Introduction:
Hey guys, been a long time lurker here. I've been pretty busy lately working on a dirt cheap unity horn project for my desktop setup. The entire horn is going to be 3D printed for ease of assembly. A lot of effort has been put into making this build as cheap as possible. I actually think this might be one of the cheapest unity horns out there. I have two goals I want to achieve for this project.
- Wide phantom image and constant directivity. I have lots of pc-monitors in-between my current speakers. The problem with this is it forces my speakers to be very far out from the listening spot (around 150 degrees). Unless you are seated perfectly between the speakers, the sound-stage collapses. It is my belief that a wide CD horn will fix this.
- Provide an affordable MEH so that more audio geeks can hear these awesome speakers. I really love unity type speakers, the linear phase, constant directivity, and acoustic filtering provide a lot of benefits that are hard to achieve otherwise. My dream is to eventually make this into a diy kit, but I'm not sure if I have the resources to make this a reality.
This all started when I discovered these little Dayton TCP115 woofers. I was working around in hornresp and pasted the driver profile by mistake. Low and behold, the simulation had an extremely wide bandwidth for such a cheap driver. The high xmax is also nice for going low so I was able to make this design a 2 way and still get decent bass.
Tweeter:
Most, unity horns seem to use compression drivers. Since I don't need high output and need a cheaper option I decided to go with something different. The XT25 is what I am using for now and it seems to work quiet well. By using a really large horn I was actually able to push the crossover frequency quiet low (around 1.2khz). I know, I know, that crossover is insanely low for this tweeter. I have more data on why this works; will post more later. Right now I am having trouble getting accurate distortion measurements to confirm my theories.
Measurements:
Around 15 degrees off axis after some quick EQ. The dips at 6k and 10k appear to be from a problem at the throat which are being fixed in the next version. The next box is also going to have a flush mounted horn and 18mm round overs on the baffle. I am hoping this will improve those ripples. You can't just EQ those ripples at any one angle. As you can see the dips move around depending on your angle off axis.
1/3 smoothing of the same graph. I don't know what the exact angles are, it's really just a proof of concept to show the directivity is being controlled. The light green line is around 60 degrees, which is why it is so low in level.
Things I still need to do:
Throat optimizations, active xover tweaking, passive xover development, take more accurate polars outside, measure below 200hz, port it (maybe?), make it pretty.
Last edited by a moderator:
The sounds a bit strange... what do you mean - the phantom image makes it possible to create the notion of a sound source not playing from an actual speaker in a stereo setup. You want that as small as absolutely possible - i.e. you test it by playing a tone in both speakers and the best result is if you hear it between the speakers and as if there was a very tine speaker there. Maybe you mean wide soundstage?
Nice build btw!! 🙂
//
@TNT
Yes wide sound stage is basically what I mean. I'm hoping to achieve a wider sweet spot so that voices and mono sources are more anchored to the center of my setup. In my experience this comes from a better polar response. Time will tell, I'm printing 2 more horns this weekend. Will keep you updated
Yes wide sound stage is basically what I mean. I'm hoping to achieve a wider sweet spot so that voices and mono sources are more anchored to the center of my setup. In my experience this comes from a better polar response. Time will tell, I'm printing 2 more horns this weekend. Will keep you updated
Looking forward to your results rk.
what printer and material are you using?
Keep the posts coming 👍
what printer and material are you using?
Keep the posts coming 👍
Okay so I think I managed to fix most of the ripples in the 10k region. Source of the issue appears to be from the chamfer on the tweeter faceplate. So I printed a new one. Also I filled the port holes with a very small amount of polyfill.
This thing just slaps on the front of the xt25. I was gonna remove the faceplate entirely but the glue they use is stubborn. Anyway, here's the results:
I added some separation to the lines to make things more readable. Measurement angles are similar to last time. I just rotated the speaker until I reached 45 degrees. I'm pretty happy with this response should respond pretty well to EQ I think.
This thing just slaps on the front of the xt25. I was gonna remove the faceplate entirely but the glue they use is stubborn. Anyway, here's the results:
I added some separation to the lines to make things more readable. Measurement angles are similar to last time. I just rotated the speaker until I reached 45 degrees. I'm pretty happy with this response should respond pretty well to EQ I think.
