Advantages/disadvantages of using LM350/LM317 linear regulator over RC filter to remove PSU ripples

[moty22]

I have made many audio amplifiers using ICs and never used a voltage regulator. None of them produced hum in the speaker.
Why are people talking about 50V supplies and 20W or 22W into an 8 ohm speaker?

Rwind said he has a 12.6V transformer. Its peak is 17.8V and the full wave bridge rectifier drops it down to 15.8V which is too low for the LM1875. With a 16V supply a graph shows that its power into an 8 ohm speaker is only 2.5W when it is already clipping a little.

Hi audioguru,
rwind has 2 transformers of 12V and he was advised to put them in series. I also advised him against regulators.
He is lucky to have you taking over the case. I'm sure you'll give the best help.

 

[audioguru]

The peak of each 12.6VAC transformer is 17.8V and the rectifiers reduce it to plus and minus 16.8V. Then the output power to an 8 ohm speaker is about 12W when clipping a little. Voltage regulators are not needed.

 

[rwind]

The peak of each 12.6VAC transformer is 17.8V and the rectifiers reduce it to plus and minus 16.8V. Then the output power to an 8 ohm speaker is about 12W when clipping a little. Voltage regulators are not needed.

Thanks, I am looking for about 10W per channel, it is mostly going to be used for near field listening for computer. I have already made the power supply with regulators, other than loosing about 1V to the LM350, is there any major disadvantage to using regulators. I mean like quality of sound, if it is detrimental I would remove the regulators but for now I am using it to filter ripple ?

 

[audioguru]

The LM1875 is a high gain circuit. Negative feedback reduces the gain to what you want, reduces the distortion and reduces hum caused by power supply ripple.
If your mains is 50Hz then power supply ripple from a full wave bridge rectifier is at 100Hz, not 50Hz. Hum caused at 50Hz would be from using an input cable that has no shield.
The datasheet for the LM1875 amplifier says its typical Power Supply Rejection Ratio of ripple on its positive supply is 95dB so if the power supply ripple is 1V then at the output of the amp it will be at 15uV which is nothing.
Therefore a voltage regulator that has a 2.25V loss is not needed.

Are you using another regulator for the negative supply? I do not know any 3A negative regulator but if you find one it will also have a loss of about 2.25V. Then the total voltage is plus and minus 14.55V so the output power will be about 8W.

 

[rwind]

The LM1875 is a high gain circuit. Negative feedback reduces the gain to what you want, reduces the distortion and reduces hum caused by power supply ripple.
If your mains is 50Hz then power supply ripple from a full wave bridge rectifier is at 100Hz, not 50Hz. Hum caused at 50Hz would be from using an input cable that has no shield.
The datasheet for the LM1875 amplifier says its typical Power Supply Rejection Ratio of ripple on its positive supply is 95dB so if the power supply ripple is 1V then at the output of the amp it will be at 15uV which is nothing.
Therefore a voltage regulator that has a 2.25V loss is not needed.

Are you using another regulator for the negative supply? I do not know any 3A negative regulator but if you find one it will also have a loss of about 2.25V. Then the total voltage is plus and minus 14.55V so the output power will be about 8W.

Thanks. Got it, ripple is not an issue with LM1875.
So now the question is does a regulated power supply add to the quality of the sound, if so then the regulator can stay, right? if not I will get rid of the regulator.

My plan id to make two 18V from 2 separate transformers-rectifiers-LM350. Then connect the 2 separate 18V dc in series and use the junction as the ground and the other 2 ends as +18V and -18V. This way I don't need a negative regulator right?

 

[audioguru]

An LM350 voltage regulator might limit the current during a loud powerful sound. You do not want that to happen.

If you have an unregulated supply voltage and the amplifier has its input level too high so that it clips, then the clipping distortion will contain 100Hz ripple buzzing.
But if you use a voltage regulator to power the amplifier (the minimum unregulated input voltage to the regulator must be at least 2.5V higher than its regulated output) then if the amplifier has its input level too high so that it clips, then the clipping distortion will not contain 100Hz ripple buzzing. Clipping distortion sounds awful, with some added buzzing from ripple then it still sounds awful.

So do not use regulators and keep the input level low enough that the amplifier does not clip.

 

[rwind]

An LM350 voltage regulator might limit the current during a loud powerful sound. You do not want that to happen.

If you have an unregulated supply voltage and the amplifier has its input level too high so that it clips, then the clipping distortion will contain 100Hz ripple buzzing.
But if you use a voltage regulator to power the amplifier (the minimum unregulated input voltage to the regulator must be at least 2.5V higher than its regulated output) then if the amplifier has its input level too high so that it clips, then the clipping distortion will not contain 100Hz ripple buzzing. Clipping distortion sounds awful, with some added buzzing from ripple then it still sounds awful.

So do not use regulators and keep the input level low enough that the amplifier does not clip.

Got it, thanks.