Speakers

What do they use in cell phones for speakers?  I need something that will produce 90 dB at 10 cm at 250 Hz and should not be so large.  The inexpensive devices I find are either not so small (under 2") or have no low end, duh.  But they have great speakers in cell phones and I know they squeeze on the buckaroonies big time.  

How do they do it?  Is it about creating speaker cabinets?  A tablet I had once was hugely heavy at one end, I think because of the speaker magnets.  Does that end up costing more than a couple of bucks? 

-- 

  Rick C.

  - Get 1,000 miles of free Supercharging
  - Tesla referral code - https://ts.la/richard11209 

Rick C <gnuarm.deletethisbit@gmail.com> writes:
> But they have great speakers in cell phones and I know they squeeze on
> the buckaroonies big time.

The transducer is pretty small but there is typically a cavity around it
in the phone, with a tuned port (= judiciously placed and sized hole in
the phone).  90db at 250hz is asking a lot from a phone though.  You
might be "hearing" undertones of the frequencies that they do reproduce.
If you have a serious requirement then better use an audio spectrum
analyzer and calibrated mic.

Th style of dinky little external, enclosed speaker generally sounds
better than phone or laptop speakers:

  https://www.sparkfun.com/products/14023

There's lots of sub-2" speakers on adafruit and sparkfun.  You could
also try something like this, depending on the surrounding device:

  https://www.sparkfun.com/products/10917

On 11/5/2020 2:56 PM, Paul Rubin wrote:
> Rick C <gnuarm.deletethisbit@gmail.com> writes:
>> But they have great speakers in cell phones and I know they squeeze on
>> the buckaroonies big time.
> 
> The transducer is pretty small but there is typically a cavity around it
> in the phone, with a tuned port (= judiciously placed and sized hole in
> the phone).  90db at 250hz is asking a lot from a phone though.  You
> might be "hearing" undertones of the frequencies that they do reproduce.
> If you have a serious requirement then better use an audio spectrum
> analyzer and calibrated mic.

Also, I believe that phone speakers are also tuned with the appropriate 
DSP to make them sound better than they would otherwise sound.

On Thursday, November 5, 2020 at 9:47:32 PM UTC-5, John Speth wrote:
> On 11/5/2020 2:56 PM, Paul Rubin wrote:
> > Rick C <gnuarm.deletethisbit@gmail.com> writes:
> >> But they have great speakers in cell phones and I know they squeeze on
> >> the buckaroonies big time.
> > 
> > The transducer is pretty small but there is typically a cavity around it
> > in the phone, with a tuned port (= judiciously placed and sized hole in
> > the phone).  90db at 250hz is asking a lot from a phone though.  You
> > might be "hearing" undertones of the frequencies that they do reproduce.
> > If you have a serious requirement then better use an audio spectrum
> > analyzer and calibrated mic.
> 
> Also, I believe that phone speakers are also tuned with the appropriate 
> DSP to make them sound better than they would otherwise sound.

Yeah, this is to produce informative tones, not speech or music.  But there are requirements for 250 Hz and a certain harmonic minimum variation.  So it will be driven by square waves at a few different frequencies. 

The 90 dB at 10 cm we mostly pulled from our butts, but it will be inside a metal enclosure with openings for other things toward the rear.  This needs to be heard across a room so it needs to be loud.  The entire front is covered with a plastic label.  We can have openings in the metal behind the label.  

I'm just wondering what it takes to get some volume in a small speaker.  Cell phones might not be 90 dB, but they are so freaking tiny too.  I did see some units that are in a small can with a port on the end, but $40!  I'd be willing to bet the speakers in the $250 tablet will reach 90 dB.  They had magnets that weighted more than the battery.  It was very end heavy, also very magnet! 

-- 

  Rick C.

  + Get 1,000 miles of free Supercharging
  + Tesla referral code - https://ts.la/richard11209

Rick C <gnuarm.deletethisbit@gmail.com> writes:
> Yeah, this is to produce informative tones, not speech or music.  

Oh, this sounds like an alarm for your ventilator.  First thing I
thought of was a Sonalert but those apparently are around 2 khz.

Anyway, maybe try this: https://www.adafruit.com/product/1739

Rather than a 250hz tone you can get a hard-to-ignore effect by pulsing
a higher frequency tone at 250 hz.

