Hello, everyone!
How can I get the most loudness from a 25 mm speaker? Which amplifier voltage/wattage, and speaker impedance, would provide the most loudness from battery power?
While I've done this with various Class D amps (from adafruit, sparkfun et al), they seem to have suffered over-heating damage (distortion, clipping, clicking) with continuous play-back.
I'll be grateful for any insight you may have.
PS I recognise that this question has a speaker-design dimention as well, but I thought that that post ought to be to a Loudspeakers forum eg Planars and Exotics'. For _here_, I am asking about the voltage, wattage and impedance.
How can I get the most loudness from a 25 mm speaker? Which amplifier voltage/wattage, and speaker impedance, would provide the most loudness from battery power?
While I've done this with various Class D amps (from adafruit, sparkfun et al), they seem to have suffered over-heating damage (distortion, clipping, clicking) with continuous play-back.
I'll be grateful for any insight you may have.
PS I recognise that this question has a speaker-design dimention as well, but I thought that that post ought to be to a Loudspeakers forum eg Planars and Exotics'. For _here_, I am asking about the voltage, wattage and impedance.
Sound waves run up to 22 feet (7 meters) across/long in the bass (50Hz).
A 1-inch air-shover has NO chance of making big waves like that.
Amplifier is not the answer.
Shove the 1-inch next to your ear like a headphone. That will work. (It does not have to fill all of space, or even the whole room.)
In my attempt at making the problem-description as brief and as encapsulated as possible, I skipped other aspects--namely, that only human voice is involved. AFAIK, the corresponding frequency range is mainly 85Hz-255Hz. (IIRC the source of the info was Wikipedia.)
But I _have_ shared your concern. Never the less, I _have_ managed to create such loudness, so the wave propagation has been sufficient in those particular tests.
Also, folded-horn speakers reach far greater distances, so enclosure-design and other factors do come in.
I agree, in general. But my needs are only to have a message discerned at eg 5 meters, a task well within its capability.
There are other constraints, but I was going to leave those to a Loudspeakers forum. For _here_, I was focusing on tackling my ignorance of the effect, on loudness, of voltage and wattage and impedance. Naively, I used 20 W Class D amplifiers, with a 9 v battery, to increase loudness; they worked, but for only a while!
Thank you for your reply. Greatly appreciated. I hope you won't be put off by the additional details.
I understand. But, extensive, logged tests have been carried out to nail down the root cause. Parts have been replaced with identical copies as well as those in the same, or other, classes. It is not the loudspeaker mechanism (aka speakers, in this context), nor the sound source (which is another componenet under-test), nor a Shut-down Mode in the amplifiers, nor other components.
The amps used to work, for a few seconds of play-back; the damaged ones no longer do so at all. The damage arose, in under 1 minute, when continuous play-back of the sound was tested.
The cause is unrecoverable damage in the Class D amplifiers.
OC16SC00-04 Peerless by Tymphany | Audio Products | DigiKey
Thank you for your reply. Greatly appreciated.
Um....THAT is a tweeter.... A high frequency transducer.
Not at all suitable for full vocal coverage.
I know. But, its range is fine for my purposes, namely playing a message; its loudness is not.
Thank you for your response.
You quote the vocal range you require "85-280Hz" or thereabouts.
Then you seek to use a loudspeaker which is designed to be used from 1000Hz and above. This driver cannot reproduce even the 2nd harmonic of the sounds you wish to use it for, let alone the fundamental. No amount of bass boost or equalization will help.
Then you complain about gross distortion, and failing amplifiers.
Seems to me that the gross distortion is caused by this I'll thought out application of a tweeter for vocal reproduction.
The result is amplifier failure, I assume, once the voice coil overheats, then burns out and produces a dead short to the speaker chassis.
Any dirt cheap full range or toy speaker would have been a better choice.
Clueless.
Then you seek to use a loudspeaker which is designed to be used from 1000Hz and above. This driver cannot reproduce even the 2nd harmonic of the sounds you wish to use it for, let alone the fundamental. No amount of bass boost or equalization will help.
Then you complain about gross distortion, and failing amplifiers.
Seems to me that the gross distortion is caused by this I'll thought out application of a tweeter for vocal reproduction.
The result is amplifier failure, I assume, once the voice coil overheats, then burns out and produces a dead short to the speaker chassis.
Any dirt cheap full range or toy speaker would have been a better choice.
Clueless.
Last edited:
To emphasize:
no vocal _range_ is specified. What voice reproduction is needed has already been successfully carried out by the speakers used, including the tweeter. It is sufficient, proven, and not the subject of this post.
