Does rounding edges matter if you are using waveguide?

[matsurus]

Hi all.

My 2-way speaker project has come to the point whrere i have bought drivers, tested them and know about size of the speaker.
(SBacoustic tw29txn-b & AT 18h52)

I will be using wavequide (jantzen 165mm) on tw29txn-b tweeter, and little selfmade wavequide on 18h52 (diy 200mm)

Thing is i would rather make from aesthetical reasons a sharp edged box, but i know that in "normal" nonwavequide elements rounding edges is benetifical.
I read that 1 inch is minium for it to have positive effect, and even more 1,75 inch is better (more depth in soundstage, reference from Lynn Olson)
----------------------------------------------------------------------------------------------------------------------------------------------------------------------

-So what is your experience? Do i need rounded edges with wavequides or not?
I feel that because wavequides shape sound direction edgerounding would be not needed, but i don't really know..

 

[AllenB]

It depends on how much you use the waveguide beyond it's edges. That is part to do with the waveguide design and part to do with how low you go.

 

[matsurus]

Hi Allen, thanks for the answer. Can you elaborate a little for a newbie?

So way i understand your response: Sound also travels beyond wavequide because of diffraction.
How much it travels is up to shape design of wavequide.

So in practical terms, yes rounding edges still would be beneficial with wavequide, if i want to go for "as good as possible".

-Do you think (in your own experience) that in this case with bookshelf there would be noticeable difference with straight edges vs rounded?
(Edges to baffle-edge will be 25mm straight, after wavequide edge, or 25mm rounded)

-What do you mean low? How low the crossover point will be for tweeter?

 

[AllenB]

There are many factors so I can only outline them. Measurements might be able to help an experienced builder. Yes, diffraction.. but if a waveguide is undersized it will spill over more. Some design to fit the box, rather than the acoutsic requirements. This thread is about a different issue but gives a little insight into problems with the waveguide size being too small. - https://www.diyaudio.com/community/threads/is-it-best-to-use-a-non-axisymmetric-waveguide.375799/

It can also be a matter of the way the edge handling is implemented, the shape. Also yes, I did mean how low in frequency.

Don't forget that the waveguide is supposed to give directivity, and if it spills over too much then it may not be doing that too well. In any case, there's always a lot of overlap.

 

[matsurus]

Thanks, i will look about it.

 

[YSDR]

In short, if the speaker baffle is larger than the waveguide, then diffraction effects can starts to present. Here is a study of the effect, and examples of how can be treated:
https://heissmann-acoustics.de/en/kantendiffraktion-sekundaerschallquellen-treiberanordnun/

 

[matsurus]

Well, i have decided to do test panels, one for rounded edges and one for straight.

I have not heard (yet) myself, but i looked trough this issue of similar comparison between sharp wavequide and baffled edges:

One user reported he owned older JBL 530 that has big wavequide for a standmount, and built joseph crowe NR:1159 speaker and compared them each other and found them similar in sound:
" I dont think the 1159's are much of a sonic upgrade. The 1159 has a slightly mellowed top end like the 530, actually a bit more mellow. I actually slightly prefer the 530 over-all as it has better bass, equal midrange, similar highs and a slightly more pleasing quality to me personally that I find hard to describe."

Interestingly, new version JBL 630 is similar construction as 1159, so this would seems to show that rounding edges would be benetifical with this "small" size wavequide.

But i will build test front baffles to hear it myself..

 

[GM]

It depends on how much you use the waveguide beyond it's edges. That is part to do with the waveguide design and part to do with how low you go.

+1, though there's eigenmodes between any pair of acoustic impedance discontinuity and why I went with at least baffle width open cell foam horn extensions way back when to also deal with the 'round the bend' (side, top wall) diffraction's.

In retrospect it always falls back to Olson's seminal work.

 

[hifijim]

As Allen B has said, it may or may not matter... Here is one example where baffle edge treatment has a big impact on a waveguide tweeter.
https://www.diyaudio.com/community/threads/purifi-waveguide-project.394174/post-7278797

j.

 

[PKAudio]

From my experience, WG + 20mm chamfering of the baffle provides perfect results if the width of the baffle is close to the dimension of the WG. In the attachment it is TXN + Jantzen WG, baffle 230mm, chamfering 20mm.
I did also measurement in 330mm baffle with 20mm chamfering and there was mild diffraction signature so in this case 30mm roundover would be better solution.

 

[mabat]

-So what is your experience? Do i need rounded edges with wavequides or not?

This is all pretty complicated stuff, but you can try it easily with a BEM simulation, prior to building anything:

https://www.diyaudio.com/community/...-design-the-easy-way-ath4.338806/post-7353694
https://www.diyaudio.com/community/...-design-the-easy-way-ath4.338806/post-7351271
https://www.diyaudio.com/community/...-design-the-easy-way-ath4.338806/post-7354055

I would say that a rounding helps in any case. The best rounding actually being a free-standing waveguide - perhaps unintuitive but true.

- It can be simulated even for an existing waveguide, if you already have one: https://www.diyaudio.com/community/...-design-the-easy-way-ath4.338806/post-7228138

 

[NeonDriver]

In my experience, sharp speaker corners are a minor factor for sound quality. Just look at all of the sharp corners present in most rooms. I also like sharp rectangle boxes and my speakers sound fine.

I would highly suggest building with sharp corners if you prefer that style. After all, style is likely to be a major factor for most people.