Look in the manual. The PDF I have, has this on PDF page 26.
F_PANEL (Front Panel Header)
The pins labeled SPKR+ and SPKR- have a spacing
of four pins total. That's intended for a case
speaker. OEM cases have a two pin spacing for
SPKR, in which case, that causes problems when the
world has two different standards for the stupid
things.
Some motherboards, will allow both two and four
pin spacings, the way they're designed, so you're covered
for both instances. The diagram in your user manual,
does not provide evidence of this feature. It only
shows a four pin spacing working. The silk screen
(white lettering) on the motherboard, may give hints
whether it's intended for both.
If the piezo has any polarity markings, please follow
them. Of the two types of piezos, one cares about
polarity and the other doesn't. The one that cares
about polarity, the internal circuit design is
unknown (so I can't guess at whether the piezo oscillator
smokes when it is reversed or not).
*******
There are two kinds of piezo devices. Transducer and
oscillator type.
Transducer, you are driving the capacitive element directly.
This requires a decent amplitude signal to get it working.
(AFAIK, piezos "need voltage" to work.)
The signal in this case would be AC for best results. It
could also work with DC_offset + AC, but like any ordinary
speaker, that affects how far the "cone" can move. The
DC would push the cone to one side, and limit the travel.
The oscillator type, they're driven by a DC source. The
device has an oscillator of its own inside, tuned for
the resonant frequency of the transducer (which gives
a loud sound, with moderate power consumption). Maybe
your smoke detector would be an example of this
type of noise source. The circuit would be designed
to deliver pure AC to the actual piezo element.
The motherboard side, I can find at least two driver
configurations. I could find an open collector circuit
(that's the kind I would do, as an "analog dummy").
I could also find an emitter follower circuit. The
emitter follower was biased from VCC3, driven by
the SPKR pin of the Southbridge, and on the output
side, the collector is connected directly to +5V.
The output voltage would be one diode drop below
the bias voltage, max. And this means, with an output value
of around 2.3V or so (0 to 2.3V signal), the output
on the emitter follower motherboard is not likely
to be enough to power an oscillator (beeper) type
piezo. The output from the emitter follower
might be too low for the job. The oscillator might
want 3.3V or so.
So my answer would be, it's possible you got caught
in the details. If you use a regular magnetic case
speaker, both motherboard circuits would work with that.
The emitter follower circuit is marginally safer,
in that if the SPKR+ or SPKR- wires are pinched
by the bare metal of the computer case (short
circuit), the bias circuit times the Hfe gain
of the transistor, limits current flow. And even
if the transistor is heroic, the transistor can
probably burn without anyone noticing smoke or
anything
PDF page 74 features emitter follower SPKR+ driver.
Max voltage on pin would be VCC3 minus one diode drop.
Min voltage would be zero (when the transistor is off).
http://www.intel.com/design/chipsets/schematics/252812.htm
PDF page 32. Analysis is more complicated with a
capacitive piezo of some sort connected, but a guess
is this makes more amplitude available than the previous
circuit. And is more likely to run a beeper of any type.
If you pinch the red SPKR+ wire with this circuit, the
red speaker wire glows red hot (ATX +5V shorted), and
all the vinyl insulation on the wire burns. Spectacular.
(A poster years ago, described such an event, pinching
that wire by accident in the computer case door.)
The design idiots don't waste money on a fuse, or adjust the
circuit to not do this! As a digital designer, this
is more my style
Although I would probably
split the load resistance, into two 18 ohm resistors,
and put an 18 ohm in each leg. Then, instead of
a spectacular "wire fire", one 18 ohm resistor goes up
in smoke
And the 5V rail hardly notices.
http://www.intel.com/design/chipsets/designex/BXDPDG10.PDF
*******
The piezo provided, you could use the part number off
it, to look up the characteristics.
Chances are, it's rated for 5V operation, if it's thrown
into a motherboard box.
You could take three dry cells, giving a 4.5V supply
and try and test it. If it is the transducer type,
connecting the red leg to +4.5V, should make a "tick" or
"click" sound. If it is the oscillator type, connecting
the red lead to +4.5V will make a loud beep. Probably
causing you to throw the batteries in the air
If
I was doing this test, I'd probably stick a limiting
resistor in line for the test. But not everyone has
multiple drawers full of resistors, to do this. I have
a fairly complete collection of 1/4 Watt and 1/8 Watt
resistors. So I'm well prepared for these experiments.
If the piezo has one black wire, that's the negative
end. If the piezo has a red wire, that's the positive
one. But they could use other, non-black wire color
for the positive one (like white or orange). But it
would be tradition for the wire to be red. (Red for
"fire" maybe
)
If both wires are the same color, there are no
polarity markings, it could be the transducer type,
and be more likely to give some sort of signal.
The oscillator one, with two different color wires,
and a diet for DC on input, could be more picky.
Have fun,
Paul
