I want to add a piezo sounder to an existing arduino project (a USB caps lock indicator - I want it to bleep when I (usually accidentally) press caps lock and peebl when I release it).

In my spares box I found a 28mm enclosed piezo sounder which works very nicely driven with square waves of suitable frequencies to sound the notes I want. But both its diameter and thickness are too big to fit in the existing box. No probs, I thought, and ordered a couple of bare 15mm piezo sounder disks from eBay.

At first I thought they weren't working at all until listening very carefully I could just make out a very faint sound.

So I opened up my 28mm one to see what magic dust it might contain. The back came off with some difficulty but the disk appeared to be secured in the shell quite firmly, maybe glued around its edges. I didn't want to bend the brass disk trying to get it out.

Is anyone familiar with these things? Do they need to be mounted in a particular way to make a decent sound? Or do I just have a pair of duds? I can accept that a 15mm one won't be as loud as a 28m one, but I should at least be able to hear it.

 

The case of a piezo audio transducer is an active part of the device just like the enclosure for a speaker. It needs to be designed to get max efficiency at the frequencies of interest. Why not just buy a smaller complete transducer?

 

The case of a piezo audio transducer is an active part of the device just like the enclosure for a speaker. It needs to be designed to get max efficiency at the frequencies of interest. Why not just buy a smaller complete transducer?

Space is extremely tight. Anyway, if something doesn't work, don't you want to know why and how to fix it?

 

The plastic housing is an acoustic resonant cavity- to get any decent sound out of these, you need a cavity.

 

Space is extremely tight. Anyway, if something doesn't work, don't you want to know why and how to fix it?

Because I am not an acoustic engineer. This is not an electronics problem.

Do you know anything about speaker design? If you have a speaker driver that is not in an enclosure, it cannot produce any bass frequencies because the sound coming from the back cancels the sound from the front. The speaker enclosure is extremely important. It has to damp the resonance and enhance frequencies below the resonance to get good sound. It is the same for your piezo disk, but at much higher frequencies because of the small size. If you want to try to design the housing, go for it, but it will be a lot of work and your results are unlikely to he as good as commercial ones you can get for a couple of bucks.

 

A piezo transducer produces sounds only near is few resonant frequencies that are fairly high pitched from 1kHz to 12kHz.
The transducer in a smoke detector resonates at about 3kHz. Smaller piesos produce higher frequency sounds. Here is the frequency response of a 34.5mm diameter one in its enclosure:

 

The plastic housing is an acoustic resonant cavity- to get any decent sound out of these, you need a cavity.

These things operate over a range of several octaves at least, so if it were resonant then the Q wouldn't be worth having. More likely a matter of acoustic impedance matching to free air.

 

A piezo transducer is used as a warning beeper. Music and speech sound awful because of the few high-Q resonances.
Even kids toys use a REAL speaker, not a piezo for saying some speech.

 

Just happen to have disassembled an alarm clock. It has a raw piezo disk glued to the the case with an air gap of about 1mm. In the center of the disk is a hole about 1-2 mm in diameter to the outside of the the case, which is maybe 1 mm thick. The back of the piezo is open to the inside of the case. You could try something like that.

 

Just happen to have disassembled an alarm clock. It has a raw piezo disk glued to the the case with an air gap of about 1mm. In the center of the disk is a hole about 1-2 mm in diameter to the outside of the the case, which is maybe 1 mm thick. The back of the piezo is open to the inside of the case. You could try something like that.

Interesting, though it'd be hard for me to replicate in my existing case. But I've found some on eBay with a low profile aluminium cover with a hole in it which I should be able to use. So the answer to my original question definitely seems to be that you need a small enclosed body of air to acoustically load the disk. Since the bare ones work very poorly as sounders I guess they're sold as transducers for drum kits and the like. Connected to an oscilloscope you get a nice pulse if you tap it with a finger.

 

Interesting, though it'd be hard for me to replicate in my existing case. But I've found some on eBay with a low profile aluminium cover with a hole in it which I should be able to use. So the answer to my original question definitely seems to be that you need a small enclosed body of air to acoustically load the disk. Since the bare ones work very poorly as sounders I guess they're sold as transducers for drum kits and the like. Connected to an oscilloscope you get a nice pulse if you tap it with a finger.

@BobTPH I wonder what size the one in your alarm clock was? The ones with the Al cover came the other day and they're better but not hugely and still not loud enough driven from an Arduino. I suspect size may be more significant than enclosure. These are only around a quarter of the area (15mm vs 28mm dia) of the original one I tried (which is too big to fit the existing enclosure). As an experiment I might try adding a transistor in order drive the two terminals antiphase rather than single ended, but fitting the extras in the enclosure would be challenging.

