Building Valkyrie: A compact high-output speaker with smooth directivity

[Aijan]

I previously built Valeria, a compact bookshelf speaker using pro-audio drivers, namely a 6.5" mid-bass and a 1" horn-loaded compression driver. The dynamics offered by this small speaker are impressive, if I may say so myself. It’s currently serving as the center channel in my home theater.

Lately, I’ve been thinking of building speakers with better bass response (a subwoofer will still be necessary though) and improved horizontal and especially vertical directivity. This necessitated using an 8" mid-bass driver, a compression driver that rolls off below 1 kHz, and a horn that can control directivity better than the B&C ME10 I used in Valeria.

The mid-bass driver I selected for Valkyrie is the Lavoce WAF082.00. It has a 2" voice coil, a paper cone, and a cast-aluminum basket. The compression driver, Lavoce DF10.171K, features a 1.75" voice coil and a polyimide diaphragm. As for the horn, I decided to use the RCF HF101, a cast aluminum horn with relatively compact dimensions, which is important to minimize driver-to-driver spacing and consequently improving vertical directivity.

I designed a compact cabinet with internal bracing that couples to the drivers, making the drivers part of the brace and reinforcing the front and rear sides of the cabinet.

The braces were made of 16 mm beech plywood. I attached the braces to the cabinet walls with 3M VHB tape instead of regular wood glue. The soft layer in between acts as constrained-layer damping to reduce resonances.

I decided to use a 3rd order LP and 2nd order HP for the crossover along with two notch filters for the compression driver. The crossover frequency is approximately 1 kHz.

It’s critical to add absorption material to manage the reflections and standing waves that form inside the cabinet. I lined the walls with felt and placed extra-thick 60 mm pyramid foam on top while leaving enough space for the drivers.

Pictured below: Valkyrie standing next to Valeria. Valkyrie’s dimensions are 24 x 42 x 21.5 cm (WxHxD), while Valeria measures 19 x 33 x 22.5 cm (WxHxD).

So, how is the sound? By itself, the bass is lacking, as expected. However, as soon as you add a subwoofer to the mix, the missing low-end is filled in, and what you get is effortless dynamics throughout the entire frequency range assuming the subwoofer can keep up with Valkyrie at higher volumes.

In listening tests, it’s immediately apparent that, compared to Valeria, the bass is considerably louder and deeper below 200 Hz. Otherwise, they’re well matched.

The impedance plot looks clean, with the exception of a resonance at around 300 Hz. This appears to be inherent to the mid-bass driver’s construction, as the free-air impedance measurements of the driver show it as well, albeit not as pronounced. The nominal impedance is 6 ohms dipping to 4.4 ohms at 210 Hz.

The frequency response is within ±3 dB from 100 to 20,000 Hz (quasi-anechoic measurement). The mid-bass to compression driver crossover frequency is quite low at 1 kHz to help make vertical directivity well behaved. A subwoofer is necessary for good bass response.

The horizontal directivity is ±45 degrees nominal, smooth down to 800 Hz. Below that frequency, the directivity quickly becomes omnidirectional. Thanks to the uniform directivity, the frequency response can be equalized without any issues.

The vertical directivity is ±35 degrees nominal, and it’s smooth down to a respectable 1200 Hz. As long as your ears aren’t below around 15 degrees of the horn’s center axis, the timbre of the speaker shouldn’t change noticeably regardless of where you sit or stand.

Last but not least, a complete build video is available below. Please let me know what you think.

 

[thewas]

Great engineering, can be used also with subs as satellite for high SPL home theatre or even small PA.

 

[Lars Risbo]

love the bracing of the motors

 

[Tim Link]

Looks great. I see that horn has a diffraction throat. My inclination is to avoid those, although maybe that's because I've listened to Earl Geddes too much.

 

[Aijan]

Looks great. I see that horn has a diffraction throat. My inclination is to avoid those, although maybe that's because I've listened to Earl Geddes too much.

Thanks! Have you noticed a significant difference between the sound of regular constant directivity horns and waveguides with oblate spheroidal throats?

 

[Sal1950]

Awesome build, I'm always impressed by folks with the woodworking craftsmanship to design and build
these projects. Measurements suggest a very linear sounding result.
Congrats!

