Spherical speaker design help needed

[Neko88]

Hi, I'm new to this (rather amazing) forum and only recently became interested in designing speakers after stumbling across the various ingenious Ikea bowl designs that people have made. I made a similar speaker from solid ash and black walnut (photos below, not very well taken, and without a proper speaker stand so they don't really do it justice).

Anyway, it sounds really good to my ears (i.e. not dissimilar to my dad's £500 Mission towers, apart from a slight tinniness. However, using my sequencer's EQ to cut by 6db at about 1500Hz removes the tinniness, and vastly improves the sound. Is this likely to be caused by the design of the enclosure?

Would making a basic notch filter, like the designs I've seen elsewhere on this forum get me a similar result without having to play all my music through Logic's EQ? Is there a better way of establishing exactly where this spike is, or should I just continue going by ear? I'm pretty happy with the post-EQ result I currently have.

Would anyone be able to help me establish the values of the components I would need? Do i just solder them together and stick them in front of the negative terminal on the speaker? I'm really struggling to make sense of the information out there. I'll try to upload a screenshot of the EQ curve I'm trying to achieve.

thanks very much - and sorry if these are newbie questions

[planet10]

What is the inside diameter of the sphere? 23 cm? Is that an FR125?

dave

[Neko88]

Hi. It is a shielded CSS FR125. The inside diameter of the sphere is tricky - in the cross section below the external diameter of the light grey circle is 28cm, and the internal diameter is 26cm -however, I left those internal ridges in the sphere as an attempt to break up the soundwaves and avoid standing wave problems I had heard people saying might be a problem with spheres.

After I had made the speaker, the only way I could think to ascertain the actual volume was by filling it with small bean-bag beans poured out from a litre measuring jug. The internal volume seems to be seven and a bit litres. The speakers don't really sound too boomy in the low end, so I'm assuming that this volume is OK for these drivers.

Thanks for replying

[cogitech]

Originally Posted by Neko88

-however, I left those internal ridges in the sphere as an attempt to break up the soundwaves and avoid standing wave problems I had heard people saying might be a problem with spheres.

I find this to be counter-intuitive. I have no idea about standing-wave theory, but I would have assumed the flat surfaces would make the issue worse, not better. I could be completely wrong, though.

[planet10]

Originally Posted by Neko88

Hi. It is a shielded CSS FR125. The inside diameter of the sphere is tricky - in the cross section below the external diameter of the light grey circle is 28cm, and the internal diameter is 26cm -however, I left those internal ridges in the sphere as an attempt to break up the soundwaves and avoid standing wave problems I had heard people saying might be a problem with spheres.

7 litres is a good volume for FR125

A sphere will usually have a single resonce related to the diameter and can be ameriolated by putting a good wad of damping at the centre. By squaring things off you have distributed them a bit, but it is most likey still the problem. The dimensions are very close to the frequency you ar ehacing to cut. I would start by suspending a dense bit of damping in the centre and at the walls, with less dense fluff in between.

You might want to drill an array of small holes in the back & damp them aperiodically. Similar to below. It will decrease pressure in the box that is supporting the resonance.



dave

[norman bates]

yes, you have to address the single (big) standing wave inside the enclosure.


And there is baffle step.
Notice the smooth roll-off of the sphere (like yours, but yours is smaller ?)
True Audio TechTopics: Diffraction Loss

and also fletcher munson curve (why a smily face sounds good to our ears on a 10-band eq).
Fletcher-Munson Curve Explanation
Fletcher?Munson curves - Wikipedia, the free encyclopedia


Just some food for thought.
Norman

[Neko88]

Dave - I've tried drilling holes in the back and damping them, but it seems to make little or no difference. I've also tried damping with lambswool, with a dense wad tied around the y-shaped brace in the centre of the sphere (sorry, forgot to mention this bracing wasn't on my cross-section) and more teased out wool elsewhere. Even if I radically overstuff the enclosure to the point that the music sounds dead, there's still quite a strong nasal-sounding resonance.

I had hoped that the brace running through the centre would be enough to avoid standing wave problems, but seemingly not. Perhaps instead of this 3-pronged propellor shaped brace, I should have left a wall between the two halves of the sphere, then drilled lots of different sized holes in it with a hole-saw?



Norman - Thanks for the link, but I found it swiftly lost me when it got more technical. Is it suggesting that to make up for the bass that is lost behind the cabinet (due to being less directional than higher frequencies), that the mids and highs should be cut using a passive circuit, until they are brought down to a compensatory level? How would I go about designing such a circuit? Would it be similar to the notch filter idea, just set much, much broader?


I'm assuming that getting rid of the standing wave internally will give a better result than trying to EQ away the problem?

[planet10]

Originally Posted by Neko88

Even if I radically overstuff the enclosure to the point that the music sounds dead, there's still quite a strong nasal-sounding resonance.

The driver cut-out is well champered on the back-side?

dave

[norman bates]

hiya, it says how to one the true audio site.
but full range speakers (unless shorting ring or copper cap) have a climbing (changing) impedence. The you get into zobels (another tricky subject).

You have something nice there.
Sitting 3' away, do you need boost ?

Norman

[Neko88]

Dave - it wasn't, but it is now, and doesn't seem to have made much difference, if any. There's still quite a nasal peak.

Norman - thanks, at least the speaker looks good - and hopefully if I keep working at it, it will sound good, too

I used my sequencer to approximate the EQ curves suggested on that page, and the speaker sounded much better balanced, less bright and more bass, a much more natural sound. I'll definitely look into compensating for the baffle loss, but it's a struggle getting my head round the technicalities of that article. Are you saying that the circuit design given there won't be compatible with my full-range speaker?