A little over 2 years ago I was checking out my local Goodwill for any miscellaneous stereo/electronics/speakers (as one does on a Saturday morning) when I came across a pretty interesting find – 8 small carboard boxes with a familiar logo printed on the side – it said Dayton Audio. A closer look revealed 2 pairs of Dayton Audio Signature Series DA175 7″ aluminum cone woofers and 2 pairs of ND90 aluminum dish cone full-range midwoofers. I couldn’t believe what I was seeing. The boxes were opened but the drivers appeared to be brand new. And they were priced at only $6.99 and $3.99 each respectively. I immediately grabbed a cart and added the 8 boxes to it and headed to the checkout absolutely giddy with my find. For a total of $44 I basically had in my possession over $200 worth of brand new Dayton Audio drivers, for which at the time had no idea what I might do with them, but wasn’t about to pass up the opportunity to potentially make something great and do it for insanely cheap.
Fast forward 2 years and I have officially completed the Goodwill-inspired 4-way tower speakers that now adorn my small garage home theater. For nearly 2 years I tossed around almost every possible speaker combination I could come up with using these 8 drivers. Initially I was thinking of building two sperate types of speakers, like a Bluetooth speaker with the ND90s and maybe Zaph’s Bargain Aluminum MTM with the DA175s. Years ago I built a pair of speakers for my friend’s home theater using the DA175s and they were great speakers. But the more I thought about what do build with them, I just never could get over the fact these speakers need a pretty decent size box if you go the ported route. Going to a sealed box to remedy the size issue only brings about a major excursion issue above about 7 watts. So building a nice smaller 2-way or MTM just didn’t get me all that jazzed. And as far as the ND90s, I’m not too big on Bluetooth-based portable speaker systems in general. Parts Express sells some decent BT amp kits but I’ve got so many BT speakers lying around that just don’t really get used at all, so again I wasn’t super excited about that option either. So the speakers sat, month after month, with no plans in sight to every build them into the sweet speakers they deserved to be.
That’s when it sort of hit me one day to combine both sets of drivers into just one speaker design. It was an odd combo, the 7″ DA175 woofers and the 3.5″ ND90 full-range drivers, I thought there’s nothing I could possibly throw together where these two speakers would sound good together. Right? But I drafted up some concepts, taking into consideration the bass limitations of the DA175s and the lower efficiency of the ND90s and the upper frequency limitations (despite claiming to be full-range), I modeled up an WMTMW setup with a soft dome tweeter that I thought looked pretty good. I figured the ND90s would play great in just the midrange section and I could cross those to a nice soft dome in a 3-way setup. Add a sub and these could play great Left/Right duty in a home theater, with the receiver set to SMALL for the fronts. A potential design was coming together here, but it was still missing something.
Slowly that idea morphed into me deciding I really wanted a full-range tower speaker, something I could use in a 2-channel stereo setup without a sub. That’s when I added a pair of 8″ woofers to the design, making it effectively a 4-way speaker in a SSWMTMW (sub, sub, woofer, mid, tweeter, mid, woofer) configuration. After modeling up a bunch of different 8″ drivers from Parts Express, I settled on the DSA215-8 aluminum cone Designer Series woofers to act as the “sub” woofer in this 4-way monster. I liked that they matched the look of the DA175 speakers, they were offered in just a standard 8-ohm version so I could wire them in parallel for a final 4-ohm load (more power, easier crossover), they weren’t expensive, keeping in line with this being a budget speaker build, and they modeled great in a tower-sized cabinet that was exactly about the size I was shooting for. I drafted up some of the artwork you see here which helped give me an idea of exactly what this concoction of a speaker would look like. And you know what? I thought it looked awesome! You don’t see many commercial speakers in a configuration like this, but I knew it had the potential to sound amazing. For the tweeter I picked out the ND20FB-4 Rear-Mount 3/4″ Soft Dome Neodymium Tweeter. This ended up being an amazing tweeter for these speakers and was also pretty inexpensive. In the end, I think spent more on new crossover parts than anything else in this build. I was getting low on all my resistor/cap/inductor values from my last few passive speaker projects, so I stocked up on just about every value I thought I might need to build the massive crossovers for this 4-way speaker.
