[jdpas29]

Hello all!  this is my first post on the board and i'm somewhat of a newbie with the PCB board mods and wiring but i have dabbled in DIY audio with car stereo and also PA/live sound for years. 

I am confused by the amount of suggested capping on the datasheet for boards like the TPA3110 and PAM8403.  It seems like a lot of work and not much space to do so on a small board (especially the PAM).  All i want is the simplest, HiFi (or close to it) sound that i can get with quality components to power a couple of speakers with minimal EMI problem. 

1.  I have noticed that apple phones in particular send a lot of EMI to the board,wires, etc.  the clicking usually shows up in the speakers.  Do you think ferrite beads are enough to deal with this problem?  Must there be a cap placed after the ferrite bead?

2. Datasheet suggests caps at voltage input, VREF to ground, VCC on the board at each power input to the chip going to ground, caps on the input signal, caps on the speaker wires themselves.  This, to me, seems like overkill.  I have noticed that many of the pictures on this site don't have any added caps or very few.  Do you guys think that some of these added caps can be skipped?  If so, which ones make the most sense to skip? 

3.  Overall, what suggestion would you have for the capacitance layout for the 3110 and the PAm8403?  The TPA3110 will be powered by wall wart 12V 2-3A or so.... the PAM8403 will be powered by 3 AA batteries.   The signal will run to a pair of Visiton 5" or similar type full range speaker.   I already have a selection of caps of different values available and i have some good ferrite beads on hand also. 

Any help, suggestions would be greatly appreciated.     pictures would be good also if you have them.

[Folsom]

The TPA3110 sure boards and such come with all the small caps on them. Don't worry about anything except the large DC reservoir capacitors. For those replace them with these or these.

The ferrite will do nothing but degrade sound. Apple makes clicks on purpose, so you know the volume without music playing.

The Walwart will be a disappointment. Try an Astron unit.

[jdpas29]

thank you very much...  what you're mentioning replacing is basically due to the quality of the caps already placed on the board right?

Would a board as small as the PAM8403 benefit from some caps maybe on the DC voltage input? 

so the popping from iphones is not really possible to be affected?

[Folsom]

The only popping on an iphone that's bad would be plugging one in when the amplifier is already on. Which, again, isn't "noise" a ferrite can do anything about. The "noise" you hear isn't in the RF range so a ferrite can't affect it.

You're replacing it with capacitors that are MUCH faster, and reduce noise significantly. They'll change the sound very dramatically in a good way.

The PAM8403 could also get one, but I don't know what size. Plus it uses SMD. There's lots of OSCON capacitor shapes and sizes, but I have to know the size to find it. I don't know what the PAM8403 sounds like so I can't tell you how much better it would get. You can only do so much to make a turd smell nice, if you know what I mean. Maybe it sounds great, so a cap swap would be awesome, but I have no idea.

[Markvdv]

...All i want is the simplest, HiFi (or close to it) sound that i can get with quality components to power a couple of speakers with minimal EMI problem...

Folsom only answers for the "HiFi" part of your question. Ferrite beads or LC outputfilters are not added to classD amplifiers for "HiFi" reasons, they are added to comply with EMI and other regulations. 3110 datasheet tells you with ferrite bead/1nF simple filter and up to 1.25 meter twisted loudspeaker cable it is within (USA) FCC regulations. If you live in Japan you could end up in prison when using amp with same simple filter, laws are a little stricter there.

So yes it will only affect audio negatively, but it does have another function.

[jdpas29]

Thank you very much, Mark.  The speaker cables i would be using would be only 1ft or less in length.  I wonder if that excludes me from the requirement....  any idea where these regs are printed?

[jdpas29]

i found it.. nm..  8 inches is the rule

[Speedskater]

The Apple phone problem is a extremely high frequency RFI interference problem. It's not a low frequency EMI problem, it needs to be dealt with using shielding and filtering.

While there is a lot more to ferrites than just clipping one on somewhere, they can be very effective at dealing with RFI problems.

Jim Brown is the go-to expert on EMI/RFI interference.
Three of his papers on the subject.

The in-depth one:
"RFI, Ferrites, and Common Mode Chokes For Hams"
skip the Ham transmitter sections.
http://www.audiosystemsgroup.com/RFI-Ham.pdf

The audio and ferrites one:
"Understanding How Ferrites Can Prevent and Eliminate RF Interference to Audio Systems"
http://www.audiosystemsgroup.com/SAC0305Ferrites.pdf

and also:
"RF Interference in Audio Systems"
http://www.audiosystemsgroup.com/AES-RFI-SF08.pdf

[Speedskater]

If a correctly chosen and placed ferrite changes the audio quality, it's only because the ferrite reduced the background interference noise level.  So feel free to pick your poison.

[jdpas29]

i see.  thank you for the reply.  i already have some ferrite beads that are recommended for audio applications.  I may need to be more specific about the resistance characteristics of the beads but...  A ceramic cap of 1nF after the bead should take care of it. 

one more question...   On these class D boards, what is the best place to take the capacitor on this EMI filter to GND as there are several different locations and different grounds on the PCB. 

and with these boards, what voltage would you suggest on this capacitor in the ferrite bead/cap filter combo?  the Pam 8403 is a 5V power.   the 3110 is 12V.

[Folsom]

I've got an idea for a better device that will also eliminate ground noise. It's part of my new volume device coming out.

SC-02-300JV from mouser, 2x



That's from Keantoken at DIYaudio Forum. The most important thing is that the box be fully metal, and with an ipod you may have to experiment to whether it's best with source or amp side being connected to the enclosure, before or after CMC's.

And seriously, I'm telling you, ferrites sound like shit. Speedskater knows oogles of information but if there's one thing you'll learn that's overly evident... despite all know engineering conception, ferrites sound like shit.

Are you actually in Japan? Because there's no reason to be concerned about the output of the 3110 unless it interfers with something at home. Realistically your wireless network is probably a much bigger pollutant in your immediate area at such low power.

[Speedskater]

That is one very strange circuit in reply #10.  A common mode choke won't work very well when you unbalance it with a resistor.  The word choices in the labels also seem to be poorly chosen.  But that may just be a translation problem.

[Speedskater]

i see.  thank you for the reply.  i already have some ferrite beads that are recommended for audio applications.  I may need to be more specific about the resistance characteristics of the beads but...  A ceramic cap of 1nF after the bead should take care of it. 

one more question...   On these class D boards, what is the best place to take the capacitor on this EMI filter to GND as there are several different locations and different grounds on the PCB. 
................................

Is this an input circuit filter or an output circuit filter?

Do you have a link to the correct schematic?

[jdpas29]

i was asking about the output circuit filter.  schematic is here:    http://www.diodes.com/datasheets/PAM8403.pdf

[jdpas29]

one more question.   

Datasheet from PAm suggests the range for VREF from .47uF to 1uF...  Strangely the Schematic has a .1uF capacitor..

any ideas why they would contradict themselves in the same document? 

I am making a small stereo portable amplifier that really won't be able to reproduce super low frequencies very well.  this cap is said to be the most important by the manufacturer.  what do you guys think would be the best choice?  should i just try a bunch and see?

any help is appreciated

[Folsom]

That is one very strange circuit in reply #10.  A common mode choke won't work very well when you unbalance it with a resistor.  The word choices in the labels also seem to be poorly chosen.  But that may just be a translation problem.

I improved labeling. And changed R value to better one.

It works well. The CMC isn't meant to provide regular CMC duties, it's just for breaking ground loop in this case (and hence noise).