With 500 mA, the (sine wave) output power is 1 W in 8 ohm, 500 mW in 4 ohm. That is enough when you play dynamic range compressed music at moderate volumes over reasonably sensitive loudspeakers, but otherwise it's rather low - unless this is meant as a headphone amplifier, then it is very high.
Paralleling LM317 current sources is useful when the supply voltage is so high that a single LM317 would overheat at 500 mA. If you multiply (+Vcc - (-Vcc)) * 500 mA * (Rth_jc + Rth_heatsink + Rth_insulation) and add that to the maximum ambient temperature, you have to stay below the maximum junction temperature of the LM317. Rth_jc and the maximum recommended junction temperature are in the LM317 datasheet. Rth_heatsink should be specified in the heatsink datasheet and Rth_insulation in the datasheet of the pad you use to insulate the LM317 from the heatsink. When you divide the current over N LM317s and put all N of them on the same heatsink, the Rth_jc and Rth_insulation terms become N times smaller.
Regarding the BUF634, is the Cgd + Cgs / (gm RL + 1) of the MOSFET low enough for the op-amp to drive directly? (RL is the load resistance of the amplifier, the other terms are MOSFET small-signal parameters.) Is it within the load capacitance range specified in the op-amp datasheet and what kind of slew rate do you get when you divide the op-amp's maximum output current by Cgd + Cgs / (gm RL + 1)?
Chances are that the amplifier will oscillate. Measures you can take to solve that are a small capacitor in parallel with the resistor to the negative input of the op-amp, or a small capacitor between output and negative input of the op-amp. If it works, the first option gives the best distortion performance. You can measure or simulate the square wave response and experimentally determine the needed capacitance to get a well-damped response, or study control theory and do a lot of calculations.
Paralleling LM317 current sources is useful when the supply voltage is so high that a single LM317 would overheat at 500 mA. If you multiply (+Vcc - (-Vcc)) * 500 mA * (Rth_jc + Rth_heatsink + Rth_insulation) and add that to the maximum ambient temperature, you have to stay below the maximum junction temperature of the LM317. Rth_jc and the maximum recommended junction temperature are in the LM317 datasheet. Rth_heatsink should be specified in the heatsink datasheet and Rth_insulation in the datasheet of the pad you use to insulate the LM317 from the heatsink. When you divide the current over N LM317s and put all N of them on the same heatsink, the Rth_jc and Rth_insulation terms become N times smaller.
Regarding the BUF634, is the Cgd + Cgs / (gm RL + 1) of the MOSFET low enough for the op-amp to drive directly? (RL is the load resistance of the amplifier, the other terms are MOSFET small-signal parameters.) Is it within the load capacitance range specified in the op-amp datasheet and what kind of slew rate do you get when you divide the op-amp's maximum output current by Cgd + Cgs / (gm RL + 1)?
Chances are that the amplifier will oscillate. Measures you can take to solve that are a small capacitor in parallel with the resistor to the negative input of the op-amp, or a small capacitor between output and negative input of the op-amp. If it works, the first option gives the best distortion performance. You can measure or simulate the square wave response and experimentally determine the needed capacitance to get a well-damped response, or study control theory and do a lot of calculations.
- Status
- Not open for further replies.