With the basic speaker configuration finally settled on, I played around in Passive Crossover Designer 8.0 (using the stock .frd and .zma files) to get an idea of what in the world the crossover for this thing would look like. Starting with 2nd order filters for all drivers, I came away with some candidate crossover frequencies that appeared to work. All the speakers would need is some attenuation to match the levels of the 8″ woofers after baffle step losses. So at least on paper, I knew I could get the response essentially flat from ~35 Hz to 20 kHz and I wouldn’t need any exotic notch filtering or any other overly complicated filter networks. Each driver was operating in what I would call its “sweet spot”, well below the cone break-up region but also above the bass region where the smaller drivers didn’t play as well either. The ND90s operate happily up to 10 kHz so cutting them off at a couple of kHz allowed the tweeter to pick up this critical region alone. And cutting off the DA175’s below 200 Hz prevented them from ever experiencing any over excursion. Also, limiting them to about 900 Hz, since they are in an MWTWM configuration and are over a foot apart from either, helped to minimize comb filtering effects, allowing the vertical off axis response of this speaker to be quite good despite the WMTMW configuration. Keeping the ND90s as close to together as possible and with a reasonably low crossover point also helps with improving the vertical off-axis response. Once I got the cabinets built I spent a good deal of time measuring all the drivers and ultimately designing the crossover real-time with REW, though PCD at least got me started in the right direction, especially when it came to picking a reasonable selection of part values from Parts Express. Inductors are not cheap, so I only bought what I thought I might use, and even then, I ended up purchasing iron-core inductors just to keep the cost down. Also purchased simple electrolytic caps for anything above about 20 uF. At 20 uF and below, they are all Dayton Poly caps.
Speaking of measuring and designing the crossover, I’m getting a little ahead of myself on this writeup. DIY speaker building has so many fun phases to it and I honestly enjoy each one so much, as each part of the design and build process carries its own challenges and rewards. Starting of with the concept phase (one of my favorites) just picking drivers, drawing up some pictures, pricing out options, I spend hours on Parts Express’s website downloading datasheets, running enclosure response simulations, trying ported, sealed, passive radiators, with EQ, without EQ, I mean I can do this for weeks, months, even years when I’m designing a new speaker. And it never gets old. It’s kind of funny because I go through my DIY Speaker folder on my computer and there’s at least 20 unfinished, un-built, half-baked speaker ideas that just never made it for every one speaker project that actually sees the light of day. If I had endless resources, money and time, I could easily build a new speaker every month and I would never tire of it. For now, I’m basically limited to about one speaker project every year, or if I’m lucky, I might squeak in two in a year, or more if it’s just simple “upgrade” project, like I’ve done a few of recently. Then there’s building the cabinet, cutting the wood, gluing it up, sanding it down, using every tool I own and spending weeks in the garage just making these bare wooden boxes. I’ve gotten more creative with some of my speakers and have been adding some real-wood veneers, or rather, plywood veneers like Birch. Wrapping plain MDF in some real wood suddenly makes the hobby go from what I did in High School to something a little more meaningful. Getting to use my flush trim bit on the router to get a perfect seam between the real wood and the MDF just feels so good. I spend a lot of time sanding and routering once I get the boxes built in preparation for my least favorite part of every speaker I’ve ever built – painting and staining. I settled on a process that works for me, which is spray painting the painted parts with a regular flat spray paint and then finishing the speaker with a Satin or semi-gloss polyacrylic. This water-based clear coat is the most forgiving of imperfections and leaves very few lines, but still has to be done right, or it can look really bad. I’ve gotten better over the years, but am by no means an expert. This goes well over an oil-based stain which is what I usually do if I’ve added a real-wood plywood to the sides. The stain goes on fine and is usually pretty forgiving of bad technique as long as you don’t put on too much, wipe the excess after each coat and do 3-4 coats until the desired darkness is reached. I’ve also found that using a pre-stain treatment on the wood before staining it works wonders for getting that even, clean, professional, stained look. It prevents that botchy look you can get with softer woods if not treated beforehand.
