Designing a dome faceted loudspeaker

Show off your FreeCAD projects here!
User avatar
DeepSOIC
Posts: 6953
Joined: Fri Aug 29, 2014 12:45 am
Location: Saint-Petersburg, Russia

Re: Designing a dome faceted loudspeaker

Postby DeepSOIC » Wed Mar 13, 2019 8:54 pm

The project got a bit delayed because weekdays came, and also because I have sent out nvme ssd from my laptop to be cloned onto new ssd, and IT service companies local to bumblef. called Peterhof are seeing nvme ssd and their eyes pop out from wondering what the heck to do with this (not quite) very new stuff. Apparently I'm the first one in the whole town throwing this kind of task at them.

I have wired up and sealed the speaker.

As is, it sounds quite horrible, with no bass and a noticeable resonance at circa 300 Hz. That is totally expected, for a lot of drivers and a very small enclosure volume.

But I know, how to deal with it: "feedbass!". So I quickly stuck a pressure sensor onto it, and connected it to my amp.. Hoping for something epic.

The bass came to exist. But the sound was yet again horrible. I had trouble understanding, what exactly is horrible about it. So I went ahead with random troubleshooting, in hopes to find something.

That something was found. The wall I stuck the sensor to was resonating sharply at 1250 Hz. That resonance was picked up by the pressure sensor, causing some spike in the frequency response.

Now with the resonance fixed, it was sounding much better. But still, it was sort of muffled, not good at all. And by playing with eq in the smartphone, I found that dragging up "7k" band was helping a lot. This got me thinking, why the hell 7k. The driver was very flat up there when I was measuring it alone. This must be coming from the structure of the dome itself. Fascinating.

I think what's happening is that the regular pattern of drivers creates a standing wave near the surface of the sphere. That wave keeps it from radiating into space. That sounds plausible, but I hope to get more insight by measuring frequency responses. But I have to get my ssd back for that.

So I hope for more updates. But for now, it seems, the second speaker for completeing the stereo pair is not going to be built.
Jee-Bee
Posts: 1950
Joined: Tue Jun 16, 2015 10:32 am
Location: Netherlands

Re: Designing a dome faceted loudspeaker

Postby Jee-Bee » Thu Mar 14, 2019 9:12 am

for measuring sound response check https://www.roomeqwizard.com/ this.
User avatar
DeepSOIC
Posts: 6953
Joined: Fri Aug 29, 2014 12:45 am
Location: Saint-Petersburg, Russia

Re: Designing a dome faceted loudspeaker

Postby DeepSOIC » Sat Mar 16, 2019 4:47 pm

Jee-Bee wrote:
Thu Mar 14, 2019 9:12 am
for measuring sound response check https://www.roomeqwizard.com/ this.
thanks, I know.


Now, finally, my laptop is back (they failed to clone the drive, but at least nothing was screwed up). Here comes the science!

This is the frequency response of the drivers I used, in an infinite-ish baffle (embedded into room wall).
19.png
19.png (233.53 KiB) Viewed 733 times
Don't pay attention to 76-hz peak and 1.6K downward peak on the responses, they are because of poor mounting. Everything else, I believe, is pretty reliable, except for no way of correcting for room response.

One plot is for on axis measurement, which represents the typical "sweet spot" listening. The other one is at 45 degrees off of the axis, which (I hope) represents overall output energy in all directions. The second one is probably more relevant for my dome speaker.

And this is the response of the dome.
6-dome v-drive.png
6-dome v-drive.png (67.84 KiB) Viewed 733 times
There is a wide peak centered at circa 450 Hz. Which I probably interpreted as "pronounced resonance" with my ears. It doesn't look quite like a resonance.

To further bust the resonance origin of this peak, take a look at the response with current-drive amplifier (i.e., one with very large output resistance, so that there is no electrical damping):
6-dome v-drive vs i-drive.png
6-dome v-drive vs i-drive.png (84.74 KiB) Viewed 733 times
The resonance peak here becomes very sharp. Now, it's quite obvious that in voltage drive, the resonance is not really a big issue.

Anyway. Feedbass eliminates resonance, because it effectively removes the return-spring action of the air in sealed enclosure. Basically, it makes the driver behave as if it weren't in the enclosure at all. This is best to compare with driver response in open baffle.
dome vs driver responses.png
dome vs driver responses.png (82.24 KiB) Viewed 733 times
Okay, that's interesting. So it looks like the dome somehow causes all output above circa 1.7 KHz to drop by about 6 db. BUT WHY?
I totally don't know, why. Ideas, anyone?

BTW, after adding a high-shelf filter to compensate for this drop, the sound of the speaker gets quite enjoyable.
jeno
Posts: 569
Joined: Sun Jun 29, 2014 10:41 am

Re: Designing a dome faceted loudspeaker

Postby jeno » Sun Mar 17, 2019 7:19 am

DeepSOIC wrote:
Sat Mar 16, 2019 4:47 pm
Jee-Bee wrote:
Thu Mar 14, 2019 9:12 am
for measuring sound response check https://www.roomeqwizard.com/ this.
thanks, I know.

