MLTL using TC9FD and Pro audio 8?

[Jerms]

I've been playing with Transmission Line software for a while now. And also Sketchup for some years.

I have a pair of 8s, from Parts Express, sort of an impulse buy as they were sold as woofers....

AT-8030 8" Treated Paper Cone Woofer with Rubber Surround 4 Ohm

Also have a pair of Vifa TC9FD18-08 from a defunct attempt at kitchen speakers.

Vifas are 84-85 sensitivity, the "8s" are apparently 93.

Creating an MLTL in TL software has been interesting.
170CM line length. Driver at 60-66CM, port at 135CM

This is where I've got to.

And in Sketchup...

Going to be using a PC to supply signal to 4 mono TPA3118s.

 

[Jerms]

I haven't shown the modelled the Vifa, but it's going in a 1/2 -1 litre enclosure of "pipe". Crossover frequency is somewhere in the 300-400hz range.

Port is also currently a pipe... What can I say, 4" pipe is cheap.....
Nothing is set in stone yet. I may mount the Vifa on top in a triangle. Also mooting a slot port rather than the pipe. Pipe is going to be easier to tune though, I think.

 

[xrk971]

This can work well as a MLTL FAST with 8in woofers. The MLTL you have shown can probably be made to have a smoother response with a bit less volume to reduce that bass peak bump at 45Hz. Try placing the 8in driver about 1/3 distance from the closed end and put the vent a little closer to the "open" end maybe 5 in or so. The TC9 can go in a Dagger (a 3 sided tall pyramid stuffed with wool or fiberglass). A Dagger has no parallel walls and sounds less boxy with no resonances. About 1.5L would be a better volume to let its soft suspension breathe more and reach the lower 150Hz region smoothly to do a XO at say, 350Hz to 500Hz. The 93dB senstitivity sounds exaggerated for a buyout woofer. Probably closer to 91dB or 90dB, in which case well matched after 4-5dB baffle step loss. Use a 1mH and 2R to 4R shelving filter on TC9FD to balance the response above 1kHz with the 300Hz to 1kHz output.

 

[Jerms]

I can shrink the volume but to keep bass sensitivity above 90, I lose 5hz+ on the bottom end.... I can extend the line to get some low end back but I still get a "scoop" in the middle. Maybe I'm aiming too low?

Yep, the woofer models as a 90-91 sensitivity driver. So unless the T&S are way out....

I may see if I can build a long tapered dagger into the top of the line. One less wall to build for it and the initial corner is dead space anyway. Though to get the voice coils lined up I may end up burying it deeper. That was one reason low mounting the TC9 seemed a good idea.

 

[Jerms]

A slight update on this....

It's evolved a little....

- TC9 will sit in 120mm diameter PVC pipe of about 1.5L.
- Rear slot port used.
- Cabinet cut from 25mm MDF. Most parts cut by B&Q for free. Router table used for exact (!?!) thicknessing.
- 8" to run from around 40hz to above 300hz.
- Baffle step F3 is 430hz.

As implied, Parts have mostly been cut and assigned to spkr 1 or 2. Lengths need a little trimming and I need a hell of a lot more clamps and glue.

J.

 

[Jerms]

I tried the fit of the PVC tube today... 120x350mm from Screwfix.com. Wall thickness is about 1mm. Cut it using the router fence as a guide on the router/saw table.

Front baffle and separator have close tolerance recesses for mounting the pipe. It feels VERY rigid when held between those two panels. It'll have an amount of flashing tape on it to reduce panel vibrations and about 20g of wool inside to damp the driver.

 

[Jerms]

Would be an edit but too late...
120mm diameter (127 actually) by 145mm PVC pipe for the Vifa equates to 1.5-1.6L for it to live in. +/- a bit.

 

[Jerms]

More piccies.... Currently waiting for parts before I can do more.

Mounting the PVC pipe. The "dimple" in the second pic has been removed.

[IMGDEAD]https://dl.dropboxusercontent.com/s/1s9ab217t49z6j8/2016-07-23%2016.56.14.jpg?dl=0[/IMGDEAD]

[IMGDEAD]https://dl.dropboxusercontent.com/s/smofkfvyop2ldu1/2016-07-23%2016.56.30.jpg?dl=0[/IMGDEAD]

Mock up of the cabinet - couldn't find a banana so chilli sauce for scale. These things already weigh significantly, without the 8s mounted.
[IMGDEAD]https://dl.dropboxusercontent.com/s/ez84xan7gpeggc7/2016-07-25%2011.29.52.jpg?dl=0[/IMGDEAD]

 

[Mayuri]

I made a similar project last year, and for me it worked really well to put the crossover frequency to the baffle step. I also rounded the baffle edges with a router. End result was a very smooth crossover region.

 

[Jerms]

That's the idea..... Though the MLTL drops woofer response into a hole at 300+.... Making an apparent transition around 430 should be a challenge.

 

[Jerms]

I've been listening to the one prototype, clamped up for a quick listen this morning.

Running REW's spectrum analyser on the laptop playing various source material from the NAS through an RPi3.... It certainly manages to find 40hz on occasion. .

The time consuming aspect was configuring ALSA (again!) on an RPi. It (ALSA) didn't like me connecting a ROUTE plugin directly to a LADSPA plugin. So I have an intermediate PLUG plugin that on the face of it does nothing..... But lets it work.

I currently have one channel playing high and one low. Using ACDf V2 written by Charlie Laub to provide low and high level 2nd order crossovers at 430hz.

J.

 

[Jerms]

<sigh>
After a bit of interneting, I discover the "Edge" baffle diffraction calculator. Plug in the numbers for my Woden Redeyes and whoa! The suggested filter frequencies are far lower than ones I got from a simple online calculator.

On the FASt speakers with the 8s, it turns out my apparent 430hz baffle step should read more like a 175-200hz baffle step - LOL.

About to glue up the second sides of both cabinets in the next day or so. Will have a play then if parts for my amps arrive from HK.

Ha! Once I re-fettle ALSA on the Pi3 that is..... And get some long RCA cables to feed the speakers.

J.

 

[Jerms]

During playing with ALSA and trying to setup the filters for these another thought crossed my mind.

One way of eliminating BSC on a FAST system is to attempt what I'm trying.... Have a rather more efficient woofer than the FR driver. In this case, circa 90dBA vs 84-85dBA.

However, selecting two drivers of around the same sensitivity but having the FR driver have twice the impedance of the woofer may go some way to addressing the same issue.

J.

 

[Jerms]

I have both of these up and running now. I have a problem. They're too d*mned loud.

Playing through CM6206LX 5.1 USB sound card into 4xTPA3118s at 19V I get listen-able volume levels with the ALSA volume control sliders set at zero!

So I'm using ALSA "ttable" both to separate out my channels and now, to massively attenuate them.