Which brings me to another favorite and basically the final part of the speaker building process, measuring and listening, and designing the crossovers. This also is one of the more frustrating parts of the design, because it’s so tedious. The endless tweaking of crossover values, pushing the response up and down, listening to music, gauging if it sounds better now with a 4.7 uF cap on the tweeter and a 4 ohm resistor or if it sounded better with the 5.6 uF cap and a 5.1 ohm resistor. But that’s just the start, with a 4-way speaker such as this, I had to decide where each speaker sounded best in terms of crossover point, summing with its neighbor, overall volume of each driver, figuring out which phase/polarity sounded best, especially when I started to venture into 18 dB/octave (3rd order) slopes, it was like, I can’t keep everything straight! Keeping good notes of every adjustment is key and of course, using REW with every measurement so that you an actually see how each crossover adjustment affects the response. I rely almost entirely on REW to get the response of each driver flat, and summing properly, before I start to listen to anything as part of the listening tests. I like to have the crossover basically where I want it visually before listening to anything. And then the listening part of things just allows me to dial in my own personal preferences such as the overall tonality of the speaker, which is influenced a lot by just how loud I allow each driver to play. Oddly enough with these speakers, those little ND90s were just screaming, despite them being the least efficient drivers of the group, I have a pretty decent sized resistor on them (3.3 ohms) just to keep them balanced between the tweeter and the DA175s. Anyway, last but not least in this magical DIY journey is the day the speakers are officially done and you get just sit back and put on some of your favorite music or watch a familiar scene in your favorite movie and just get to enjoy the creation that you and your little hands have put together. The months of design work, and then weekends in the garage cutting, gluing, painting and staining, to weeks behind a laptop screen staring at REW plots all day, to finally come out on the other side with something you can be truly proud of. And in some cases, something that is really one of a kind. And something you will enjoy for years, even decades to come, as I recently just turned the corner on 25 years of owning my Swans M3s, which I consider some of the nicest speakers I’ve ever built. I’m hoping that I get to enjoy these speakers for at least that long.
So back to the Goodwill Budget 4-way speaker project, I’ll sort of just go through really quickly some of the design choices and build process and crossover design for these speakers. And I realize that this speaker project was really just for me. I expect zero people to attempt to build this speaker. While I do provide enough design guide information that you could pretty much build any of the speakers on my site, I realize that this is truly a unique combination, inspired only by finding half of these drivers at my local Goodwill, and just wanting to make something of them. I never would have come up with this design on my own if I hadn’t walked into Goodwill that day. And to the person who donated these speakers probably a few days before, I imagine you possibly decided that whatever speaker project you had planned just wasn’t going to happen, so you dropped them off hoping, just maybe, they’d find a good home. I want you to know that they did find a good home and that I am enjoying them every chance I get to listen to them. And they sound amazing. Right now these are easily my favorite speakers of any speaker I’ve ever built. Anyway, thank you stranger, your donation did not go wasted. I did also buy those Goldwood GT-336 1″ titanium dome tweeters that were probably yours as well, I measured them, they were not great, so I didn’t use them. Don’t feel bad, they were a terrible sounding tweeter. Which makes me kind of wonder, what were you going to do with all those drivers anyway?
