This project got its start because we wanted to enjoy outdoor music at home during the pandemic. I needed to think about what the requirements would be for an outdoor speaker in our setting in the US Upper Midwest. This is what I came up with:
Sealed - keep dirt and critters out.
Small footprint - the deck isn't that big and now has more furniture.
Inexpensive - it can't be a life changing loss when the next Tornado, Derecho, crazed pet, or over-served houseguest takes out one or both speakers.
Enclosure Design Concept:
My initial thought was to use a 5" x 5" composite fence post as the primary enclosure, but as I started to think about how I wanted to assemble everything, I decided it would be easier to just use some of the Birch plywood I already had languishing in the garage. I had a 5' by 30" sheet of 1/2" material, so I cut eight 5" x 30" strips, giving an enclosure 5" x 6" x 31" after adding end caps (about 9.8 liters).
In order to tolerate the weather, my finishing approach was to put a couple coats of deck stain on the sides and contrasting black stain on the endcaps. I then covered all panels with a couple coats of marine polyurethane. The bottom end cap is 7" x 8" to allow for more secure mounting.
Driver Selection:
I was looking for either an aluminum or polypropylene cone woofer with rubber surround and with an OD of 5" or less and low cost (approx. $25 or less). The HiVi M4N has a 1 piece aluminum cone with rubber surround, works well sealed, and is $13 each. It doesn't have huge excursion, so we'll need two. For a tweeter, I have some Peerless DX20BF00-08's that came with a Solen sale order a while ago. I'm not sure if the fabric dome will handle the elements, but they are small and cheap with the added bonus of slight horn loading, which will help keep them out of harms way.
Crossover Design (including measurements and simulation):
I measured all drivers on the tweeter axis at 1 meter using 2 channel measurements to capture absolute phase. The impedance measurements for the woofers were a little messed up, as the cabinet wasn't stuffed and wasn't well sealed due to the wires needing a place to exit. These drivers fell pretty easily into a crossover point of about 2.75 kHz, which should be reasonable for both drivers. Here is where it wound up.
The 0.3 uF cap in parallel with the woofer inductor creates a notch around 9 kHz, which helps some with the M4N breakup. These are gated measurements with a pretty short window (about 3.5 ms), so anything below about 300 Hz probably isn't very accurate. Above that, the response looks to be within about +/- 2.5 kHz. Initial listening went well enough to justify assembling everything and doing final measurements.
Enclosure Design Details:
If one were to model a pair of M4N's in about 9.5 liters sealed, there would be something like a 3 dB peak. Clearly, this will need some significant stuffing or an aperiodic vent, which isn't really what we're looking for in an outdoor speaker. I took a strip of ultratouch R13 denim insulation, pulled it apart in roughly two half thickness strips, and attached that to the interior rear wall. I then stuffed the balance with several handfuls of morning glory fiber fill.
When these were assembled, the first measurement I took was an impedance measurement to get some insight into the alignment.
The peak impedance around 20 ohms just above 2 kHz agrees well with the sim. The resonance between 200 and 300 Hz is gone, and we now have a single sealed box peak around 95 Hz (the smaller peak between 40 and 50 Hz is a measurement artifact I see in most of my other impedance sweeps and can be ignored). The 95 Hz peak is now pretty small. I am able to model a similar impedance peak with a Ql of 20 and Qa of 2 in WinISD. This gives the a Qtc around 0.89 and an F3 around 80 Hz. The peak is clearly under control.
Final Frequency Response Measurements:
The first measurements were the final on-axis measurement coupled with individual driver responses with the crossover applied. All of these were taken on the tweeter axis at 1 meter.
Key features of this are a 2.77 kHz crossover point indicated by the low point in the reverse null and 5 dB between the highest and lowest response point. The final response matched the sim pretty closely, which is always reassuring.
I then took a set of horizontal off axis responses.
I thought the most interesting observation was the off axis response starting to drop off above 1 kHz. I would not have expected that to start at such a low frequency for a 4.5" woofer. Coincidentally, there is a rise in the on axis response in the same area, so maybe that is a good thing...
There is a dip around the crossover point that gets a little more pronounced off-axis. Otherwise, things seem pretty well-behaved.
Home made Spinorama
If any of you follow the Audio Science Review forum, you may have noticed that I submitted a couple of my designs (the Helium and perfectionist mod of the HiVi DIY 3.1) for review over there. I picked those because enough of both have been built that they might be of wider interest. This was, in a way, a substitute for me for DIY events in 2020. It was also an opportunity to try to up my design and measurement game. With that in mind, I started messing around with VituixCAD. I did a more extensive set of horizontal off axis measurements than I usually do to support this to produce an approximate set of spinorama data.
Here are some results.
The power response is nearly ideal up to the crossover point, there is a dip in the crossover region, then the small horn/waveguide leads to a little too much output from 3-6 kHz and too little above 10 kHz. This didn't stand out as a large issue in listening tests.
The polar response clearly shows the dip around the crossover point. It would be improved by a lower crossover point, but that wouldn't work well for the DX-20 from a distortion/power handling perspective. The DIY 3.1 had a similar issue, but was did very well in distortion testing. I think that is a reasonable trade off here as well.
Observations and Conclusions:
All-in-all my wife and I have enjoyed listening to these. There could be more bass, but they can play pretty loud without falling apart and the driver cost is about the same as what is in the C-Notes.
I did bring these to the 2021 DIY Iowa gathering hosted by Midwest Audio Club in Ankeny Iowa. By that time, the speakers had been outside for about a year and the finish was starting to deteriorate considerably. The speakers did hold their own in that setting among a number of larger, more expensive designs.
I think it is possible that, in addition to outdoor use, these would be applicable to home theater. The height is very similar to that of a 55" to 60" HDTV and they are only 5" wide and 6" deep, which might allow on wall or near wall mounting in addition to flanking the screen on a shallow surface. The 80 Hz sealed F3 (12 dB/octave) would integrate well with most subwoofers.