Piezo VS voice coil tweeters

Voice coils, with their endless crossover problems, must be superior to piezo-driven tweeters, since most, if not all of you use them. Is their non-linearity and "scratchy high frequencies" their only downfalls?

Apart from not needing a xover, are there any advantages to using piezo tweeters? How is development of piezo's going? Are they getting better? Will they ever reach close to a par with voice coils in tweeter applications?
secateurs said:
Voice coils, with their endless crossover problems, must be superior to piezo-driven tweeters, since most, if not all of you use them. Is their non-linearity and "scratchy high frequencies" their only downfalls?

Apart from not needing a xover, are there any advantages to using piezo tweeters? How is development of piezo's going? Are they getting better? Will they ever reach close to a par with voice coils in tweeter applications?

IMHO, a lot of the bad rap piezos get is because of how they have often been used, not in their inherent nature.

A piezo absolutly NEEDs an XO. Jon Risch (somewhere on AA or at his site -- URL not at hand) outlines how to best approach the XO. If you are using the cheap horns, they need external damping (with something like ductseal), the cones benefit from a little treatment (damar, puzzlecoat, or a layer of aluminum art foil), and the internal back chamber needs a little damping. The paper cone can be removed from the horn entirely and used with some loss of effiency (and the horns colorations) -- what PHY-HP uses on their VERY expensive 12" is, from the outside, indistinguishable from the paper cone i have pulled out of Motorola piezo horns. And then there is Audax's big-buck ovoid shaped piezo dome that by all accounts is stunning.



diyAudio Moderator Emeritus
2001-09-18 2:33 am
Connecticut, The Nutmeg State
I have used piezos, and Motorola sent me a packet of frequency response charts years ago. They also sent an Application Note with advice about crossovers.

Many of these piezos have a nice response except for one or two nasty peaks that need taming. You can tame peaks on a magnetic speaker with inductor/capacitor parallel networks, but piezos are different.

I have searched, but I am surprised that there isn't some sort of online Piezo Fan Club where more competent electronics hobbyists than myself have worked these networks out, and with other ideas on how to use these nifty drivers.

Piezos seem to have a lot going for them. Pity there isn't more interest in them.
Risch copy & paste

(cut and paste, most of the below information credited to JON RISCH)
The Parts Express catalog suggests putting a 20-Ohm resister inline with any Piezo tweeter to make it a more stable load for an amp.Will this not also attenuate the tweeter? If so, and if I need further attenuation, can I simply add more resistance? Is there a rule of thumb for how much attenuation I will get with further resistance, or a way to compute this number?

The recommended resistor is to help protect the amplifier from oscillating due to the raw capacitance that is a piezo driver. Adding resistance in series with a piezo will actually roll off the highs a bit, adding more will roll off the highs noticably. To attenuate a piezo, add a series cap, which creates a voltage divider with the capacitance that is the piezo drive element. Most piezo elements run in the 0.1 to 0.26
uF range, so a cap of the same value as the piezo element will attenuate it 6 dB.

Piezo's can be crossed over, and to great advantage. I have often thought that some of the bad rap piezo drivers have is due to the "you can use them without a crossover" fallacy. Yes, you _can_ use them without a crossover, but just because you can get away with it, does not mean it is optimal.

Since most piezo's are used in inexpensive systems, the cost of adding in "unecessary" components is often never even considered.

How to crossover a piezo:
Add a resistor in parallel, and the driver can be made to look like a current driven device to any outside components, such as a crossover cap. However, to keep costs and power dissipation down, 8 ohms is way too small of a value. The impedance of most piezo's is still quite high at 20KHz, so use a 22 ohm resistor, this makes any series crossover cap smaller and less expensive, and the resistor dissipates less energy. Use of an 8 ohm parallel resistor will also tend to lose you a little bit of output level.