So the design starts off with a pair of SD215A-8 woofers wired in parallel and mounted in a 72-liter enclosure tuned to ~32 Hz with a pair of rear-mounted 3″ x 8″ PVC ports. They are crossed over with an 18 dB/octave passive crossover at around 200 Hz. The cabinet is decently braced and damped with 1″ acoustic foam and poly-fil stuffing. The two DA175 woofers are also wired in parallel and mounted in a separate sealed cabinet of approximately 9 liters. The enclosure is fully insulated with 1″ acoustic foam panels and pink fiberglass insulation as fill. The crossover is a bandpass 12 db/octave set at around 200 Hz and 900 Hz. The drivers have a series 1.5 ohm resistor to level match them with the woofers/mids/tweeter. This frequency range represents the lowest distortion region for this driver. After doing several full-range sweeps with REW, this driver does not play well below 100 Hz, with distortion rising to almost 20%. But between 200 Hz and 900 Hz, this driver has an excellent distortion curve and a very flat response and in general just seemed to be happiest in this fairly narrow region of the audio spectrum. Also, keeping their upper frequency limited to only 900 Hz prevents (or at least minimizes to some degree) comb filtering due to the drivers being almost 18″ apart and their centers. This is roughly equal to 1 wavelength at 900 Hz so moving off axis up or down will create gaps or nulls in the response (also called lobing). However, they should not be super significant and certainly could be worse if the cutoff frequency were any higher or the drivers were any further apart. The ND90’s start at 900 Hz and play up to 2.5 kHz also with 12 dB/octave filter slopes. From the distortion plots you can see that they really have a hard time below 200 Hz. They are okay from 900-2,500 Hz, but not super great as far as distortion goes. What they make up for in this region is just an awesome flat response. These suckers are super flat from 200-5,000 kHz and operate great in just the region I have them running in. They are mounted in another separate sealed enclosure of about 1.6 liters. They are filled with pink fiberglass insulation as well as 100% cotton batting material.
Of note, the ND90s and the ND20FB tweeter are mounted to a custom 3D-printed baffle that sits flush-mounted to the cabinet. My reason for doing this was I wanted to flexibility to design different midrange/tweeter combos if the ND90/ND20FB combo didn’t work out. I thought it would be kind of cool to have a part of the speaker have this unique interchangeable piece, that way I would have the option to swap out parts either as the design ages or if I just want to mix things up again down the road. The tweeter is flush mounted and sits directly between the two ND90s with almost zero clearance between them. This was in essence to get the ND90s as close to each other as possible, to reduce the effects of comb filtering. Since these drivers are crossed pretty high, up at 2.5 kHz, and are separated by several inches, there will be some comb filtering effects in this region. To minimize this, the drivers are placed as close together as possible, and then only allowed to operate up to a point that we can get this little 3/4″ tweeter to be happy and not a single Hz higher. I would have crossed these lower, if I had a more capable tweeter (distortion on the tweeter really rises below about 3 kHz), but I think we got away with it for now. The drivers are exactly 5.5 inches apart at their centers which corresponds to 1 wavelength at 2,500 Hz. Ideally this should be maybe 1/2 wavelength (<1/4 ideal) but 1 wavelength is still acceptable enough (see vertical off-axis response below). The tweeter is also a 12 dB/octave filter set at 2.5 kHz and has a 4 ohm series and 20 ohm parallel L-pad to not only balance the tweeter with the mids but to shape the top end of the tweeter just a touch (looked better than just using a series resistor alone). The tweeter then takes us all the way up to 22 kHz (my measurement cutoff) and is super flat up to about 12 kHz before there is a subtle rise in the response. I don't really consider this very audible, at least not by my ears. If anything, it may add a hint of airiness to the sound which I really don't mind. A notch filter could bring that flat pretty easily, but I don't mind it, and I like that the crossover is simple on the tweeter with only 4 elements. Besides, once you get about 10 degrees off axis, this higher-frequency rising response flattens out almost perfectly.
