Cheap TPA3118D2 boards, modding them and everything that comes with it

Jump to Latest
#2,563
I just purchased several from China off of ebay that were fakes. None even worked and one passed supply voltage to the speaker and I didn't notice until the voice coil started smoking (too late). It didn't make any weird sounds, just a pop upon start up that didn't sound unusual.
 
#2,566

Attachments

  • index.jpeg
    index.jpeg
    174.3 KB · Views: 170
  • Like
Reactions: availlyrics
#2,568
Yes, but not too different from the one on the left here which is THE BEST in my opinion of the cheap chinese TPA3118 boards. The one in this picture is real. Note the bigger and correct TPA chip. Note also the differences in the SMD resistors and caps. They are in different locations on the input side of the chip vs the fake one in my previous post.

EDIT: to point out the components circled DO NOT apply to this post... that was marked up by someone else. This is not my photo.
 

Attachments

  • Side by Side w Notes.JPG
    Side by Side w Notes.JPG
    270.7 KB · Views: 224
  • Like
Reactions: availlyrics
#2,569
I wouldn't claim any are real, but the ampboard on the right is the normal "sanwu" clone one, with this time the tpa lettering in about the correct location. But never mind, even the tpachips with incorrect printing did measure ok earlier. The 28pin is rebadged 3110 or 3110 clone most likely, they made ampboards with 3110 too, but maybe sold too little ???? But I would connect DC to the speakeroutput automatically LOL
 
#2,570
The left one in the picture above (with 330uH inductors) is one I've ordered many of and they sound better. They have 35V caps instead of 25V. They also seem to last pretty well, whereas the small board on the right with 100uH inductors do not seem to last well (at least running on 24VDC).
 
#2,574
The board you like will measure worse, the design is ugly to me, a bit messy with input signal from the rear needing to get to the front side of chip and voltage input at the front needing to get to the rear side of chip, the outputfilter you like is a filter for 4 ohm but for a older lower switching frequency design, the 33uH inductors will add more distortion and the filter probably alters phase in the audible frequencies a little more. I don't know what c8 and c9 would do, but there is no close to powersupplypins decoupling, design is ugly, measured performance will be bad. Just my personal opinion looking at that funny board🙂
 
  • Like
Reactions: availlyrics
#2,575
Yes, what a strange thing to put the input on the back just to make it go to the front and to put the power on the front just to make it go to the back. Why?

I feel like the TPA3250 is now the new 3116/18. Prices have dropped down to the $15 region and is there any reason to continue with 3116?
https://a.aliexpress.com/_mL7KEZ4

D7026D25-9CB9-4443-B613-8052C0FB66FB.jpeg

Inputs on the front and power and speakers on the back, as it should be.
 
#2,576
irribeo said:
The board you like will measure worse, the design is ugly to me, a bit messy with input signal from the rear needing to get to the front side of chip and voltage input at the front needing to get to the rear side of chip, the outputfilter you like is a filter for 4 ohm but for a older lower switching frequency design, the 33uH inductors will add more distortion and the filter probably alters phase in the audible frequencies a little more. I don't know what c8 and c9 would do, but there is no close to powersupplypins decoupling, design is ugly, measured performance will be bad. Just my personal opinion looking at that funny board🙂
Click to expand...
I tested both the big and small boards back and forth (with a 4 ohm speaker load) and the bigger sounded much more "musical" to me. Less digital and plastic. Of course you can't quantify that using test equipment and that's just my opinion.

One thing you CAN quantify is that the little boards use 25V caps. There's no reason to do that since the 3118 is designed to run at up to 24VDC. I feel much more comfortable with the 35V caps on the bigger boards.
 
Last edited:
#2,577
You are funny too. You had 28pin tpa3118's destroying your speaker and you feel comfortable with a plastic wrap on a can showing 35V. You do realise even we can replace that plastic wrap within seconds, now think how more difficult it is to print tpa3118 on a differently labelled 28pin chip and even that is what happens, they sand down the surface and re-laserprint them, telling you to relax because it is a 32pin TI chip with all kind of protection circuits baked in 🙂
 
#2,578
I have used the small 25v board and had a few of them fail already. The big board with the actually 60W 3118 chip, I've had zero failures.

Could be coincidence. I suppose, I could test the failure point for both caps if it really was in question. I think though, it's much more likely to fake the chip (the heart of the board) than voltage rating on the caps. But... that's just me.
 
#2,579
Did any of the 4 25V caps on any of the failing ampboards fail ? Destroying a speaker is failing, because tpachip has protection that should trip when too much dc is on output, chip should start trying to restart pop pop pop pop when it fails every time to restart with protection tripping. Most common external to ampchip fails are cracked little ceramic caps, tension the pcb a little and crack it can on any of those tiny buggers, also pop pop pop, then frequency will tell a little about the size/value of the cracked ceramic, we tend to throw away the ampboard when one of those 0,1 cent parts are damaged some how, we can replace the ampboard so cheaply and we expect the ampchip on the board again to be as 3118 as before 🙂
Ampboards with tpa3118 sourced from TI or Digikey/Mouser fail in same manner, cracked ceramics or damaged smd film caps for the $130+ ampboards
 
You must log in or register to reply here.
Top Bottom