Please help me with LM3886 PCB

[aparatusonitus]

Please i do not have sch file, i have only .brd file, please check and just i need value for every parts because i have already make 60 pcs board for that, my pc formatted that's why i lost all file only brd file i have now.

R1 = 1k*
R2 = 22-27k (should be in parallel with 100p or more)*
R3 = 1k*
R4 = 22-27k/0.6W min. (should be in parallel with 20p or more)*
R5 = 27k**
R6 = 2R7/3W min. non-iductive
R7 = 10R/1W min. (should be in paralell with air wound inductor L = 0,7uH)
C1 = 2u2/10V non-polar electrolytic
C2 = 100n
C3 = 100n
C4 = 100n
C5 = 22uF/10V non-polar electrolytic
C6 = 2200uF/35V
C7 = 100u/35V**
C8 = 2200uF/35V

* R1 = R3; R2 = R4
** R5 + C7 with -28VDC= Vee = time const. of 2.7 sec

P.S your pcb sucks

 

[tomchr]

If you take speaker return to pcb the speaker return signal will modulate the ground.

Sense the feedback signal across the speaker output (or just inside the L||R output network). This will always result in the lowest contribution to the output signal of the various ground currents. If you connect the signal ground, feedback ground, and power/speaker ground at the reservoir caps far from the amp circuit, you will leave the feedback and speaker ground voltage drops uncorrected by the feedback loop and they will be added to the amp's input signal. This will result in higher THD (i.e. higher error) from those ground currents. In particular, the feedback current can be rather gnarly, so keep the feedback ground impedance low. Doug Self speaks to this as well in his power amp design book (and mid-1990ies Wireless World articles, which became the book).

You can see my experiments here: LM3886 PCB vs Point-to-Point (with data)

And read my analysis with simulation results here: LM3886 chip amp grounding.

I haven't seen an amp design yet where they didn't say return speaker to main smoothing caps.

I guess you haven't seen any of my amps then. Or Hypex's amps (as mentioned above). How do you think I achieve 0.0000xx % THD in my designs (or Bruno achieves 0.000xx % in his)? This does not happen by happy coincidence or serendipitous luck. It happens by careful analysis of the circuit - including the layout parasitics.

The recommendation to take all grounds to the reservoir caps is from the days we didn't know better. Now we do...

Now for the original poster: It should be pretty easy to reverse engineer that board. Follow the traces and draw the schematic. Then have a look at the LM3886 Data Sheet from TI. From there, it should be plug-n-play. It's unfortunate that whomever laid out that board missed the opportunity to beef up the supply lines. That could easily have been done with copper pours. At least they left the solder mask off those traces, so they can be beefed up in assembly.

Tom