Once amplifier failure occurs, it does so with other speakers as well. This, I should underscore, is merely the most basic of the innumerable tests carried out. Another is to change the source to be PWM beeps, which again reproduces the failure.
I'll rephrase the original question--and that is all that I am seeking here:
What is the effect, on loudness, of increasing input-voltage to an amplifier?
What is the effect, on loudness, of increasing the output-wattage of an amplifier?
How does speaker impedance affect loudness in relation to the above two (input voltage, and wattage) variables?
Knowing the above, I can focus on an amplifier with the lowest possible input-voltage, and the lowest output wattage for a given speaker impedance. Thank you for answers to the above three questions.
no vocal _range_ is specified. What voice reproduction is needed has already been successfully carried out by the speakers used, including the tweeter. It is sufficient, proven, and not the subject of this post.
Once amplifier failure occurs, it does so with other speakers as well. This, I should underscore, is merely the most basic of the innumerable tests carried out. Another is to change the source to be PWM beeps, which again reproduces the failure.
I'll rephrase the original question--and that is all that I am seeking here:
What is the effect, on loudness, of increasing input-voltage to an amplifier?
What is the effect, on loudness, of increasing the output-wattage of an amplifier?
How does speaker impedance affect loudness in relation to the above two (input voltage, and wattage) variables?
Knowing the above, I can focus on an amplifier with the lowest possible input-voltage, and the lowest output wattage for a given speaker impedance. Thank you for answers to the above three questions.
What you need first is efficiency, search for a driver with higher value of dB/W for the band of interest.....the tweeter you have chosen is totally not fit for purpose for "85Hz-255Hz" and using it in this band of interest is certain to cause failures of both driver and amplifier.
Dan.
Impedances correct? Power supply voltage, AND power ability?
Are these "good" amplifiers or $3 cheapies?
The amplifiers should not fail. The small speaker should melt first. So something IS wrong but you only say "various" amplifiers with vague "distortion, clipping, clicking" symptoms (which could even be bad wiring).
Let's forget your doomed speaker and LOOK AT the amplifier including its power supply and wiring.
I'll keep this in mind. Thanks.
This range is not of concern to me! I mentioned it in only a response to someone else's post. Even the mention of "voice" was meant to convey that I am not seeking great fidelity in the reproduction; so long as the message is discerned, it'd be fine. And the discernment has been _verified_ _experimentally_.
Let's hope that this will be the last time I will need to mention this:
The tweeter plays the message perfectly well. It is discernible, even when loud.
Please elaborate on how a speaker can cause permanent failures in an amplifier, such that changing the speaker, and even the sound, will continue to exhibit the same symptoms.
Thank you for your response.
Not sure what you're asking. In fact, these were _my_ questions--the very reason for this thread!
In most tests, there is a 9 V input-voltage to the amp, a 4 Ohms impedance of the speaker, and the amp is a 20 W Class D. Fine enough for Ohm's Law, but I expect that physical hardware might not like it, which is why I started the thread.
Good ones. Various adafruit, sparkfun etc Class D amps. There's even a non-Class D amp with a significant heat sink.
Real-world tests fail both of the above assertions. I expect that there are unstated--and I don't know what they may be--details here which make this different from your expectations: for one thing, there is just not enough power available to "melt" the tweeters, whereas the Class D amps are far more friable!
Indeed, loose wiring was an early culprit, but I now check for it.
Some amps used: Adafruit MAX9744; TPA2005D; pam8320.
which was my point! Here's hoping that this thread can now focus on the amp.
What I was hoping to hear was whether I could use amps with a lower wattage. It does not take much to generate loudness if the speaker is larger eg a 1-2 W amp can be very loud with a 3-inch speaker. Unsurprisingly, loudness diminishes with the cone size, but I did not expect to have to use a 20 W amp for a 1-inch speaker. (Yes, I _have_ read that the relationship is not linear.)
Thank you for your response.
Accepted band for optimal speech transmission and minimal bandwidth and minimum power requirement is 300Hz to 3.4kHz.
This standard has been in use by telephone industry for the last 100 years.
When a driver operated out of its bandwidth it no longer acts a transducer and instead as an inductance/resistance.
Excess power is dissipated in the coil directly as heat, causing winding wire insulation breakdown causing a short circuit or partial short circuit load presented to the amplifier and can cause amplifier failure.
What exactly are you trying to do ?.
Dan.
Thank you. But I am not concerned with RF transmission along a copper wire for long distances, using modulators/demodulators and circuit-switching.
