A loudspeaker project, finally finished.

[Jonte Knif]

I realize I don't contribute much to this forum. So excuse me for showing what we have been dealing with lately.
This is a 3-way speaker with active X-over between bass and the rest. Bass amp is integrated to the speaker. Active X over is a separate box which includes also a 2 band parametric EQ for taming some room problems. X-over frequencies are 300 and 2k. The lower X-over is acoustically 1st order. With the proper placement of acoustic centers this creates phase linear response in that range.  Mid and treble (ribbon from RAAL) are placed in a wave guide. The ribbon, which is fairly long, has acoustic lenses to broaden the vertical radiation pattern.

We finished the speaker only barely in time on Wednesday, and launched it at Helsinki Hifi fair yesterday.

I can tell more about the design if anyone is interested.

-Jonte

 

[rob_gould]

My heart does a double beat when I see Jonte Knif has posted on the forum ;-)

The knifonium still sets a bar that's never been bettered in certain ways imho.

I'm not much of a HiFi guy Jonte, but these speakers look extraordinary and I'd love to hear more about them  ;D

 

[scott2000]

Wowza

 

[shot]

Jonte is back! 
It's always exciting to have you here Jonte! You should come by more often!

I'm wondering why you chose to have upper crossover at 2k?
In 3 way systems manufacturers choose to go higher with crossover to avoid issues in this critical frequency range.
I was under assumption that choice of 3 way over 2 way system is actually to avoid crossovers at around 2k that 2 way systems have. 


Luka

 

[PRR]

I'n not the designer, and am quite rusty.

> why you chose to have upper crossover at 2k?

The two 5" speakers are quite far apart vertically. While they sum on-axis, the off-axis (down and up) are starting to comb-filter by 2K, and could be quite snarly by 5KHz.

 

[Jonte Knif]

PRR is right. The upper X-over should in this construction be as low as possible. The compromise is made between distortion from the ribbon and vertical radiation pattern.  Another reasons to not go above 2,5-3 kHz with the X-over are caused by the rigid membranes. Horizontal directivity gets too high and the membrane break up mode at 10k will raise 3rd harmonic distortion from the mid driver at 3,3k.
Just as a side note: It is funny how the "most sensitive range of hearing" is used in the industry.  While the threshold of hearing is most sensitive at 4kHz, the phase or impulse "coding" of hearing gets pretty bad above 1kHz. So, manufacturers can choose what ever is most convenient to explain the compromises they have done. Loudspeaker design is nothing more than the ultimate exercise in compromises 

 

[ruairioflaherty]

Loudspeaker design is nothing more than the ultimate exercise in compromises 

Absolutely true.

Speaking of compromises what drew you to the WMTMW layout?

 

[Jonte Knif]

Some reasons:
To fight the floor bounce irregularities and vertical room modes
To get more SPL with the same distances between elements
To fight the tendency of perceived source height to vary at different frequency ranges

Some problems it causes are obviously size, cost and complex vertical radiation patterns.

 

[Spiritworks]

The woodworking, and the visual presentation, is astounding. If they sound as good as they look, my oh my...….

 

[JohnRoberts]

PRR is right. The upper X-over should in this construction be as low as possible. The compromise is made between distortion from the ribbon and vertical radiation pattern.  Another reasons to not go above 2,5-3 kHz with the X-over are caused by the rigid membranes. Horizontal directivity gets too high and the membrane break up mode at 10k will raise 3rd harmonic distortion from the mid driver at 3,3k.
Just as a side note: It is funny how the "most sensitive range of hearing" is used in the industry.  While the threshold of hearing is most sensitive at 4kHz, the phase or impulse "coding" of hearing gets pretty bad above 1kHz. So, manufacturers can choose what ever is most convenient to explain the compromises they have done. Loudspeaker design is nothing more than the ultimate exercise in compromises 

I know enough to know what I don't know about loudspeaker design  (I haven't made that kind of sawdust since the 1970s).

The midrange is important because human hearing is optimized for human speech range. Phase/amplitude errors in that region are more easily perceived when listening to familiar (known) vocals. A lot of loudspeaker distortion is not very audible, in some cases LF distortion is almost considered normal or preferable.

Indeed loudspeakers are the hard part of the sound reproduction chain, and don't get me started on stereophony.

Pretty boxes, good luck with the show.

JR

 

[boji]

Thanks for posting.  The wood that blends into the waveguide is very pretty. Have you any frequency response graphs you're willing to share?

 

[Jonte Knif]

Since the  local big anechoic chamber is currently being refurbished we don't have 100% reliable measurements. I can how ever share this picture from last Sunday, when we tried to get some useful data in sort-of-free-field environment.

 

[JohnRoberts]

Don't spend too much time inside the anechoic chamber... you can actually hear yourself think.  ;D

JR

 

[boji]

Ok you're nutz! lol did you have another scissor lift for the mic?

 

[Jarno]

Since the  local big anechoic chamber is currently being refurbished we don't have 100% reliable measurements. I can how ever share this picture from last Sunday, when we tried to get some useful data in sort-of-free-field environment.

LOL, that is an interesting measurement setup!

 

[gyraf]

Around here, we usually just fly the mic and have speaker pointing upwards - easier on lifting..

 

[JohnRoberts]

Around here, we usually just fly the mic and have speaker pointing upwards - easier on lifting..

Boundaries near both are important. Of course you can factor half-space into SPL measurements, for frequency response free field is probably useful.

JR

PS: I found a reprint from 1929 bell laboratories record article (in my dad's notebook) showing an apparatus with a rotating microphone for indoor speaker measurements (Back then they called microphones transmitters). No doubt to average out room modes.

 

[boji]

https://www.youtube.com/watch?v=7pNy-paH25o

 

[scott2000]

;D that's funny.... And I didn't even listen..

 

[Jonte Knif]

Around here, we usually just fly the mic and have speaker pointing upwards - easier on lifting..

You will not get proper representation of free field in the lower midrange, where diffraction around speaker edges has a role and the distance from ground to drivers is substantial.  Could be useful with soffit type mounting though.

In our case a setup with the speaker laying on ground and mic placed on ground too will be useful up to perhaps 400Hz. No need to fly the mic. A real half space measurement.

How ever, this is not enough, because the critical range in this speaker is around its lower X-over, which is a wide area due to the 1st order slopes.