Are PCB of LM3886 correct?

[gengcard]

I made a PCB follow as Carlosfm's circuit but I am not sure whether some parts are put in correct positions. Please help me to check the PCB. Thank you.

[gengcard]

This picture is shown the circuit is used.

[Mick_F]

The supply bypass caps are way too far from the chip. Also, It is preferrable to separate signal and power ground. I am not sure about the use of a ground plane anyway - I prefer star grounds.

Mick

[CJ900RR]

Originally posted by Mick_F
The supply bypass caps are way too far from the chip. Also, It is preferrable to separate signal and power ground. I am not sure about the use of a ground plane anyway - I prefer star grounds.

Mick

I seccond that...

[jackinnj]

Originally posted by Mick_F
The supply bypass caps are way too far from the chip. Also, It is preferrable to separate signal and power ground. I am not sure about the use of a ground plane anyway - I prefer star grounds.

Mick

No ground plane -- you can use a 100 - 150 mil trace to connect the reservoir caps to the star.

[gengcard]

Thank you everyone for comments.

Can you please tell me the adventage of star ground?

[zenon]

can you please post the original file, I'm trying to learn the eagle software.

[gengcard]

Hi zenon,

Do you have LM3886T in your library? If no, I'll attach my library to you.

[CJ900RR]

Originally posted by gengcard
Thank you everyone for comments.

Can you please tell me the adventage of star ground?

Yes. Read the datasheet for the LM3886, page 19

[jackinnj]

Originally posted by gengcard
Thank you everyone for comments.

Can you please tell me the adventage of star ground?

The star ground prevents "current loops" which occur on a ground plane. Current loops exist since -- the resistance of the copper foil isn't perfectly uniform and the distances between points of current flow are variable -- when current flows you have an electromagnetic effect -- radiation of energy -- small in the case of analog circuits -- very large in switch mode power supplies -- these differential currents get magnified by the amplifier chip and "whoops" you have a problem. There was a neat article in EDN magazine, the July 6th issue by a fellow from Analog Devices with pictures which better describe what my verbiage fails to do.

Try to keep all your grounds on one side of the printed circuit board -- i usually have the power traces on the bottom, signal traces on the top. If you have to make a connection between two layers on a power or ground trace you might want to widen the trace and punch several vias through so that there's enough copper on the plated-through via to prevent it from acting like a fuse.

You want to be careful in not running the ground from the input pin too far to the star -- by means of capacitative coupling it can pick up RFI.