buzz of harddisk/DVDRW audible through speakers

[Christophe_work]

Hi,

I've designed a small amplifier for a PC. It's built around the LM4952.
The PSU is a PICO-PSU:
**broken link removed**

This Pico-PSU also supplies the SATA harddisk and DVDRW drive.

The problem I experience is that I hear the noise/buzz from the motors of the harddisk and DVDRW throught the speakers when I put the gain louder.

What can I do to solve this?

Here's a picture of the setup:
**broken link removed**

 

[sunderwood]

I'd recommend you use shielded wires on the power cables to the hard disk and the DVD drive. You can also add more capacitance close to them...

 

[Christophe_work]

I've tried that, that didn't work..

 

[FvM]

Ground loops may be another reason. You should supply the amplifier by way of trial from a separate supply (not a PC supply with GND conected to protective earth!). If this stopps the interferences, you should think about a pseudo differential input at your amplifier, isolating the input signal ground from supply ground.

 

[Christophe_work]

Ground loops may be another reason. You should supply the amplifier by way of trial from a separate supply (not a PC supply with GND conected to protective earth!). If this stopps the interferences, you should think about a pseudo differential input at your amplifier, isolating the input signal ground from supply ground.

When I supply my amplifier with a separate DC supply, all the buzz noises are gone.

What do you mean by pseudo differential input?

The analog input grond signal is put in a star configuration, it's only connected from the PC motherboard to the LM4952 large ground tab, as stated in the datasheet.

**broken link removed**

 

[FvM]

Unfortunately the LM4952 has no separate input common (for pseudo-differential connection) or true differential input feature. The next best option is to disconnect the amplifier to PSU ground connection and connect it to the mainboard directly. Or use a daisy-chain from ATX supply connector. This should at least reduce the interferences.

P.S.: It could be,that the interferences aren't cause by ground loops, but simply AC voltage superimposed to the supply. In this case a supply LC filter at the amplifier should help.

 

[Christophe_work]

Unfortunately the LM4952 has no separate input common (for pseudo-differential connection) or true differential input feature. The next best option is to disconnect the amplifier to PSU ground connection and connect it to the mainboard directly. Or use a daisy-chain from ATX supply connector. This should at least reduce the interferences.

P.S.: It could be,that the interferences aren't cause by ground loops, but simply AC voltage superimposed to the supply. In this case a supply LC filter at the amplifier should help.

Hi, I will try connecting the amplifiers ground pin to the PC motherboard.

I allready tried a LC filter in the 12V line, this didn't help.. My guess is that you have to have a way too large coil (L) to make a low pass DC filter..

I'll keep you posted

 

[Christophe_work]

Unfortunately the LM4952 has no separate input common (for pseudo-differential connection) or true differential input feature. The next best option is to disconnect the amplifier to PSU ground connection and connect it to the mainboard directly. Or use a daisy-chain from ATX supply connector. This should at least reduce the interferences.

P.S.: It could be,that the interferences aren't cause by ground loops, but simply AC voltage superimposed to the supply. In this case a supply LC filter at the amplifier should help.

Hi,

the best results (= noise is gone) I have when I don't connect the POWER_GND lead, in other words the amplifiers ground = the audio ground. So only the 12V lead from the PSU is connected and the ground from the PSU is not connected.

But when I put a big cap over the 12V line to ground, all the buzz noises are back. So that means that the cap couples the DVD noises from the 12V line to the audio via it's analog ground!

=> Is it OK to wire the amp like this (with no GND connected, or only the audioground as ground) ?

 

[FvM]

when I put a big cap over the 12V line to ground

Which ground?

Your test seems to show, that the probelm is caused by ground loops, as suspected. Not connecting the ground uis O.K., if you don't overload the thin ground connection of audio signal connection.

 

[Christophe_work]

when I put a big cap over the 12V line to ground

Which ground?

The analog ground..

The datasheet states that you put a 10µF tantalum cap close to the chip on the supply. However by doing this, you actually couple noise from the 12V line to the AGND line, thus the audio signal...

Added after 40 minutes:

Another question:

On the LM4952, there is a DC-gain pin. By adjusting the DC voltage, you can set the volume.

Now, what would be best, (or doesn't it matter):

- giving max. input to the amp and setting the gain to an acceptable level
- resistor divide the input signal in order to have no distortion at maxium gain and setting the gain to max (20db)

 

[Christophe_work]

Hi,

here's the schematic:

**broken link removed**

Any remarks are welcome; also tips for layout and routing