I spent hours in the garage measuring each driver with many combinations of parts until I came up with what you see here. I measured the drivers individually, I measured them combined with other drivers, I played around with crossover points, resistor values, slopes, I did everything I could with that crossover to give these speakers the most neutral, flattest, best tonality I could come up with. I listened to a bunch of music along the way to make sure the measurements matched what I was hearing. I did off-axis measurements both horizontally and vertically to make sure there were no weird dips or gaps. The final crossover here is basically the result of over 100 measurements of the drivers and the crossover being tweaked ever so slightly until they sounded just about perfect. You can see some of the iterations I went through for each driver in the plots. Though it’s hard to really tell the whole story. I don’t show all the summed combinations that show how each driver summed with its neighbor, which is also something I considered for each measurement. I wouldn’t pay too much attention to the response below 100 Hz. Room modes are playing a pretty big role in this region here, so I wouldn’t trust the graphs down that low. Listening to them at the listening position (about 2m) they sound very flat and do not have a domineering upper mid-bass region like the plots would suggest. This isn’t Klippel here, this is just an ECM8000 mic and REW, about as basic as it gets. But still good enough to get the job done. At this point in the build process the crossovers are just a pile of parts on the ground connected with dozens of alligator clips. To get a real sense of the way everything will sound, I have to solder the crossover together and mount everything down. So that was the next step and how the crossovers exist as of today. They are still outside the enclosures, but are fully soldered together and give me the option to listen for a few months to how they sound, I can make any adjustments still pretty easily, before committing them to their own little boards and mounting them inside the enclosures. After a few hours of some good jamming though, I don’t think I can make these speakers sound much better.
So a few other notes about some of this speaker design that are worth mentioning. I 3D printed a set of 3″ PVC end-caps which allowed me to make the holes for the 3″ PVC pipe as if I were going to permanently install them into the cabinet (with glue) but so that I wouldn’t have to. These end-caps actually allow me to remove the ports to adjust the tuning as needed. The end-cap is glued to the port tube with just hot glue but is screwed into the cabinet to secure them in place. Then once I eventually settle on the final tuning I like, I can remove the PVC end cap and glue the ports in place for a permanent (and cleaner) install. My main reason for wanting to do this is I quickly realized that there were many tuning options that appeared to work fine for this speaker. But I wanted to have a good listen in my listening space to see which one sounded the best in-room. Right now they are tuned to 32 Hz and I think they sound great here. While I would have been able to add port length if I had permanently installed them (without the 3D printed end-cap) there wasn’t a good way to get pipe cutter or jigsaw or oscillating tool in there to cut the ports shorter if I wanted a higher tune. So these little 3D printed bad boys offer a semi-permanent install, but with the option to remove the ports and adjust the tuning as needed later on. They aren’t super pretty however and they really only work because the ports are in the back and you don’t see them. I came up with a few better options after printing these that including a nicer taper, or flare, and not quite as bulky, but these are good enough for now. I also came up with other ideas such as 3D printing the entire port and ditching the PVC pipe all together. Once I started designing the entire port to be 3D printed, I realized there are endless opportunities for port diameters, internal port flaring, cabinet mounting options, etc. Note that port air speed with these dual 8″ woofers and dual 3″ ports at 100W is about 20 m/s, so under normal listening, there should be very little opportunity for port chuffing. In my listening tests, despite there being zero flare, there is no audible port noise.
Once I had the 3D printer all hot and ready, I designed these little speaker cable inserts which allowed me to easily install the 14 AWG speaker wire through the backs of the two sealed enclosures without needing to just hot glue the crap out of the wire to keep the box sealed. The insert perfectly fits the outer dimensions of the 14 gauge wire (without splitting it into two wires) and then fits into a regular 3/8″ hole. You can see the pictures of it below. The insert was glued into the cabinet, but the wire fits so snuggly that it doesn’t need to be glued to keep the enclosure sealed. This way I can still remove the speaker wire, or make it longer or shorter, once I get the crossovers finished up. I know you can just drill a big hole and shove the wire it and just glue it sealed, but that’s what I had done on my Swans M3 speakers and honestly it was just pain, and unnecessary. This littler insert was simple and effective and it looks clean, and still allows the speaker wire length to be adjusted if needed. Plus they took like 10 minutes to print. I did have to print a few of them in different sizes until I got the exact diameter where the speaker wire was just snug enough. But still, it was super easy and cost me almost nothing to make.
