IC gets really hot

Hello
I'm trying to use, togher with an arduino mega, one of those cheap chinese mp3 player (ipod shuffle clones).
This is its pcb (picture taken from the web, not my actual PCB, but almost identical to mine)

I connected 5V and GND from my arduino to this PCB (I connected 5V on the USB socket pin).
Everything works (led blinks, I can hear sounds, skip songs, etc), except the fact that the IC on the player PCB gets really really hot.
This doesn't happen when I connect the mp3 player to a USB port on my PC, skipping the arduino (the same USB port that supplies the arduino).

Unfortunately I can't find any info or datasheet about the IC (furthermore I have others of those very players and they all have different IC names, but the PCB is identical).

PS: if I put 3.3V instead of 5V on the player, it won't even power up.

What could be the reason?
thank you for the help

I connected 5V and GND from my arduino to this PCB (I connected 5V on the USB socket pin).
Everything works (led blinks, I can hear sounds, skip songs, etc), except the fact that the IC on the player PCB gets really really hot.

Are you 100%v sure you’re connected to +5V and ground? Do you have a multi-meter to check the voltage?

Some of these things also have a virtual ground for the headphone output, and that can cause overheating/excessive heating if you connect the virtual ground to an actual ground. I don’t see any “big” output capacitors, so I’m guessing it uses a virtual ground. If it’s a virtual ground issue, we need to see the actual datasheet (or schematic). And if you bought on eBay from someone who can’t provide documentation, you got what you paid for…

Do you have headphones (or a speaker) connected? Does it get hot with no audio connection? Usually, these things are not designed to directly-drive an 8-Ohm speaker, and it might get warm/hot when driving headphones.

The photo (which you say isn’t yours) shows red to - and black to +, which is “unconventional”. But, the electrons don’t know the wire color and if the power is reversed it won’t work at all.

Overheating could be caused by reversed power, a shorted output-pin on the chip, or excessive voltage.

Some of these things also have a virtual ground for the headphone output, and that can cause overheating/excessive heating if you connect the virtual ground to an actual ground. I don't see any "big" output capacitors, so I'm guessing it uses a virtual ground. If it's a virtual ground issue, we need to see the actual datasheet (or schematic). And if you bought on eBay from someone who can't provide documentation, you got what you paid for..

Do you have headphones (or a speaker) connected? Does it get hot with no audio connection? Usually, these things are not designed to directly-drive an 8-Ohm speaker, and it might get warm/hot when driving headphones.

I had it connected to a PAM8403 amplifier PCB, like this:

I connected the mp3 player output jack to this amp. pcb (tip to LIN/RIN, ring to GND).
I removed the GND connection and not it doesn't get hot!
so how should I connect the player to the amp then?

I removed the GND connection and not it doesn't get hot!
so how should I connect the player to the amp then?

You need a "signal" ground... Without a ground you'll get a "vocal removal" effect with stereo sources where the left & right channels are subtracted, removing everything in the "center". And mono sounds (sounds that are identical in the left & right channels) will be completely (or almost completely) cancelled.

You need to add 3 capacitors in series in-between the sound board and the amplifier (left, right, and audio ground). The capacitors will block DC voltage & current while allowing the AC audio signal through.

Depending on the input impedance of the amplifier and the frequency range you need, you may be able to use 0.1uF capacitors. If you are not getting enough bass, use 1uF (or greater) electrolytic capacitors with the +end toward the audio board (where the voltage is coming from). It's only a couple of volts, so don't worry about the voltage rating on the capacitors.

It's OK to connect the power grounds of all of the boards together. It's that audio-output virtual ground that's causing the trouble.

(I used the wrong terminology... It's a "virtual ground", not a "floating ground".)

NEXT TIME, don't buy stuff if you can't get the datasheet! :wink:

DVDdoug:
You need to add 3 capacitors in series in-between the sound board and the amplifier (left, right, and audio ground).

ok, one thing that I didn't understand is, the third capacitor, the one for the audio ground, should be placed from the audio ground of the mp3 player (electrolytic "+" here) to the power ground on the amp pcb (electrolytic "+" here) or where else?

DVDdoug:
NEXT TIME, don't buy stuff if you can't get the datasheet! :wink:

you're right, but a requirement for this little project was to be as cheap as possible :slight_smile:

Now, I should add a SPI digital potentiometer as volume control.
(Microchip MCP41100, datasheet: http://ww1.microchip.com/downloads/en/DeviceDoc/11195c.pdf)
I guess I should place it between the MP3 player and the amp board, and not between the amp and the speaker.
If so, where should I connect the third lug of the potentiometer (that is supposed to be tied to ground in order to reduce volume)?
I guess I should tie it to the audio ground before the added capacitor, right?

Would I get some other kind of problem using that digital potentiometer?
thank you

ok, one thing that I didn't understand is, the third capacitor, the one for the audio ground, should be placed from the audio ground of the mp3 player (electrolytic "+" here) to the power ground on the amp pcb (electrolytic "+" here) or where else?

The + side of all 3 capacitors goes to that sound board with the "weird" output. All 3 of those pins (including virtual ground) are "biased" at about +2.5V. If a 0.1uF cap works, those ceramic capacitors are not polarized so it doesn't matter which way you wire 'em.

Now, I should add a SPI digital potentiometer as volume control.
(Microchip MCP41100, datasheet: http://ww1.microchip.com/downloads/en/DeviceDoc/11195c.pdf)
I guess I should place it between the MP3 player and the amp board, and not between the amp and the speaker.

Right! The volume control goes at the input of the amp.

If so, where should I connect the third lug of the potentiometer (that is supposed to be tied to ground in order to reduce volume)?
I guess I should tie it to the audio ground before the added capacitor, right?

It goes after those capacitors... Where the DC bias has been filtered-out and you have a "normal" audio signal. The pot's ground goes to true-ground (the amp's ground).

Those capacitors go right to the board's output to "clean up" to signals (and ground) before anything else is connected.

Would I get some other kind of problem using that digital potentiometer?

I don't know about that particular digital pot. Usually, the only issue is if you don't get the timing right and you get a short full-volume burst between volume-changes. It usually happens fast, so it doesn't sound like a "volume blast", it usually sounds like a click or a pop.

You may also find that you'll need to program-in a "curve". Our hearing is approximately logarithmic, so when you set the pot to 50%, it will sound like you are at nearly 100%. Mechanical volume controls use an [u]audio taper[/u] so at the half-way mechanical setting, the electrical output is 10 or 20% (I forget exactly).

That's easy to deal with in software.... Just take smaller steps (or fewer steps) when you're at low volume, and make bigger steps (or more steps) when you are at higher volume settings.

thank you now it works flawlessy,
I added the digital potentiometer.
The problem now is that even with the potentiometer full "closed", I can hear music low, it's not mute.
I measured resistance and is about 80 ohm when it is supposed to be tied to ground (the digital potentiometer is 10K, 8 bit, so it should be normal).
how can I avoid that?

edit. I noticed that bypassing the DC cap that I put in order to remove the virtual ground problem, it becomes almost mute (but still audible)

Dimitree:
I measured resistance and is about 80 ohm when it is supposed to be tied to ground (the digital potentiometer is 10K, 8 bit, so it should be normal).
how can I avoid that?

How about...
more resistance in series with the "pot" (between the audio source and the pot's high side)?
Top volume may be compromised, but maybe running it all the way up now puts in "Distortionville", so - no big loss (Arf!).