| samsagaz |
Hi guys, im using the Schematic that i found in the Datasheet of the LM3886, its how i draw it
http://www.proyectosdiy.com/_tmp/sch.pdf
and i draw the PCB using just one side in that way
http://www.proyectosdiy.com/_tmp/pcb.pdf
can someone please tellme some tips about the Sch and PCB?
i dont know if can add something to the schematic to get it sound better.
or if can fix something in the PCB, im now on this :)
Thanks in advance |
|
|
| Dino-Indo |
Hi Samsagaz,
May I propose the attached PCB Layout..
I didn't built it myself... purchased local from an Indonesian supplier here.. but probably you could try to implement it...
It sounded really good... I'm using some Riken MetalFilm 0.5 Watt resistors and some Elna RE3 blue capacitors 1000 uF/50V (for power supply smoothing caps on C 9 and C 10).
And instead of placing 0.1uF MKM poly cap on C7 and C8 also installed additional Elna RE3 blue capacitors 1000 uF/50V on C7 and C 8 position and then putting the 0.1uF cap below of the PCB at the feet of these additional 1000uF caps...
Good Luck!! |
|
|
| sneih |
hi dino i'm confuses about caps position on C3 at your picture attached!
that's show polar like + - - +
is it right position? |
|
|
| Dino-Indo |
Yes..:D you are right..
I installed them back to back.. so it's become non polar (?)
Hehehe..
But it does sound good... |
|
|
| sneih |
u mean: - + , -+ ?
thx!! |
|
|
| Dino-Indo |
Nope... just follow the silk screen/top face of the PCB (the green one).. So the lay out will be + - - +
:D
Good luck!! |
|
|
| neutron7 |
If you dont need the extra few watts,
if you can still get them, the Lm3875 in my OPINION has a much smoother sound than the 3886 |
|
|
| sneih |
| quote: | Originally posted by Dino-Indo
Nope... just follow the silk screen/top face of the PCB (the green one).. So the lay out will be + - - +
:D
Good luck!! |
thx
btw i also have this pcb n still in progress
my problem:
honestly i have 3-4 part that they included but not in pcb....
n i dont know how to install them
ex:
- C4: 47pF-50pF = that seems not in the pcb..
- C5,9,10: 100uF/50V = where i install this part? replace with 1000uf?
- R4, R6: ?? |
|
|
| samsagaz |
i will build LM3875 too, i wanna test diferents GC.
and want to build my own PCB in all the cases.
But i just dont know if the Datasheet schematic are the best choice, ir can be improved.
Regards |
|
|
| Puffin |
samsagaz. Build a simple gainclone first, get to know the pin layouts intimately. Then start to experiment with different resistor feedback values, cap values etc.
I have just spent the best part of 2 days trying to make my second bridged gainclone (first was with LM3886's, this time with LM3875's). No sound, then lots of noise, then no signal input, rebuilt the thing 5 times ! Just got it working and it sounds fantastic (of course) I can go and eat now !. |
|
|
| sneih |
| good luck for u!! fell happy right? |
|
|
| samsagaz |
| quote: | Originally posted by Puffin
samsagaz. Build a simple gainclone first, get to know the pin layouts intimately. Then start to experiment with different resistor feedback values, cap values etc.
I have just spent the best part of 2 days trying to make my second bridged gainclone (first was with LM3886's, this time with LM3875's). No sound, then lots of noise, then no signal input, rebuilt the thing 5 times ! Just got it working and it sounds fantastic (of course) I can go and eat now !. |
well, i already made an GC, but not from the sctatch like im trying now :) |
|
|
| Dino-Indo |
Hi Sneih..
Please find attached the schematic that I got from Mr. Alex P of langsungjadi.com...
Hopefully it would be useful for your reference,
Kind Regards,:D |
|
|
| Edits |
| quote: | my problem:
honestly i have 3-4 part that they included but not in pcb....
n i dont know how to install them |
R4 and C4 put in series are feedback. You don't have to use those. I think you use a single feedback resistor (in this case R5) OR try R4+C4 in place of R5.
Other components you mention are clearly marked on the image attached by dino-indo. As of C9 and C10 you can try different values: starting with 100,120,220,1000 even 2200 uF |
|
|
| sneih |
| quote: | Originally posted by Dino-Indo
Hi Sneih..
Please find attached the schematic that I got from Mr. Alex P of langsungjadi.com...
Hopefully it would be useful for your reference,
Kind Regards,:D |
| quote: | Originally posted by Edits
R4 and C4 put in series are feedback. You don't have to use those. I think you use a single feedback resistor (in this case R5) OR try R4+C4 in place of R5.
