A PVC cab is the most economical small volume chamber, likely a 3" diameter. 3D printing might be smaller.
A note though- if you make the chambers super small, the Qtc at the resonance of the chamber will be VERY high. This in turn will need to be notched or suppressed heavily. For an example, I used a 4" in a 1ltr sealed and damping lined chamber that made the Qtc VERY high at 200Hz. Since I had a xover at 400, I had to suppress it to maintain the rolloff and suppress the boost at 200Hz.
If you are using DSP, this is likely to be fixed simply. If not, it requires 2 or 3 large value passive components to get it done.
A note though- if you make the chambers super small, the Qtc at the resonance of the chamber will be VERY high. This in turn will need to be notched or suppressed heavily. For an example, I used a 4" in a 1ltr sealed and damping lined chamber that made the Qtc VERY high at 200Hz. Since I had a xover at 400, I had to suppress it to maintain the rolloff and suppress the boost at 200Hz.
If you are using DSP, this is likely to be fixed simply. If not, it requires 2 or 3 large value passive components to get it done.
A couple of things here.
1. Wolf is absolutely right that they should not occupy the same air space in the enclosure.
2. With them facing 90 degrees off axis you will run into some issues. I can simulate this though if you tell me how back one is from the other and how over. So I would need x axis offset, y axis offset, and z axis offset. To simulate it properly you would need a 90 degree off axis measurement of that W3 driver
3. With it downfiring you'll want to cross it way lower. Around 120 Hz. Any high you deal with too much reflections and the off axis will screw with your FR.
4. Wolf is also right that you will need some pretty hefty inductors if you want to cross very lower. DSP is another animal entirely and also doesn't come cheap so that's a choice you will have to make
What was your intended XO frequency?
Great, I shared the dimensions in a previous reply with the screenshots. I was going to crossover about 600Hz but may need to be lower with them down firing as you say. I wouldn't cross over any lower than 300/400 I think or the full range end up doing most the work and would end up needing a larger internal volume which would take away from the TB's space. I get there's some tradeoffs but that's to be expected for something so small trying to pack so much.
PVC is a good idea, I will probably print something in the same clear resin I intend for the enclosure - another trade off as padding and acoustic material would no doubt improve the sound but would negate the enclosure being see-through. Will check the Qtc calcs - I have the loudspeaker design cookbook which I'm sure covers these calcs but not with me where I am currently unfortunately.
JCR,
I have some experience with 3d printed enclosures that I can share. This is actually how I started out. I was making folded horn designs with full range drivers.
Always use PETG when printing enclosures. Use a lot of walls and a lot of infill. You need them heavy. The reason is that PETG rings out around 400-600 hz depending on infill. PLA and ASA ring out near 1khz. You definitely do not want a resonance in this range. You can't get rid of the resonance entirely but you want it as low as possible. PC does pretty well too but PETG is still king of low resonance.
I've seen it done where people leave an empty outside wall and then fill this with a mixture of plaster of paris and wood glue. This does dual duty of making the enclosure heavy and lowering the resonance frequency. It makes your box bigger though. This is the best way with 3d printed enclosures but obviously is a lot more complex.
I have some experience with 3d printed enclosures that I can share. This is actually how I started out. I was making folded horn designs with full range drivers.
Always use PETG when printing enclosures. Use a lot of walls and a lot of infill. You need them heavy. The reason is that PETG rings out around 400-600 hz depending on infill. PLA and ASA ring out near 1khz. You definitely do not want a resonance in this range. You can't get rid of the resonance entirely but you want it as low as possible. PC does pretty well too but PETG is still king of low resonance.
I've seen it done where people leave an empty outside wall and then fill this with a mixture of plaster of paris and wood glue. This does dual duty of making the enclosure heavy and lowering the resonance frequency. It makes your box bigger though. This is the best way with 3d printed enclosures but obviously is a lot more complex.
