I’ve had a pair of Swans M3 speakers since around the year 2000 – that was 24 years ago as of this writing – they were one of the first pair of really nice speakers I had ever built at that time and were also some of the first speakers to adorn my budding web page called Audio Innovation. I have enjoyed these speakers over the years and they have truly served me well. Whether they were set up for dedicated two-channel listening or as part of a home theater, they have always sounded great. Until a few months ago I started to notice some buzzing from one of the speakers. Upon further inspection I noticed that one of the woofer’s voice coil was rubbing and creating sort of a scratching sound. It was not super obvious, so I didn’t bother with it too much. A while later I noticed that a second woofer was also starting to rub. Okay, so that makes two. Not to mention this also happened in a different pair of speakers I had built around the same time with the W6 woofers. Which I rebuilt with a pair of Dayton Audio RS180s.
I started to consider what it might take to replace the four W6 woofers with something more modern, something newer, something better. As I was researching woofers, I thought I would go ahead and disassemble the speakers and check everything else, make sure there wasn’t something else I needed to replace. I ran a few FR sweeps in REW and that’s when I realized that both of the F5 midrange’s surrounds had broken away and completely separated from the frame! At least 1/3 of the entire surround was completely floating and not connected to anything. Okay, so that has to be fixed. That’s when I turned my focus to the RT1C-A ribbon tweeter. I could see something weird inside the grill on one of them, so I pulled it apart and to my surprise found that the long rectangular magnets on both the frontside and backside had completely disintegrated, turned to basically magnetic dust. Upon disassembly of the other tweeter, the same fate had fallen it, although not quite to the same degree. So if you want to know how long a pair of Swans M3 speakers last, it’s basically less than 24 years. The woofers were toast, the midranges had fallen apart and the tweeters had returned to their natural magnetic form. Dang, this project this got a lot more complicated, but also, a lot more interesting.
So how do you rebuild a pair of Swans M3 speakers in 2024? As it turns out, quite easily actually. HiVi still manufactures and sells both the F5 midrange and the RT1C-A ribbon tweeter. Sadly the W6 woofer is no longer in production. But this is a great start as 1/2 of the drivers are still available, they can be replaced outright without changing anything with the cabinets or the crossover. I had drawn up some different options with other midranges and other tweeters, but felt it changed the classic Swans M3 look that was just so classic of the time back in 2000. If I really wanted to use completely new drivers, I should just go ahead and build a whole new pair of speakers. I wanted to fix these speakers while still maintaining that great look that I’ve grown so accustomed to over the last 24 years. The trick was going to be finding a good woofer that matched the old woofers, but also matched the look and feel of the speakers overall. I tossed around the idea of theDayton Audio RS180s, which I have already in at least 2 other pairs of speakers, so I was sort of tired of building speakers with those drivers again. The newer Signature Series with the flat aluminum dish cones looked like great alternates, but they only offer them in 4 ohm versions. Which would have meant wiring them in series, for an overall less efficient speaker, meaning a completely new crossover and not just for the woofers, but for the mids and tweeters as well. I wasn’t in the mood to completely redesign this thing, so I narrowed my search to only 8-ohm drivers, just like the W6s, so I could wire them in parallel for a 4-ohm load and at least keep most of the crossover in tact (though I would later end up tweaking the crossover anyway, while keeping most of it unchanged).
I wasn’t dead set on maintaining all HiVi drivers initially, until I narrowed my search again to just HiVi drivers and that’s how I landed on the M6N-B 6″ black aluminum/magnesium woofers. These things looked pretty awesome! A quick simulation in Unibox showed they were a perfect fit for a 47 liter boxed tuned to 42 Hz. I wasn’t a huge fan of the stamped steal basket, but otherwise it checked all the boxes. Not to mention they are actually one of the less expensive woofers in the 6-7″ variety. I mocked up a drawing of them paired with the F5 and the RT1C-A and they looked great together, one of the best combinations I had drawn. So with that, I made the decision to buy (4) M6N-B woofers and a new pair of RT1C-A tweeters. I elected not to buy new F5 midranges, but instead bought a small bottle of rubber-infused CA glue and repaired the ones I already had. I bought some rubber feet to clean up the bottom of the cabinets and a few crossover parts so I could tweak the crossover a bit as needed. Since now 24 years later I actually have the ability to measure these speakers and can make changes to the stock crossovers that were originally designed for these speakers.
About this time, I had removed all of the drivers and all of the crossover parts from the cabinet, I decided to refinish the cabinets to make them match some my other speakers I had built recently, like my center channel. This meant sanding down the entire cabinet and refinishing and re-staining them. I ended up going with a darker sort of Walnut look, and repainted the front/top/back in all black. I finished up everything with a few coats of a satin polyacrylic to give the speakers that nice and soft sheen that looks great on any speaker, since it’s not too flashy, but isn’t too dull. It’s sort of the Goldilocks of sheens, it’s just right. With the cabinets all cleaned up, I dropped in the drivers with the crossovers outside of the cabinet and that’s where the fun really began.