....
Okay, that's interesting. So it looks like the dome somehow causes all output above circa 1.7 KHz to drop by about 6 db. BUT WHY?
I totally don't know, why. Ideas, anyone?
BTW, after adding a high-shelf filter to compensate for this drop, the sound of the speaker gets quite enjoyable.
Hi DeepSOIC,
speakers tend to emit with increasing frequency more and more directional. If you compare the frequency measurements on axis to 45 degree you'll see it. So when you measure the dome you'll get on low frequencies the sound pressure level from all speakers, but with increasing frequency only the spl form one or two. So the cut-off on higher frequncies is "normal". And the other thing is: it is the same for your ears ;)
To get meaningful measurements do only wire one speaker to the equipment to see what happens. From what I see from your posts I think some problem could be that you get a lot of direct reflections from inside wall. Did you made damping measurements? Some acoustic fibre inside? Could also decoupling the "dome" from the speakers, at least a bit. That is always a problem with lightweight cabinets.
For sheer curiosity: in which way did you wire the speakers? Are they 4 Ohm nominal or 8 Ohm?

cheers
jeno

edit: is your equipment able to measure impendance over frequency?
User avatar
DeepSOIC
Posts: 6953
Joined: Fri Aug 29, 2014 12:45 am
Location: Saint-Petersburg, Russia

Re: Designing a dome faceted loudspeaker

Postby DeepSOIC » Sun Mar 17, 2019 9:32 am

jeno wrote:
Sun Mar 17, 2019 7:19 am
If you compare the frequency measurements on axis to 45 degree you'll see it.
I see that it begins at 10k. The drop of my concern begins at 2k. That is a heck of a difference. Although, I didn't measure the driver alone at more extreme angles, that may play a role here.


jeno wrote:
Sun Mar 17, 2019 7:19 am
in which way did you wire the speakers? Are they 4 Ohm nominal or 8 Ohm?
4 ohm nominal. Wired like this (to achieve total resistance close to 4 ohms; see impedance plots for actual values):
wiring.png
wiring.png (23.12 KiB) Viewed 655 times

jeno wrote:
Sun Mar 17, 2019 7:19 am
is your equipment able to measure impendance over frequency?
Oh, how did I forget this valuable tool?! Here we go!
impedance.png
impedance.png (59.94 KiB) Viewed 655 times
Ohms scale is on the right side (dunno why REW puts decibel scale for these measurements at all). "single driver" is for driver without enclosure.

Unfortunately, I see no signs of problems like standing waves.


jeno wrote:
Sun Mar 17, 2019 7:19 am
Did you made damping measurements? Some acoustic fibre inside? Could also decoupling the "dome" from the speakers, at least a bit. That is always a problem with lightweight cabinets.
Damping measurements? I don't know what it is. Did you mean impedance measurement?

There is gyroid infill inside. Somewhat denser than in the screenshot from slicer I posted earlier. Oh, totally forgot, there is a pic of the enclosure without drivers, posted earlier in this thread :oops: . No other damping. What do you mean by "decoupling the dome from the speakers"? I don't understand.
jeno
Posts: 569
Joined: Sun Jun 29, 2014 10:41 am

Re: Designing a dome faceted loudspeaker

Postby jeno » Sun Mar 17, 2019 11:09 am

DeepSOIC wrote:
Sun Mar 17, 2019 9:32 am
jeno wrote:
Sun Mar 17, 2019 7:19 am
If you compare the frequency measurements on axis to 45 degree you'll see it.
I see that it begins at 10k. The drop of my concern begins at 2k. That is a heck of a difference. Although, I didn't measure the driver alone at more extreme angles, that may play a role here.
jeno wrote:
Sun Mar 17, 2019 7:19 am
in which way did you wire the speakers? Are they 4 Ohm nominal or 8 Ohm?
4 ohm nominal. Wired like this (to achieve total resistance close to 4 ohms; see impedance plots for actual values):
jeno wrote:
Sun Mar 17, 2019 7:19 am
is your equipment able to measure impendance over frequency?
Oh, how did I forget this valuable tool?! Here we go!
impedance.png
Ohms scale is on the right side (dunno why REW puts decibel scale for these measurements at all). "single driver" is for driver without enclosure.
Unfortunately, I see no signs of problems like standing waves.
jeno wrote:
Sun Mar 17, 2019 7:19 am
Did you made damping measurements? Some acoustic fibre inside? Could also decoupling the "dome" from the speakers, at least a bit. That is always a problem with lightweight cabinets.
Damping measurements? I don't know what it is. Did you mean impedance measurement?

