Hello out there.
I have a specific problem for a non-typical audio problem.
We want to use a piezo speaker for an alarm system an are restricted in size for the PCB and electrical components.
But we need a lot of sound preasure in a specific frequency range of 3 - 4 kHz.
Unfortunately there are not many piezo devices delivering more than 100dB in a distance of 10cm (~80 dB in 1m).
My question is: "How Do I maximize the audiable SPL of a given piezo speaker for an alarm system?"
1. So the idea arose of adding a horn channel into the extruded plastic case to enhance the coupling of the acoustic impedance.
2. What audio waveform will generate most attention?
For 1. I have the practical problem, that typical piezo drivers are not delivered with a thiele small parameter set while the horn simulation programs I know need these parameters.
Any Ideas of how to make a rough simulation?
Thanks in advance!
Michael
I have a specific problem for a non-typical audio problem.
We want to use a piezo speaker for an alarm system an are restricted in size for the PCB and electrical components.
But we need a lot of sound preasure in a specific frequency range of 3 - 4 kHz.
Unfortunately there are not many piezo devices delivering more than 100dB in a distance of 10cm (~80 dB in 1m).
My question is: "How Do I maximize the audiable SPL of a given piezo speaker for an alarm system?"
1. So the idea arose of adding a horn channel into the extruded plastic case to enhance the coupling of the acoustic impedance.
2. What audio waveform will generate most attention?
For 1. I have the practical problem, that typical piezo drivers are not delivered with a thiele small parameter set while the horn simulation programs I know need these parameters.
Any Ideas of how to make a rough simulation?
Thanks in advance!
Michael
I had one of those piezo sounders with two discs and an internal driver circuit. After years usage, it was damaged by a power surge, and the transistors died. I replaced them with transistors I had lying around (BC237, BC337) and the sound output was SIGNIFICANTLY louder after the repair. So they are not - by default - as loud as they can be.
I think you are better to use a dedicated piezo alarm sounder with built in driver circuit rather than a piezo sounder and build your own driver. There are many of these. This one, for example, is 103dB @1 metre:- High Power Buzzer : Sounders : Maplin
basicly if want it LOUD you have to drive it with a higher voltage.a square wave is very anoying and the driver circuitry would be much less complicated.jer
@wakibaki
Thanks for the tip, but: My maximum diameter is 22mm with a maximum heigth of 10mm. If the device could be bigger, I would not have these problems 🙁
Further more, I only have 2.4 - 3V from two AAA alkaline batteries.
So I already made a driver circuit that is producing a square wave with 30Vpp.
Most datasheets specify a maximum Amplitude arround 30V. Sofar no manufacturer could tell me what happens if I go beyond these limits. I assume eighter of the two effects a) brakedown of the dielectic barrier b) crack in the material due to mechanical stress.
Increasing SPL will mean going far beyond these limits (Upp x2 will boost by 6dB).
As we plan to construct a extruded plastic case where it is easy to incorporate a horn channel - this would be the preferred way.
Any ideas on this?? Thanks a lot.
Thanks for the tip, but: My maximum diameter is 22mm with a maximum heigth of 10mm. If the device could be bigger, I would not have these problems 🙁
Further more, I only have 2.4 - 3V from two AAA alkaline batteries.
So I already made a driver circuit that is producing a square wave with 30Vpp.
Most datasheets specify a maximum Amplitude arround 30V. Sofar no manufacturer could tell me what happens if I go beyond these limits. I assume eighter of the two effects a) brakedown of the dielectic barrier b) crack in the material due to mechanical stress.
Increasing SPL will mean going far beyond these limits (Upp x2 will boost by 6dB).
As we plan to construct a extruded plastic case where it is easy to incorporate a horn channel - this would be the preferred way.
