My "audiophile" LM3886 approach

Jump to Latest
#2,841
Don't forget the startup protection, there might be turnon thumps.
Click to expand...

This is a very good point, and really required. Turn-on behavior is a bit nasty until things get settled.

It should be easy to modify your current layout for the TO-220 package for the 78xx/9xx, which will relieve any concern about SOA.

Your grounding is probably the weakest part of the layout (islands), but will probably be okay.
 
#2,843
Originally posted by Dxvideo
So you wrote just before my PCB design posting?
Click to expand...

even before... it was before Brian answer! :xeye:


Then if you examined my layout, what can you say?
I applied two seperated gnd plate one is signal GND the other is dirty..
I'v tried to keep signal lines as short as possible and feeback lines are also very short (as I could make of course).
The only possible problem that I can see is, my input capacitor is very big (but its a really good cap and I dont want to leave him!) Do you think it causes any problem?
Click to expand...


as Brian said... at a first glance more or less it looks good.

The input cap. is out of the NFB loop(s), so it's less of a problem... except for possible coupling with other parts of the circuit, which may create spurious feedback paths (and thus possibly oscillations). To avoid any trouble (and save some PCB space too), you may mount it off-board.

The large output currents are creating strong EMF: try keeping the output (and output sense resistor) as much as possible far away from anything else, and in particular from the input and input components (notice that overall the amplifier is non-inverting: any coupling of the output with the amplifier input would likely result in positive feedback and oscillations...).

Of course there must be a balance here, since the series sense resistor is inside both feedback loops and can't be placed far from the rest...
 
#2,845
Mauro uses DC protection [and Served me already several times, always external causes not the circuit]

18mW is for sure Ok. Though maybe You wanted to say 180mW ?! Which is too much for TO92. Is it too much waste of room putting there a standard 78XX?

Don't turn down too much the PS, don't forget that this circuit produces some "waste" of power, You will need all that is available.

But You can try higher than 12V for the LM, also Mauro does this recently.

Ciao, George

Edit: oops, You all type too fast..😀
 
#2,848
Joseph K said:
One more: maybe You will be sorry at a later moment for putting the input film cap right below C4. I mean maybe one time You might want to try to bypass also C4 with a film cap, from below..
Click to expand...

Dear Joseph,
You mean my input cap may cause some bad feelings?
Its a MKP capacitor and not cheap! In fact its an audio grade one.
But may be youre right, this is a strange design and may be a MKP doesnt suit for this design.
So you say; Take the input cap off board and try some different types of film caps (like MKT or Wima MKS 2,4 etc..) Is it?
 
#2,850
No, I'm not talking about the input cap [I don't have it, like Brian has mentioned above]

I'm talking about the feedback cap, C4. It can be useful to leave space for bypassing it with a film cap, but You have placed the input cap below it, so now it's not possible.

Ciao, George
 
#2,855
Yes, I meant that, but this solution is good, also.

One last: don't skimp on the size/ quality of the bypass caps for the LM. for me higher voltage ones [fatter] worked well. Also, seems like Mauro in the original had managed to push them closer to the chip. But this is really just a detail, it's your decision.

Ciao, George
 
#2,858
I See.
The current electrolytics on LM318s rails are panasonic FM type (very low ESR) 100uF. As his datasheet, the current ripple of him is 470mA for 100Khz (half for 50hz)..
And OK..
I can move them closer to the chip (may be on underside 😀 as usual)
And for LM3886 I am using 1500uF also FM types (which have 3300mA ripple)..
 
You must log in or register to reply here.
Top Bottom