[454Casull]

News:

2009/07/30 - tested power supply, assembled prototype driver mount
2009/07/16 - assembled power supply for active crossovers, need to test

Goals:

• bang for the buck
• 110 dB @ 1 m peak music SPL
• relatively high sensitivity
• semi-constant power output

Isometric view




Front view




Rear view (tweeter rabbet in place)



I guess the first thing I should mention is the strange look of the cutouts. They're like that because when they come off the CNC, I want to have a little support under the router's base plate on the inside for the edge radiusing. I can knock out the insides and clean up the holes after the edges have been radiused.

The drivers are:

• (2x) Vifa DQ25SC16-04 (dipole/bipole)
• Eminence Alpha-6C (dipole)
• Eminence Beta-12A-2 (dipole)
• Dayton RSS390HF-4 (mono)

I wanted drivers that were a) cheap, b) sensitive, and c) low distortion. The first three satisfy a) and b), and the tweeters measure very well. The Eminence drivers do not have much measurement data floating around online, but the other mids and midbasses I checked didn't have as good a combo of high sensitivity and low cost. Still, they could use some work, and I think I'm going to remove the dust caps and slice the cones when I get them in. The Dayton measures very well but is expensive, especially up here in Canada. Still, I like them a lot. I may decide to use the Jensen drivers that are at a certain sale, though. They're cheaper and more sensitive, which means I can get away with a weaker amp and power supply.

The approximate crossover points are 150/200 Hz, 600 Hz and 1800 Hz. The system will be fully active, of course.

The RSS390 will probably go into sonotubes (~6 cubes each plus stuffing) on the floor, located on the lateral side of each TMW. Monopole operation allows for better power matching to dipole mids than with dipole bass, though I may try dipole bass in the future. The displacement requirements can get rather pricey, though. In any case, I will probably add more monopole subs to smooth the in-room response.

The thickness of the baffle is 2.25", or three 3/4" MDF boards laminated. This is just enough to fit the two tweeters back-to-back. May decide to have this made from one piece of hardwood or aluminum, though. The other dimensions have been selected so as to allow each driver to operate under its dipole peak... mostly. The tweeter face plate doesn't allow for the minimum distance, so according to Edge, even at the low end the tweeter is operating above the dipole peak. This won't be as big a deal since I'm planning to run the rear tweeter as low as -3 dB down.

I don't know how I am going to support the baffle.

Spreadsheet for calculations. This is what I used, along with Edge, to determine the excursion and power requirements for each driver.

[mcgsxr]

Interesting design, to be sure.  Passive I assume?

You could mount the whole baffle on a piece that runs parallel to the floor - forming an L with the drivers mounted on the vertical piece of the L.

Just a thought.

Good luck, and keep us in the loop!

[opnly bafld]

The system will be fully active, of course.

[454Casull]

All-in-one spreadsheet

[pedroskova]

What, exactly, do those wings accomplish?

[mcgsxr]

Sorry about missing the active part earlier - I noticed that one of the criterion for picking the drivers used, was efficiency, so that had me thinking passive 3 way.

As for the posted design, I would continue the baffle under the woofer, to the floor.  It will help bass response I suspect.

Nice otherwise, though if being cheeky, I would also leverage the angled baffle you initially showed, I have found success in the past with angled OB's. 

[454Casull]

As I said before, I want each driver to be operating under its dipole peak. A perfectly circular baffle gives me a peak at ~700 Hz. Extending the baffle will only drop that frequency.

[454Casull]

What, exactly, do those wings accomplish?

Nothing. They're simply what gets cut out of the bottom of the front baffle. Well, that and they add a lot of strength.

[454Casull]

I think I've come across a better design. Details to come tomorrow.

EDIT: My God, it's amazing. Hint: one of the items required for the design will also be useful for professional-level kitchen knife sharpening.

EDIT2: Does anybody know why Linkwitz prefers a shelving low-pass as opposed to a regular low-pass for OB rolloff EQ?

EDIT3: If somebody can get me the 3D CAD files for the Alpha-6C and the Beta-12A-2, I'll knock out the design right away.

[454Casull]

The top part of the new design

[454Casull]

Here we are...

http://www.linkwitzlab.com/frontiers_2.htm#N

All fastening hardware is 1/2-13 threaded rod with the exception of the tweeter pod fastener, which is a 3/8-16 socket head cap screw. Flat washers and lock washers are not shown.

The column is 1.5" outside square steel tube, wall thickness 0.065". The bottom will be welded to a flat base, probably 1/2" thick steel, with gussets (perhaps a long triangular brace as well) for strength and stiffness. The clamping bars and blocks can be aluminum, though I plan to put some rubber on the inside of the clamping bars to prevent damage to the magnet.

The same system can be used to mount the midbass. Depending on the size of the magnet on the Beta-12A-2, it might need to be upsized.

Mounted this way, the Alpha-6C without a baffle has a dipole peak of ~2 kHz, meaning that it can easily cross over to a 1" tweeter, which in this case happens to be the DQ25SC16-04 run dipole on the pod shown. The Beta-12A-2 has a dipole peak of ~850 Hz when run without a baffle, but I will add a small annular baffle to decrease the dipole loss at the low end.





