Hi everyone,
With this post i would like to share some things i learned while trying to improve the sound of a Bluetooth speaker. I have done this in the context of a project for my bachelor mechanical engineering. We had to create a small project ourselves, preferably about something that you didn't have any previous experience with. Sound is only briefly touched upon in our study and only in the sense of the dynamics of vibration.
I wanted to try to do a enclosure design inspired by HexiBase's video's on YouTube. Little did i know that this is pretty hard to do.
I started out with a lot of research about different speaker enclosures, but I wont bore you with that.
I wanted to try to improve the Anker Soundcore 2. A cheap Bluetooth speaker which sounds okay in my opinion. It uses a passive radiator paired with 2 1.5" 5 watt, 3 Ohm speakers in a small water-resistant package. At first i wanted to make a transmission-line style speaker, but i soon found out that it is important to know the properties of the drivers that are being used. As mentioned before the Soundcore 2 uses 2 small 5 watt speakers, but this is all the info i could find. The speakers have some numbers printed on them, but i cant seem to find any info on them via Google.
The transmission-line housing was out the window at this point, so i settled for an open baffle housing. This would be a very easy enclosure to design and it wouldn't be necessary to know all the important properties of the drivers. In CAD I designed a very simple square shaped baffle with geometry in the middle to perfectly fit the drivers. The baffle was 200x200mm, because this would comfortably fit on the bed of my Ender 3 printer. The housing's were printed in white PLA with 20% infill. Overall the prints took 13 hours.
When these were printed i assembled them and spliced open an old USB cable to get some wire to extend the original wires. When this was done i tested them to see if everything turned out the way i wanted to.
Then i wanted to test the "redesign" to compare it to the original Soundcore 2. Unfortunately i don't have access to an anechoic chamber or the professional equipment needed to perform these correctly, but i reckoned that i could do the 2 measurements with the stuff that i have on hand. Since both measurements will be flawed in approximately the same way i could at least compare the two and see the impact of the changed enclosure design.
Some of the reasons why my measurements could be flawed are:
The software i used for this was Arta. This is a pretty extensive software which you can use to test all kinds of stuff related to audio, but i only used the option to test the frequency response. Arta offers a paid version and a demo version. The demo version doesn't let you save files, but it still lets you export data to a excel which was enough for me, so luckily it didn't cost me anything. For the microphone i used a nameless boom microphone that i had laying around. The results i got were very useable and allowed me to learn some things about my new open baffle speakers.
As open baffle speakers have a pretty big bass rolloff i wanted to predict the frequency at which this would occur. As the baffle was 200mm wide my calculations said that the bas would start to rolloff at 135 Hz. However, as you can see the bass rolloff starts way earlier at around 240 Hz, the reason for which I'm not quiet sure. What I found quiet interesting to see is that the upper range treble seems to be better presented with the open baffle design. Something I didn't test, but I heard quite noticeable was the improved soundstage with the open baffle speakers.
If you would like some more info on this small project feel free to ask them. Any tips on how i could have done things differently of better are welcome!
With this post i would like to share some things i learned while trying to improve the sound of a Bluetooth speaker. I have done this in the context of a project for my bachelor mechanical engineering. We had to create a small project ourselves, preferably about something that you didn't have any previous experience with. Sound is only briefly touched upon in our study and only in the sense of the dynamics of vibration.
I wanted to try to do a enclosure design inspired by HexiBase's video's on YouTube. Little did i know that this is pretty hard to do.
I started out with a lot of research about different speaker enclosures, but I wont bore you with that.
I wanted to try to improve the Anker Soundcore 2. A cheap Bluetooth speaker which sounds okay in my opinion. It uses a passive radiator paired with 2 1.5" 5 watt, 3 Ohm speakers in a small water-resistant package. At first i wanted to make a transmission-line style speaker, but i soon found out that it is important to know the properties of the drivers that are being used. As mentioned before the Soundcore 2 uses 2 small 5 watt speakers, but this is all the info i could find. The speakers have some numbers printed on them, but i cant seem to find any info on them via Google.
The transmission-line housing was out the window at this point, so i settled for an open baffle housing. This would be a very easy enclosure to design and it wouldn't be necessary to know all the important properties of the drivers. In CAD I designed a very simple square shaped baffle with geometry in the middle to perfectly fit the drivers. The baffle was 200x200mm, because this would comfortably fit on the bed of my Ender 3 printer. The housing's were printed in white PLA with 20% infill. Overall the prints took 13 hours.
When these were printed i assembled them and spliced open an old USB cable to get some wire to extend the original wires. When this was done i tested them to see if everything turned out the way i wanted to.
Then i wanted to test the "redesign" to compare it to the original Soundcore 2. Unfortunately i don't have access to an anechoic chamber or the professional equipment needed to perform these correctly, but i reckoned that i could do the 2 measurements with the stuff that i have on hand. Since both measurements will be flawed in approximately the same way i could at least compare the two and see the impact of the changed enclosure design.
Some of the reasons why my measurements could be flawed are:
- No anechoic chamber; The resonances and reflections of my room affect the measurement
- No calibrated microphone: I don't know the frequency response of the microphone, so the microphone colors the results
- My room doesn't have a lot of space behind the speakers, so the open baffle speakers suffer more from the reflections than the Soundcore 2
The software i used for this was Arta. This is a pretty extensive software which you can use to test all kinds of stuff related to audio, but i only used the option to test the frequency response. Arta offers a paid version and a demo version. The demo version doesn't let you save files, but it still lets you export data to a excel which was enough for me, so luckily it didn't cost me anything. For the microphone i used a nameless boom microphone that i had laying around. The results i got were very useable and allowed me to learn some things about my new open baffle speakers.
As open baffle speakers have a pretty big bass rolloff i wanted to predict the frequency at which this would occur. As the baffle was 200mm wide my calculations said that the bas would start to rolloff at 135 Hz. However, as you can see the bass rolloff starts way earlier at around 240 Hz, the reason for which I'm not quiet sure. What I found quiet interesting to see is that the upper range treble seems to be better presented with the open baffle design. Something I didn't test, but I heard quite noticeable was the improved soundstage with the open baffle speakers.
If you would like some more info on this small project feel free to ask them. Any tips on how i could have done things differently of better are welcome!
Are you familiar with Linkwitz's site/models? Your roll-off point seems optimistically low for that width baffle, assuming a normal measuring distance, limited additional apparent baffle extension due to some other geometry that's not described, etc.
https://www.linkwitzlab.com/models.htm
https://www.linkwitzlab.com/models.htm