Last edited by a moderator:
Must have maxed out the build plate, we'll done, I would expect warping at that size? Lotsa brim or raft?
I may have missed it but what did you calc and design the horn with?
Okay I made a couple upgrades to the model. First is a phase plug to the mids, I'm able to crossover a bit higher now. This took a lot of strain off the tweeter and it sounds a lot better now. I think it also tamed some resonances in the midrange chamber, but that's just speculation. The next thing I did was build the tweeter chamfer (that plate from earlier) into the model. This gave me a little bit of area to add some throat profiling without a complete horn redesign.
Here's some pics of assembly, I'm just using super glue. I read somewhere to use isopropyl alcohol to clean up excess glue. That was a terrible idea, it left white stuff everywhere. Gonna stick with sanding in the future.
Oh yeah and if you look closely there is a sneek peek of my SH60 clones in the middle pic😉
Oh yeah and if you look closely there is a sneek peek of my SH60 clones in the middle pic😉
Been listening to this in mono a lot with my minidsp. I really happy with the sound considering the price I paid for the drivers. One regret I have is not spending more on the woofers for this project. Hornresp indicates that the TCP's perform well, but the reality is that I am not happy with the bass. Not sure what I was expecting from 2 $15 woofers. For rock n roll they are awesome, but I listen to a lot of electronic music so these things really need a sub, no surprise really.
To avoid spending too much money on more amp channels and another minidsp I really want to add a passive crossover to this. Then I can use the dsp to manage the subwoofer crossover. This is were I am at:
I still need to take proper polars outside. Been busy but I will get to it. I'm doing my best to keep the component count as low as possible. Having a high component crossover on a $90 speaker doesn't make much sense to me... Also there isn't that much room in the box. Ignore sensitivity, the level is accurate but I have no idea what the voltage is. Hornresp suggests its around 80-85 dbs at 2.83V.
It has been very difficult to keep the impedance up. Both the woofers and the tweeter have a massive rise in the 500hz-4000hz region. The grounded inductor and resistor are for impedance flattening, but at the cost of lowering the dip from 3.6 to 3.3ohms.
Anyways I'm not the best at passive crossover design so if anyone wants to chime in with ideas that would be awesome.
To avoid spending too much money on more amp channels and another minidsp I really want to add a passive crossover to this. Then I can use the dsp to manage the subwoofer crossover. This is were I am at:
I still need to take proper polars outside. Been busy but I will get to it. I'm doing my best to keep the component count as low as possible. Having a high component crossover on a $90 speaker doesn't make much sense to me... Also there isn't that much room in the box. Ignore sensitivity, the level is accurate but I have no idea what the voltage is. Hornresp suggests its around 80-85 dbs at 2.83V.
It has been very difficult to keep the impedance up. Both the woofers and the tweeter have a massive rise in the 500hz-4000hz region. The grounded inductor and resistor are for impedance flattening, but at the cost of lowering the dip from 3.6 to 3.3ohms.
Anyways I'm not the best at passive crossover design so if anyone wants to chime in with ideas that would be awesome.
@AllenB. Which ones are you referring to? If you are referring to the woofer circuit, the grounded inductor and resistor are used to create a high shelf on the woofer. It has a large peak that is from the horn loading. Here is what happens if you short them.
The other on the bottom left are just to flatten the impedance peaks if you where using something like a tube amp. I will probably omit them.
The other on the bottom left are just to flatten the impedance peaks if you where using something like a tube amp. I will probably omit them.
So I'm getting a lot of messages requesting the STL files. I just wanted to say I plan on releasing them soon on thingiverse and bambulab for a small fee. I'm making some tweaks to the file still as I make crossover tweaks/changes. It has been a struggle to get this printing reliably, with high infill, while still only using a single roll of filament. Each horn takes a full day to print, plus more time to assemble and measure.
Once I am done with the final design, I plan on releasing a full write-up with crossover and dsp settings so that everyone can build their own. Thanks everyone for your patience.
Once I am done with the final design, I plan on releasing a full write-up with crossover and dsp settings so that everyone can build their own. Thanks everyone for your patience.