• Welcome to ASR. There are many reviews of audio hardware and expert members to help answer your questions. Click here to have your audio equipment measured for free!

Small 2-way speakers with linear on-axis and power response characteristics (Scan Speak and SB Acoustics drivers). H&V off-axis measurements included

A relatively minor thing I noticed but didn't see in this thread - the fantastic crossover boards @wineds created use slightly different numbering schemes for the capacitors from those shown in the first post. In this case, C2, C3, and C4 are mixed up a little. I'm sure other users noticed this because they aren't terribly interchangeable due to the physical size difference. That or they were smart and just referenced the nice layout wineds provided in their readme.


For some reason I was referencing the first post.

In any case, I recevied the rest of the crossover components yesterday and mocked up the Bevenbi caps from Cinergy and the Mundorf resistors from Madisound. They fit! Some of the caps are a little tight (I seem to recall the 3.3uf one being a bit 'long', as was the 18uf cap) but they'll work. One thing to point out though if you go this route, some of the inductors would be a bit tight. If you go with other brands of inductors that are thicker you might run into issues. I just went with the Solen coils for this one so it'll be fine (they aren't shown because the leads are very long and I didn't want to cut them or bend them too much).

IMG_0933.JPG


Note: There is nothing wrong with going all-Solen here. I was just trying to find alternatives in the U.S. in case international shipping+duties/tariffs remain costly. If you're ordering Solen coils it'd be kind of silly to not just buy the rest of the parts from them as well since they are competitively priced.

In other news I'm almost done with the prints I'll use for testing. The baffle was straightforward since I'll use the chamfered one I had previously designed but only printed as a test run in PLA. This one is PETG, 6 wall loops, 25% gyroid infill. Woofer fitment is a little tighter than I'd like and I'm waiting on final hardware, but this is it. Printed face-up, this printed with no supports required. I'm not a huge fan of the top surface finish, the 'final' baffle might be printed face down to get some texture, though that would require some supports to do.

IMG_0928.jpg


Rear of the cabinet was slightly redesigned and printed face-down to test the aforementioned textured finish. No matter which way this was printed it would require supports so it seemed a good test subject. I chamfered the corners slightly to break up the hard edges a bit (this is purely for aesthetics) and added a place on the 'inside' for installation of the THD board. As with the baffle, this is PETG with 6 wall loops and 25% gyroid infill.

IMG_0931.jpg


IMG_0926.JPG


Finally, the first cabinet wrapped up after nearly 3.5 days and 2.2kg of filament. Also PETG, 6 wall loops and 25% gyroid infill. I started the 10% version immediately (it'll take about 2 days and 1.7kg of filament, IIRC).

Front:

IMG_0929.JPG


Back:

IMG_0930.JPG


I have started redesigning the cab, baffle, and rear a bit to simplify the prints. First, someone mentioned that some might prefer a cleaner look on the front, and since the rear is removable too we could just run bolt through the backside of the baffle to attach it to the cab. I initially disagreed because I like the look of hardware, but I'm now thinking a bit differently. This also impacts the spacing of the bolts on the rear since they originally matched the front - this is no longer a requirement so they can be spaced out further. I also added a cross-brace to the front because the box prints face down, so why not (I'm still thinking about making this brace removable, similar to the other braces). Lastly, I removed the groove for the TPU seals, at least temporarily, if not permanently. I'm thinking about these junctions a lot and may introduce some new pieces. We shall see.

By the by, I'll be making several baffles available. Baseline (as shown below), roundovers on vertical edges, chamfers on all edges, and faceted vertical edges with chamfers on top/bottom. Just want to give builders a choice. I will try to take measurements of each eventually.

1784123229022.png


1784123243986.png
 
A few other things in CAD I did this morning...

1784132350938.png


And this...

1784132367081.png


Which is maybe better viewed as this...

1784132480744.png


I'm not great with curves in Fusion so I'm just playing around at this point, but someone had asked about something like this previously.
 
Thank you, I liked it too! So much so I printed a quick version last night. I think this is promising and perhaps worth further investigation once the cabinet construction is sorted out. Don't mind the 'fuzziness', that's just a finish I was playing with. I think something like this might need to have a few extra layers of top 'skin' so it could be sanded smooth. I'll need to look into that if we like this baffle. Curving the baffle inherently curved the corners too, so I'd need to modify the cabinet corners to match.

IMG_0936.JPG



IMG_0937.JPG


IMG_0938.JPG
 
Looks impressive and eye catching. I'm really curious how this will work with the crossover as it's been tailored to the measurements taken on a rectangular panel with sharp edges.
 
