P8 - Ballon MG18SK / XT25 MMT

Ballon: [ba-LAWN] Bounce. Ballon is the light, elastic quality in jumping in which the dancer bounds up from the floor, pauses a moment in the air and descends lightly and softly. (www.abt.org)

MG18 - XT25 Ballon   Ballon construction

This was my first MMT project. I chose the Vifa MG18SK woofers because they were on sale and because I am convinced larger woofers have richer mids. I also want to build a couple of projects with currently available drivers. The box has two chambers which put the woofers into different sized boxes. I made a false back in the upper chamber that drops down, making the boxes closer in volume. This false back also added a lot of stiffness. The upper box is around 27 liters and the lower box around 33 liters. You can see the layout in the darker picture above. I tuned the ports to the same frequency using the PE variable length port. I found these little stars some time ago and they are great.

One port is in front and one in back. Note the amount of bracing. The more speakers I build, the more I think bracing is very important. Even more important than dead accurate box size. You can be off a couple of liters without problem. But a box vibration will corrupt the sound. I am also convinced tuning a port on the higher side provides much better lows than trying to go low. These are tuned to about 38Hz. Before I fooled with them, they were around 30Hz and the bass was a little muddy.

I really learned a lot on these:

1. I found an old rule in speaker building. If you put two like drivers next to each other, they will invariably have an interaction based on the center-to-center distance between the drivers. The rule is that this interaction will be in the least desirable place for the given drivers. In this case, there is a suck out that starts about 1900. By 2800, there is a hole that looks like a reverse null. So the XO cannot cross above that frequency and needs to be steep enough to kill any rise after the suck out. If the drivers were smaller the needed XO point would be higher. This is not a problem per se, but a note to understand the issue. I would guess an MTM have less suck out in exchange for other problems. The suck out or interaction would be at a lower frequency because the drivers are farther apart, but you add greater lobing. So you exchange lobing for suck out. DeApolitto covers this in a paragraph in the Cookbook.

2. I really like working with these Vifa MG drivers. I tried maybe five crossover designs to see which I liked best and all were fairly easy and provided good results. I tried series, 2.5-way and parallel designs with variations. I was all set to use the series design so a HT amp could power them easily, but realized in process that I was attenuating the tweeter down maybe 8db. That hinted the woofers in parallel were pretty inefficient, probably in the 82-83 range. So moving to the series configuration brought the efficiency up to the point the tweeter does not need a series resistor at all and the parallel resistor is mild. Also, if you make them parallel, you only need ¼ the inductor of a series configuration. The down side is they are 4 Ohm.

3. I tried the 2.5 way configuration everyone seems to think is so good. While I could get decent measurements from the configuration, I did not like the way the system sounded. I think because the upper and lower woofers had such different XOs, they were not moving in parallel. Not just rolling off the lower woofer SPL, but changing the sound between the upper and lower. Perhaps this was some sort of XO generated phase shift. The series configuration sounded the cleanest, then the 2-way parallel, then the 2.5-way. The series and parallel were very close if not identical.

4. The XT25 tweeter is very nice to work with and sounds great. There are two things to know about this tweeter. First, I don't think it sounds best directly on-axis. Anywhere a little off is very good. There is like a narrow beam or something that comes straight out. Almost nobody listens directly on axis so not a big issue. The second thing is that it needs a steep roll off. My first XO had a second order acoustic on the tweeter and I kept hearing some distortion below the XO. At first, I thought it was coming from the woofers. I adjusted LspCAD to record a lower SPL and found the tweeter had a hump below 1000 Hz about 20db down, which had been just off the scale. I moved to a third order acoustic and the bad sound went away. Search the MAD forum for discussions on the humps below 1000Hz with this tweeter.

5. This speaker has a fair amount of lobing (a change in the sound on the vertical axis, in this case, above the tweeter). I pulled another non-MMT speaker and found the lobing was more than I thought but not as much as the MMT. This is supported on the sound wave renderings on the main index on this site. So vertical ear placement is important on these speakers. A work around for this would be to build a sloped baffle. Once the lobing happens it is not that bad. So using a sloped baffle should allow some relief but would require a different XO. I suspect I will get comments from the golden ears at the DIY events about a dip at XO, but will try to get them at the right height.

All this supports one of my main theories in this hobby: you can make a speaker measure well that does not sound good. With all the low cost measuring equipment on the market, we all are doing a lot better, but we still need to rely on our ears. Also, once again I found a flat MLS did not sound as good as a one slightly tilted downward. This could easily be an anomaly in my testing rig, but my graph shows a 2db crater from about 1000-6000 Hz. I would drop the tweeter lower (and will after you all comment it is too high) but am avoiding a series resistor.

On a side note, I took the unit over to Dennis's with my first XO to get a comparative reading. This was one of the configurations where I measured a fairly flat MLS (with a 2db drop at XO) but did not like the sound. I wanted to see what Dennis got. We hooked it up in his living room and, after three seconds, from about 200 degrees off axis and 10 feet away, Dennis says, "you have a slight drop around XO". We then took the unit upstairs and he was dead on with the MLS. How does this guy do that? Drives me crazy. The guy is older than me and should be deaf by now.

Ballon MLS

Ballon bass network

ballon trebble network