This looks insane and mentions medical devices as an application area:
  https://www.sparkfun.com/products/16454

Many other ones here:

  https://www.sparkfun.com/categories/347

On Friday, November 6, 2020 at 12:58:04 AM UTC-5, Paul Rubin wrote:
> Rick C <gnuarm.deletethisbit@gmail.com> writes:
> > Yeah, this is to produce informative tones, not speech or music.  
> 
> Oh, this sounds like an alarm for your ventilator.  First thing I
> thought of was a Sonalert but those apparently are around 2 khz.
> 
> Anyway, maybe try this: https://www.adafruit.com/product/1739
> 
> Rather than a 250hz tone you can get a hard-to-ignore effect by pulsing
> a higher frequency tone at 250 hz.
> 
> This looks insane and mentions medical devices as an application area:
>   https://www.sparkfun.com/products/16454
> 
> Many other ones here:
> 
>   https://www.sparkfun.com/categories/347

Why do people post unresponsive advice?  

"So it will be driven by square waves at a few different frequencies."

I came onto the project after being a part of another such project which merged with this one.  On that project I used a single frequency sounder for a power state alarm.  This project has dug more deeply into requirements and found that medical equipment uses a series of tones in patterns that indicate the type of equipment and severity.  250 Hz is the minimum frequency.  It is not essential for the fundamental to be produced evenly with the harmonics.  There is a fairly wide range, something like &plusmn;15 dB and the harmonics of a square wave fall off.  So there may need to be some frequency shaping, but I'd like to avoid that other than perhaps a simple enclosure to improve the low end range of the speaker. 

Single frequency sounders are not at all useful. 

I'm just trying to get some insight into speakers that might be available.  Clearly phone and tablet and even PC speakers do a pretty good job of producing such tones.  Maybe I should search for laptop speakers.  

-- 

  Rick C.

  -- Get 1,000 miles of free Supercharging
  -- Tesla referral code - https://ts.la/richard11209

Rick C <gnuarm.deletethisbit@gmail.com> writes:
> Why do people post unresponsive advice?  
> "So it will be driven by square waves at a few different frequencies."

Pulse the high frequency sounder on and off at 250 hz, was the
suggestion.  Does that not get a 250 hz fundamental?  I'll try it (in
software) when I get home.  Similarly with other pulse frequencies.

Lots of small speakers are available on those sites I linked, including
some with little enclosures that should acoustically amplify the sound.

If you think a laptop or phone speaker sounds ok, then the first thing
I'd do is measure its frequency response.  Just recording a sample with
a decent microphone and using Audacity's built-in spectrum analyzer
feature is a reasonable start.  If you can measure the SPL, that is
great, but otherwise check whether it's still loud enough from say 20
feet away.

On Friday, November 6, 2020 at 5:59:54 PM UTC-5, Paul Rubin wrote:
> Rick C <gnuarm.deletethisbit@gmail.com> writes:
> > Why do people post unresponsive advice?  
> > "So it will be driven by square waves at a few different frequencies."
> 
> Pulse the high frequency sounder on and off at 250 hz, was the
> suggestion.  Does that not get a 250 hz fundamental?  I'll try it (in
> software) when I get home.  Similarly with other pulse frequencies.
> 
> Lots of small speakers are available on those sites I linked, including
> some with little enclosures that should acoustically amplify the sound.
> 
> If you think a laptop or phone speaker sounds ok, then the first thing
> I'd do is measure its frequency response.  Just recording a sample with
> a decent microphone and using Audacity's built-in spectrum analyzer
> feature is a reasonable start.  If you can measure the SPL, that is
> great, but otherwise check whether it's still loud enough from say 20
> feet away.

The higher frequency tone is not the spec.  The spec is a fundamental with some specific harmonic content that is not pulsing a 2k tone at 250 Hz.  The goal is not to be hard to ignore.  The goal is to provide information that quickly conveys information of nature, severity and how quickly the matter needs to be attended to without having to read the front panel.  The attendant can then respond appropriately if they are doing something else.  I expect the effect on the patient was also considered. 

This article has lots of good info even if the tables are crap.  Some parts are poorly organized, but the info is all there. 

http://admin.altran.it/fileadmin/medias/IT.altran.it/Images/Publication/TechnologyReview/Technology_Review_n._8_-_Ottobre_2012_P.Sessa.pdf

This one has audio files of the sounds we will be making. 

http://www.anaesthesia.med.usyd.edu.au/resources/alarms/

We have looked at this pretty thoroughly at this point.  I just need to find a bleeding speaker.  I have about two inches square panel space and it looks like I won't be getting any more.  

When I took my first stab at this I thought "sine waves" and so coded an NCO driving a PWM.  When I saw the article I realize sine waves were not right and thought saw tooth for the harmonics.  So the LUT was out and the phase accumulator was nakedly feeding the PWM.  When I looks at the audio file in an audio editor and I realized they were using square waves, I took just the top bit of the phase accumulator, but still need the PWM to set amplitude.  