Any ideas on this?? Thanks a lot.
well,the theory behind a horn is to couple the waves to open space more efficiently, by a transformation of impedence hence an acoustic transformer,there are lots of studies and dsicussions on this subject i am no expert on this and i'm still learning,however i do know the theory behind it.most of the design criteria i've read so far envolves known driver parameters and i have the same issues as you.what parameters do you use if you are not using a dynamic driver?but i can asure you this,by using a horn this will increase the efficency but by how much is determened by horn length ,throat size and flare rate.there are few programs out there for free that help to determin those demensions.i have a few of them but i have not spent enough time yet to learn them.i once as a kid built horn with frequency cutoff of 1000hz i remember the the throat was 1" square and about 6" to 8" long with a exponetial flare and i drove it with various sized p.m speakers of 1 1/2" to 4" in diameter it seemed to work ,but being 12 or 13 i did not have the equipment to fully test it.in my thinking (at the time)it was supposed to make the speaker sound better,WRONG,what went in came out.but what i do remember is that it helped the dispersion of the sound waves making it seem a little louder only i had no way to measure it.the formula i used was simple and was in a very old audio handbook i got from the school libary and think i still have it somewhere if i can find it i'll look it up and post it.i hope that helps you out. jer
What do I exactly need to calculate/optimize a horn?
As far as I understand a horn transforms the acoustic impedance that the driver sees (without horn) to a different one. In principle I could roughly estimate that the horn and piezo are fourpoles (one transforms voltage/current to preasure/velocity, the other preasure/velocity to preasure/velocity).
Is there any tool that can roughly estimate a horns impedance without the driver?
Thiele small parameteres will not be easy for a piezo with a small chamber arround it. But acoustic impedance shoult be not a big deal to measure.
As far as I understand a horn transforms the acoustic impedance that the driver sees (without horn) to a different one. In principle I could roughly estimate that the horn and piezo are fourpoles (one transforms voltage/current to preasure/velocity, the other preasure/velocity to preasure/velocity).
Is there any tool that can roughly estimate a horns impedance without the driver?
Thiele small parameteres will not be easy for a piezo with a small chamber arround it. But acoustic impedance shoult be not a big deal to measure.
correct.that is the same problem i'm having also, as i want to use a horn with small electrostatic driver.and i have been searching for the answer (and only found 1 or 2 oline calculaters);if a piston(speaker) is moving back and forth x distance at y frequency with a surface area of z what is the sound pressure level in db. everybody mistakes that with power requirment and driver efficiancy i do not care about how much power is required i want to know if an object with z area is moving x at rate of y how loud is it. i believe (although im no math major)with those known factors one could determin the impedence and apply to a horn, as the flare rate is the transformation ratio and the length determens the cut off frequency.thats how i seem to understand it.if anyone knows more abuot this feel free to jump in anytime.do a search on math cad and you will find a guy named martin king whom has done some very exstensive reaserch on the subject and has written several programs in mathcad to detremin such factors.good luck. jer p.s. if you find more answers i would like to read about them,meanwhile if you can't find the caculators i've mentioned(if you can use them)i will post them for you,good luck!
As the piezo I plan using is already encapsulated (like a driver) and has an opening of aprox 2mm my naive approach would be taking a growth function for a horn and just creating a throat from 2mm to my max surface area possible.
I know there are different formula for hyperbolic, exponential, spherical etc. horn types.
The questions I have are:
1. Does the shape of the horn (symmetric vs asymmetric) has a big influence on the impedance and thus the frequency response?
2. Can I (when yes, how can I) roughly calculate what the gain of a horn will be?
3. Is it possible to make a rough driver model for the piezo from mechanical dimensions? I know the diameter, the size of the preasure chamber, can weigth the piezo ceramic, maybe measure the exitation on a certain frequency.
Thanks for your help!
I know there are different formula for hyperbolic, exponential, spherical etc. horn types.
The questions I have are:
1. Does the shape of the horn (symmetric vs asymmetric) has a big influence on the impedance and thus the frequency response?