I took around a hundred or so measurements with REW with the stock crossover and various minor iterations on the stock crossovers. Currently I have a small supply of inductors, caps and resistors, so I wasn’t in a position to design a completely new crossover from scratch. But I did have the ability to make some adjustments, so that’s what I did. In the end I ended up with a slightly more complex crossover on just the midrange, but a slightly less complicated crossover everywhere else. I dropped the crossover point to the new M6N-B woofers by increasing the parallel cap from 68 uF to 125 uF and removing the series 3.3 ohm resistor, to really bring that frequency roll-off down quicker. The tweeters got a bump from 1 to 2 ohms on the series padding resistor and I got rid of the contour circuit which just made the peaking around 6 kHz worse. This dropped the overall volume a bit better but still left them pretty hot to maintain that sweet bright sound that I had grown accustomed to with these speakers over the years. I dropped it lower in my listening tests with both 3 and 4 ohm resistors, and while they measured arguably flatter, they also sounded flatter, boring, and I just couldn’t get on board with re-doing as much of these speakers as I had, the money I had spent, and the time already invested, to make them sound dull and lifeless just for the sake of a graph.
So with those two very minor updates out of the way, I focused on the midrange. This is where I really wanted to make some deliberate adjustments to the tonality of these speakers, because they always sounded very bright in the midrange that I never fully got used to. I always told myself I would fix them if I ever got around to getting measurement gear to be able to do so. After my first real measurement of the stock Swans M3, albeit 24 years later, I could finally see what I could always hear. These speakers had two prominent dips in response with one of the highest midrange responses I’ve ever seen standing out above everything else. Not only is the midrange coming in hot, it’s suppressed above and below at the crossover points, making it stand out just that much more. There’s too much overlap with the woofers, and they are out of phase, so there’s a null there that shouldn’t be, and then there’s another null above the midrange to the tweeter, from just not crossing over high enough to meet it. The overall tweeter by itself isn’t actually too off. In hindsight the midrange was probably always supposed to be wired out of phase, but not according to HiVi’s original schematic. Note that now the tweeter actually needed to be also wired out of phase to match the phase of the midrange since it was flipped to match the woofers.
So I brought up the crossover point to the mid by reducing the 150 uF cap to 100 uF and then flipped the phase to the midrange. This brought the woofers and midrange in alignment at the crossover point so they summed much better. Not to mention I had dropped the crossover point of the woofers as well. That took care of the woofer-to-midrange response, but the midrange was still loud. That’s where the L-pad came into play. A 1 ohm series resistor paired to a 25 ohm parallel resistor dropped the entire midrange response by about 2 dB. That placed the midrange much better in-between the tweeter and the woofers. But there was one last thing that was still bugging me. There was this peak right at about 1.8 kHz that still gave the midrange kind of loud presence. So I slapped in a parallel RLC notch filter with a 0.1 mH inductor and 68 uF capacitor. I didn’t even bother adding a resistor, as I wanted it as deep as I could get it. And that did completely did the trick. The midrange now fell in line and was finally sitting in a much more quiet spot within the frequency range of the whole speaker. The tweeter also became a bit brighter at this point too, relatively speaking, but I honestly just thought it sounded so good I just left it. I gave it a good listen with a bunch of different music and settled on it as the final design. You can check out the schematic for the crossover here. Now is this crossover perfect? No, it’s not. But given that we kept almost 90% of the original HiVi crossover in tact, I’d say we made a reasonable dent in improving the overall response and tonality of the speaker, while still maintaining some of its original signature sound.
With the tweaking stage done, I soldered up all the new parts on the original crossovers and mounted them back into the cabinets. This was a chore as I found out, since wiring this speaker up with all of these separate crossover boards was kind of a nightmare. But I persevered and got the crossovers soldered up, installed into the cabinets, I added some extra poly fill for good measure, got all the drivers installed and moved them back in my family room where they would become the permanent front Left/Right speakers of my small home theater. The Swans M3 speakers live again! The new tweeters sound fantastic, the F5 midranges with the CA seem to be working okay, though they look a little messy right where I glued them, you have to get pretty close to notice, and lastly the new M6N-B woofers look completely at home in this setup. If you didn’t know they weren’t part of the original speaker, you never could tell. They blend right in and look amazing. Despite neither the midrange nor the woofers being flush-mounted, they still look high end to me. The main thing is having that RT1C-A tweeter nice and flush mounted, and you don’t really even notice that the others speakers are not. I haven’t really got a chance to run them through their paces with movies and stuff, but so far what I have watched sounds awesome. These are truly a great speaker. And the fact that you can still build them almost true to their original form in 2024 is pretty cool. Tell me what you think? Did you ever build a pair of Swans M3 speakers? How are they holding up? Would you build this speaker again in 2024? Thanks for reading!
Picture Gallery
REW Plots