There is gyroid infill inside. Somewhat denser than in the screenshot from slicer I posted earlier. Oh, totally forgot, there is a pic of the enclosure without drivers, posted earlier in this thread :oops: . No other damping. What do you mean by "decoupling the dome from the speakers"? I don't understand.
Hi
19.png
19.png (136.17 KiB) Viewed 639 times
For me it starts to differ at point 1, but difficult to say because of the "gap" at 1.7 khz. Could also be a deflection effect (wavelength around 20 cm, means an obstacle or a edge 10 cm away for speaker center). For me for shure at the latest at point 2 (5.5 khz) the graph starts differ more. Anyway that are just my 2 cents.
What do i mean with coupling? The membran of the speaker is a moving mass. Therefore forces work on the chassis and your PLA/ABS cabinet. You can try to separate the speaker from the cabinet with some elastic material to eliminate or in reality reduce the transfer of forces.
The other story: the speaker also sends soundwaves to the inside of the cabinet. These were reflected and pass through the membran with a certain delay. Maybe the cause for the waschy sound at your first trials. Try some acoustic wool to reduce the cabinet reflections.
The impendance-graph looks ok for me. The peak at 300 hz is the built-in resonance and is shows up very clear in 6dome v-drive vs i-drive chart.

This is just wild guessing ;)
cheers
jeno
User avatar
DeepSOIC
Posts: 6953
Joined: Fri Aug 29, 2014 12:45 am
Location: Saint-Petersburg, Russia

Re: Designing a dome faceted loudspeaker

Postby DeepSOIC » Sun Mar 17, 2019 2:18 pm

jeno wrote:
Sun Mar 17, 2019 11:09 am
What do i mean with coupling? The membran of the speaker is a moving mass. Therefore forces work on the chassis and your PLA/ABS cabinet. You can try to separate the speaker from the cabinet with some elastic material to eliminate or in reality reduce the transfer of forces.
OK, I understand now. This may help with that range, but will likely cause issues elsewhere. Here's why. On lower frequencies, the pressure of air inside gets pretty high. That pressure will push the driver and the dome apart. That is, we get much higher transfer of force, amplified by the ratio of basket outline area to cone area. In fact, I've seen this effect used to an advantage in some portable speakers: they mount the driver on a surround, and the driver becomes a passive radiator. Pretty cool! Here's one:
WP_20180317_18_29_05_Pro_reduced.jpg
WP_20180317_18_29_05_Pro_reduced.jpg (521.07 KiB) Viewed 614 times
WP_20180316_21_16_08_Pro_reduced.jpg
WP_20180316_21_16_08_Pro_reduced.jpg (644.47 KiB) Viewed 614 times

The effect is so strong, the thing starts jumping on the table even at pretty small volume. This thing has the enclosure made of steel, so it's very heavy. Yet it jumps.
User avatar
DeepSOIC
Posts: 6953
Joined: Fri Aug 29, 2014 12:45 am
Location: Saint-Petersburg, Russia

Re: Designing a dome faceted loudspeaker

Postby DeepSOIC » Sun Mar 17, 2019 2:39 pm

jeno wrote:
Sun Mar 17, 2019 11:09 am
at 1.7 khz. Could also be a deflection effect (wavelength around 20 cm, means an obstacle or a edge 10 cm away for speaker center).
It might very well be it. I have an edge 12 cm off center.
WP_20190304_19_52_05_Pro_reduced.jpg
WP_20190304_19_52_05_Pro_reduced.jpg (509.08 KiB) Viewed 608 times
jeno
Posts: 569
Joined: Sun Jun 29, 2014 10:41 am

Re: Designing a dome faceted loudspeaker

Postby jeno » Sun Mar 17, 2019 3:41 pm

DeepSOIC wrote:
Sun Mar 17, 2019 2:18 pm
jeno wrote:
Sun Mar 17, 2019 11:09 am
What do i mean with coupling? The membran of the speaker is a moving mass. Therefore forces work on the chassis and your PLA/ABS cabinet. You can try to separate the speaker from the cabinet with some elastic material to eliminate or in reality reduce the transfer of forces.
OK, I understand now. This may help with that range, but will likely cause issues elsewhere. Here's why. On lower frequencies, the pressure of air inside gets pretty high. That pressure will push the driver and the dome apart. That is, we get much higher transfer of force, amplified by the ratio of basket outline area to cone area. In fact, I've seen this effect used to an advantage in some portable speakers: they mount the driver on a surround, and the driver becomes a passive radiator. Pretty cool! Here's one:
...
The effect is so strong, the thing starts jumping on the table even at pretty small volume. This thing has the enclosure made of steel, so it's very heavy. Yet it jumps.
Yes,
that is the effect I talked about. The con is that it overlays the "primary signal" unpredictable in frequency and pressure level. The best cabinet for a speaker is the one which you do not hear i.e. it is dead concerning acoustics.
btw nice audio lab :D is the speaker on the picture a 5" Focal?

cheers
jeno
User avatar
DeepSOIC
Posts: 6953
Joined: Fri Aug 29, 2014 12:45 am
Location: Saint-Petersburg, Russia

Re: Designing a dome faceted loudspeaker

Postby DeepSOIC » Sun Mar 17, 2019 4:33 pm

jeno wrote:
Sun Mar 17, 2019 3:41 pm
is the speaker on the picture a 5" Focal?
Nope. It's R2R Audio's one, made here in saint-petersburg, by a few guys in a small lab (almost a garage).