FTRS MTM and Bass Bin

rjj45

This design/build has been going on over 6 months, but I thought I'd share it here in pieces.
I bought a pair of the TangBand W4-1320SIF mid-woofers a long time ago, and enjoyed them near field on my desktop, but always saved them for something better. Early this year, another pair came up on the forums, and the game was on. Decided to use a Fountek NeoCD 3.0 from a Mini-Statement tear down, and the top was set for a reasonably high-end MTM. 
Name = Fountek Tangband RS = FTRS speaker
Decided to make a kerfed baffle wide radius with hardwood tops and bottoms.
Kerfing was really not that hard. I think I will do it again. Solid SQ benefits.
test kerfed piece - 18mm baltic birch


Baffle and 2 sides piece


Glue up- I locked the baffle down to the table, used 90 degree pieces clamped to the sides.
Came out pretty good, but not perfect. I did like using Gorilla Glue to fill the kerfs. 
Very, very strong once it had dried.



The sides came out a little uneven. 
After a some thought, I realized that I could clamp them to a 90 degree edge and run them through the table saw.


worked perfectly!


more tomorrow.

dcibel

That looks pretty killer

rjj45

I'd had bought some 2 ply oak veneer for the project, but realized too late that I "might" have problems using it over a radius curve. My front baffle radius was about 2.5 inches, and it glued down, albeit with a lot of cracking noises. At this point, I was worried that I'd screwed the pooch.


Really, really  worried. Figured that worst case, I might have to veneer over the cracks and bumps.


Unbelievable, a bit of 120 sanding followed by 220 cleaned it all up. Solid white oak top and bottom.




I love the look of fumed oak with Danish oil (Stickley finish). Put the boxes in a tube with some 5% ammonia for about 48 hours. Here is the result compared with "virgin" veneer.


The real magic happens when the oil goes on. Love the contrast between the veneer and the solid oak ends.



And in short order, started to work on the crossover.

6thplanet

Those look sweet 

PWRRYD

One of my favorite drivers (W4-1320SIF).

I like your older Fluke 80X0 series meter.

rjj45

Confession time. So tedious, but so necessary.
So I'm building this MTM with 2 W4-1320SIF mid woofers, and thinking that somehow I'm going to get accepable bass from them, even though the Fs=75.  After the boxes had been built (all the work detailed above), foro some reason I thought I'd check the Unibox model again, only come to find that some "doofus" had entered 6mm for Xmax! (obligatory WTF moment).  
Well, I'd gone to far to turn back now, so I needed a "bass bin" for the MTM "top".
I had Qty 6 of the RS180s on the shelf, so that was decided. I'm a new fan of MLTLs, so I designed a folded MLTL with 2x of the RS180s on each side. When I got the specifics laid out in HornResp, I realized that only 1 internal divider could change the basic BR box to a MLTL.

 

which became the MLTL


Unibox showed goodness in the basic design:


as did HornResp for the same volume and tuning.



Since I had 2 almost identical boxes, I ran a bunch of measurements.
Not a whole lot of difference. The MLTL wavelet seemed a bit cleaner 200-400Hz (probably due to the stuffing), and possibly extended a tad lower than the BR. 

Just for grins, here is the close miked MLTL port FR comparing an unstuffed TL line (in red) versus a stuffed line (in blue)

rjj45

I"ll wrap this up - couple of things of note.

It took me long time to dial in the crossover, got many versions that sounded good, and measured pretty good, but had some hills or valleys that I didn't like. All of the versions really sounded pretty good, but nothing to get excited about. I DID find that if I didn't rolloff the W4-1320s around 200Hz or so, they would start sounding stressed when I cranked it a bit. And I found that I liked the Fountek NeoCD 3.0 a lot when I had a proper 3rd order HP on it.

Something funny happened one day - the speaker "came alive" - after months of listening, suddenly I really liked them, and the sound image was very wide and deep. Of course, I kept that version!
Seems like the polars looked very good also, probably due to the large radius baffle.