For most piezos, use of a 22 ohm resistor, and a 4-4.7 uF cap will allow the response to be identical to what it was in stock form, but rolls off the lows at 6 dB/oct below 1 kHz or so. This actually increases the power handling of the piezo, as it is voltage limited. Exceed the voltage used to pole (polarize the piezo element during manufacture) the unit, and it will loose sensitivity, and eventually burn out. Most pro grade piezos will handle 35 volt transients, and 28 volts continuous, which are 150 watts and 100 watts into 8 ohms respectively.

Add in a capacitor and 22 ohm resistor, and the power handling could effectively be quadrupled, as the LF voltages are not imposed upon the unit, just the HF voltages.

Piezo's crossed over in this manner don't sound as harsh and spity, and tend to be quite a bit more reliable. Many of the piezo units have a mild peak just before they roll off in the LF, so making the series cap a little smaller can actualy flatten response, and provide even more protection and smoother sound. For the smaller piezo units that cut off at 4-5 kHz, a series cap of 1.5 uF will do the trick, larger units that go down to 3 kHz can use a 2.2 uF, and the large compression driver units meant to be mounted on a horn need about 5 uF, as they do not peak, and any higher would lose the sloping output even more.

Attenuation, HF roll-off AND the crossing over can all be done at the same time. To attenuate, place a cap in between the piezo and the 22 ohm resistor that is shunting across the unit, then if HF roll-off is desired, use a series resistor in this location too. Then the series crossover cap should be in front of the 22 ohm shunt.

Looking from the amp, first the series crossover cap, say 4 uF, then the 22 ohm shunt from hot to ground, then a series cap of about 0.15 uF for 6 dB attenuation, and then a series resistor of about 30-50 ohms to tame the very top end, then the piezo itself.



diyAudio Moderator Emeritus
2001-09-18 2:33 am
Connecticut, The Nutmeg State
Thank you, Dave, that was very helpful. I will do some experimenting with those guidelines soon. Also, thank you for your T-Line site.

I still wonder if you could use networks using an inductor, to tame specific peaks. Maybe by using a parallel resistor? Hmm.

LOL, as good as this information is, I still think that piezos deserve their own Fan Club. :)


diyAudio Moderator Emeritus
2001-09-18 2:33 am
Connecticut, The Nutmeg State
Just came across this. Wonder what is down the road for piezoelectrics.


A NASA technology that could make everything from
audio speakers to heart pumps smaller and more efficient has
been recognized as one of the 100 most significant
technological advancements of the past year.

Dubbed THUNDER, for Thin-Layer Composite-Unimorph
Piezoelectric Driver and Sensor, its potential applications
could be applied in electronics, optics, jitter (irregular
motion) suppression, noise cancellation, pumps, valves and a
variety of other fields.

Researchers at NASA's Langley Research Center,
Hampton, VA, have taken advantage of a well-known phenomenon
exhibited by piezoelectric materials. These materials
generate mechanical movement when subjected to an electric
current, as in a speaker or actuator, and generate electrical
charge in response to mechanical stress, as in certain types
of sensors.

The Langley researchers, a multi-disciplinary
materials integration team, have succeeded in developing and
demonstrating a piezoelectric material that is superior to
commercially available piezoelectric materials in several
significant ways. It is tougher, more durable, allows lower
voltage operation, has greater mechanical load capacity, can
be easily produced at a relatively low cost and lends itself
well to mass production.

The first generation of THUNDER devices are being
fabricated in the lab by building up layers of commercially
available ceramic wafers. The layers are bonded using a
Langley-developed polymer adhesive. The process results in a
prestressed device with significantly improved performance.
In addition, the process is controllable and results in
highly uniform pieces of hardware.

In the ideal fabrication process -- which researchers believe
is technically feasible -- piezoelectric ceramic materials
would be ground to a powder, processed and blended with an
adhesive before being pressed, molded or extruded into wafer
form. The result would be increased ability to tailor
properties, more flexibility in choosing methods of
manufacturing and increased amenability toward mass

THUNDER wafers could be any practical size from areas
of a few square millimeters to several square meters and
thicknesses of fractional millimeters to several millimeters.

Memoranda of agreements have been signed with six
companies to develop THUNDER technologies and related
commercial products. Approximately 15-20 more companies are
negotiating agreements.