2. Can I (when yes, how can I) roughly calculate what the gain of a horn will be?
3. Is it possible to make a rough driver model for the piezo from mechanical dimensions? I know the diameter, the size of the preasure chamber, can weigth the piezo ceramic, maybe measure the exitation on a certain frequency.
Thanks for your help!
a horn will not add gain it only helps to couple the sound wave to the to the open area in other words ,it helps to disperse the sound waves into a larger spherical wave front rather than a smaller one reducing the the conical beaming effect in doing this it may seem louder off to the sides due to better dispersion.but it will not add any energy to make it louder.you can't get some thing for nothing.you can also use a horn to concentrate all of the energy to smaller spot thus lessening the dispersion into a narrower conical beam.is this what you mean?yes this can intensify the sound but by not much, at adegradation of dispersion which isn't what you want for an alram if you need it louder then the rule of thumb is for every doubling of equivelant drivers (each driven at the same power level as one driver is driven)equals an increase of 6db.so therefor 2 drivers =+6db,4 drivers =+12 db,8 drivers=+18db,16 driver+24db,and so on.does that help? jer
Hi geraldfryjr,
I think you are only partially right.
Imagine a bass driver (moving cylinder) at a low frequency. When you leave the case arround it away it will perform an acoustic short circuit. That means the mechanical energy will be result in a massive air flow but only from the front to the back of the driver.This causes massive turbulances and thus only mechanical losses with only little soundpresure going into the long distance field.
When closing the back of the bass driver with a case the mechanical movement is dampened from two sides where the dampening from the back (compressing air in the case) will be the stronger effect (acts like a mechanical spring).
To my understanding - increasing the energy transferred from the moving cylinder to the air even better, a better coupling of the acoustic impedance is neaded.
I learned that in acoustics it is possible to make an equivalent electrical circuit for the acousitc problem.
e.g. sound preasure is equivalent to a current and velocity to a voltage.
When given a cylinder moving forth and back with a fixed velocity (fixed frequency and excursion) this would be equivalent to having an AC voltage source. To maximize the power drawn from that source, it is neccessary to increase the current by decreasing the impedance of a connected load.
In my understanding this is what happens when a horn is added to a acoustic driver. By optimizing the horn it will change the acoustic impedance to increase soundpreasure output for the same excursion of the diaphragma.
But how does this help with my problem? Thanks for input!
I think you are only partially right.
Imagine a bass driver (moving cylinder) at a low frequency. When you leave the case arround it away it will perform an acoustic short circuit. That means the mechanical energy will be result in a massive air flow but only from the front to the back of the driver.This causes massive turbulances and thus only mechanical losses with only little soundpresure going into the long distance field.
When closing the back of the bass driver with a case the mechanical movement is dampened from two sides where the dampening from the back (compressing air in the case) will be the stronger effect (acts like a mechanical spring).
To my understanding - increasing the energy transferred from the moving cylinder to the air even better, a better coupling of the acoustic impedance is neaded.
I learned that in acoustics it is possible to make an equivalent electrical circuit for the acousitc problem.
e.g. sound preasure is equivalent to a current and velocity to a voltage.
When given a cylinder moving forth and back with a fixed velocity (fixed frequency and excursion) this would be equivalent to having an AC voltage source. To maximize the power drawn from that source, it is neccessary to increase the current by decreasing the impedance of a connected load.
In my understanding this is what happens when a horn is added to a acoustic driver. By optimizing the horn it will change the acoustic impedance to increase soundpreasure output for the same excursion of the diaphragma.
But how does this help with my problem? Thanks for input!
yes i understand this,sometimes i don't explian myself very well.however i don't know the equations to horn design but if you do some searching on this forum your bound to find someone with more knowledge and more experience than i however if i find any more info that help you as well me,i will post it here as our situations are similar.good luck.jer
I think a plastic horn is indeed the best solution to increase efficiency. Yes, simulating will be a problem due to lack of data. I would suggest to just skip that step. Give the horn an exponential taper, fit the throat area to the piezo surface, and maximize the mouth area and depth, given the constraints of the product.
very well put, you might also consider using two or more drivers as well with a horn depending on your size constrants since you already have a 30v driver circuit as it would give some extra db's as well. jer
I guess my main problem will be the that the horn should exit in an 90° angle to the sound exit of the piezo driver.