Thanks for the advice. I used to do 3D printed speaker kits a few years ago. This one I'll get PCBWay to print for me though as I'll be using UTR-8100 clear resin. It has a similar mass density to PETG and being resin it'll be solid infill (only way to make it clear). I'll be printing the enclosure probably around 6mm but adding rib features and bracing. Should make for a solid enclosure.
I saw the video DIYPerks did with the plaster of paris, cool idea and something I'll probably try on a larger speaker one day.
There's a lot to resonance but generally the thicker the material and the greater the density the lower the resonant frequency (why MDF is popular). Also bracing will be key and avoiding sharp corners and edges so 3D printing will lend itself nicely here. To keep the design clear I will probably apply padding for dampening on the back wall (where all the PRs are) and behind the front baffle as this won't impact the see-through aspect for the majority of the speaker and I'll make the sides thicker to compensate.
JCR,
It seems you know what you're doing as far box building. You don't need any help there.
I haven't gotten into resin printing yet. As soon as a project requires it I'll be buying a Form 4 for the shop.
It seems you know what you're doing as far box building. You don't need any help there.
I haven't gotten into resin printing yet. As soon as a project requires it I'll be buying a Form 4 for the shop.
Ah form 4 will be awesome! I've not printed in resin before so will share the results when I get round to building it.
I like the response you got with your crossover. Were you able to model the effect of the offset axis? I think I'll give your crossover a try so long as I can be comfortable knowing they won't be out of phase to a noticeable degree.
JCR,
I did run it assuming a 90 degree off axis response. It didn't really change anything because the XO frequency is so low that the off axis doesn't change too much there.
With most lower range drivers the off axis doesn't change too much low down. Here is just one example. See how much it changes at the high end hut how little it changes at the low end?
You can run that series XO I showed you. That one should knock out that huge high frequency notch from the full rannge drivers. If you don't knock that down somehow they will sound very harsh.
I did run it assuming a 90 degree off axis response. It didn't really change anything because the XO frequency is so low that the off axis doesn't change too much there.
With most lower range drivers the off axis doesn't change too much low down. Here is just one example. See how much it changes at the high end hut how little it changes at the low end?
You can run that series XO I showed you. That one should knock out that huge high frequency notch from the full rannge drivers. If you don't knock that down somehow they will sound very harsh.
Couldn't see an attached image here if there was one. Thanks for the info though. I figured should be okay low end but I was planning on crossing over a lil higher than 250Hz so I can keep the full range driver enclosure size as small as possible (bearing in mind the Qtc values).
What equations did you use to calculate your component values as I'd be interested in having a stab at raising the crossover to around 500Hz but keeping that HF notch smoothed out?
I mapped the drivers out in WinISD allowing for 0.3L for the Dayton Audio RS75-4 with the crossover at 550Hz:
Qtc is higher than ideal for RS75-4 but I'm going to have to compromise somewhere trying to keep this thing so small. This already involves making it larger as I have 1.7L for the TangBands plus the two 0.3L enclosures surrounding the full range drivers so a total of 2.5L by the time PCBs and the like are in there.
I forgot the image. Let me show it here.
As you can see, the off axis doesn't start to take effect until around 800 Hz. If you're using a first order XO then you need to worry about it leaking over the higher end. With traces none of this matters a ton. You are just getting within the ballpark.
The first 3 way I made was with traces. It's frequency rspsosne is far from flat but I think it sounds amazing (possibly an IKEA effect).
It's doing with the big issues like that giant peak with the full ranger, crossover over high enough to not exceed the xmax of the tweeter, dealing with woofer cone breakup, all of this matters more than a dead flat frequency response in my opinion.
I let VituixCAD do the heavy lifting as far as choosing component values. You'd be doing yourself a favor to start using that simulator for you XOs. It's pretty much dead on. Made by someone much smarter than I. I will attach the FR from the last two way I made. The one I used a series XO on.
As you can see, the off axis doesn't start to take effect until around 800 Hz. If you're using a first order XO then you need to worry about it leaking over the higher end. With traces none of this matters a ton. You are just getting within the ballpark.
