Shop OBEX P1 Docs P2 Docs Learn Events
loud but small, thin audio speaker - Page 2 — Parallax Forums

loud but small, thin audio speaker

2

Comments

  • Tracy AllenTracy Allen Posts: 6,664
    edited 2011-12-20 10:17
    I wanted to try out that Tang 1" full range pair but they are out of stock until February at Parts Express, and I couldn't find an alternative source. The 2" ones look good too, but are just too big. I've had to fix on that enclosure pictured, which will accommodate 40mm width internally. The CUI speaker will I think meet the requirement that it produce a sound comparable to, better than, an iPhone. That is interesting about the exponential cone. I'd guess that it would have to be fairly deep too, depth equal to the width or something like that, right? We'll experiment, but I don't know if it will be practical to fabricate something like that.

    One reason I like the class D amplifier TPA2006 is the simple interface circuit, which would as usual have to be preceded by a lowpass filter from the Prop, maybe duty mode differential. The efficiency curve attached wrt a class AB ampifier is quite striking, it seems.
    TPA_2006.png


    efficiency.png
    864 x 338 - 55K
    348 x 383 - 32K
  • EmptyBitEmptyBit Posts: 72
    edited 2011-12-20 21:25
    Tracy,

    I tried a few hole pattern grids and did not like the burr it put in ABS at each hole. Deburring it was a pain. I opted for minimal cut time and rugged strength, not so much art. The NEMA box I am using is made of some glass filled resin and very brittle. You might be able to get away with a simple cut in grille.

    It is my understanding that there only needs to be enough air flow based on the speaker throw. Highly doubtful the pattern will make that much difference with such small speakers unless there is severely restricted flow. I have even contemplated our organizations logo in shadow highlights, just for the fun of it.


    My application requires speech fidelity in the mid-range and amplified enough for manufacturing environments. Low watts and lots of dB’s fit the bill. Sure, symphonic quality will take more refinement.

    I may go this route with future projects. I think I was quoted around $8-12ea on one size here. W or W/O speaker.

    http://www.uniboxinfo.com/speaker.html

    I like the feature they can also offer some speaker box sizes with this flanged wall mount. All you can do is inquire for your size.

    http://www.uniboxinfo.com/wallmount.html
  • Phil Pilgrim (PhiPi)Phil Pilgrim (PhiPi) Posts: 23,514
    edited 2011-12-20 23:33
    Tracy,

    It's too bad that, with the Class D amp, you have to filter the digital output to analog first, just so it can be converted back to digital to drive the speaker. I did some work with a MOSFET driver and with an H-bridge motor driver in Class D arrangements to drive a speaker. You can read about them here:

    I'm not necessarily recommending either approach, BTW, since I don't know what kind of sound quality your app demands.

    -Phil
  • Tracy AllenTracy Allen Posts: 6,664
    edited 2011-12-21 09:33
    I'm leaning to stick with the simple circular punch, which leaves nice clean edges on the ABS. The speaker seems to work fine with the SS woven wire mesh stainless steel screen. The screen sheet I had on hand from an earlier project probably has a 30 x 30 mesh with about 40% open area. I see that McMaster even sells those as 2" diameter precut disks. It is quite stiff enough when glued to the back of the opening (1.375" d), with the speaker held back by a gasket ring. Having the speaker in the closed box does improve the sound on some clips but not on others; overall having it in the enclosure makes it qualitatively less "tinny".

    That's an interesting pair of threads Phil, thanks for pointing those out. TI's class D chips derive their constant-frequency PWM at 250kHz from a comparator action between the audio and a ramp. That can deliver vanishing narrow duration pulses to the speaker at the zeros and at the highs, but at least it would not have the rich subharmonic content of the Prop's duty mode. One way around that on the Prop might be to use the technique of overlapping two COG counters to make constant frequency PWM, and modulate phsa-phsb instead of frqa with the audio amplitude. An XOR gate would have to interpose in your Micrel circuit, to provide conditional inversion.

    I have a couple of the TI class D chips on order, just to try it out. It is indeed too bad to have to go from digital density to analog and back to PWM, and I am still wondering if there might be issues with aliasing or other hits on quality. I have a class AB in the same package too as the fallback if there are quality issues with the class D. It is adequate, so why tinker? One system feature of those chips that I do really like is the shutdown pin.
  • Phil Pilgrim (PhiPi)Phil Pilgrim (PhiPi) Posts: 23,514
    edited 2011-12-21 11:00
    Tracy,

    If you need to wring more apprarent loudness out of a given speaker/amp combo, you can always resort to dynamic range compression:

    It's a pure software play that's more gentle than raw clipping, and you can choose a compression level to suit the audio material.