Thank you, and I agree! It'll be an interesting experiment for sure. I'm not sure I can do the same level of sweeps you did but I'll give it a shot at some point. Individual drivers and with the crossover installed. Probably a few weeks away from that as I'm prioritizing the cabinet construction right now. I'm waiting on some more TPU to arrive then I should be good to go.
 
Been following this thread for a minute and I’m excited to see where the 3D printed variant of this goes. Would be fantastic to get a version of this that tests as well as the original but without the need to resort to wood. Does anyone have a few spare PCBs?
 
Been following this thread for a minute and I’m excited to see where the 3D printed variant of this goes. Would be fantastic to get a version of this that tests as well as the original but without the need to resort to wood. Does anyone have a few spare PCBs?
I'm most interested to see how the box performs compared to @XMechanik's original measurements. I'm considering making a 'flat' test baffle to use in place of the chamfered one I already made. I'm guessing the steep chamfers do just about nothing but they do narrow the baffle a bit. It is my hope we can achieve at least similar performance.

I am also considering making a wooden cabinet to test in my own setup to ensure consistency. If I do that, I could also make a spare wooden baffle I could connect to the 3D printed cabinet to see if there is any difference. That's more effort though, and it is kinda hot outside these days. We'll see.
 
Here’s a little food for thought. Rather than a flat gasket, perhaps a round seal could also work. This is not a speaker, but it gives you an idea how I used an EPDM seal in a groove. Got the seal from Amazon. I figure if it’s good enough for a crank case it should be good enough for a speaker box.
IMG_3442.jpeg
 
Here’s a little food for thought. Rather than a flat gasket, perhaps a round seal could also work. This is not a speaker, but it gives you an idea how I used an EPDM seal in a groove. Got the seal from Amazon. I figure if it’s good enough for a crank case it should be good enough for a speaker box.

Do you mind linking me to that seal? I'm interested in trying a few things out. I have some neoprene sheet here I was going to try to have my printer cut at some point too.
 
Do you mind linking me to that seal? I'm interested in trying a few things out. I have some neoprene sheet here I was going to try to have my printer cut at some point too.
Sure. I chose 3mm and used an 1/8 cutter with groove depth slightly over half the diameter. Worked so well, I'm looking for ways to use the method!
 
First prints of the flexi-braces I'll test. I went with 10% and 15% honeycomb infill for these. I'm not sure the 'top' of the brace would offer much in the way of damping as currently designed but the lower section should (maybe). I can tell the difference between them but I have no idea what would be 'good'. I think I might go ahead and make 20% at least...maybe 25% too. This is 90A TPU by the way, so there are both softer and firmer types available. I don't think going softer would be beneficial, but 95A might be interesting and I think I have a little left from another project. Both test boxes are done as well, so I'm almost ready for some numbers!

IMG_0968.jpg



IMG_0969.JPG



IMG_0967.JPG


IMG_0970.jpg
 
Last edited:
Here’s a little food for thought. Rather than a flat gasket, perhaps a round seal could also work. This is not a speaker, but it gives you an idea how I used an EPDM seal in a groove. Got the seal from Amazon. I figure if it’s good enough for a crank case it should be good enough for a speaker box.
Also @technodanvan
Another interesting option is non-curing butyl tape or rope typically used in roofing and windows. It comes in round beads and flat tape and is the same butyl used in CLD sheets like Dynamat and Kilmat. Using it as a seal interface makes the junction a partial CLD damper even with the orthogonal joining and screws acting as short circuit couplers. It is significantly harder to disassemble depending on the surface area connected, but the standard "glue" tricks apply such as heat (within reason for the filament to handle, so none for PLA), using wedges along the edge, "piano" cutting wire / floss, etc.

Could also use it to CLD the braces.
 
Hey Aaron, that is a possibility as well for folks that don't want to deal with TPU. I need to get a stethoscope to see (hear) if I have any leaks but I really don't think I do. I have some other ideas for sealing and 'suspending' the woofer/tweeters that would require TPU though. I think I'll stick with 3D printed parts where I can.

Some assorted pictures of where I'm sitting. Completed speaker is the 10% infill cabinet with no bracing and...it mostly seems to measure fine. Once the baffle and rear were bolted on it really seems pretty darn solid. Impedance graph shown at the end. At least one of the blips is a standing wave since I have no inner treatments (no stuffing or foam of any kind). Still struggling to learn REW so I'm working on that. Also discovered my braces were too tight so I'm working on the redesign there.

IMG_0979.jpg


IMG_0980.jpg


Got a tall heat insert press to ensure I installed them straight. Not a requirement really but for $50 it came with a bunch of inserts too. Need to work on the dimensions for these things as I probably could have gone a little tighter tolerances for the 1/4-20 inserts. They hold fine as-is so no big deal.