Other components you mention are clearly marked on the image attached by dino-indo. As of C9 and C10 you can try different values: starting with 100,120,220,1000 even 2200 uF |
thx friends 4 ur support |
|
|
| yangsmm |
| I made LM3886 Power Amp. Modules. It support stereo output. |
|
|
| yangsmm |
I made LM3886 Power Amp. Modules. It support stereo output.
Oh.. Picture didn't attached on the previous post, So I attach the picture. |
|
|
| Dino-Indo |
Hi Edits..
Really appreciate your post...
Guess you also built one using the same PCB?
:D :D
Hi Yang!
Nice PCB !! and truly simple layout...
How much does it cost?
Wonder whether I can purchase it from Indonesia?
:D |
|
|
| Rave17 |
| quote: | Originally posted by yangsmm
I made LM3886 Power Amp. Modules. It support stereo output.
Oh.. Picture didn't attached on the previous post, So I attach the picture. |
The lay-out of the PCB looks really clean :) Where did you get it from, made it yourself? |
|
|
| yangsmm |
| When I assembly the LM3886 Power Amp, It was to find the heatsink for it. I can't make the hole for bolts. so I paid to drill the hole. and I am going to try to make upgraded power amp with LM2886. So why don't we try together? |
|
|
| Edits |
| quote: | | Guess you also built one using the same PCB? |
Yes dino, I used them as my first GC, tweaked them until broken beyond repair:D
And congratulation for the BT! who knows may be I was one of the judge? or one of the paticipant sitting next to you?:D :D |
|
|
| Dino-Indo |
Whohoho Edits.. Many Thanks! :D :D
Now.. I'm still trying hard to remember who was sitting next to me during the Blind Test.. :bawling:
Also tried hard to remember who was the juror for that event.. :bawling:
But, I got to admit that it was fun for me to be able to participate in that blind test... truly a memorable experience.. especially for me.. a newbie DIY :D :D |
|
|
| samsagaz |
Hi guys, i build this schematic, test it and dont work :( no sound at the output, just an HUMMM.
Can someone please check if everything are OK?
http://www.proyectosdiy.com/_tmp/sch2.pdf
Thanks in advance. |
|
|
| pacificblue |
Hola, Samsagaz.
The schematic shows all GND connections to have the same potential. Maybe you have produced a ground loop somewhere. Many threads in this forum deal with separate grounding for signal and power, and with star grounding concepts. If you cannot find your solution there, please post photos and/or your PCB layout.
Do you get hum with anything connected to the input? Or with the input floating? Or shorted?
Does the hum go away, if you lift GND from PE?
Do you support C2 and C3 with more capacitors outside the PCB? Or do you really only use 220µF per rail? |
|
|
| Puffin |
I Should learn to look more closely "my first LM 3886GC".
Is the picture of the PCB from the top or underneath ?
From the schematic you posted it looks as though Pin 3 should be connected to pin 10, not pin 9 |
|
|
| pacificblue |
As far as I understand your PCB you have put the star grounding point on the PCB at the bottom of your drawing (bigger drilling), is that right?
The ground plane forms a closed loop around the PCB. You could try to interrupt it somewhere, e. g. behind the chip. At the moment you have two paths from the ground connection to Ri. A short path over the right side, and a long path over the left side running on the top over the chip to the right and then down again. You might get the hum from there.
You can check that before cutting the PCB. Lift Ri or Ci, which ever you have connected there, from the ground below IC pin 11, where it seems to be connected now, and connect Ri/Ci directly to the star point with a wire. Make sure the wire doesn't connect to the ground anywhere else. If the hum goes away, you need to cut the loop as I wrote above.
If the hum doesn't go away, you may still cut the loop and it will probably help, but then there might be some further cause for the hum. |
|
|
| ionomolo |
| quote: | Originally posted by samsagaz
I get the HUM with any input, if i get it floating no HUM.
im using just 220uf per rail, the Power Supply have like 30000uf
this is the PCB Design.
http://www.proyectosdiy.com/_tmp/pcb.pdf |
If the hum is not present with the input floating then there is a ground loop with your source through safety earth. Try feeding it with an ipod/mp3 player (of course not connected to a computer by usb) and check there is no hum with the ipod paused (if it's switched off the output becomes high-impedance and acts like an antenna. |
|
|
| samsagaz |
| quote: | Originally posted by Puffin
I Should learn to look more closely "my first LM 3886GC".
