12 Watt amp with piezo tweeter coupling question

[aliyesami]

I am planning to use the LA4450 12W power amplifier to drive a piezo tweeter . the input to the amplifier is 23Khz square wave .
I am putting a 8 ohm 50 W resistor in parallel with the piezo tweeter to satisfy the amplifier load requirement.
will this setup work ?
I have attached the schematics as jpg file.

 

[FvM]

I am putting a 8 ohm 50 W resistor in parallel with the piezo tweeter to satisfy the amplifier load requirement.

Where do you see the requirement?

 

[KerimF]

The piezo tweeter (GT-1005) is rated for an amplifier delivering 50W RMS likely to a load of 8 Ohm.
Therefore the RMS rated voltage of the tweeter is:
P = V^2 / R
V = SQRT ( P * R )
V = SQRT ( 50 * 8 ) = 20 Vrms
or
Vpeak = 20 * 1.414 = 28.28 V

The current is relatively small.

 

[aliyesami]

the datasheet is showing 20Watts for RL=4ohm and 12 watts for RL=8ohm.

- - - Updated - - -

The piezo tweeter (GT-1005) is rated for an amplifier delivering 50W RMS likely to a load of 8 Ohm.
Therefore the RMS rated voltage of the tweeter is:
P = V^2 / R
V = SQRT ( P * R )
V = SQRT ( 50 * 8 ) = 20 Vrms
or
Vpeak = 20 * 1.414 = 28.28 V

The current is relatively small.

which tweeter can i use for 20W 4ohm load ? as the amplifier chip i bought is rated for 20w at 4ohm.

 

[KerimF]

I noticed that your Vcc (single supply) is equal to 26.4V

Therefore, assuming the amplifier output is rail-to-rail, the maximum peak voltage (of the square wave) on the tweeter is:
Vp = Vcc / 2 = 26.4 / 2 = 13.2 V
So Vp << Vpeak that we found earlier which is 28.28 V.

Since the equivalent impedance of a piezo tweeter is mostly as of a capacitor, adding a damping resistor in series with it may be necessary to avoid ringing (its value depends on the amplifier). You can start with 10 Ohm 5 Watt. If you have a scope you can monitor the tweeter current by displaying the voltage across this series resistor. Don't forget protecting your ears first

Added:
Since your IC has two outputs, you can use both of them to drive ONE tweeter. Their inputs should be 180° out of phase. In this case , Vp = Vcc = 26.4 V which is a bit lower than Vpeak. Here the output power is multiplied by 4 (since the voltage is doubled) and you don't need a capacitor in series with the tweeter.

Please note, there is no need adding a resistance as a load as long the amplifier is stable (no self-oscillation and no ringing at each rising and/or falling edge).

 

[jiripolivka]

I noticed that your Vcc (single supply) is equal to 26.4V

Therefore, assuming the amplifier output is rail-to-rail, the maximum peak voltage (of the square wave) on the tweeter is:
Vp = Vcc / 2 = 26.4 / 2 = 13.2 V
So Vp << Vpeak that we found earlier which is 28.28 V.

Since the equivalent impedance of a piezo tweeter is mostly as of a capacitor, adding a damping resistor in series with it may be necessary to avoid ringing (its value depends on the amplifier). You can start with 10 Ohm 5 Watt. If you have a scope you can monitor the tweeter current by displaying the voltage across this series resistor. Don't forget protecting your ears first

Added:
Since your IC has two outputs, you can use both of them to drive ONE tweeter. Their inputs should be 180° out of phase. In this case , Vp = Vcc = 26.4 V which is a bit lower than Vpeak. Here the output power is multiplied by 4 (since the voltage is doubled) and you don't need a capacitor in series with the tweeter.

Please note, there is no need adding a resistance as a load as long the amplifier is stable (no self-oscillation and no ringing at each rising and/or falling edge).

All the above considerations are correct but one important point is missing. The ultrasonic transducers are high-impedance devices. To get a reasonable output, I found the best way to use a tapped output transformer at power amplifier output, so the transducer sees say 1...10 kOhms from signal source. I used a 555 pulse generator and a power-transistor amplifier, both run from +12v VDC.
I adjusted the "system" response by tuning the resonance while for indication I used the same transducer connected to a scope, some 1-meter away. The resonance is quite sharp. When I used the scope probe (10x) I measured over 100 V p-p on the transmitting transducer.

So your power amplifier with 12 V into 8 Ohms will work but to optimize transducer operation, I would advise inserting a 10:1 transformer for impedance matching.

 

[KerimF]

All the above considerations are correct but one important point is missing.

I agree with you. If Vcc is relatively low it is better using a tapped transformer.

If the tweeter has to transmit low-rate binary data, the bandwith at the ultrasonic frequency (carrier) of the resonant system should be wide enough at both sides, RX and TX, by using damping resistors.

I measured over 100 V p-p on the transmitting transducer

Your piezo transmitter seems being rated at a higher power (higher peak voltage) than of the GT-1005 here.

Kerim

 

[audiomik]

Running the LA4450 into a purely capacitive load with a square-wave will most likely cause problems.
Not the least of which can be overshoot of the Amplifier output, as well as it presenting a very adverse VI phase angle to the output stage(s) of the IC.