And as far as the 3D printed midrange/tweeter baffle goes, I have to give mad props to my 16-year old son who designed the whole thing in Fusion and did an awesome job. Up until this point, I had never designed anything in Fusion, or anything in 3D for that matter, I was 2D CAD guy at heart. I’ve been designing speakers with 2D software for 20+ years. So he came in clutch and drafted up this baffle piece all on his own. Once I started talking about how I thought it would be cool if there were a way to rear-mount the ND90s and maybe make something that could allow other drivers to be installed into the cabinet, I was initially just going to make it out of wood, he was like I can design it and 3D print it, it can look like anything you want. I was like let’s do it! He had just recently purchased a Bambu Lab P1S printer which he absolutely loves so he was taking advantage of an opportunity to get to use it. This printer is a beast. We maxed out the print bed with this thing too and made the baffle insert take up almost the entire thing. I’ll add some pictures of the design overall, but this idea of 3D printing parts and incorporating them into DIY speaker builds has really got me thinking outside the box of what can we do with this tech. Beyond just 3D printing waveguides, which I know people have been doing for years, I think there’s a lot of opportunities for novel ideas to take over and improve areas of the speaker building landscape. I’ve dabbled with other speaker combinations that fit this interchangeable baffle that range from single 4″ drivers and other tweeters as well as a design that houses Dayton Audio’s massive AMTPro-4 tweeter. The options are endless for what you can do here. Of course it’s possible to change out the 7″ and 8″ woofers for something fancier/nicer down the road too, since they are not flush mounted and just have a basic cutout which could probably fit just about any other 7″ and 8″ driver out there. In the end, this speaker cabinet has the potential to home any number of other driver combinations down the road. And for me that is kind of exciting. I know as my tastes change, or as the drivers may wear out over the years, or if I come into a pile of money one day, the option to be able to replace them while still keeping the nice cabinets (which I spent weeks building, sanding, painting, etc.) is just so appealing. Combine that with the fact that the crossovers can always be adjusted and tuned, the port tuning can be tweaked, means that this speaker project can evolved endlessly and never has to be completely done done. And I don’t mean tearing it apart tomorrow, I mean years down the road. I’ve created essentially a tailorable workspace in a speaker cabinet that allows more adjustability and tunability than I’ve ever incorporated into a speaker before. It’s possible to almost take this speaker back to the drawing board one day and re-do the entire driver selection and not have to build entirely new cabinets to do so. Anyway, for now I have no intention of changing the design. These speakers sound phenomenal and I am absolutely enjoying them to the fullest as they stand right now.
I think I’ll end this post here for now and let the picture galleries describe this whole process from start to finish. This was a really fun speaker build, I think they turned out great, they look great, they sound great, and are working dual duty for music in just 2-channel mode as well as for movies in a 5.1 setup. They may not be for everyone, but given the random find at my local Goodwill one day, I sort of took it on as a challenge to make something of this interesting combination of drivers. The speakers were not super expensive either, but I never tallied up the total. Sometimes it’s not about how much you spend or how much you saved but just enjoying working on the speaker along the way. And for me this was a really fun build. Notice in the pictures that since they play mains duty for my garage home theater, I do have a 12″ TC Sounds sub that I can play along with them when listening to music. And even though they’ve got plenty of good bass all on their own, man this setup sounds good with the sub. It’s basically flat from 20-20,000 Hz. And it can play crazy loud and it just sounds so good. Most of my music playback has just been via YouTube an a Firestick, but it sounds great and gives me the option to listen to a wide variety of music from artists I don’t normally have in my rotation. Including a bunch of audiophile jams that can really help showcase a good pair of speakers. And not to be outdone, I’ve got a killer subwoofer project to compliment this setup that’s in the works for later this year. Something along the lines of multiple smaller sealed subs with active EQ for a truly immersive experience. Until then, thanks for stopping by! I hope you enjoyed the read and enjoy the pictures below. Hopefully if nothing else this will inspire you to think what else can be built in the world of DIY speakers! And leave comment below, let me know you were here! I love hearing from strangers who happen up my blog and enjoy it. Thanks!
Design Drawings and Box Models
In-process Build Pics
Just the 3D Printed Bits
Final Completed Pics in Theater
REW Measurements