The first 3 way I made was with traces. It's frequency rspsosne is far from flat but I think it sounds amazing (possibly an IKEA effect).
It's doing with the big issues like that giant peak with the full ranger, crossover over high enough to not exceed the xmax of the tweeter, dealing with woofer cone breakup, all of this matters more than a dead flat frequency response in my opinion.
I let VituixCAD do the heavy lifting as far as choosing component values. You'd be doing yourself a favor to start using that simulator for you XOs. It's pretty much dead on. Made by someone much smarter than I. I will attach the FR from the last two way I made. The one I used a series XO on.
Attachments
I was using VituixCAD but didn't know it could determine values for you? Perhaps I'm using an older version as I note your screenshots say REW V5. Looks like I need to do some more research. Do you tell it the crossover frequency you want and it determines the values? This is what my window looks like:
In the XO view on the bar by the components is a tab called LIB. Click that tab. That will get you your basic parallel components.
With a series though, you're on your own.
You get a feel for it after some practice. Always keep an eye on the filters graph. This shows you what you are bucking/boosting.
With a series though, you're on your own.
You get a feel for it after some practice. Always keep an eye on the filters graph. This shows you what you are bucking/boosting.
REW is a different program. REW is what we use to take the measurements of the drivers. This is our real world measurement. Using REW you can take the measurements of the drivers you have in your actual box and then use those measurements to make the XO
Hi, just to go back to this: If I have an enclosure for the full range drivers is it better if there is a division/divider (shown in red) so that the left and right channels aren't sharing the same volume of air? Since it's frequencies above 550Hz I wonder if it matters so much? Any divider will subtract from the overall volume of air even if only slightly. I will also add dampening material in here to artificially reduce Qtc.
For the application, I'd say no. For critical listening, you can get crosstalk that way.
Separate them. It is best practice.
Also, don't leave just a flat plate behind them. That is the worst for reflections. You're 3d printing it right? So put all sorts of wierd shape elements on that plate on the inside that faces the driver backing to break up the sounds waves. Sort of like how they those diffraction/dispersion walls. KEF has actually started doing this. You should be able to find an article about it and copy the design
Okay call I'll put a thin 2-3mm wall in the centre of that part but I wonder if this raises the Qtc more than if they shared the enclosure even though it's a small overall volume difference when divided per driver.
Would this be necessary if I intend to fill the space with wool? Wouldn't this also break up the sound waves? Don't mind texturing the inside as you propose though, would be fun and add a bit more rigidity to the part too. I'm guessing you mean something along these lines:
Yeh I figure cross talk isn't such an issue when the speaker is only 340mm wide as L/R will hardly be separated. I did move the drivers slightly further apart as you suggested though.
JCR, the divider is going to remove minimal space. It won't even matter.
Sound waves pass through the stuffing. That is the point. If they didn't, it would just reflect off of the stuffing. The idea is you want the waves to NOT hit the back of the cone. So once they pass through the stuffing they bounce back at an angle, diagonally through the stuffing (therefore increasing the thickness of the stuffing as it has to pass through it at an angle), and they don't hit directly at the cone. Hopefully they hit another couple walls and go through more stuffing before finding their way to the cone. By that point they have such littler energy that it doesn't matte. This is best case scenario. It all works together
Sound waves pass through the stuffing. That is the point. If they didn't, it would just reflect off of the stuffing. The idea is you want the waves to NOT hit the back of the cone. So once they pass through the stuffing they bounce back at an angle, diagonally through the stuffing (therefore increasing the thickness of the stuffing as it has to pass through it at an angle), and they don't hit directly at the cone. Hopefully they hit another couple walls and go through more stuffing before finding their way to the cone. By that point they have such littler energy that it doesn't matte. This is best case scenario. It all works together
Ah okay, got you - thanks. Will work on that and the rest of it. Still a ways to go!