Is the picture of the PCB from the top or underneath ?
From the schematic you posted it looks as though Pin 3 should be connected to pin 10, not pin 9 |
This PCB are from Top. the schematic show that Pin3 are connected to pin 9, not 10. and in the PCB are connected to Pin 9, not 10.
| quote: | Originally posted by pacificblue
As far as I understand your PCB you have put the star grounding point on the PCB at the bottom of your drawing (bigger drilling), is that right?
The ground plane forms a closed loop around the PCB. You could try to interrupt it somewhere, e. g. behind the chip. At the moment you have two paths from the ground connection to Ri. A short path over the right side, and a long path over the left side running on the top over the chip to the right and then down again. You might get the hum from there.
You can check that before cutting the PCB. Lift Ri or Ci, which ever you have connected there, from the ground below IC pin 11, where it seems to be connected now, and connect Ri/Ci directly to the star point with a wire. Make sure the wire doesn't connect to the ground anywhere else. If the hum goes away, you need to cut the loop as I wrote above.
If the hum doesn't go away, you may still cut the loop and it will probably help, but then there might be some further cause for the hum. |
The ground is the big Hole at bottom like you say. So you say that is not good to use an full plane gnd, right? will be better to cut the PCB to separate the Gnd?
i will try what you say.
Btw the HUM that i get are without sound, just HUM, no input.
Btw, i dont have anything connected to Pin 11 right now :S in the datasheet i found that Pin11 are NC.
| quote: | Originally posted by ionomolo
If the hum is not present with the input floating then there is a ground loop with your source through safety earth. Try feeding it with an ipod/mp3 player (of course not connected to a computer by usb) and check there is no hum with the ipod paused (if it's switched off the output becomes high-impedance and acts like an antenna. |
i thinked that the PCB are wrong, but if noone noticed anything strange, so will try to play with the GNDs, to see if can improve it. |
|
|
| pacificblue |
It is a matter of taste, if you prefer planes or not. The important thing is that you don't produce a closed loop. You don't need to separate the grounds, only break the loop. A small slit somewhere is sufficient to do that. Behind the IC seems to be the best place for that.
You can of course separate the grounds entirely, if you like. But that would mean a total redesign for your PCB. Then you would probably not be working with planes any more, but with two star-shaped grounds.
Planes have their advantages as well. They reduce resistances, take higher currents and improve cooling (the last very little, but still), and somethimes even provide a kind of shielding. So a small cut through the path on top of the IC should be sufficient. But try first by lifting Ri/Ci from the current ground connection to the star point, if that is already the solution. Ionomolo has a point, if the hum goes away with the input floating, there is a loop with the source.
Pin 11 is NC. You can connect it wherever you like or leave it floating like you do now. |
|
|
| samsagaz |
i cut the PCB to recomve the loop gnd and still getting the same problem :(
i connected an ipod and still get the same problem, no sound from speakers just an HUM.
i think that the problem is the schematic, im using something wrong, or maybe the IC are dear :S |
|
|
| pacificblue |
Do you only have one PCB or do you have two (for stereo)? If you have two, are both humming? Does the hum still go away, when you have nothing connected to the input?
The strange thing is that the hum is so strong, you don't even get a music signal. Sounds like you have a feedback loop from output to input that you interrupt, when nothing is connected to the input.
Although it is somehow a good sign, because it shows that the IC is still working. A broken IC has either only DC at the output or nothing at all.
Back to the loop. Are your input and output connectors mounted in the amp case already? And are they isolated from it?
I once had a similar behaviour, loud hum and no music, when I had a bad connection from Ri/Ci to ground (R2/C1 in your drawing). Maybe you can check that, too. |
|
|
| samsagaz |
i have just one PCB; i will made a few now, just to test if the IC are wrong, the output of the amplifier are always DC, if i turn on the vol of my Pre, i get the same RMS value :( no increased the volts :(
will made this new board and will check it something are wrong...
what happend if i change the Ci capacitor? its not an electrolytic one, right? i tryed with an 0.22uf cap and still getting the same problem. |
|
|
| samsagaz |
| quote: | Originally posted by pacificblue
Do you only have one PCB or do you have two (for stereo)? If you have two, are both humming? Does the hum still go away, when you have nothing connected to the input?
The strange thing is that the hum is so strong, you don't even get a music signal. Sounds like you have a feedback loop from output to input that you interrupt, when nothing is connected to the input.
Although it is somehow a good sign, because it shows that the IC is still working. A broken IC has either only DC at the output or nothing at all.
