Is a crossover circuit really needed for this sound reinforcement application

[tspringer99]

I'm building a pair of speaker enclosures to be used for sound reinforcement for a rock band. Each enclosure contains an Eminence Gamma 15 speaker and a piezo horn. The Gamma 15's are rated at 300 Wrms and have a useable frequency range of 40Hz to 4kHz. I don't have specs for the piezos, but I think that they can handle at least 150W. The enclosures are tuned and ported to give a low-frequency (F3) cutoff of about 50 Hz.

I am questioning the need for a crossover circuits between the piezos and woofers. Here is my reasoning:

As I understand it, a piezo driver does not require a high-pass filter to protect it as other tweeters do. Also, a piezo driver wired in parallel with the woofer should not effect the impedence of the circuit significantly. The Gamma 15 can be used as a speaker for bass guitars, where it could be used without a tweeter at all, so I assume that can absorb some high frequency energy input (e.g. slaps against frets) without damage.

If I don't put in a crossover circuit, one consequence is bound to be a frequency response curve that is not very flat because the frequency response of the piezo and speaker will overlap some, but I can equalize that out. There shouldn't be any impedence problems though, so if the piezo is safe without a high pass filter, the only other danger I could see would be inductive heating in the Gamma 15 as it dissipates the energy from frequencies above 4kHz that would be dissipated in the inductor of the low pass filter if it were present. The capacitor in the hiqh pass filter is cheap but an inductor that can handle 300W is fairly expensive, so I am especially interested in opinions on whether the inductor is really needed.

So what do I lose, and what do I risk, if I don't put in the crossover circuits? Or have I missed the point entirely?

[leadbelly]

Originally posted by tspringer99
So what do I lose, and what do I risk, if I don't put in the crossover circuits? Or have I missed the point entirely?

Go for it. (1) It's been done before with positive results. (2) Nothing will break if you try it. (3) You can add L-pad or crossover later.

[AGGEMAM]

You should under no circumstances try to use a passive crossover to filter the low frequencies of a piezo tweeter.

A piezo is a capacitive load and does not work as other drivers do.

But you should have the 10 ohms load resistor there at least. It's there to protect the amp, not the tweeter.

[tspringer99]

To leadbelly: Thanks for the info and the encouragement.

To aggeman: Thanks to you too. You may have prevented major heartbreak. I had planned to experiment with doing exactly what you warned me not to do. But I'm not sure that I understand the physics behind your comment. Do you mean that if I put a capacitor in series with the piezo that I could damage my power amp? Is the problem that the amp will see the sum of the capacitance of the piezo and the capacitor? If the capacitance is high enough it would act as a shortcircuit to high frequencies and draw too much current I suppose...

[dooper]

You can find some very good piezo info here:

http://www.diyaudio.com/wiki/index.php?page=piezoXO

For a sound reinforcement application, you might look at the KSN1188A driver and matching horn. Mike Klasco did an interesting article for Speaker Builder magazine, and it is available in PDF format here:

http://users.tpg.com.au/users/gradds/piezos.htm

With a simple RC network in front of it, it might just do the job for you, until you can get into some more serious HF components, such as a JBL horn / driver combination, an active crossover and another power amp.

dooper

[tspringer99]

I read the web page that Duper pointed me toward, and now I'm convinced that I should at least put a 20 ohm resistor in series with the piezo driver OR put the 20 ohm resistor in parallel with the piezo driver and put a capacitor in series with it. From what I read, both of these approaches would protect the amplifier from oscillating due to ineraction with a purely capacative load (which is a risk factor). The series resistor used alone rolls-off the high frequency also. The use of a parallel resistor and a series capacitor roll of the low frequencies. If I go for the second approach, will a 5W 20 ohm parallel resistor and a 100V 4uF series capacitor do the trick?

[Ap]

You dont need to 'protect' the woofer from high fre'qs - there is very little power in the upper range & that is not the function of the xover in this application.
I would add an inductor to the bass just to help combat the woofers naturally rising reponse & also help with beaming issues.
This is a typical problem of PA speakers & is why so often you see them used 4+ per side in an array.

[phase_accurate]

The use of a parallel resistor and a series capacitor roll of the low frequencies. If I go for the second approach, will a 5W 20 ohm parallel resistor and a 100V 4uF series capacitor do the trick?

This would definitely reduce stress on the piezo and reduce distortion.

Regards

Charles

[AGGEMAM]

Thanks dooper for providing the link. I had forgotten about that.

tspringer99, it depends on your piezo driver but as a general rule 2.2 uF would probably be the best choice as that would filter away frequencies below 2KHz and flatten the response slightly.

The Gamma drivers are indeed good in a bass guitar setup but that doesn't mean they don't sound really bad in the highest part of the midrange when used for music. So I really suggest as others have that you filter away the top with a small low resistance coil. Use a 0.33 mH as a good starting point (if your speaker is 8 Ohms, half if it's 4 Ohms).

[Andy G]

for a Gamma 15, I would recommend using a KSN1188 on a widish horn flare..

roll off the top end of the Gamma with a coil about .6mH,
then on the KSN 1188 put a 4.7uF then a 22ohm 10w+ resistor across then about 10 ohms leading to the + of the piezo. You may want to leave the 10 ohms out, depending on how bright you want the top end.

If you use a ksn1142 style piezo, use a coil of about .5mH, a cap of 3.3uF then 22ohms across, and probably about 22ohms to the + of the piezo.