THUNDER technology will be honored by Research and
Development magazine at the annual "R&D 100 Awards" banquet
October 14 in Philadelphia, PA. The R&D 100 Award, formerly
IR-100, is presented annually by the magazine to the
innovators of the 100 most technologically significant new
products of the year. Selections are made by a panel of
scientists and engineers after studying new technologies from
around the globe. Langley has competed in the program for 27
years and has earned 31 awards.

End quote.

LOL, the first letters of "Thin-Layer Composite-Unimorph
Piezoelectric Driver and Sensor" don't spell THUNDER, but who's complaining?

PS: Looks like someone is doing something with it. Apparently this material has 10 times the "wiggle ability" of present products, which I believe is lead zirconate and titanate.
All for the Piezo Fan Club

I'd be in on a piezo fan club! I personally think they're great, which is one of the reasons i want to incorproate one into my first speaker project.

That development by NASA sounds good. I wonder when it will reach the commercial audio market....

Thanks Dave for that information. Sounds good. I think i'll use some of your suggestions. As i am currently doing a course in Electrical Engineering (first year), i may soon be able to design my own low-cost crossover to use, too.
Audax produces an extremely high quality piezo tweeter, using a different method to the usual piezo tweeters.
They use a thin piezo polymer, coated on both sides with gold for the electrical conductivity, and mounted so it forms an elliptical dome about 25mm x 50mm. It is direct radiating, with no horn or flare, apparently the moving mass is very low and has frequency response way past 35kHz with no dome breakup.

World Audio Design www.worldaudiodesign.co.uk used to sell them individually, but they only sell them now with other drivers (other Audax mids and woofers) to form their kits.

Cheers, Adrian
Piezo vs electromechanical tweeters

At first glance, a piezo tweeter would appear to be ideal for the HF section of a speaker system because it does not require a XO.
However by far the majority of drive manufacturers concentrate on the "standard" EM types which would seem to empasize that Piezo's are less than ideal.
What are the shortcomings of Piezo Tweeters?
My first DIY speaker is intended to be a good quality but cheap set of small speakers for use with a PC, initially powered by the PC (in other words direct from the sound card) but later on via a small external amp (say 11W). I hope to make the speakers have a good sound but do not see the reason for making them "Hi-Fi" per se as onboard PC sound cards are not that good to start with. Hence the addition of a crossover would seem "over-kill" and an unecessary compexity.
However, not-with-standing this project, how is the quality of Piezo tweeters for projects that would be considered "Hi-Fi"?

All comments appreciated!

Motorola did the work on the recent crop of piezos that have been around for the last 20-30 years. I think CTS now have the rights and market them. The driver that is probably the most popular is the small horn tweeter which begins to be useful above about 4k Hz. It is quite efficient. They used to make a similar one that was marketed as a "Hi Fi" model. It had a smooth throat to the horn where as the other one has quite prominent ribs in the mouth of the horn. As I understand it the "Hi-Fi" version is smoother in response but less efficient. People have reported that they sound better with a cross over although, as you say, they don't technically need one. Alan Smith pubilshed some work in the JAES about 30 odd years ago that examined the tranient behaviour of horns and he made the point that they were better with a decent slope x-over. They have good power handling and high output.
As to their limitations. Some people just won't use horns! That is not so much a limitation as a faith statement. Secondly, as is pointed out in this forum when ever they are mentioned you can't use normal x-overs. A Piezo is a capacitor. The Motorola paper (Bost, I think in the JAES) will advice you to attenuate response with other capacitors in series. Small values such as 0.1uF from memory. Resistors tend to taper the high frequency responce. 20 or 30 ohms being typical. That will tame the top end a bit and ensure a more stable load for your amp.
If you use the small version you need a bass-mid that will go to 4 K Hz. Since the original one was produced they ahve lowered the useful range and there are both round and rectangular versions that will go down to 2 K now. Generally Hi-Fi shops don't stock them so you will get more joy from public address/sound reinforcement people.
I am glad you are thinking using them as I think they have some potential. Let us know how they go.
Jonathan Bright said:
...As to their limitations. Some people just won't use horns! That is not so much a limitation as a faith statement....