Back to the loop. Are your input and output connectors mounted in the amp case already? And are they isolated from it?
I once had a similar behaviour, loud hum and no music, when I had a bad connection from Ri/Ci to ground (R2/C1 in your drawing). Maybe you can check that, too. |
The Input and Output are not connected to the AMP, im just using some wires to test it.
The new PCB looks nice right?, i really like it :) |
|
|
| samsagaz |
Now i build other with the new PCB and still getting the same Error, i cant get an 22uf capacitor just 10uf (electrolytic ones), is possible to use other value on CI?
Thanks |
|
|
| pacificblue |
Ci is most probably an electrolitc capacitor, because better capacitor types are very big and expensive at 22µF. You can use different values for Ci, and have to find your best compromise. The bigger Ci, the lower the bottom roll-off frequency. 0.22µF is not a practical value, it should be 22µF or bigger. With 0.22µF and 1k you get a lower-3dB at 723 Hz. (fc=1/(2xPixRixCi)
So you have DC at the output? How much? And do you also have DC at the input?
Is the humming accompanied by a clicking sound like very fast switching noise? Then you hear the spike protection system trying to protect the IC. Can you put any capacitor in as Cin (between input and Rb or R1 for you) to see if blocking DC from the input helps? You can use any value for the test. If the test is successful, use 1µF or bigger. Must be good quality, MKS or better. |
|
|
| samsagaz |
| quote: | Originally posted by pacificblue
Ci is most probably an electrolitc capacitor, because better capacitor types are very big and expensive at 22µF. You can use different values for Ci, and have to find your best compromise. The bigger Ci, the lower the bottom roll-off frequency. 0.22µF is not a practical value, it should be 22µF or bigger. With 0.22µF and 1k you get a lower-3dB at 723 Hz. (fc=1/(2xPixRixCi)
So you have DC at the output? How much? And do you also have DC at the input?
Is the humming accompanied by a clicking sound like very fast switching noise? Then you hear the spike protection system trying to protect the IC. Can you put any capacitor in as Cin (between input and Rb or R1 for you) to see if blocking DC from the input helps? You can use any value for the test. If the test is successful, use 1µF or bigger. Must be good quality, MKS or better. |
well, this dont will help me, i used diferents inputs and always the same hum :(
i will check again the schematic, but cant found anything wrong :( |
|
|
| pacificblue |
| You used different inputs?:confused: |
|
|
| samsagaz |
| Yes, i tryed with my PC and with an Ipod |
|
|
| pacificblue |
Ah, I see. But can you measure DC at the input, when you connect anything? Or can you put a capacitor in line with the input to see, if the hum goes away (Cin in the datasheet)? For the test the value doesn't matter.
Could it be that your input connections are reversed? |
|
|
| samsagaz |
will try with other Input wires, i will build another wire tomorrow, i cant figure whats wrong with this amps :(
Can someone tellme some values that i can measure to check if everything are OK?
Regards |
|
|
| samsagaz |
i really dont want an Point To Point amp, i want to use an PCB that works :(
I cant figure whats wrong will try with other components values. |
|
|
| pacificblue |
You can find many working PCBs on the web. Buy or copy any of them. If it doesn't give you the same problems, you know for sure that your PCB didn't work. Then you can still improve yours.
If you have the same problems with the PCB from the web, you know that you have to find the solution outside of the PCB.
The advantage of the point-to-point solution is that you can do the same test without any PCB. No problems with p2p means, improve your layout. Problems with p2p means, search the fault outside of your PCB. |
|
|
| samsagaz |
I CANT BELIAVE IT!!!! i get it working! both PCBs works great!!! :S
the problem was the input wire :S was in short!!!! DAMN DAMN DAMN
now i get it working and works GREAT !!! :D
well, now i have two design (2 pcbs), and want to optimize the components values / brand to get the best possible sound.
Hope you guys can help me on that !
will play with PSU later, now i just want to optimize the size of the PCB and the components. like i told before, i think that the new PCB are a little small, and maybe will get into troubles to solder the output wires and stuff like that.
Can someone tellme if exist some kind of guideline to build the pcbs, and if always small are better? :D |
|
|
| samsagaz |
wel,l now im trying to build the Deluxe version of my LM3886 Gainclone, im using 1 trafo / channel (maybe will try with two to see if something are improved.
Will use Caddock MK132 for feedback, Riken for gain setting, and for smoth my transformer will use some Panasonic Caps. Wich any can improve the amp?
Thanks |
|
|
| sneih |
hmmm congratulation....
please take a photo of your new GC hehehe |
|
|
|