IMG_0982.JPG


IMG_0983.JPG


IMG_0984.jpg


IMG_0985.jpg


IMG_0986.jpg


IMG_0987.JPG


IMG_0988.jpg


Only the woofer is hooked up here as I don't think the tweeter (or crossover) is necessary for this part.

IMG_0990.JPG


DATS impedance reading. Tuning for the empty box is about 60 Hz, a little higher than what I was shooting for, but it IS empty after all. We'll see what bracing and a little stuffing do at some point. Blips were at 280 Hz and 680 Hz, or thereabouts. The 680 Hz blip is probably a standing wave at 10", which is roughly the interior height of the box. The 280 Hz blip might be a port resonance or something. Unsure, open to thoughts.

260722 - Impedance - 10% - Empty.png


Edit: Regarding the ~280 Hz blip. I see there is something going on in the low 300 Hz range in Post #1 and in Amir's review. Might be the same thing?
 
Last edited:
Admittedly I do not know how to use REW very well - I'm barely a novice, so please bear with me. Part of the rationale for tackling this project was also learning the software, so if anyone has some pointers I'm all ears. I don't know how to apply gating so that's something that would help a lot.

Green: QNF response of the woofer, mic distance around 8-12" away.
Red: Nearfield response of the port.

10% No Brace Port + QNF Woofer.png


QNF + Distortion. Probably should have cut off more of the left side of this one.

10% No Brace QNF Woofer Distortion.png
 
Admittedly I do not know how to use REW very well - I'm barely a novice, so please bear with me. Part of the rationale for tackling this project was also learning the software, so if anyone has some pointers I'm all ears. I don't know how to apply gating so that's something that would help a lot.

Green: QNF response of the woofer, mic distance around 8-12" away.
Red: Nearfield response of the port.

View attachment 546984

QNF + Distortion. Probably should have cut off more of the left side of this one.

View attachment 546985
Is the port port length a multiple of the port diameter? The second amplified resonance doesn't look too good. The main resonance frequency has 2x the port length as wavelength. The flare only helps with chuffing but not by reducing resonance. The flare actually 'reduces' the port length. This is maybe why you end up having a higher tuning frequency.

P.s. can you show the graph down to 20hz, so we can see the port output?
 
Is the port port length a multiple of the port diameter? The second amplified resonance doesn't look too good. The main resonance frequency has 2x the port length as wavelength. The flare only helps with chuffing but not by reducing resonance. The flare actually 'reduces' the port length. This is maybe why you end up having a higher tuning frequency.

P.s. can you show the graph down to 20hz, so we can see the port output?

Hey thanks for the reply! The port diameter is variable in this case as I used the port optimizer to model it. The minimum diameter is 38mm (center of port) while the ends are 57mm. Length is 125 mm...so the answer might be yes, in this case? 57mm is approximately 1/2 of 125 mm. This was intentionally modeled at 55 Hz by the way, based on some earlier comments in this thread that suggested the standard port was tuned a little low.

I need to reread augerpro's studies. I could try printing this with holes in it, my understanding is that sometimes helps with resonances. I could also print some or all of it with TPU to make it a bit flexible - I think a manufacturer does that for some reason. Again, need to read up on it.

Below is a revised graph showing the port (red line) and the QNF response for both the 10% infill cabinet and 25% infill cabinet. I did not go out of my way to match volume or anything between tests, but I think the takeaway is that the woofer response is pretty unaffected by infill density, even without any bracing present. There might be a slight difference around 30 Hz but that's pretty quiet anyway.

I would have shown the 25% port response too but I apparently didn't save that data this morning and cannot retest until later today. I do not believe there was anything remarkable in it though.

Combined QNF + Port 10% and 25%.png


Something that may have improved though is distortion down low, but it's slight even when we're talking 100+ dB. This is probably where I should have taken a few tests at lower volumes and tried to level match a bit. I also don't think my mic interface is set up correctly for distortion measurements, so there may be some from that. It should be low though and probably inconsequential for this woofer.

Distortion: 10% Infill cabinet, no bracing.

10% No Brace QNF Woofer Distortion 110dB.png


Distortion: 25% infill cabinet, no bracing (note the response is a few dB louder than the above, so naturally some distortion is higher as well)

25% No Brace QNF Woofer Distortion 110dB.png


Edit:

Amir's port measurements also show two peaks, though at a bit lower frequencies and the first is not nearly as pronounced as I'm seeing.



Mechano23 Klippel Port.png


Edit again:

Amir's distortion measurements may also be useful. Doesn't really look like the port resonance shows up much, but the bump I have in the impedance plot occurs here at just over 300 Hz (at least, I think they're related).

Mechano23 Klippel Distortion.png
 
Last edited:
Back
Top Bottom