3d Printed Speakerenclosure 2 Way

[Salt]

It's not necessarily the tweeter's fault. I see this behaviour using REW with Asio4All and 44.1 kHz sampling rate.

 

[Ktacos]

I do not get this behavior at all using asio4all, or at 44.1khz or a combo of both. I have never even heard of such a thing happening. It really looks like OP just didn't filter the response correctly. It appears the builder did not take into account the additional gain in the lower response of the tweeter given by the waveguide.

It seems pretty clear the builder just put a simple high pass filter on the tweeter, but more response shaping is required.

Here is the builders raw tweeter response traced and filtered in two ways. One with a simple highpass (green) and a more appropriate filter network (red). As we can see a single highpass is not sufficient and lines up with the response the builder is getting.

Here are the filter tasks to give an idea of what the difference in filters is.

To give some advice to the builder, an easy way to design filters is to use REW's EQ function. You can simply create a target curve for your desired xover slope and response. Here is an example of a 1500hz 2nd order filter task.

Here are the filters it generated and the predicted response. This is just an example, you can also linearize the drivers response down lower and apply your high pass filter later. Hope that helps.

 

[D!sco]

...Or it's the in-room response and not a nearfield, and this is appropriate roll-off for the sound power? It's also likely that the waveguide is contributing, but EQ and tuning can determine the end result before a crossover is built.

If OP wants any advice on optimizing responses with REW I'm here for it. Want to share your filters and the full VCAD/REW screens?

 

[Ktacos]

...Or it's the in-room response and not a nearfield, and this is appropriate roll-off for the sound power? It's also likely that the waveguide is contributing, but EQ and tuning can determine the end result before a crossover is built.

If OP wants any advice on optimizing responses with REW I'm here for it. Want to share your filters and the full VCAD/REW screens?

OP stated his measurements were from 25 inches away on axis. There would not really be any roll off from room interaction/distance when measured at that distance. Even if it was farfield, there would not be a such a sharp droop in HF.

The info I've provided makes it pretty clear what the issue is. Having worked with many waveguides it was easy to deduce.

Want to share your filters and the full VCAD/REW screens

I did not save the vcad filters, they were quick and just an example. The REW screenshots are basically the full screenshots, there's not really any other info to share. A filter set is generated in the last screenshot OP can try, however they are better off generating their own with the steps and info I've provided.

 

[Terplover]

Sorry, was pretty busy the last few days.

Some strange things are going on – the difference between L and R speaker is quite big.

I measured the left one again today with a better setup and hope to get to the right one at the weekend along with some troubleshooting.


I’m almost 99% sure I disabled every filter I could.
Attached is the latest REW file from the left speaker including the alignment.
I wanted to share a corrected full measurement (woofer + tweeter), but my MiniDSP didn’t take the filters for some reason.

Measured with right Calibration File on Tweeter Axis.
Used 48 kHz and Java Drivers
100% without any Filters.

30cm on Axis Tweeter. Without any Filter or Crossover.

 

[Terplover]

Doesnt look that bad. 1/2 finished. measured 30mm on tweeter axis

 

[Terplover]

Disassembled the second speaker and sanded the tweeter mounting area again.
Voila – now I’ve got two almost identical speakers running with the same filters. I’m calm now

Crossover turned out nearly perfect with just LR 24/oct at 1500 Hz for tweeter and woofer. Added a high-shelf to compensate for the horn’s behavior, the rest I just fine-tuned with EQ.


Some more tweaking and the optical finish will follow, but the project is basically done.
Was a fun build – and I definitely learned way more than I expected when I first started this thing.

 

[Fabsen]

Looks so nice!
Would it be rude to ask for the printing files? Or any advice how to get started with Fusion?

 

[Terplover]

Looks so nice!
Would it be rude to ask for the printing files? Or any advice how to get started with Fusion?

@Barnaby1 Asked for the files too.

Updated them slightly so you can use screws to clamp the pieces together. Don´t overtighten them or Layers will break.

6x M5x50-60 screw
6x M5 Nut
12x10mm washers

Connector Part at the back is fully closed

Would be happy to see some pictures if you finish them

 

[Terplover]

Looks so nice!
Would it be rude to ask for the printing files? Or any advice how to get started with Fusion?

Fusion isn’t that hard to use it’s just hard to really master.
Basically, you just draw sketches and then raise them into 3D. With that approach you can model almost everything in this project, except the horn.
I honestly don’t know how to calculate or generate something like that either, so I just copy-pasted it.

There are plenty of YouTube tutorials to get you started, and it’s free to use so just give it a try

 

[Ktacos]

I may pursue this print myself, I'd be curious if you could share an impedance measurement if you have that capability.

 

[Terplover]

I may pursue this print myself, I'd be curious if you could share an impedance measurement if you have that capability.

Unfortunately, I haven't done any impedance measurements yet, and don´t know exactly how to do it.
I'll definitely keep it in mind if I find the time or continue working on the speakers.

 

[Yevhen]

The tweeter filtering looks wrong, unless you want it to drop down 10db in the top end.

Taking into account the amount of absorbing materials on the walls and ceiling, I would indeed aim for the flat response in HF, no tilt.

 

[Digital_Thor]

These were at the Austrian hifi exhibition... fully 3D printed 2 way for 2k(think it was euro).
Sounded good. Chesky Audio.

 

[somebodyelse]

These were at the Austrian hifi exhibition... fully 3D printed 2 way for 2k(think it was euro).
Sounded good. Chesky Audio.
View attachment 538083

Is that the LC1 or LC2? Erin measured the LC1 and it was terrible. They're raw prints too - no surface treatment, visible banding and ringing. I'd expect better all round for 2k.

 

[Redacted]

Award yourself the still-unofficial "ASR Developer Badge"!

 

[Digital_Thor]

Is that the LC1 or LC2? Erin measured the LC1 and it was terrible. They're raw prints too - no surface treatment, visible banding and ringing. I'd expect better all round for 2k.

Can't remember, but I remember clearly the designer saying that he knew it did not measure great, but it still sounded entertaining and several people liked it, my friends included.
Further, hearing the Dutch C6, I was initially a little disappointed with the 'open' midrange and tweeter/midrange integration, even though it properly measures much better. Room and setup means a lot, so exhibition sound is tricky. They both had impressive bass for their size though, with Dutch being definitely better in the lower octave.
Privately I use KEF and think it's better overall than both mentioned.
I believe Chesky Audio will develop over time.