LM3886 Layout ok?

Status
Not open for further replies.
Jump to Latest
#43
Done - for the "real" tracks...

What would be if I would move the power-ground into the star-ground?
And is anyone able to tell me how often the copperwire has to be wound around the resistor for R/L? ANd does it have to have a diameter of 1mm? or would 0.8mm would work too?
 
#46
Maxxtr0 said:
And is anyone able to tell me how often the copperwire has to be wound around the resistor for R/L? ANd does it have to have a diameter of 1mm? or would 0.8mm would work too?
Click to expand...

I used 0.8mm, 10 turns tightly on the body of 10ohm/3W resistor.

Here's how I did the layout. Although the traces are only 86 mils in width, they are covered with around 0.5mm solder layer and thy withstanded an unwanted shortcircuit between the negative rail and ground. The chip leges melted, the tracks were fine.
 

Attachments

  • GC.PNG
    GC.PNG
    54.7 KB · Views: 574
#52
They do specify an inductance but in all truth, its really not that critical. If you have the means to measure inductance i'd be interested to find out what my suggestion of just winding the turns around the 2W resistor produces.
 
#53
No sorry, at the moment I do not have the necessary measurement device...
But do you maybe wanna give ma an answer to my question about the "GND-pad"? 😀 So that I finally can finish my PCB 🙂

Edit: So now - this is it:

An externally hosted image should be here but it was not working when we last tested it.


Everything ok now?
The two single vias are for the aircoil! I hope it won´t affect the amp, as though it´s pretty near?!
 
Last edited:
#54
Rz top end must connect to the chipamp output. Disconnect that trace that goes down to the speaker +.
If you want to add an extra R+C after the L//R, then do that across the output terminals. This catches the back EMF and/or interference before it gets too far into the chassis/feedback loop.

Are you planning to add RF attenuation on the back of the PCB? You have omitted it on the front.
Your latest layout has returned to putting the feedback resistor across the chip pins and similarly the +ve supply to pin5.

Keep the separation of the decoupling Ground from the PCB star ground. This will facilitate a multi-channel amplifier. Use or remove that link you have shown.
 
#55
You mean an extra R+C, altough there already is Rz and Cz?

And yes of course, Rf and the connection for Pin5 will be made on the chips pins 🙂

An externally hosted image should be here but it was not working when we last tested it.


Edit: Now I´ve understood - >after< L//R 😀 So does this mean, that my Zobel-Network is senseless??
 
Last edited:
#56
Hi,
most designers and builders use a Zobel (R+C) at the PCB output.
Many also add in L//C to create the Thiele Network of R+C & L//R.
A few and I am now a convert to this, use a Pi network, where an extra R+C are added to give R+C, L//R, R+C.
This still gives a nice load at both low and high frequencies but does a better job of attenuating back EMF and interference.
 
#59
The mute pin is there to allow the user/operator to use the mute to defeat start up noises or to silence the amp at will.
Connecting a permanent resistor as you have shown defeats the mute.
The amp will attempt to turn on as soon as the Low voltage threshold has been passed. That may generate unwanted noise at the output of the starting amplifier.
 
#60
Maxxtr0 said:
Good 🙂 Wich values do you recommend for R+C between the outputs? 2.7/100, too?

P.S. So... we´re finished now? Nothing more to correct in my layout? 🙂
Click to expand...

for the Pi filter I reduce the two Cs to 47nF.
The first R ~1.5* speaker impedance.
//R ~ 0.5* speaker impedance.
Last R ~speaker impedance.

i.e. for 8ohm speaker
12r + 47nF
4r //1uF
8r + 47nF

At very high frequency when the L is effectively infinity and the Cs are short circuit; the load is ~6r.
At low frequency when the Cs are near infinity; the speaker sees 1uF in series with it's own impedance.
 
Status
Not open for further replies.
Top Bottom