    -Phil
  • Tracy AllenTracy Allen Posts: 6,664
    edited 2011-12-21 11:38
    Phil, that could definitely be an asset. I am using a modified version of Ray Allen's wav player, and it looks like there is time in the loop at 44kHz sample rate to include the compression. I already have a volume level (3dB steps) associated with each sound clip, and each file could also have a compression setting or flag. I'll give it a try. We really need to monitor ambient sound level and adjust the punch-through accordingly.
  • CircuitsoftCircuitsoft Posts: 1,166
    edited 2011-12-21 12:23
    The 2" ones look good too, but are just too big.
    Based on the datasheet, you could get by with 30.2mm from the outside surface of the case to the inside of the opposite side. If the box were anywhere near 5mm thick, that could work, especially if you were to shave out the other side of the box just a little. Better yet, cut a hole in the opposite side the right size to stick the magnet through, and glue it. That will strengthen the box and reduce resonance in the enclosure.

    These HiVi 1" drivers are on sale today. There's a fuller review available at Zaph|Audio
  • Toby SeckshundToby Seckshund Posts: 2,027
    edited 2011-12-22 05:37
    I just thought, if you use a modulated Tesla coil then only a couple of small holes will have to be made. It will keep the fingerds at bay too.

    On other commercial possibilities, perhaps a modified Tazer could announce itself, and save the policeman all of that tedious shouting.
  • Tracy AllenTracy Allen Posts: 6,664
    edited 2011-12-22 11:29
    A plasma speaker could be pretty awesome for snap crackle and pop and to keep fingers at a distance. But I can't afford a nuclear power plant to operate it. There was a guy at UPEW year before last from the Portland science center, using Props for exhibits that encouraged public interaction, and it was very interesting to hear the lengths to which he had to go to kid-harden his engineering.

    Circuitsoft, I've ordered a couple of those speakers from part-express to try out, but this is going to have to be a work of "good enough" and not "optimal". I guess the CUI speakers are not on the audiophile's radar (sonar?). The Zaph comparative ratings are interesting. What he says about that HiVi B1S: " The rear chamber for this is critical, and must be shaped as described here for the best effect: drill a 1-1/4" hole through 3/4" MDF. Then round over the back side with a 1/2" roundover bit. Finally, densly stuff that area with Acousti-stuf and glue another board over it to seal, while allowing holes for the wires to pass though. It took experimentation to get this response, and without it the response curve was terrible." Yikes!
  • Toby SeckshundToby Seckshund Posts: 2,027
    edited 2011-12-23 03:38
    There is a huge difference between what an audiophile (with access to analytical equipment) would deem unacceptable and what is good enough to convey an adiquate communication. 90% of consumer devices are served perfectly well by a lump of papier mache. Mylar or metalic cones would be better for exposed use though.

    I have some homemade loudspeakers in the living room, they are not using high end units and are constructed from non real wood flooring sheets. They had to suffer the demands of HER that they didn't look like MDF and were not "too big" (as if any LS could be too big!!!! ). I tried various back loadings and ended up with a form of tapered tube with as much internal bracing as possible. At least the results are a vast improvement over the plain reflex boxes that came with the unit. With a small plastic box as the enclosure the back loading will be severe and cut down on efficiency, but could flatten out the response a bit.
  • Tracy AllenTracy Allen Posts: 6,664
    edited 2011-12-23 09:53
    Likewise an audiophile with no analytical equipment other than a good ear, or ostensibly so. Our goal is something that can compare to the volume and quality of a smartphone, and that is not a particularly high bar. It is just taking me a larger, more expensive speaker to get there than one would expect to find in a phone. All surfaces of this 10 cubic inch box vibrate like crazy to the touch when the speaker is at full volume. The detail is pretty good though. For example, one clip is garden chimes, and in a quiet room one can hear the crickets deep in the background.
  • Toby SeckshundToby Seckshund Posts: 2,027
    edited 2011-12-23 11:43
    presumably, the crickets are on the recording ...
  • CircuitsoftCircuitsoft Posts: 1,166
    edited 2011-12-27 16:48
    Circuitsoft, I've ordered a couple of those speakers from part-express to try out, but this is going to have to be a work of "good enough" and not "optimal". I guess the CUI speakers are not on the audiophile's radar (sonar?). The Zaph comparative ratings are interesting. What he says about that HiVi B1S: " The rear chamber for this is critical, and must be shaped as described here for the best effect: drill a 1-1/4" hole through 3/4" MDF. Then round over the back side with a 1/2" roundover bit. Finally, densly stuff that area with Acousti-stuf and glue another board over it to seal, while allowing holes for the wires to pass though. It took experimentation to get this response, and without it the response curve was terrible." Yikes!

