TPA3116 Board running hot?

[Jon86]

I bought one of those cheapy little TPA3116 50W+50W boards off ebay, and after testing it for a few hours last night off two 12v lead acids, it doesn't seem to be doing a great job.
It all works fine until about 20-30 minutes when it starts distorting horribly and once it's left to cool off it works fine again. I can't understand why as we were only running it at around 1W. If it's left to idle for a while the inductors get a bit toasty, and the chip also warms up quite a bit.
My first thought was the fact that I might be killing the caps as they're being run at half a volt more than their W.V, but they don't seem to be hot or drawing any extra current. I'll replace them anyway, but I just wanted to know if I should try and solve the problem now or just return the board and get another.
Correct me if I'm wrong, but the maximum supply voltage for this chip is 30, and I'm running it at 25.5, so surely it shouldn't have any problems?
Output is going into two 6R speakers.

http://www.ebay.co.uk/itm/TPA3116-Inductor-Circuit-Design-Class-D-Digital-Power-Amplifier-Board-50W-50W-UK-/291152081620?pt=UK_AudioTVElectronics_HomeAudioHiFi_Amplifiers&hash=item43ca0432d4

[IO390]

The max supply for the TPA3116D2 is 26V, so it may be getting a bit unhappy. Perhaps reduce it a bit and see if that makes a difference.

That is odd, though, as I've played with the mid-sized TPA amps a lot (really impressed by them) and I can't make the things heat up no matter what happens (until they blow up of course).

[Jon86]

But at 25.5? Seems a bit strange... Especially the inductors...
I'm idling it at 22v now to see if that makes a difference, but surely if the absolute max is 30 then it should be fine anywhere below that?

[mij59]

Hi,

Maybe the heat sink is too small, the one used in the data sheet is a lot bigger.
You could forse  some air over it and see what happens.

[IO390]

The inductors warm up eventually because of the current passing through them - that is normal.

If I were you, give it a go on one lead acid and see if it works.

BTW I'm making another batch of TPA3132 25W/channel amps in a few weeks. 1x1.5" board - £15 if you're interested.

[Jon86]

I really need a bit more than 25W/Channel for this project, but I'd still be interested in a board. Send me a message when you're getting close to production 

Okay so I tried running it at 22V and it's drawing 60mA, so 1.3W is being dissipated into that tiny heatsink... So it still all gets pretty toasty. Yeah somehow I don't think this board was designed to handle it running at 50+50W at full supply voltage 

I'm looking at this one next:
http://www.ebay.co.uk/itm/TDA7492-2x50W-HIFI-D-Class-Digital-Amplifier-Board-Amplificador-Bordo-Radiador-/390885570549?pt=LH_DefaultDomain_186&hash=item5b029877f5
Seems like it has a much more substantial heatsink on it, and going by the datasheet it looks like it would be happier at 25.5V
Anyone had any experience with this amp? I'm guessing the quality/frequency response is roughly the same?

[Tinkerer]

Well darn, I think most people could have said the heat sink is the problem there. Not that it matters if you go with a new one, but have you tried measuring the temperature and comparing to the datasheet curves?

[Jon86]

I don't know where my thermocouple is at the moment, but yeah that must be the problem, if it's getting hot to the touch with no input, then running it at a couple of watts for a little while must be pushing it over its thermal cutoff. Time to try and return this heap of shit, 100W my arse!

[Jon86]

I reckon there are some dry solder joints lurking on that board somewhere too, when I pick it up and put my finger on some of the components it starts making popping noises...
Yeah, f this board.

[IO390]

Your best bet might actually be to just design your own.

I'd recommend staying with the TPA3116 as it's a good chip. Based on similar things I've done, total BOM would be a tenner max with PCBs from Itead or wherever. If you want decent sound quality, all these ebay boards are pretty hopeless. Nowhere near enough via stitching for a start.

Also ferrite beads and a RC snubber on the output sounds far better than LC filters from my experience.

Oh and I saw in the datasheet that the max recommended voltage for the TPA3116 when running a 4 ohm load is 21V, so that may be the problem.

[Jon86]

Yeah I was originally going to design my own, but I don't really have the tools to do proper SMD soldering yet, at least not at that kind of pitch. I just want a quick fix to this problem really, I won't be using it an awful lot anyway.

[IO390]

All you need is a pair of tweezers and a heat gun (carefully controlling the temperature of a toaster oven works best though).

Keep it bigger than 0805 and it's easy as.

BTW, what is this project? Does the sound quality matter at all and how much power do you actually need? Also how many units will you be using?

[mrflibble]

On switchers from ebay the cheap inductor is candidate numero uno to be swapped out and replaced by a better quality inductor.

[Jon86]

I want to get a good few watts out of it, probably about 25 per channel, but obviously I don't want to be pushing the amp too much and I want plenty of headroom. It's just a little portable speaker to take outside with me, well not really little it's half a metre wide and weights a good few kilos. The sound quality isn't really a big deal considering it's just an MDF box with a couple of cheap speakers in it, but it sounds pretty damn good compared to phone speakers.

Could the inductors be what was causing the problem then? I'm sending this board back anyway, but strangely I can't seem to be able to recreate that problem, I ran it at quite high power for a few hours earlier and it didn't seem to care, and the heatsink didn't get above idle temp anyway. Maybe just a dodgy connection somewhere? Or I guess it could be that it wasn't thermal and something else was going on?

[David_AVD]

Amplifiers getting hot without a lot of signal is sometimes due to supersonic oscillation.  I'm not sure if that could apply to your (class D ?) amplifier though.