I can add some filtering in the analog domain to shape the frequency response, but I'd prefer not to.  I could do the same in the digital domain, but again, prefer not to.  My focus is on finding a speaker that won't require filtering of the signal. 

-- 

  Rick C.

  -+ Get 1,000 miles of free Supercharging
  -+ Tesla referral code - https://ts.la/richard11209

On Fri, 06 Nov 2020 14:59:48 -0800, Paul Rubin
<no.email@nospam.invalid> wrote:

>Rick C <gnuarm.deletethisbit@gmail.com> writes:
>> Why do people post unresponsive advice?  
>> "So it will be driven by square waves at a few different frequencies."
>
>Pulse the high frequency sounder on and off at 250 hz, was the
>suggestion.  Does that not get a 250 hz fundamental?  I'll try it (in
>software) when I get home.  Similarly with other pulse frequencies.

The centuries old trick how to generate the lowest notes on a small
pipe organ is to sound two short pipes simultaneously in which the
frequency difference is the same as the low desired note. This may
require significant audio levels.

>Lots of small speakers are available on those sites I linked, including
>some with little enclosures that should acoustically amplify the sound.

The wavelength at 250 Hz is about 1.3 m, so you must keep the waves
generated on opposite side of the cone separated to avoid acoustic
"short circuiting". One way is to use a baffle so large that the air
wave path from the front to the back side of the cone is at least half
a wavelengths. If this is too large, use a sealed box construction to
keep the waves separated. The air in the closed box dampens the cone
movement, so that more power can be fed into the speaker before
hitting the excursion limit.

On Friday, November 6, 2020 at 11:47:28 PM UTC-5, upsid...@downunder.com wrote:
> On Fri, 06 Nov 2020 14:59:48 -0800, Paul Rubin
> <no.email@nospam.invalid> wrote:
> 
> >Rick C <gnuarm.deletethisbit@gmail.com> writes:
> >> Why do people post unresponsive advice?  
> >> "So it will be driven by square waves at a few different frequencies."
> >
> >Pulse the high frequency sounder on and off at 250 hz, was the
> >suggestion.  Does that not get a 250 hz fundamental?  I'll try it (in
> >software) when I get home.  Similarly with other pulse frequencies.
> 
> The centuries old trick how to generate the lowest notes on a small
> pipe organ is to sound two short pipes simultaneously in which the
> frequency difference is the same as the low desired note. This may
> require significant audio levels.
> 
> >Lots of small speakers are available on those sites I linked, including
> >some with little enclosures that should acoustically amplify the sound.
> 
> The wavelength at 250 Hz is about 1.3 m, so you must keep the waves
> generated on opposite side of the cone separated to avoid acoustic
> "short circuiting". One way is to use a baffle so large that the air
> wave path from the front to the back side of the cone is at least half
> a wavelengths. If this is too large, use a sealed box construction to
> keep the waves separated. The air in the closed box dampens the cone
> movement, so that more power can be fed into the speaker before
> hitting the excursion limit.

I'm going to hate asking the mechanical engineer to add a baffle.  The guy will do all manner of things to make his stuff work, but anything the electronic guys want to do he pushes hard on every penny.  He's the project lead, btw.  I pointed out the speakers can be secured with a fastening plate on the back with a hole to go around the speaker magnet, secure and positions the speaker without issue.  It needs to be plexi or other plastic to make sure it doesn't contact the terminals which are very close.  His suggestion was to use a double sided sticky washer thing.  I'm about ready to punt on this and focus on other work.  It's not a job and I'm not going to fight about this stuff anymore.  

I'm also a bit ticked when I realized that while we have a public document showing our names and images of everyone contributing, in a recent write up about the project it sounded like this was purely his project.  I suppose that is more the writer than him, but his is the only truly public face on it.  I may have mentioned we are one of very few such projects that actually seem to be getting somewhere at all.  In fact, other projects are expressing interest in our alarm FPGA design as a common entity. 

I just wish I could use the durn VHDL language the way it is intended.  VHDL-2008 is now 12 years old and parts are still implemented poorly if at all.  I'm trying to use a simple aggregate construct and the simulator I'm using balked at one form of it, now it's balking at every form.  It's hard to get work done when you have to fight the tools.  Worse, because I'm using free tools from an FPGA vendor, Aldec doesn't even want to hear about the bugs. 

Stupid Aldec! 

-- 

  Rick C.

  +- Get 1,000 miles of free Supercharging
  +- Tesla referral code - https://ts.la/richard11209