T
Tom
Everyone,
In my design I need to drive headphones with nominally 10 ohm impedance
with an audio signal. The supply for the audio is +12V single supply
(this is created from the main 2.5V-3.3V battery supply using a DC/DC
converter). The output swing does not need to be all the way to the
rail, something like 7..8V peak-peak is acceptable. The input signal is
0..2.5V.
The signals are synthesized sine waves and tone bursts coming from a
DAC, with frequencies from 250Hz to 1kHz. Not music or speech.
Important is that the quiescent power consumption is low (<10 mA if
possible). A shutdown input would be nice but not essential (I can shut
down the DC/DC). Distortion is not so important for this application,
obviously within some reasonable limits (the sine wave shouldn't end up
as something close to a triangle or square wave). Several % THD is
perfectly acceptable.
Important for this application is that the gain of the amplifier is low
to keep a decent signal to noise ratio. If the gain is too high I would
have to attenuate the 0..2.5V DAC signal too much on the input, causing
a penalty in signal to noise ratio (especially at small signal
amplitudes). Ideally the gain would be 4..5.
Currently I am using the LM386, with a gain of 9, which is the minimum
stable gain (by tying a 10K resistor from pin 1 to pin 5). It has low
quiescent power consumption (4mA) and sufficient output swing and
drive, as well as low distortion. However it has one huge disadvantage:
the noise is WAY too high. Even if I short the input pins, the output
noise is VERY audible in the headphones (it sounds white). The noise is
coming from the LM386 and not elsewhere in the circuit, because if I
attenuate the LM386 output, the noise attenuates also (but the signal
also of course).
The LM386 datasheet conveniently does not seem to have any mention of
noise (non-)performance at all!
So, basically I'm looking for a better audio amp, similar to the LM386
but with lower noise. I have found several amps with quite low noise
but either they have too much gain (e.g. 40 dB) or they cannot drive a
10R load. It should also be compact (SMD preferred, but a small DIL
package is also acceptable) and as mentioned with a low quiescent power
consumption. A buffer amp (gain 1) would also be OK, then I will simply
do the required x4..x5 gain in a separate low-power amp stage before.
Does anybody have any suggestions?
greetings,
Tom
In my design I need to drive headphones with nominally 10 ohm impedance
with an audio signal. The supply for the audio is +12V single supply
(this is created from the main 2.5V-3.3V battery supply using a DC/DC
converter). The output swing does not need to be all the way to the
rail, something like 7..8V peak-peak is acceptable. The input signal is
0..2.5V.
The signals are synthesized sine waves and tone bursts coming from a
DAC, with frequencies from 250Hz to 1kHz. Not music or speech.
Important is that the quiescent power consumption is low (<10 mA if
possible). A shutdown input would be nice but not essential (I can shut
down the DC/DC). Distortion is not so important for this application,
obviously within some reasonable limits (the sine wave shouldn't end up
as something close to a triangle or square wave). Several % THD is
perfectly acceptable.
Important for this application is that the gain of the amplifier is low
to keep a decent signal to noise ratio. If the gain is too high I would
have to attenuate the 0..2.5V DAC signal too much on the input, causing
a penalty in signal to noise ratio (especially at small signal
amplitudes). Ideally the gain would be 4..5.
Currently I am using the LM386, with a gain of 9, which is the minimum
stable gain (by tying a 10K resistor from pin 1 to pin 5). It has low
quiescent power consumption (4mA) and sufficient output swing and
drive, as well as low distortion. However it has one huge disadvantage:
the noise is WAY too high. Even if I short the input pins, the output
noise is VERY audible in the headphones (it sounds white). The noise is
coming from the LM386 and not elsewhere in the circuit, because if I
attenuate the LM386 output, the noise attenuates also (but the signal
also of course).
The LM386 datasheet conveniently does not seem to have any mention of
noise (non-)performance at all!
So, basically I'm looking for a better audio amp, similar to the LM386
but with lower noise. I have found several amps with quite low noise
but either they have too much gain (e.g. 40 dB) or they cannot drive a
10R load. It should also be compact (SMD preferred, but a small DIL
package is also acceptable) and as mentioned with a low quiescent power
consumption. A buffer amp (gain 1) would also be OK, then I will simply
do the required x4..x5 gain in a separate low-power amp stage before.
Does anybody have any suggestions?
greetings,
Tom