Also as the piezo Drivers have falling Impedance with frequency, a parallel resistance in order to moderate the capacitance should most likely be of a higher value than 8 Ohms. If you can measure the Impedance of the piezo at your working frequency, then select your parallel resistance for a combined 16 Ohms load.
Otherwise, try 47 Ohms in parallel to start with and reduce your output series Capacitor to 4.7µF - these values may need to be changed on test. This reduction in the output capacitor value additionally will reduce having LF Voltage spikes below the working frequency range of the tweeter which might damage it.
The Zobel correction may not be required for your application as you have a largely capacitive load, however look for instability after disconnecting it.

So long as you retain the minimum gain for the Amplifier IC sections, a Bridge-mode solution will be possible using the NF2 input of channel 2 which may well give you your required Sound Level.

hope this assists
Mik

 

[aliyesami]

sorry i couldnt follow up on this thread for long but picking it up again .
I need the answer to this question badly :
can i use the piezo tweeter (GT-1005) with LA4450 amp ? if not which piezo tweeter can i use with this amp ?

thanks

 

[audiomik]

the GT-1005 Piezo should be fine with your LA4450.
Just be aware that this 'Tweeter' won't like signals with a lower frequency than 3.5kHz from it's published specification, it may fail prematurely with lower frequencies!

hope this assists
Mik

 

[Audioguru]

Most piezo tweeters sound like a whistle. An awful sound with many peaks in the frequency response.
Oh, are you the guy who wants to shoo away cats? Your idea will also shoo away little kids, women and other "animals".

The audio amplifier IC does not needs an 8 ohm load! It is simply rated to drive an 8 ohm load. An old amplifier with vacuum tubes and an output transformer needed to be loaded properly. The IC amplifier datasheet shows a 0.15uF capacitor in series with a 2.2 ohm resistor as its high frequency load to prevent oscillations.

The LA4450 amplifier is rated to deliver 12W per channel at an awful 10% distortion to an 8 ohm speaker. But cats do not care about distortion.

 

[aliyesami]

hi Mike !
I wont be feeding any other freq to this tweeter other than 23Khz generated by the microcontroller, so then I should be fine?
at 7 watts am I endangering any human ears ?

- - - Updated - - -

hi Audioguru!
I am a bit confused , isn't 23Khz an ultrasonic freq which shouldn't be audible ?
since they are selling these cat repllents in the market which operate on 23Khz there must be a way to do this and that's what I am after

and I wont use the 8 ohm load.

 

[Audioguru]

When I was a kid until I was about 22 years old I could clearly hear 23kHz and many "ultrasonic" burglar alarms. Your 23kHz into a piezo tweeter would have been painful for me and other people.
Kids and women can also hear 23kHz.

The cat repellents were designed by deaf old men who cannot hear anything above 3kHz like a telephone.

 

[aliyesami]

this opens an interesting debate . I doubt the alarms in the market are mislabeling their products , so if they say they are inaudible and operating at around 23Khz how are they achieving this ?

 

[Audioguru]

this opens an interesting debate . I doubt the alarms in the market are mislabeling their products , so if they say they are inaudible and operating at around 23Khz how are they achieving this ?

Almost every link in Google says that a healthy average human can hear from 20hz to 20kHz. Some people can hear higher and some people cannot hear as high as 20kHz.

Many hi-fi tweeters roll off above 18kHz. Ear doctors test to only 8kHz.

 

[aliyesami]

what is this bridge mode solution ?

 

[audiomik]

what is this bridge mode solution ?

This is when two Amplifier Channels are used to drive a single Load.
The Channels are used with one running with an inverted output signal to the other so that there is twice the Output Voltage but at twice the minimum Load Impedance for one Channel. The Load is then connected between each Amplifier Output.

With 12 Watts/4 Ohms available from a single channel of your chosen IC, = ~7 Volts, you will then in bridged mode have something like ~14 Volts of output which equates to ~24 Watts for an 8 Ohm load.

Your suggestion of the IC Amplifier type is capable of this and with the Load Impedance being relatively high should work fine.
There is quite likely a circuit to do this in the Application notes of the Amplifier IC.

hope this assists
Mik

 

[Evan Humphreys]

If audiomik or KerimF are still following this post, I have a related question: When driving piezo tweeters like this from commercial audio amplifiers, is there a good way to tell if a resistor will be required to prevent "self-oscillation and ringing"? Either something that could be determined from the schematic of the amplifier, or a good way to test whether it is happening (without damaging the amp)? My application is to play music, not square waves or cat scaring.

 

[chuckey]

These tweeters are an oxymoron, they are quoted at eight ohms but every one knows they behave like a capacitor, so they can only be eight ohms at one frequency. So the thing to do is to test them and find out what their impedance is at the operating frequency. As its only the real part of their impedance that is going to dissipate power, the series capacitive impedance should be tuned out by a series inductor. With careful design and using the transducer as the output items in a PI network, the impedance can be changed to eight ohms resistive with a useful current magnification.
Frank

 

[FvM]

the series capacitive impedance should be tuned out by a series inductor

That's right to increase the output for a single frequency application, e.g. a pet repeller. But Evan Humphreys wants to use the piezo tweeter for audio, and in this case a series inductor isn't appropriate.

Regarding amplifier stability problems, most audio amplifiers have sufficient phase margin to drive piezo tweeters, I have used them occasionally and newer experienced problems. But I must confess that I don't like their sound, I prefer classical dynamical tweeters like a IBL ring radiator or a ribbon tweeter.