Which pcb is better (LM4766)

[linket]

Hello!
I decided to build and amp using LM4766. Everything seems ok - the parts are cheap, scheme is easy, but one day I accidentaly found some guy's PCB and similar schematics to the auxiliary circuit in pdf. Here it is:

http://majki.freeweb.hu/elektro/kapc...onal_2x40w.htm

Is it risky to do an amp using his pcb? Or do I need to go on a pdf variant? I found some differences like the boucherot cell in pdf stands before the coil, and in this pcb - after. Thanks for
comments.

Here's pdf for comparison:
http://<br /> <a href="http://www.a...M4766.html</a>

[Leolabs]

If u r not sure,then try the pdf version.

[SpittinLLama]

I'd put it before the coil.

[rajudiy]

i assembled this board

http://majki.freeweb.hu/elektro/kap...ional_2x40w.htm


and the working was very good

rajudp

[peranders]

rajudiy, you can't copy an URL with just copying the visible text. You have to right-click and choose something like "copy link location".

[sangram]

linket/rajudiy

IME, 4766 is quite pathetic when it comes to the sound. Takes a lot of burning in, the sound for the first hour sounds shouty and closed in, gets better after about 15 hours. I expect some more improvement. However, channel separation specs are anyway very low, so unless you do pure dual mono with 1 chip per channel in bridged/parallel/biamping, I doubt very much this chip can be used for hifi. The sound kind of reflects that shortcoming (and many others) and I feel that even TDA 2030 beats it on pure SQ terms, though not on power.

It's dead silent and easy to work with though on the positive side, I've used P2P and though it's a little messy to look at, there's no noise / hum on the output. If proper heatsinking is done, the chip can take a lot of abuse without getting toasty.

Far as that modified schematic goes, zobel should be before the coil. Also the feedback resistors seem to have very high values, this is a subject for some discussion, because we almost always use smaller values for the feedback resistors. Accordingly the value for the blocking cap must increase as per the equation in the data sheet. I think the amp is designed to accomodate polyester/metallised caps in that position, looking at the PCB layout. This is not needed and a normal electro works fine as it is only a blocking cap and should not affect the signal.

Also muting can be combined into one single resistor/cap and does not need to be separated for both channels. The national PDF gives the equations on how to calculate its value.

Zobel values will need calculating. Some people use 1 ohm, some people use 100 ohm. There's no 'right' value and it depends on your speakers, cable, amp layout etc.

And that PCB... looks mighty thin traces. I wold not trust it one bit, actually. Even though the amp is low power, the traces look like they will need some buffer wires definitely.