When asking manufacturer of piezo horns I got the answer - we are not simulating, we do try and error....
- Having more drives will be a cost and space issue and thus not a solution for me.
+ Maybe I can start milling a horn structure into an aluminum block an measure the result with a microphone.
When asking manufacturer of piezo horns I got the answer - we are not simulating, we do try and error....
- Having more drives will be a cost and space issue and thus not a solution for me.
+ Maybe I can start milling a horn structure into an aluminum block an measure the result with a microphone.
just curious,i can understand this "- Having more drives will be a cost and space issue and thus not a solution for me." but i don't understand "+ Maybe I can start milling a horn structure into an aluminum block an measure the result with a microphone." i know that r&d is not cheap.but wouldn't it be easier and cheaper to use somthing like wood,balsa wood,or plasitic block.or use some sort of flexible sheet to build the horn with that can be easily reshaped until you find the proper diminsions your after?i once thought of using modeling clay to get my shape and then cast the horn in plaster or wax in order to test it and if i wasn't satisified with the results, the material could be recycled and be reused for a new design.just some thoughts. jer
As Gerald says, you'll probably have to "try it and see".
Kavermei's suggestion of an exponential flare sized to fit the available space sounds good. For experimentation, just roll a conical horn out of stiff paper and attach it to the transducer with sticky tape.
Your 90 degree bend should go as close as possible to the piezo driver. Make it a curved bend, not hard right angle. This might be hard to mold in plastic. Consider making the horn straight, and mounting the transducer to the horn rather than to the PC board.
Try and get the horn throat to fit the opening in the transducer as accurately as possible. Errors in size or join accuracy will have quite a large effect.
For inspiration, take a look at "personal alarms". These are small devices that you carry and are set off by a "panic button" or by pulling out a pin. They often have small plastic horns moulded into the case to increase the loudness of the piezo alarm.
Kavermei's suggestion of an exponential flare sized to fit the available space sounds good. For experimentation, just roll a conical horn out of stiff paper and attach it to the transducer with sticky tape.
Your 90 degree bend should go as close as possible to the piezo driver. Make it a curved bend, not hard right angle. This might be hard to mold in plastic. Consider making the horn straight, and mounting the transducer to the horn rather than to the PC board.
Try and get the horn throat to fit the opening in the transducer as accurately as possible. Errors in size or join accuracy will have quite a large effect.
For inspiration, take a look at "personal alarms". These are small devices that you carry and are set off by a "panic button" or by pulling out a pin. They often have small plastic horns moulded into the case to increase the loudness of the piezo alarm.
@geraldfryjr
Aluminum was the material of chioce as the guys responsible for milling won't let some non metal parts on there milling machine. Wood is easier to cut but they won't process it...
@Don Hills
The PCB should not be manufactured in china - this can only be achieved by using all parts as SMD without any additional cables. Due to this constraint I could only use a driver with 90° exit (but there are only few and not loud enough).
The hint with the alignment is good - this will make the plastic case a little more complex but this will make somebody else some headache 🙂 .
Aluminum was the material of chioce as the guys responsible for milling won't let some non metal parts on there milling machine. Wood is easier to cut but they won't process it...
@Don Hills
The PCB should not be manufactured in china - this can only be achieved by using all parts as SMD without any additional cables. Due to this constraint I could only use a driver with 90° exit (but there are only few and not loud enough).
The hint with the alignment is good - this will make the plastic case a little more complex but this will make somebody else some headache 🙂 .
yes, i understand : "Aluminum was the material of chioce as the guys responsible for milling won't let some non metal parts on there milling machine. Wood is easier to cut but they won't process it..." as i was refering to the r&d process and not the end product. jer
- Status
- Not open for further replies.