    Zaph is pretty picky with his sound. The fact that he listed the B1S as usable at all is pretty amazing.
  • CircuitsoftCircuitsoft Posts: 1,166
    edited 2012-01-08 20:09
    Anyway, how is this project going?
  • Tracy AllenTracy Allen Posts: 6,664
    edited 2012-01-09 11:32
    Thanks for your continuing interest. I was (as usual) sidetracked during holidays and travel. Now I'm back to it, mechanical and power supply issues at the moment.

    For the audio I'm waiting for breakout circuit boards that can accommodate either the TPA6204 (class AB) or TPA2006 (class D). I'll try one and the other to see if there is a noticeable difference in quality or other factors. Both of those chips are set to take differential inputs and to produce differential (h-bridge) outputs to the speaker. On the input side, I plan to use duty-differential mode from the COG counter.
    duty-differential-audio.png

    That mode starts off both counters at 50%, so it helps reduce initial pop, also it produces a cleaner, balanced signal in other ways as well.
    318 x 95 - 8K
  • CircuitsoftCircuitsoft Posts: 1,166
    edited 2012-01-09 12:04
    I would give both pins a 10nF cap to ground rather than to each other. That way, you won't have common-mode high frequency signal components going into your amplifier. Even if it should filter them out, the common-mode-rejection may not be as good at higher frequencies.
  • Tracy AllenTracy Allen Posts: 6,664
    edited 2012-01-09 14:14
    That's a good point. I thought about it more and I think I see what you mean. In the following scheme:
    better Vcm rejection.png

    would you say Vcm will come from currents in the bond wires etc. and will be applied roughly equally to p5 and p6, and will be averaged by the two RC circuits with respect to Vss?

    That Vss point where the capacitors attach would be the reference common for the amplifier both power and signal. My intent is to do a first stage of filtering right at the Prop pins, and then the differential signal on a fine shielded twisted pair, and additional filtering with caps to Vss at the amplifier input. The amplifier would be close to the speaker, which is located a few inches from the Prop.

    The CMRR of these amplifiers is stated to be around -65dB at 217 Hz, but it doesn't say about higher frequencies. Their main focus seems to be power supply rejection as would be found in a GPRS phone.

    Incidentally, when running two neighboring pins in duty-differential with RC filtering, the DC offset between the two seems to be around 30mV, with spread in the matching depending on the prop and the pin. So I'll AC couple the signal into the amplifier chip in order to avoid DC offset or accidents in the speaker.
    307 x 111 - 8K
  • CircuitsoftCircuitsoft Posts: 1,166
    edited 2012-01-09 14:32
    The Prop produces a lot of high frequency energy in DUTY mode. Duty Differential may make things a little nicer, but you still may have skew issues between p5 and p6. With the separate filter on each, you obviate that problem, and probably bring any common-mode noise into the frequency range that the amplifier can compensate for.
  • LawsonLawson Posts: 870
    edited 2012-01-20 10:02
    This blurb from Tom's Hardware seems appropriate http://www.tomsguide.com/us/Speaker-Thin-Flexible-CES-Portable,news-13906.html The comments make it sound like it's an electrostatic speaker. Sounds bloody good in the YouTube video.

    Lawson
  • Tracy AllenTracy Allen Posts: 6,664
    edited 2012-01-21 15:22
    A wearable speaker! That's something, just when I was getting used to normal people walking down the street talking to themselves.

    I wonder how it works. I have one of the piezo speakers from PUI, which is 22mm square and only 1mm thick. The volume was nowhere near as good as magnetic speakers of the same size. It specifies 74 dBA at 3Vrms drive, and that is about what I was seeing, compared to 90+ for other speakers I was trying. Quite a difference. I don't know where one would use something like this, and the same principle would not seem to translate easily to something wearable and "able to play sound at a volume and quality that is just as good as conventional speakers."

    IMG_2106.jpg
    630 x 551 - 82K
  • ericballericball Posts: 774
    edited 2012-01-23 06:55
    Many, many (many) years ago I bodged together a pair of headphones made from piezo speakers from two identical digital watches. Zero bass response and not very loud, but ghods was the treble crisp - they really showed off the clarity of the new-fangled portable CD player I had.
  • dneill468dneill468 Posts: 2
    edited 2013-09-05 13:52
    I was just wondering if someone could help me regarding a project.
    I need a very small speaker with a good frequency range for music playing.
    the speaker needs to be able to run off a 3volt supply.
    can any one suggest a good quality small speaker which may have a tiny built in amplifier to produce a quality sound.
  • CircuitsoftCircuitsoft Posts: 1,166
    edited 2013-09-05 17:45
    Can you specify requirements?

    What is Small?

    What is Good Quality? 1%THD? 0.01%THD? 100Hz-10Khz? 20Hz-20KHz?

    How loud do you want it to get?

    What is your price range?