 

[Tim Link]

Thanks! Have you noticed a significant difference between the sound of regular constant directivity horns and waveguides with oblate spheroidal throats?

I've never heard an oblate spheroid horn. I've also never listened to a diffraction throat type horn. So honestly I have no experience based opinion about it.

 

[Head_Unit]

love the bracing of the motors

I was thinking the same, especially the sticky-looking damping plug.

 

[Sal1950]

love the bracing of the motors

Me Too
Did someone say Valkyrie ?

 

[Wolf11Man]

It would be interesting to see the the response and impedance plot with less foam damping material. To me it look as if the speaker may work better in the 70-250 Hz region with less stuffing of the foam type. Usually just covering the walls is enough and the cabinet dimensions and braceing do not look critical.

 

[Lars Risbo]

damping material has the least effect on the walls (zero velocity). taking down standing waves in the 200Hz region requires heavy foam in the interior of the box (pressure minimum and velocity max)

 

[Wolf11Man]

In theory you are absolutely right, but this is no large box and the internal braces diffuse. I expect to see more output of the vent with less foam inside. Damping material on the wall reduces reflections without absorbing too much energy. Just stick your head inside a large enough box and compare with and without the walls covered. Then you will understand what I talk about. Too much damping can reduce the wanted effect of the resonator.
This is not scientific talk but the result of simple experience. In the end the measured result is what counts. Without measuring you can not know if there is enough, too little or too much inside the box. Last, I don't think foam is a good fill for the volume of a vented cabinet. If you use a sub, closing the vent and filling the whole thing with something like Rock Wool may work better. The impedance curve will show the result quite well.

 

[Lars Risbo]

In theory you are absolutely right, but this is no large box and the internal braces diffuse. I expect to see more output of the vent with less foam inside.

Damping material on the wall reduces reflections without absorbing too much energy.

Damping requires absorption of energy.

Just stick your head inside a large enough box and compare with and without the walls covered.

And then compare to placing the damping material and n the interior.

. Too much damping can reduce the wanted effect of the resonator.

true but this more about placing the damping material the right places. if you keep an unbroken air channel between the driver and port then the added port losses are minimal. The port requires high air velocity so damping material must be kept out of the direct path. all other places you can fill up. This combines killing of standing waves without destroying the desired port resonance.

This is not scientific talk but the result of simple experience. In the end the measured result is what counts. Without measuring you can not know if there is enough, too little or too much inside the box. Last, I don't think foam is a good fill for the volume of a vented cabinet. If you use a sub, closing the vent and filling the whole thing with something like Rock Wool may work better. The impedance curve will show the result quite well.

yes indeed. I do model fitting on the measured impedance curves and extract port loss Q to check the bass alignment.

you can see more here : https://purifi-audio.com/blog/tech-notes-1/spk16-reference-design-12

 

[Aijan]

After unscrewing and reattaching 16 screws four times (fun times), I measured the impedance with and without foam, keeping the felt in place. It appears that foam effectively smooths the internal resonances between 300 and 400 Hz with minimal loss of efficiency. It may not be obvious from the pictures, but there isn't actually that much foam inside.

 

[Wolf11Man]

Your foam seems to work quite well, as the resonance peaks are a little lower in frequency than without. I had so called acoustic foam that reduced the volume and pushed it up, so your inside damping looked over the top to me. I'm surprised it works that well.
I thought the bump around 160Hz could be less pronounced with less foam. Anyway, this is why we measure speaker constructions, to see what may work better and to proof we are wrong or right with our ideas. Sorry for making you do something useless...

I prefer fiber materials, like acrylic or stone wool. Often both, so the acrylic wool works like a filter for Rock Wool particles.
If, just for an experiment, you put different (small) amounts inside a vent, you will be surprised about the changes in the impedance curve even from very tiny pieces. I did a lot of experimenting while trying to understand the Dynaudio Variovent and its stuffing, which is a controlled leak in theory.

 

[Aijan]

Sorry for making you do something useless...

No problem. It's just that I think the horn and the mid-bass driver could have done with less mounting holes.