Volare

Add Headings and they will appear in your table of contents.

When my younger son started college, he had a nice location for a flat screen TV and pair of speakers in his dorm room except that it is a relatively shallow ledge along one side of the room. I built the "Hafnium" speakers for him a number of years ago, but those are over 10" deep and rear ported which just didn't work for that situation situation. I sent him with the Rhodiums, but there are shelves full of drivers, so there was no reason this problem couldn't be solved with something a little more substantial.


One thought I had was something with a similar form factor to the Advent Baby, which was about 16" tall, 11" wide, and 6" deep. Given that these are going on either side of a TV, I also considered something taller and thinner and taught myself enough rudimentary sketch up skills to do a visual comparison.

We both preferred the tall, skinny form factor. The next step was to pick some drivers. I have a pair of SB Acoustics SB13PFC25-4 woofers that are happy in about 7 liters tuned to 60 Hz. For tweeters, we have a fair amount of baffle area to work with, so it seemed that something with a full-sized flange would work well aesthetically. I thought about Peerless/Vifa/Tymphany XT25 or DX25 variants, but I also had a pair of Seas 29TFF/W sitting on the shelf as a 5 year old door prize. I hadn't seen them used much, so thought this would be a good opportunity.

True to the initial Advent Baby, the concept includes black baffle, sides, and rear with woodgrain top and bottom panels. Given the ledge these needed to occupy, a rear port was probably not an option - so the design includes a front port below the woofer.  The required enclosure tuning for the SB13PFC25-4 can be achieved with the fairly ubiquitous Goldwood PT-F415 1 1/2" by 4" flared port sold by many vendors - so that is what was used here.  The cabinet was constructed using 3/4" MDF with a 7" baffle width, 6" cabinet depth and 18 1/2" height.  End caps were made using 3/8" thick bamboo cutting boards and cut to 7" x 6" rectangles, resulting in a final overall height of 19 1/4".  The cabinet was finished with 3/8" roundovers on the 4 vertical surfaces.  The sides were finished using an old standby, black Duratex.  Once assembled, my next step was to measure the drivers in the box.



My tweeter measurements agreed pretty well with the manufacturer's quasi-infinite baffle measurements except for an on-axis dip at 3 kHz and a little more output at 60 degrees off-axis above 5 kHz.  Both differences are probably baffle-related.

For the woofer, baffle step seems to have flattened things out between 350 and 1000 Hz versus the manufacturer's infinite baffle measurement, but now there is a pretty big hump around 1.3 kHz in the Volare woofer response. The fact that these were designed to be placed near a rear wall on a ledge led me to do a front port, so I wonder if that is contributing to the hump. That required more investigation.

Tweeter impedance is pretty close to the manufacturer's curve. Fs is a little higher, as is the minimum impedance above Fs. The measured Fs peak is lower than spec.

The in-box woofer impedance shows a tuning frequency around the target 60 Hz using the PE 1.5" by 4" flared port. There's a little blip around 350 Hz that may require more investigation. The impedance minimum around 300 Hz and impedance at 20 kHz matches the manufacturer's curve quite well.

With the initial sanity check measurements done, I loaded those into Xsim and tried to come up with an initial crossover that had pretty flat on-axis response so I could get a feel for how the drivers would sound.  I came up with a second order electrical low-pass on the woofer with a notch to address the 7 kHz peaking and a third order electrical high-pass on the tweeter.  

The woofer circuit doesn't require a lot of extra inductance as the driver has some natural BSC built in (higher output at 200 Hz than 1 kHz on the "infinite baffle". It looks like sensitivity will wind up at 85 dB (2.83V at 1m), which is pretty good for a small two way. The SB13 is not a bass monster, but is both clear and non-fatiguing through the mids. The 29TFF/W has similar qualities.

My son and I thought this sounded pretty good, but it wasn't the final answer for a couple of reasons.

