LM4651 datasheet(13/20 Pages) NSC

13 / 20 page

Application Information (Continued)

applications. In bridge operation, each output sees C

BYP.

This causes the extra factor of 2 in the formula. The alterna-

tive to C

BYP is a capacitor connected between each output,

O, and VO2, and ground. This alternative is, however, not

size or cost efficient because each capacitor must be twice

BYP’s value to achieve the same filter cutoff frequency. The

additional small value capacitors connected between each

output and ground (C

1) help filter the high frequency from the

output to ground . The recommended value for C

1 is 0.1µF to

1µF or 2% to 20% of C

BYP.’

Modulation Frequency Optimization

Setting the modulation frequency depends largely on the

application requirements. To maximize efficiency and output

power a lower modulation frequency should be used. The

lower modulation frequency will lower the amount of loss

caused by switching the output MOSFETs increasing the

efficiency a few percent. A lower switching frequency will

also increase the peak output power before clipping because

the over modulation protection time is a smaller percentage

of the total period. Unfortunately, the lower modulation fre-

quency has worse THD+N performance when the output

power is below 10 watts. The recommended switching fre-

quency to balance the THD+N performance, efficiency and

output power is 125kHz to 145kHz.

THD+N Measurements and Out of Audio Band Noise

THD+N (Total Harmonic Distortion plus Noise) is a very

important parameter by which all audio amplifiers are mea-

sured. Often it is shown as a graph where either the output

power or frequency is changed over the operating range. A

very important variable in the measurement of THD+N is the

bandwidth limiting filter at the input of the test equipment.

Class D amplifiers, by design, switch their output power

devices at a much higher frequency than the accepted audio

range (20Hz - 20kHz). Switching the outputs makes the

amplifier much more efficient than a traditional Class A/B

amplifier. Switching the outputs at high frequency also in-

creases the out-of-band noise. Under normal circumstances

this out-of-band noise is significantly reduced by the output

low pass filter. If the low pass filter is not optimized for a

given switching frequency, there can be significant increase

in out-of-band noise.

THD+N measurements can be significantly affected by

out-of-band noise, resulting in a higher than expected

THD+N measurement. To achieve a more accurate mea-

surement of THD, the bandwidth at the input of the test

equipment must be limited. Some common upper filter points

are 22kHz, 30kHz, and 80kHz. The input filter limits the

noise component of the THD+N measurement to a smaller

bandwidth resulting in a more real-world THD+N value.

The output low pass filter does not remove all of the switch-

ing fundamental and harmonics. If the switching frequency

fundamental is in the measurement range of the test equip-

ment, the THD+N measurement will include switching fre-

quency energy not removed by the output filter. Whereas the

switching frequency energy is not audible, it’s presence de-

grades the THD+N measurement. Reducing the bandwidth

to 30kHz and 22kHz reveals the true THD performance of

the Class D amplifier. Increasing the switching frequency or

reducing the cutoff frequency of the output filter will also

reduce the level of the switching frequency fundamental and

it’s harmonics present at the output. This is caused by a

switching frequency that is higher than the output filter cutoff

frequency and, therefore, more attenuation of the switching

frequency.

In-band noise is higher in switching amplifiers than in linear

amplifiers because of increased noise from the switching

waveform. The majority of noise is out of band (as discussed

above), but there is also an increase of audible noise. The

output filter design (order and location of poles) has a large

effect on the audible noise level. Power supply voltage also

has an effect on noise level. The output filter removes a

certain amount of the switching noise. As the supply in-

creases, the attenuation by the output fiter is constant. How-

ever, the switching waveform is now much larger resulting in

higher noise levels.

THERMAL CONSIDERATIONS

Heat Sinking

The choice of a heat sink for the output FETs in a Class D

audio amplifier is made such that the die temperature does

not exceed T

JMAX and activate the thermal protection cir-

cuitry under normal operating conditions. The heat sink

should be chosen to dissipate the maximum IC power which

occurs at maximum output power for a given load. Knowing

the maximum output power, the ambient temperature sur-

rounding the device, the load and the switching frequency,

the maximum power dissipation can be calculated. The ad-

ditional parameters needed are the maximum junction tem-

perature and the thermal resistance of the IC package (

JC,

junction to case), both of which are provided in the Absolute

Maximum Ratings and Operating Ratings sections above.

It should be noted that the idea behind dissipating the power

within the IC is to provide the device with a low resistance to

convection heat transfer such as a heat sink. Convection

cooling heat sinks are available commercially and their

manufacturers should be consulted for ratings. It is always

safer to be conservative in thermal design.

Proper IC mounting is required to minimize the thermal drop

between the package and the heat sink. The heat sink must

also have enough metal under the package to conduct heat

from the center of the package bottom to the fins without

excessive temperature drop. A thermal grease such as

Wakefield type 120 or Thermalloy Thermacote should be

used when mounting the package to the heat sink. Without

some thermal grease, the thermal resistance

CS (case to

sink) will be no better than 0.5˚C/W, and probably much

worse. With the thermal grease, the thermal resistance will

be 0.2˚C/W or less. It is important to properly torque the

mounting screw. Over tightening the mounting screw will

cause the package to warp and reduce the contact area with

the heat sink. It can also crack the die and cause failure of

the IC. The recommended maximum torque applied to the

mounting screw is 40 inch-lbs. or 3.3 foot-lbs.

Determining Maximum Power Dissipation

Power dissipation within the integrated circuit package is a

very important parameter. An incorrect maximum power dis-

sipation (P

D) calculation may result in inadequate heat sink-

ing, causing thermal shutdown circuitry to operate intermit-

tently. There are two components of power dissipation in a

class D amplifier. One component of power dissipation in the

LM4651 データシート(PDF) 13 Page - National Semiconductor (TI)

LM4651 データシート(HTML) 13 Page - National Semiconductor (TI)

Html Pages

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20

データシートダウンロード