There is theory, then there's data/experiment. The latter has proven sufficiency _for the purpose at hand_.
Inductance and resistance cannot be equated. I don't know that frequency change (that is, one outside a BW) results in excess power! And I generally question the coil-short's specifics.
The scenario above will be manifested by speaker damage, and perhaps also amplifier damage. The former has not been shown--to wit, when the speaker is moved to another board, to play a different sound, it performs well.
I am grateful for your response. Thank you for taking the time and sharing your knowledge. Is there another speaker you can recommend that'd be at/under 25 mm? I've tried a small surface-transducer (which BTW does show over-heating, but that's because its wattage was rated too low for my input
), but they need too large a surface. Is there a small enough folded-horn design that I can DIY? Are there unconventional designs which I can explore?Play a message. But the source is in a tiny, confined place.
Tell me detail more about the application and I can tell you how to go about it.
Dan.
First, you are "in the wrong place", though that is not clear. Nearly all the chat in this "Audio" forum is "High Fidelity", or nearly-so, or slightly-so. We avoid signal damage (yes, the guitarists apply selected signal damage).
The telephone 300Hz-3KHz (which is not "RF") is for the wires and amps. Until mid-century the handset fidelity was much less. Around 1900 it was only a couple peaks near 1KC and 1.5KC. This WAS enough to be intelligible and make the telephone companies huge. It is astonishing how much you can damage speech and still get a message across.
Yes male speech runs to 85Hz, but a 300Hz bass-cut bothers nobody but Sinatra and my mother. (Singers often want full-balls sound; my Mom can't tell me from my brothers on the phone because she goes by our different base-pitches.) Even 300Hz is generous. The phone does that so men don't sound thin. Just to communicate, a 600Hz bass-cut does almost no harm and 900Hz is still fully intelligible.
There's nothing needed past 3KHz. Just spit and whistles. (OK "sss" sounds benefit somewhat with 3.5KHz.)
Speech is so redundant that even quite narrow-band slices of it get the message through.
Speech will tolerate HUGE distortion. Clipped to utter square-wave. Transmitting just the zero-crossings sounds horrid but after a moment it is quite clear.
I have thrown intelligible speech 100' in a large loud reverberant space with 1/10 Watt of power. It was very un-natural yet communicated very well. I used a 4-foot horn for efficient throw, so this is not your solution. But key tips are: lose everything below 600Hz, and clip the snot out of it to get the most punch from the least Watts. (It helps to have a rising response before the clip-limit, and a falling response in the speaker so it does not ice-pick; but with 1" speaker there is not much you can do.)
The TPA2005D1 is rated 5.5V max; I would expect damage at 9V.
The "Stereo 20W" board will not make that much on 9V. Single-ended gives about 2 Watt in 4 Ohms. BTL gives 8W in 4 Ohms.
Aside from over-volting, it is hard to hurt such chips. I remember killing '709s like they were penny-candy (but they were $9). In self defense, nearly ALL chips now are very very well protected for heat and current. They will cut-out, but work again when they cool.
I *still* want to know the power source. A small wall-wart, or a few penlite cells, will crap-out under the load of a many-Watt amplifier.
Yes, use a smaller amplifier. As some hi-fi guy says, the First Watt is the most important. One Watt was the standard for good kitchen radios and phonos for 50+ years. With a 4-inch speaker, these would annoy neighbors. Your 1-inch tweeter is 1/16th as good at moving air, so it won't be so "loud"; still it should be loud enough to show the concept reliably and justify maybe a 4-Watt amp to hit the delivery target.
The telephone 300Hz-3KHz (which is not "RF") is for the wires and amps. Until mid-century the handset fidelity was much less. Around 1900 it was only a couple peaks near 1KC and 1.5KC. This WAS enough to be intelligible and make the telephone companies huge. It is astonishing how much you can damage speech and still get a message across.
Yes male speech runs to 85Hz, but a 300Hz bass-cut bothers nobody but Sinatra and my mother. (Singers often want full-balls sound; my Mom can't tell me from my brothers on the phone because she goes by our different base-pitches.) Even 300Hz is generous. The phone does that so men don't sound thin. Just to communicate, a 600Hz bass-cut does almost no harm and 900Hz is still fully intelligible.
There's nothing needed past 3KHz. Just spit and whistles. (OK "sss" sounds benefit somewhat with 3.5KHz.)
Speech is so redundant that even quite narrow-band slices of it get the message through.
Speech will tolerate HUGE distortion. Clipped to utter square-wave. Transmitting just the zero-crossings sounds horrid but after a moment it is quite clear.