    There are a lot of trade-offs in choosing an audio solution for any particular application.
  • Tracy AllenTracy Allen Posts: 6,664
    edited 2013-09-06 08:16
    Seeing this thread again reminds me to thank everyone for their excellent help and to give an update on how it turned out. I ended up with a speaker not so thin, but it is one that fit well in the enclosure and was readily available. The speaker is the CUI CMS040 rated nominally 5W, 40mm x 40mm and 4.5mm thick. Big magnet. Not small at all, nothing like whatever it is that is in an iPhone. Having it mounted in the enclosure makes a huge difference in the sound quality, as was pointed out several times in this thread. The speaker is mounted behind an opening in the enclosure, behind a fine stainless steel mesh and a rubber gasket. I found later on that the opening in the front could have been substantially smaller without much effect on the sound quality. Inside the box behind the speaker is a LiPoly battery+charger, and at the other end of the box is the data logger with the propeller board (OWL3pro), rgb LED, SD card and XBee. The wiring tightened up from batch to batch, and the audio amplifier (TPA6204) ended up mounted on a post just above the logger. From the back, there are slots for access to the SD card and for the Prop Plug, and an RS232 cable and barrel plug to mate with the Dylos air quality monitor, and a barrel socket for external power.
    FreshAirDylos.jpg
    sidePAk1interior.jpg
    sidePak123interior.jpg
    sidePakBack.jpg
    dc1700-freshAir-row.jpg
    576 x 541 - 76K
    720 x 540 - 139K
    720 x 540 - 118K
    576 x 432 - 67K
    900 x 675 - 181K
  • PublisonPublison Posts: 12,366
    edited 2013-09-06 08:30
    That IS pretty neat Tracy.

    I love "Happily Ever After" threads. :)

    Jim
  • CircuitsoftCircuitsoft Posts: 1,166
    edited 2013-09-06 11:25
    Tracy,
    Since this thread originally "Aired", TangBand has come out with a new product that you may find useful: T1-1931S Speaker Module. It is a 1" driver, passive radiator, and tune enclosure in a single piece that's easy to integrate into your application.
  • dneill468dneill468 Posts: 2
    edited 2013-09-07 01:53
    ok thanks
    small is maximum size of 20mm by 20mm by 3-4mm
    around 300Hz to 20KHz
    up to 80 decibels
    maximum of 5 pounds per unit
    is there an associated amp for the speaker or in fact does the speaker have a built in amp?
    the size of the unit is important in this design so a speaker with a built in amp would be space saving.
  • Tracy AllenTracy Allen Posts: 6,664
    edited 2013-09-07 09:51
    dneill,
    One that I evaluated for this project was from Kobitone, 22mm diameter. link. I see it is now eol. I don't see anything in the Kobitone line currently at Mouser that meets your specs. I think circuitsoft can explain much better I can why this is, but you really do have to get a sample of whatever and try it out in your enclosure and with your electronics to see if it will have acceptable sound quality and loudness. In my case there was room for 40mm, and that plus the heavier magnet and conservative power rating made a great qualitative difference, no matter what it might seem on paper.

    Kobitone 253-CE221-RO, 22mm x 3.5mm, 8Ω, 0.75W:
    Screen shot 2013-09-07 at 9.39.42 AM.png
  • Tracy AllenTracy Allen Posts: 6,664
    edited 2013-09-07 10:02
    Circuitsoft, thanks for the link to TangBand. I've bookmarked all the links you've provided. My own applications are pretty utilitarian, but my son who works with me is a musician and fanatic about microphones and recording and performance spaces, and his wife, also a musician is studying acoustic engineering. Their ears are much more discriminating than mine!

    The circuit I ended up with uses the TPA6204 as mentioned earlier in the thread. It is differential both at the input and BTL at the output. From the Prop it is driven via duty-differential. Filtering happens in two stages. The first is differential with a 1nF mylar capacitor close to the Prop pins, and the second stage addresses the common mode issue that you brought up. One nice thing about the differential drive is that it avoids the pop that occurs with some single-ended amplifiers when the power turns on / off. A prop pin controls power to the chip.

    TPA6204_FA.png
    509 x 268 - 21K
  • zachthebomb13zachthebomb13 Posts: 1
    edited 2014-02-16 10:14
    Hello all,

    I'm brand new here (this is my first post), and I came across this thread this morning. I'm looking for some small speakers, and you all seem really knowledgable. Let me start this post by saying I know almost nothing when it comes to speakers, amps, etc., but I am a quick learner. What I'm trying to do is build a tablet computer case that has integrated front facing speakers. I want to use something that doesn't need an amp to use it, and will work by just plugging into the head phone jack, like these sony passive speakers:
    http://amzn.com/B000PGAX9W

    They seem to be just a little to thick, however. I thought about using small powered speakers, but I don't know what all is needed for those to work. I'm assuming the signal would have to go into some kind of amp, which would have to be powered by some kind of battery, and then into the speakers. Any ideas? Thanks!

    Zach
Sign In or Register to comment.