My_Ref Fremen Edition - need help on PCB evaluation

[AndrewT]

What is of more importance to you?
That it is all on one PCB making it easy to assemble and package with virtually no wiring.
or
That it appeals to many other Members.

I doubt the two are compatible.

[ClaveFremen]

Originally Posted by AndrewT

What is of more importance to you?
That it is all on one PCB making it easy to assemble and package with virtually no wiring.
or
That it appeals to many other Members.

I doubt the two are compatible.

I do agree Andrew, the two alternatives are hardly compatible...

But if I must separate the PCB I think it's better to dump the DC protection that it's easier to replace with kits available from eBay or with a custom designed PCB that can accomodate DC protection for both channels rather than spread the virtual ground between two boards...

[soongsc]

Didn't you already get all of it on there? I though it was just a matter of shifting things around. Seems to be a three partition board, power supply, small signal and high power with protection. Hmm I have not used Eagle PCB, maybe I should give it a try.

[ClaveFremen]

Originally Posted by soongsc

Didn't you already get all of it on there? I though it was just a matter of shifting things around. Seems to be a three partition board, power supply, small signal and high power with protection. Hmm I have not used Eagle PCB, maybe I should give it a try.

I've tried all day, even changing the orientation of the whole board... to have both the DC protection and LM318 PS on sides of the bridges and smoothing caps a slightly bigger PCB is needed (the DC protection takes 1.5*4.0-5.0 cm).

The only way I found so far is the full PCB (0.9.8).

BTW I'm confident this complete configuration would fare well, the 'U' shaped radiating area could be a false problem...

In fact in the little 'U''s branch under half of the LM3886 no important currents should flows.

The LM3886 GND pin is only a reference ground for the mute circuit...

Give a look at the attached zoom picture and help me determining if important currents radiates.

Would that currents be lower if the ground plane under the LM3886 is removed?

[soongsc]

I think most current are going to come from the V+ V- lines through the output 0.47 ohm resistor to the speakers. The OUT line being so close to the LM318 probably is not the best. So what I think is that, on your schematic, anything that is left of the limiter zener should be as far away from the high current section because the LM318 gain is high. All components feeding the inputs should be furthest away. It seems that if the LM3886, c1, c2, c4, 0R47, OUT terminal are moved a bit to the side, they can be further away from the LM318 loop (are those surface mount resistors or limiter diodes beside the OUT trace?)

[ClaveFremen]

nFirst of all thanks a lot.

Originally Posted by soongsc

I think most current are going to come from the V+ V- lines through the output 0.47 ohm resistor to the speakers.

I do Agree.

Originally Posted by soongsc

The OUT line being so close to the LM318 probably is not the best. So what I think is that, on your schematic, anything that is left of the limiter zener should be as far away from the high current section because the LM318 gain is high.

I'll check if it's possible.

I could rotate R3 at the expense of a much longer feedback but distance wouldn't change much.

And/or I can reduce C4 size back to 7.5mm lead spacing gaining 3mm to move R3 toward LM3886.

The original TP layout (excerpt attached) it's not much different in this regard but, obviously, one of my goals is to design a better PCB.

Originally Posted by soongsc

All components feeding the inputs should be furthest away.

Aren't they?

Originally Posted by soongsc

It seems that if the LM3886, c1, c2, c4, 0R47, OUT terminal are moved a bit to the side, they can be further away from the LM318 loop

I thought about it.

It could be possible reducing C1, C2 size back to 10mm diameter gaining 0,5 cm.

But it's worth it?

Originally Posted by soongsc

(are those surface mount resistors or limiter diodes beside the OUT trace?)

If you refer to the two ones near the 10pF caps those are MELF resistors.

PS

Rereading the post I see now that some minor component moving is possible... stay tuned... thanks again!

[soongsc]

Got out my magnifying glass, D1, D2, and the zener are too close to the OUT. If the component/trace runs perpendicular to the high current trace, then it is less effected. Also, field strength reduces by the square with distance.

You can decide where to put you name after the layout is complete.

[ClaveFremen]

Originally Posted by soongsc

Got out my magnifying glass, D1, D2, and the zener are too close to the OUT. If the component/trace runs perpendicular to the high current trace, then it is less effected. Also, field strength reduces by the square with distance.

Ok

Originally Posted by soongsc

You can decide where to put you name after the layout is complete.

You can trust me or not but the presence of credits on PCB from the start was a precise request by Mauro...

[ClaveFremen]

Originally Posted by ClaveFremen

Rereading the post I see now that some minor component moving is possible... stay tuned... thanks again!

Is it any better?

Take in account that unless Vpp on limiter is higher than 28V that circuit should behave like an open circuit.

[soongsc]

Originally Posted by ClaveFremen

Is it any better?

Take in account that unless Vpp on limiter is higher than 28V that circuit should behave like an open circuit.

Even though that might be true, but even a hanging open circuit with only one end connected could effect the sound. That is why the "ground side electrons" thing works. It seems you could still move the limiter to the right.

I wonder if it's a good idea to have the resistor heat sink against the cap, and also it seems that the real heat sink would interfere with the OUT connector because the relation between the legs and the heat sink is different. What might be a way to solve it is rotate the resistor clockwise by 90 deg and rotate the heat sink counterclockwise 90deg and correct the relation between heat sink and resistor leg. This will give you more room around the OUT connector.