 

If you stepup the signal voltage from the Arduino that drives the piezo then it will be louder. A bridged (antiphase) amplifier will double the signal voltage and a transformer (maybe needing a current amplifier) can give almost any signal voltage.

 

@BobTPH I wonder what size the one in your alarm clock was? The ones with the Al cover came the other day and they're better but not hugely and still not loud enough driven from an Arduino. I suspect size may be more significant than enclosure. These are only around a quarter of the area (15mm vs 28mm dia) of the original one I tried (which is too big to fit the existing enclosure). As an experiment I might try adding a transistor in order drive the two terminals antiphase rather than single ended, but fitting the extras in the enclosure would be challenging.

I think the one in the alarm clock is 15mm. Definitely not 28.

How are you driving it? GPIO and ground? If so, use two GPIOs set opposite to each other to get twice the driving voltage.

 

Twice the driving voltage produces an output that is +6dB that is only a little louder. 10 times the voltage sounds twice as loud since the sensitivity of our hearing is logarithmic.

 

I think the one in the alarm clock is 15mm. Definitely not 28.

How are you driving it? GPIO and ground? If so, use two GPIOs set opposite to each other to get twice the driving voltage.

Yes, GPIO and ground, using the Arduino tone() function. This leverages one of the chip's internal timer/counters so it wouldn't seem to be possible to get antiphase outputs as I could with a software loop.

Don't you hate being proved wrong by a mere computer? Just to be sure, I asked ChatGPT to give me a code snippet to generate two antiphase 1KHz signals on two different GPIO pins, for 0.1 sec, and it's done just that, using one of the timers!

But the simpler solution (breadboarded) of adding a transistor to generate the antiphase gave a louder sound but still not loud enough by quite a lot.

But @BobTPH - if it's a quartz alarm it presumably runs off a single 1.5V cell, or maybe two if it's radio controlled. That's going to give considerably less drive tha I'm giving mine with an Arduino running at 5V. These piezo disks all look the same, but maybe there are different qualities.

 

But @BobTPH - if it's a quartz alarm it presumably runs off a single 1.5V cell, or maybe two if it's radio controlled. That's going to give considerably less drive tha I'm giving mine with an Arduino running at 5V. These piezo disks all look the same, but maybe there are different qualities.

The alarm clock is line powered with 9V battery backup. So it likely uses a higher driving voltage. Do you have any specs on the piezo element?

 

I have a couple of radio controlled quartz alarms, one, the alarm sounds at 4.096KHz (evidently tapped off the divider fron a 32.768KHz crystal) which these days I can easy sleep through with my loss of hearing at higher frequencies. I took it apart, and the sounder is a black cylinder approx 1cm dia and 1cm tall. One terminal is marked +, implying maybe it contained a transistor, yet connecting it to a 1.5V cell it just made a little click, indicating that it doesn't contain an oscillator. It made a very acceptable noise when connected to the Arduino.

So if it doesn't contain a transistor or an oschillator, I thought why wouldn't it work reverse polarity, albeit perhaps less efficiently? I reversed it. No sound. I reconnected it the right way round. Still no sound! Somehow I'd killed it. So now no reason not to tak it apart.

Turns out it was electromagnetic after all!

Not sure how I killed it - maybe demagnetised the permanant magnet? Yet it was definitely o/c.

However, the 28mm enclosed disk I have which works nicely is definitely a piezo as you can see if you take the back off.

 

generate two antiphase 1KHz signals

The small piezo might not produce much sound at the "low" frequency of 1kHz. It is much louder at its resonant frequency that might be 4kHz or 5kHz.

EDIT: Now we know that there is no piezo. The coil and magnet tiny speaker needs a fairly high current to drive it. Without a schematic in this thread we do not see if the speaker is fed through a low value capacitor that is suitable to feed a high impedance piezo but is useless to feed a low impedance coil and magnet speaker.

 

The small piezo might not produce much sound at the "low" frequency of 1kHz. It is much louder at its resonant frequency that might be 4kHz or 5kHz.

The 27mm enclosed disk I have works just fine, but is too big for the applicarion. My hearing is 45dB down at 4kHz.

 

My hearing is 45dB down at 4kHz.

A hearing loss of -45dB at 4kHz is normal for 70 years old men. My hearing is also normal for my age and I got hearing aids when I turned 69 which was 9 years ago. Now my hearing is young again and the noise reduction feature on the hearing aids works well.