I have thrown intelligible speech 100' in a large loud reverberant space with 1/10 Watt of power. It was very un-natural yet communicated very well. I used a 4-foot horn for efficient throw, so this is not your solution. But key tips are: lose everything below 600Hz, and clip the snot out of it to get the most punch from the least Watts. (It helps to have a rising response before the clip-limit, and a falling response in the speaker so it does not ice-pick; but with 1" speaker there is not much you can do.)
The TPA2005D1 is rated 5.5V max; I would expect damage at 9V.
The "Stereo 20W" board will not make that much on 9V. Single-ended gives about 2 Watt in 4 Ohms. BTL gives 8W in 4 Ohms.
Aside from over-volting, it is hard to hurt such chips. I remember killing '709s like they were penny-candy (but they were $9). In self defense, nearly ALL chips now are very very well protected for heat and current. They will cut-out, but work again when they cool.
I *still* want to know the power source. A small wall-wart, or a few penlite cells, will crap-out under the load of a many-Watt amplifier.
Yes, use a smaller amplifier. As some hi-fi guy says, the First Watt is the most important. One Watt was the standard for good kitchen radios and phonos for 50+ years. With a 4-inch speaker, these would annoy neighbors. Your 1-inch tweeter is 1/16th as good at moving air, so it won't be so "loud"; still it should be loud enough to show the concept reliably and justify maybe a 4-Watt amp to hit the delivery target.
Last edited:
I knew that. But, there _were_ threads tackling one of my chips eg
Why does amp distort even when fed voltage well below is range?
With a folded horn? Either way, the nature of the room helped.
Thought so.
I have no idea how I am supposed to do this to a voice file!
It was not at 9v. The latter was for other amps.
I feared so. Even the 9 V battery hardly provides anything near that, as judged by a multi-meter.
I expected so, but tests indicated otherwise.
It varies with the amp. It's either a bread-board power-supply, or a battery, or via USB+FTDI.
This is the first response to the original question asked. Thanks! A lower wattage, you seem to be confirming, ought to work--which is what my tests indicated early on--but that it needs more than a 1-inch speaker--which, again, my tests indicated early on. The question then became: How can I squeeze more loudness out of a 25 mm cone? IOW, how can I get _a_ loudspeaker mechanism--of whatever description, whether a surface transducer or another device devised--to generate larger air-waves?
Thank you greatly for your response.
As everyone else is saying the limitation will be the driver, not the amp.
Lets take the 1'' driver on parts express with the greatest possible volume displacement: the AuraSound Cougar NSW1-205-8A. It has 5.31 cm2 surface area and +- 4.5mm xmax so 2.4cm3 volume displacement at xmax.
2.4cm3 of volume displacement will get you:
(SPL @ 1m)
85hz : 67 dB
120hz : 73 dB
170hz : 79 dB
240hz : 85 dB
255hz : 86 dB
And that driver is only rated for 20W at peaks so it will melt long before the amps start to distort.
If those SPL figures are good enough then great, get that driver or equivalent and a small say 20W amp and you are set. It will probably not sound that good but it won't melt and it will produce sound. If you want louder or less distortion then get a bigger driver with a higher possible volume displacement.
Lets take the 1'' driver on parts express with the greatest possible volume displacement: the AuraSound Cougar NSW1-205-8A. It has 5.31 cm2 surface area and +- 4.5mm xmax so 2.4cm3 volume displacement at xmax.
2.4cm3 of volume displacement will get you:
(SPL @ 1m)
85hz : 67 dB
120hz : 73 dB
170hz : 79 dB
240hz : 85 dB
255hz : 86 dB
And that driver is only rated for 20W at peaks so it will melt long before the amps start to distort.
If those SPL figures are good enough then great, get that driver or equivalent and a small say 20W amp and you are set. It will probably not sound that good but it won't melt and it will produce sound. If you want louder or less distortion then get a bigger driver with a higher possible volume displacement.
Last edited:
I'll look up its specs, but I'm assuming that the 4.5mm is the maximum outward displacement of the cone, resulting in
5.31 sq. cm * 0.45 cm = 2.4 cu. cm.
I don't know how you got the numbers below (ie the exact formulas for the relationships between volume displacement, SPL and dB) but I'll take them on faith:
Hmm. An extremely informative post. Thanks so much.
I see.
I'll give it a try.
I've continued testing, though with more focus on the speakers. And, indeed, I've not, so far, confirmed the presence of damage in them.
Greatly appreciated.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.