Not only does this stupid-looking system allow for easy magnet mounting (the only type of mounting that it does, actually), it allows one to easily adjust the physical offset between drivers. Once I get everything set up, I'll simply slide the mount back and forth to find the position with the deepest null.

[454Casull]

Actually, I was planning to add a slotted plate underneath the tweeter so that it could be moved forward. Even with the mount all the way in, the physical offset between the front tweeter and the mid will be too great.

Too lazy to implement it for you guys, though.














Just kidding.





The two screws up top are 3/8-16x1" hex head cap screws. The socket head cap screw that previously went through the tweeter pod has been removed.

[454Casull]

Improved the cable routing in the tweeter pod.

[454Casull]

Improved pod mounting

[454Casull]

With midbass mount:



The lack of comments is quite discouraging. 

[mcgsxr]

I actually don't know what to say.  You have tremendous CAD skills, and the design is very cool.

I am still trying to grasp the reasoning for the design - not that I am questioning your reasoning, I am trying to get why you want to design it the way you are.

If you can bring this design to life, that is one cool effort!

[Rudolf]

The lack of comments is quite discouraging. 

Your CAD skills are discouraging too.

I like the no-baffle approach and the X-over points look well-thought-out.
Just one comment: The front-to-rear distance of the tweeters is part of their effective baffle width. You want them to be as close together as possible. Would you dare to cut-off part of the cooling fins to get the Vifas closer together? At least you should provide cut-outs for the tweeter pod fastening rod, so you don´t waste any space.
If the tweeter pod is made from any heat-conducting metal, you possibly could abandon the cooling fins altogether and mount the tweeters on some back plate inside the pod.

Rudolf

[454Casull]

I actually don't know what to say.  You have tremendous CAD skills, and the design is very cool.

I am still trying to grasp the reasoning for the design - not that I am questioning your reasoning, I am trying to get why you want to design it the way you are.

If you can bring this design to life, that is one cool effort!

Remember that a requirement for one of my goals (constant power output) is having the drivers operate under the dipole peak, at the very least, and preferably with as much "headroom" as possible. This means a minimum baffle size, especially since a 1"+6" which is otherwise a fantastic pairing will not match properly if baffled unless the XO is low (~1500 Hz, which I don't want to do). Larger mids will certainly not work. So I decided not to use any baffles at all and just run them as they are. EDIT: I may need to add small rectangular baffles to smooth the response above the dipole peak. Then I had to decide how to mount them, and I figured it would be simplest to friction-clamp the magnets. This system then has four side benefits:

• eliminating the baffle/basket/motor mass-spring system [SL] which produces resonances
• easy, reliable driver setup with cap screws and nuts
• extremely easy physical offset adjustment, which is better than electronic delay [JK] and in this case is simpler too
• simple driver swapping - e.g. replacing the Beta-12A-2 with a TD12S/TD15S (I wish)

The ease of construction just depends on how good you are with basic metal fabrication.

As for the tweeters, I wanted to run them back-to-back, of course, but I was afraid of the terrible diffraction peaks and nulls that happen with circular baffles, which are what you get with the face plates of the DQSC. The Neo3-PDR is a great driver but in this application has several caveats, and the DQSC is both cheap and good, so I decided to stay with it. What I did then was basically design a rabbeted spherical pod to contain the two tweeters. What you see here is the minimum size that will fit the face plate (it's actually the face plate, not the depth, that is the minimum size factor). The complexity of the pods will probably make me have them manufactured using SLS (a type of 3D printing) or a similar method, probably out of brass or bronze.

The lack of comments is quite discouraging. 

Your CAD skills are discouraging too.

I like the no-baffle approach and the X-over points look well-thought-out.
Just one comment: The front-to-rear distance of the tweeters is part of their effective baffle width. You want them to be as close together as possible. Would you dare to cut-off part of the cooling fins to get the Vifas closer together? At least you should provide cut-outs for the tweeter pod fastening rod, so you don´t waste any space.
If the tweeter pod is made from any heat-conducting metal, you possibly could abandon the cooling fins altogether and mount the tweeters on some back plate inside the pod.

Rudolf

I hope that first comment was a compliment.

Yes, I had to play around with a lot of numbers. Putting everything into that spreadsheet was a huge benefit.

I understand that the spacing between the tweeters is part of D. I wouldn't mind cutting the fins, as the tweeters will not be receiving significant amounts of steady-state power, but as I mentioned above the limiting factor is the face plate. If I turn down the face plates I can shrink the pod up until the fins butt together, at which point I will have to shorten the fins as well. This is not out of the question and will certainly lead to better performance. I just need to figure out how much I can turn the face plates down before screwing the tweeters up; the screw holes are something I'd like to keep too, but I can simply glue them to the pod with a weak glue just the same.

What exactly do you mean by "provide space for the fastening rod"? As far as I can tell there is no interference from having either the cap screw or the bar in place.

Thanks for the comments, all.

[454Casull]

With mid: (model taken from http://www.swxdesign.com/content/view/165/27/)

[Bob in St. Louis]

Ohhh, it's a magnet mount!! 
I gotcha now. I was wondering what all the extra "stuff" on the structure was for.
Silly me. I should have recognized a magnet mount. 

Cool design! Can't wait to see it in a tangible form. That's going to be pretty neat.

Bob