Most horns have a measurable limitation that is: severe linear distortion. Even simple computer models such as Hornresp usually show sharp notches in the frequency response and phase response. This indicates severe ringing. Similar ringing occurs naturally with various musical instruments, so some people won't be able to hear it except after prolonged listening as the ears finally learn which resonances are characteristic of the speaker, not the song. That's when listening-fatigue sets in.

It's definitely not a "faith" but a human reaction to avoid certain types of speakers that are consistently bad. I'm not saying that "all" horn speakers sound bad, but certainly all of the ones I've heard belong in that category. People don't seem to understand that designing a perfect horn-shape is the easy part. The hard part is designing a stable amplifier with reduced output impedance. I'd in fact be willing to design a horn speaker, if it wasn't so difficult to design a tailored amplifier for it (think: stable negative output impedance with highly reactive feedback, ouch!).

Anyway, who said that piezo tweeters had to be hornloaded? I think Audax makes an elliptical direct-radiating piezo tweeter.

There was also once a tweeter available from Motorola that didn't have a horn. Maybe it is still made by CTS as well.
As already suggested, level matching and a simple crossover can greatly improve the quality of a piezo tweeter.

The Audax tweeter mentioned above used a piezo polymer and was used together with an "air-cored" step up transformer. I have never heard it but it showed good measurements. Unfortunately it is not officially available anymore.


Re: Piezo vs electromechanical tweeters

polsol said:
What are the shortcomings of Piezo Tweeters?

My thoughts, non-technical, purely subjective.

Piezo tweeters are great for in your face rock and roll where realism is not paramount. They're inexpensive, easy to produce and have balls. They are suited for a small PA system where a large compression horn is not practical. I would not use them in a house system.

I find their sound harsh and lacking in depth, ie: imaging

That being said, I have used them in many systems over the years, but always with the understanding that you are getting bang for your buck, not something to set the world on fire. You can replace a piezo ten times before you reach the cost of a good compression unit.

There once was a long throat piezo that sounded OK. It may be the one Jonathan is speaking of, ie: the HiFi model.

This may sound a bit strange but the best I've ever heard a piezo sound was when a choke coil, yes, a choke coil was wired in series. About a 5mH if memory serves me correctly. It boosts the high end, much the opposite effect one would expect, and the tweeter actually starts sounding OK.

If your intent is to use them for computers speakers then it's hard to advise you. They may or may not meet your needs. It depends on how much you will be listening to them. If you listen a lot then I'd consider stepping up a bit.

This is the tweeter I use in my computer speakers. Cheap and not bad sound.


small horn
Cal is right. I've got a circuit in the files somewhere. The coil had a high DC value, 12 ohms I think and it was preceded by a cap' of a few uF. At first sight it lookled fairly conventional but the high DC coil was needed. Again, Charles is correct there was a non-horn model but I've only seen photos and that was a while ago. I think these units are interesting because of the high impedence and high efficiency. It lend themselves to low power class A amps' with an active Xover. You could have an amp that was designed for a couple of watts into 30-40 ohms. V.clean and no great demands in the way of current etc. Yes CM there are problems but I should have explained that the ringing you mention is precisely what Smith's work identified and addressed in his paper. If your interested I'll dig it up. I think it is from the 1960's.
Ben Duncan is a great believer in horns but he advocates active systems where each horn is used over a v.narrow range say, 2 octaves.
I have a pair of cheap & nasty Philips speakers which use an ordinary-looking peizo sounder (the flat brass plate kind) as the tweeter. They sound quite reasonable - certainly not as harsh as the typical horn loaded peizos, and have no crossover.
Might be interesting to experiment with these sounders, with and without home-made horns.
I'm old enought to remember when some of the better record players used peizo tweeters, which were usually about 2-3 ins. square X 3/4 in thick, and sounded pretty good to me at the time.