 Having established that the design had some promise, I went ahead and did polar measurements at tweeter height at 15 degree intervals from 0 to 180 degrees so that could be plugged into VituixCAD for a full horizontal polar simulation.  On the Power & DI chart, I'm targeting a slope of -0.9 dB/octave, which is shown by the pink line for the estimated in-room response, which is the orange curve.  This shows, in addition to the issues mentioned above, a little too much off-axis response between 2-5 kHz and ragged response on and off axis from 1-2 kHz.  The limited dispersion at 10 kHz and above is likely a byproduct of the large-ish 29mm dome, which we'll mostly have to live with.  With these measurements in place, I started to look for opportunities to improve on the initial crossover.

My next step was to add a resistor to the shunt leg of the tweeter network, run the optimizer included with VituixCAD, and make a few more tweaks to help smooth out the estimated room response. I found I could get everything lined up pretty well except for that nasty little peak around 1.3 kHz. Other than that, and the general narrowing of the tweeter axial response above 7 kHz, this is pretty ideal behavior.  The crossover point is now in the vicinity of 2.5 kHz and the minimum impedance is up to approximately 4 ohms, which will be friendlier to both the tweeter and any amplifier driving these.

This really made me want to try to attack the 1.3 kHz peak. A notch filter would be an option, but I'm not sure this peak would be reproducible if someone were to build this in, for example, the Denovo 0.23 cu. ft. flat pack - so my next step will be to look at the port behavior a little more closely to see if there is a way to help there.

I took a nearfield measurement of the port output. There are some pretty strong resonances coming out at 1.2 and 1.5 kHz. Here it is plotted along with the woofer response.  It seemed likely that at least one of those peaks is a pipe resonance from the port. I was aware of a few techniques to deal with this.


Bill S. and Jon Hollander have designed traps teed off from the port to suppress resonances.

I have read a few accounts of using a perforated port to dissipate resonances, with foam wrapped around the port.

I have also read about KEF using a "soft" port in order to damp resonances.

I'm already using the Parts Express flared port that is press-in, so I think it will be difficult to integrate a tee into those. I'm curious about the perforated port as it would not take up additional space. Also, if I screw that up too badly, I can still cut off the tube from the flare and fashion a soft port out of foam sheets, or a paper towel roll (which I know from experience works with the PE port I have).

My first attempt was to drill four 1/8" holes at approximately halfway down the port (which is supposed to act on the amplitude peak of primary standing wave). That didn't do as much as I would have liked, so I doubled the number of holes and also added eight holes at about halfway between the first set of holes and the exit (to deal with the next standing wave). It did something...  The yellow curve is the unmodified port, the blue curve has 4 holes near the midpoint of the port, and the teal curve has 8 holes near the midpoint and another 8 holes around the quarter length point.


Here is a picture of the configuration that produced the teal curve.  The paper from Philips that I read about the technique wrapped foam around the outside of the port after the perforations were made. I had a thin sheet of some wool-like damping material, so I tried that for the measurements above. It seemed to do the trick and meet the intent.

To make it easier to take the port in and out for these experiments, I put the terminal cup cutout right behind the port exit. I was a little worried about that creating some issues, so I made a little damping pad for it out of denim insulation and fiberglass window screen...



Let's see if this has any effect on the measured woofer response.

Color coding is the same as the port graph. Yellow is port with no modification. Red has 4 holes at the midpoint. Teal has 8 holes at the midpoint and 8 holes at the quarter point. This might be as good as I can do as the manufacturer's curve (shown in my second post in this thread) has a little peak (looks like a couple of dB) in the same frequency range.

Below are measurements taken on the completed speaker with the port and terminal cup modifications.  The peak at 1.3 kHz associated with the front port resonance is no longer a major issue in the frequency response measurement, which was taken at 0.5 meter with a 5 msec gate time.  Tuning frequency went up 1-2 Hz - not too bad. There is a small blip around 350 Hz, which I still assume is a cabinet mode.  Estimated in-room response and power response fall off nice and smoothly, as the sim indicated. Dispersion narrows a little more than ideal at high frequencies, but, as mentioned earlier it is a 29mm dome.

All in all, I am happy with how these turned out. They throw a wide soundstage and did well head-to-head against the Rhodiums, which I consider to be a pretty solid test.  Here is a final picture in something approximating their intended application.