Audio: Simple power amplifier

012/Jan/30

each radio receiver need a audio amplifier, a LM386 is a simple choice. but some time transistor in a junk box do the work well and you might want it. for experiment, a mini box integrated a audio power amplifier and a speaker is convenient way for testing

Decibels (good to know)

before started i need know what is Sound levels Decibels (db). most people will notice 1db volume change, 3 db is a moderate change in volume, 10db hearing approximate doubling of volume. and sound level V.S. real-world situations:

0 db is the threshold of hearing

Whisper : 15-25 dB

Background noise : about 35 dB

Normal home or office background: 40-60 dB

Normal speaking voice: 65-70 dB

Orchestral climax: 105 dB

Live Rock music: 120 dB+

Pain Threshold: 130 dB

Jet aircraft: 140-180 dB

Popular Push Pull Audio PA

here is a popular design for this purpose. most of this kind of amplifier use 2 diodes replace T5. but T5 is more flexible and easy to adjust for bias.

*the suplly voltage is optimized for 2 Li-Po battery, 7.2V to 8.4V.

*work point is marked

*output power to a 4R speaker about 4Vp-p, 500mW

The PCB is optimised for size, it's very small, 1.5mm x 4.5mm. easy to put to a hand phone size box, i.e. the popular minty box.

performance and problem

*Gain loss : voltage stage T2 had an open circuit 20 times voltage gain, but when connect to output push pull stage with a 8ohm speaker, the gain drop to around 13 , a 4ohm load make it drop to 9. it's cause by the out-stage input Z is very low and gain stage at a high output Z, not matched.

*Current Gain compression and up swing: when T3 is driven to open hard, current flow to speak is 2Vpeak to 4R load, the peak current will be 2V/(4R+1)=400mApeak, large current gain of the transistor is severely reduced, per data-sheet, about 40x, so base need 10mA peak to driven the T3. then 10mA will drop on R1 about 2.7Vpeak,this so much limit the output swing. and 400mApeak will drop 0.4V on 1R resistor[R5]. (updated @7/17 2012)

*Efficiency: 2 series Li-Po but only get around 2V out put swing(4V-pp).

*distortion : not so care about it. and sound so so.

* DC couple make it hard to bias.

Improvement

bootstrapping

Nerwin Mora mention of a bootstrapping method to enlarge output swing, which push the top swing much reach to the rail.

how this works:

on low swing C2 is charged by R1, then the up swing, the middle point voltage became high plus C2's voltage provide R2 a bias exceed the VCC, make it up swing reach to rail is possible.

*out put swing maximize : bootstrap or driver stage via the constant current source

* more current Gain via Sziklai pair

a Sziklai output stage is worthy to try which low distortion (<100dBc), high thermal stability, and the most of important is high current gain suitable for high power output. use R=270R to start your experiment

Google Sites

Report abuse