Bottom line, these are my first real reference speakers from scratch.
Smooth and strong from deep bass to treble. They handle all my stress tracks without any artifacts. <grin>

FR (does not include bass bin)



Distortion




Polar

PWRRYD

Wow those are too cool!  I am using my single pair of W4-1320SIF in place of the PLA200 planars.  Similar look to your project with oak veneer and mahogany.

I hope I get to hear your project sometime.  They look fantastic and I know how good those TB W4's sound!

S7910

Those look awesome!

dcibel

Great looking speakers, I'm surprised as well that you didn't sand all the way through to remove those crack lines, but they turned out great.

With the polar response, if you really want to show off how well those curves perform, load all your responses as you normally do, but then select the "normalize" option, then select the on axis response file to normalize. This will remove the frequency response from the equation, showing only the relative difference between the on and off-axis measurements. Purely polar response without any influence from the "natural" frequency response.

rjj45

dcibel said:

Great looking speakers, I'm surprised as well that you didn't sand all the way through to remove those crack lines, but they turned out great.

With the polar response, if you really want to show off how well those curves perform, load all your responses as you normally do, but then select the "normalize" option, then select the on axis response file to normalize. This will remove the frequency response from the equation, showing only the relative difference between the on and off-axis measurements. Purely polar response without any influence from the "natural" frequency response.

Thanks for the polar normalize tip....
Yikes! Do I have a problem? What might I do to correct this? The spike is about 2.5Khz above the NeoCD Fc.

rjj45

Guys, 
Many thanks for the kind words!
Hope to meet some of y'all in person sometime.

re: W4-1320SIF
I tore down my pair of mini-statements because I couldn't stand that W4-1337 anymore.
I bought the 1320s years ago on a sale, and used them full range on my desktop for a year or so.
Frankly surprised how clean, powerful and detailed they can be when used with highish HP.

I didn't want to admit it, but I didn't like the NeoCD in the mini-statements either, but didn't really know why.
 I found that using a proper 3rd order crossover on the NeoCD really opened it up. It had been sounding stressed
in the min-statements. As usual, you don't know how a driver really sounds until you play with crossovers for a while.

At several points in the crossover process, I was using several extra components on each crossover, but things came together and I was able to achieve minimal component counts and hit the responses I needed.




-Don

Nicholas_23

Wow , beautiful . Love these. Im hoping to make it to Iowa DIY this year. It would be nice to see these in person . hint hint nudge nudge.

rjj45

There is a slight chance I could make MAF in Dayton - about 10 hours on the road for me.
Iowa is considerably farther, sorry...

dcibel

rjj45 said:

dcibel said:

Great looking speakers, I'm surprised as well that you didn't sand all the way through to remove those crack lines, but they turned out great.

With the polar response, if you really want to show off how well those curves perform, load all your responses as you normally do, but then select the "normalize" option, then select the on axis response file to normalize. This will remove the frequency response from the equation, showing only the relative difference between the on and off-axis measurements. Purely polar response without any influence from the "natural" frequency response.

Thanks for the polar normalize tip....
Yikes! Do I have a problem? What might I do to correct this? The spike is about 2.5Khz above the NeoCD Fc.

If you compare your normalized polar response to the other plot you created above, you can see that the issue here is merely an on-axis dip in the response, that doesn't occur off-axis. It is promenent in the normalized plot because you have normalized everything to the on-axis response, so you see a gain in the off-axis responses here.

Anyway...it's very narrow band, I wouldn't worry about it. If you like the sound as-is, then there's nothing to fix

rjj45

Wow. it all makes sense (except that hole in the on axis response) The sound overall is super smooth, and does not give any hint of harshness or bloom on an cut I've played.  Thanks so much for the technical support. Another tool for my build/tune process.


Yeah, I gotta assume that this was some momentary glitch in Omnimic, since it doesn't appear in any other FR measurements. But there it is in the 0 degree polar FR. 

4thtry

Very nice speaker project.  That does not look like a very big ripple (6.5kHz).  Try panning your head from left to right as you listen to  your favorite stereo track.   See if you can hear a change in the stereo image as you move from a position where one speaker is directly on axis while the other speaker is slightly off-axis.