bits, I did try to get some but ended up using what I had. The "normal" HD router bits cut trough aluminium like butter. The biggest problem I had to deal with was the exploding (literary) bearings due to the high temps. I'm glad I always used plenty of protection for my eyes, ears, hands etc... I've had bearing parts flying around me more than once... :eek:

Bearings.jpg


I tried using different bearing types with ceramic roller balls and they lasted only a fraction longer.

I did cool as much as I could with WD40. The cutters did a marvellous job after I replaced the function of the bearings with a static "following pin". I used a jig saw with metal blades to do the rough cutting, followed by the HD router bits.

Beugeldetail.jpg

close up of the following pin following the template (the upper part)

Beugel.jpg

The construction I made to replace the function of the bearings

Even the polishing is a hand job (using my drill) after a careful router job making the fillets.

polish-6.jpg


I could see myself do a job like the Magico Mini, though not with the baffle thickness they used :D. My baffle was a twin aluminium plate by design to be able to use damping between both layers and have a good solution to mount the speakers..

BaffleDetail-diy.jpg

The two baffles, one alu colour and one in black

transition.jpg

Seamless transition from speaker cone to baffle to enclosure in this cross section

This was my first router job ever... I had never used or owned one before this project. First time for using epoxy as well. The original plan was CNC cutting. A lack of funds made me pursue this DIY route.
 
Last edited: