I've noticed some audio amplifiers, like this one, need +/- 35VDC:
http://www.national.com/mpf/LM/LM3886.html
and some amplifiers are happy with 9 to 18VDC, with respect to ground:
http://focus.ti.com/lit/ds/symlink/tpa1517.pdf
Besides the obvious difference in power, how come one needs +/- power,
and the other is happy with 0-X power?
In the second case, for "no audio signal", is the speaker suspended
halfway between rest and max, resulting in only half of the available
speaker movement range?
Michael
Some of this is illusion. You can have a dual supply for an op-amp
or audio output amplifier, and then the "resting point" is at zero
volts.
But you can have the same circuitry running off a single supply, but
then you need to bias the opamp or output amplifier so it rests
somewhere in between the positive voltage and ground, often right
in the middle. It gives you the same effect, the only difference
is really where you measure the voltage from.
With a single supply opamp or audio amplifier, your DVM's negative
lead going to ground would show the output voltage to be 1/2 the
supply voltage (we're generalizing here). But, if you put your
negative lead to the point that biases the opamp or output amplifier,
and then measure to the positive supply (ie put your positive lead
on the positive supply bus), you'd find half your DC voltage. Connect the
positive lead to the ground of the amplifier, and you'd read a negative
voltage, equal to the value you saw at the positive line. (This is
assuming the opamp or output amplifier is biased at the half-way point.).
How do you get a negative voltage when you aren't powering it with a
negative voltage?
It's all a matter of reference point. In both cases, you have your
ground lead at a point somewhere between the supply lines. Both
basically give you the same effect.
But of course, if you bias the amplifier input at half the supply
voltage, you only get the swing of the supply voltage (likely a bit
less). ANd the resting point of the output will be half-way between
ground and the positive supply. You can't connect that to the speaker,
the DC voltage will cause problems. Return the other speaker lead to
the same bias voltage, and that DC component goes away. Of course,
then that DC voltage at 1/2 the supply voltage will have to handle
the current of the output stage. Or, you can go the normal way,
putting a coupling capacitor at the output, which strips off the
DC component.
You will need higher voltage if you want more power, and at
that point it likely is worth adding a negative supply rather
than double the voltage and dealing with the DC component on
the output.
As someone pointed out, another method of getting rid of the
DC component is to feed the other side of the speaker with
an out of phase audio signal:
Audio ---------------AMP1 --
| speaker
|--inverter----AMP2---
Both leads of the speaker sits at the same "resting point",
ie half the supply voltage. So it's not seeing a DC component.
But since there is an inverter stage between the audio signal
and the speaker, the audio signal to the speaker is different
on each speaker lead, and the audio comes through.
Michael
