Colored 2way Bookshelf Cheapies

[LOUT]

This was my first 2way build, so it has let me try a few different things while being something of a work in progress to see what I can improve.

They use the Dayton PC105-4 midwoofer and TD20-4 tweeter in ~7liter/0.25ft boxes with slot-port tuning around 55hz (0.5"T x 5.5"W x 6.75"L).
The boxes are quarter-inch plywood for the top/bottom/sides/port and half-inch for the front and back, giving more surface area to glue the thinner panels and screw in speakers+terminals. I was curious about cabinet resonance and other possible problems from using 1/4" plywood without bracing, but resonance hasn't been nearly the issue I'd feared except when cranking them with sweeping sine waves. I may glue in a dowel or two in the future.
The thin plywood allows these to have modest 10" x 10" x 6" outer dimensions, and they're very light-weight.
For coloring, I thinned some acrylic paint with a lot of water and rubbed it on with a few paper towels (cheaper/rougher paper towels tend to drop less lint), first testing color on cut-off scraps.


The original crossover (pictured below) was 4part series designed to be simple and very cheap. It allowed a single 0.5mH inductor to somewhat fill the job of two (bringing the build cost even lower), while a couple capacitors and a single 4ohm resister are all that's left. The end result leaves the woofer's highs more present than ideal, and the phase overlapping is not good.
The newer parallel XO uses a 1.0mH coil and 33uF cap for the LowPass while the HighPass is a 2-4ohm resistor (depending on preference) and 6.8uF cap with a series-notch shunt between the HP and tweeter using a 1.0mH coil and 10uF cap. This shows much better phase overlap which sounds a little more consistant and detailed in the high mids. Although the change can be subtle to my untrained ears for music, I think I prefer this newer parallel XO.

I've been using the free program VituixCAD for crossover design after seeing a video from KirbyMeetsAudio a little over a year ago. It also has a nice box/woofer sim but I never found a port/air-speed modeller so I still need to use WinISD for that although WinISD seems generally less flexible and a lot more glitchy.


These get loud and can produce a healthy amount of musical bass. They start to sound sloppier in the lows when playing near maximum volume with really low bass below tuning, but they are surprisingly forgiving about it and nearly invincible against over-excursion. I wonder if the PC105's xmax is actually closer to 3-4mm?

[LOUT]

Original, series crossover which I'm not sure I'd recommend:


At the time PE sold 0.5mH ironcore inductors which saved a couple dollars for each speaker (they were around $2-$3 each instead of the $4-$5 for inexpensive aircore). I'm pretty sure the wood, parts, drivers and everything altogether for these totalled around $60-$70/pair.

I think they should be able to reach about 96-98db/1m peaks from 40watts after -4-6db BSC when crossed to a subwoofer. Used without a sub, music with really deep lows can make them run out of steam earlier; for example Lorn 555-5555 sounds happier about 6-9db down closer to 5-10watts while cranking it up starts to make the ends of bass notes sound extra fat and "boomfy". The ridiculous pc105's still refuse to bottom-out or do anything super ugly even when pushed with far worse, so I can't help but love them.

[LOUT]

Updated, parallel crossover:

Along the way I had a much steeper XO that modelled and sounded good, but I really wanted to try to keep cost and complexity down. This simpler version doesn't give up much, mostly letting the woofer roll-off a little more gradually and letting the tweeter play a hair lower.
The new XO can be made for a little over $6/each using Dayton/PE parts...a decent fit for these affordable drivers.



The modelled graphs look good at least to my inexperienced eyes, and the XO has been behaving well so far in the real world too.
I was worried about playing the TD20 with such a shallow highpass (1st order with a small notch), but I tried listening to it with the woofer disconnected at this slope as well as even larger cap values playing sine sweeps and some music while slowly increasing the volume. With a 10uF or larger cap, while cranked all the way up and sitting within a few inches I can barely hear it starting around 300-400hz. Down there I think it might have a faint but very musical distortion with a following note, but it's too quiet to hear from a normal distance let alone with the woofer playing. The current XO's 6.8uF cap seems to be small enough that I'm not noticing any distortion nor harshness. I think it's technically a 6000hz 1st order XO, so that might be why it's behaving nicely.

Do tweeters generally let you know if you're playing them too low either by breaking pretty fast or by sounding obviously distorted?

[ceiol]

Looks pretty good! Nice work.

[LOUT]

Thanks.

[LOUT]

I'd like to add that I specifically built the tweeter's series-notch-shunt around a 1.0mH SolidCore inductor with the idea that someone building five or more of these for an inexpensive surround-sound can then save $4-$8 from PE's buy-10-discount on parts.
But anyone who prefers AirCore inductors could easily switch some parts around on the tweeter's series-notch-shunt to save a couple dollars with a larger capacitor and smaller inductor if they're already using cheap/NPE caps...If the tweeter's caps are poly, then the smaller cap and larger coil is still the cheaper option whether its AirCore or SolidCore.





The tweeter also tames down nicely with a 4ohm resistor VS 2ohm. I don't mind the 2ohm for television, but for music I think it's too sibilant.

[Regore]

Very nice build for your first. That woof is almost a full range. Shame to cut it out at so low a frequency.
Also just curious ? What is your LCR filter tuned to ?

[LOUT]

Very nice build for your first. That woof is almost a full range. Shame to cut it out at so low a frequency.
Also just curious ? What is your LCR filter tuned to ?

Thanks. Hearing any kind of interest in them from someone else after this much time is kind of flattering.

I couldn't get a satifying amount of BaffleStepCorrection with a two-part LowPass any other way, but I think adding 2-3 more parts should let it cross over as high as 4-6khz (maybe an 0.5mH followed by a 2ohm+6.8uF to ground, then another 0.5mH for a 3rd order LP). That would probably let it work really well with one of the Dayton ND13 or ND16 tweeters using a simple R+C HighPass

The notch on the tweeter (really just a cap and coil...the resistance shown is just the natural DCR of the coil) is centered around 1600-1700hz. It gets rid of the tweeter's frequency peak around FS and seems to make it a lot friendlier for crossing low without sounding super harsh.

[Chris Roemer]

(just countering regore's statement here)

Personally, I think 6k is awfully high to cross ANY tweeter, (even in a 3 or 4-way) how high the next driver below COULD play doesn't really enter into it (much).
There are (or should be) OTHER concerns, like beaming.

[LOUT]

With a gentle roll-off on the tweeter (1st or 2nd order HP depending on the tweeter's own acoustic roll-off) it looks like the PC105 might be alright crossing as high as 5Khz, at least in the simulation out to 45degrees. But I agree that just because it's possible doesn't mean it makes sense to do it...I imagine most tweeters are totally fine playing lower which should help avoid vertical axis lobing and woofer beaming (like you already mentioned).