Amplifier noise/heat up the wazoo (TDA2009)

[Rattanee]

Hey guys, I'm in a bit of a pickle here with this amplifier circuit. I've a few TDA 2009 chips on hand, hence my choice to use these.

I've basically built the reference schematics listed in the datasheet. I have two of these chips on a single board, one in a stereo setup, and the other in bridged mode. The sound source is the line level out of a PC, which, on the pcb, gets to the stereo amp, and gets a mono mix through some resistors into the bridged amp which goes to a subwoofer. It is powered from the PC powersupply's 12V rail.

The setup does work to some degree, but something's not quite right. I'm getting an odd white-noise, and all sorts of HDD and fan-noises in the signal, along with severe distortion even at low volume levels. The amp chips also get quite hot, something I wouldn't have expected from these guys, seeing how they're fairly low power. Pretty sure these symptoms will be somehow connected.

I already built a simple stereo amp with these chips, which did work well, but also suffered from HDD/Fan noise, it didn't heat up near as much though, and didn't have any distortion.

Does anyone have any tips what the culprit could be ?

I'll post my schematics in a bit (soon as I get to my workshop puter)

Thanks in advance!

 

[Rattanee]

So I've attached the schematics. Pardon the slight jumble with U2, the datasheet flips the inputs and outputs upside down and I noticed that after I drew this so it was a quick fix.

Basically the inputs go direct to U1 for a stereo setup through a pair of 50k trimmers to adjust volume. The inputs are tapped through two resistors to combine the channels for the bridged amp, through another 50k trimmer to adjust the volume. The output of the second amp has a header J6 for an inductor and an added capacitor C19 that comprise a second order low pass filter as this is to drive a subwoofer.

The input headers are wired directly to a jack that goes to the soundcard.

All component values are accurate to the PCB except C19 which shouldn't matter for these problems.

Anyone has any ideas?

 

[Rattanee]

Addenda: When the jack is not connected to the PC the amp is completely silent, and there's no heating either. Anyone know what the heck is going on here ?

 

[Rattanee]

Further addenda: I read somewhere that the input ground and power ground should be separated. My PCB has a large ground plane, so it's not that easy to cut traces, but I managed to separate the ground pins of the amps and the power ground pin, and tied them together, thus leaving the rest to the PC ground. When it's powered on, plugged into the pc, I get an almost rythmic on-and off static. So that's a no go...

 

[Rattanee]

Okay I'm stumped. Another idea I found said to disconnect the ground from the PC chassis, which I did, and it got a -hint- less noisy, but it's still nowhere near listenable. (After all, the amp itself is powered from the PC ground as well, which I'm fairly certain is also the chassis ground. It also introduced rolling 'noise' on the video signals, which makes me certain of this.) It's still heating up really quickly even though it's not being driven any hard. Guess the last thing left is a separate powersupply, which I was hoping to avoid to keep costs to a minimum.

 

[Rattanee]

So I'm open to ideas on how to make this work... I tried battery-powering the amp, it works silently at first, but there's a lot of clipping noise at peaks in the music, and as it heats up gradually, a sizzling white noise begins to come on. There's something not quite right here...

 

[davenn]

PC power rails are "VERY DIRTY" .... full of all sorts of noise
if you really need to use them then you need to have some really good filtering
on the power rails right at the amplifier board.
If you have the amp board mounted inside the PC case, you may also find you need to put
it in a metal box for shielding to stop it picking up induced noise

have you heatsunk the chips ??
if not, that would be a really good place to start

Dave

 

[Rattanee]

Hey, I've actually moved it off of the PC power rail and battery powered it from my cordless drill's 12v battery pack. It's not mounted inside the pc case either.
Certain types of noise -have- vanished from battery powering it, but I can still hear hdd spinup and whatnot faintly, and white noise popping in and out seemingly randomly.
I've both TDA's bolted to an aluminum heatsink I stole from a dud PC powersupply. The heatsink gets so hot I can smell it. But the amps still operate, just not corectly.

I pulled the inputs off, and from a cold start, the amp is dead silent for a few minutes, then starts producing random burps of whitenoise and the occasional quiet high frequency fading in or out. Very strange noises to be honest. The heatsink is barely warm when this starts up.

 

[davenn]

I've both TDA's bolted to an aluminum heatsink I stole from a dud PC powersupply. The heatsink gets so hot I can smell it. But the amps still operate, just not corectly.

that's a major part of your problem that needs to be solved
a much, much biggger heatsink with LOTS of fins

sort out your heatsinking, then we can look at other noises

Dave

 

[Rattanee]

It's a fairly low powered amp, I wouldn't think it'd need such a huge heatsink? I have much bigger sinks as well but I felt they'd be overkill. Plus the sink isn't particularily warm when the amp starts acting up.

Anyways I'll bolt them onto a larger sink tomorrow and see if that helps.

TBH I'd think that without any input they shouldn't really be getting more then barely warm even without a sink?

 

[davenn]

The heatsink gets so hot I can smell it. But the amps still operate, just not corectly.

that's a major part of your problem that needs to be solved
a much, much biggger heatsink with LOTS of fins

what part of that dont you understand ?

 

[Rattanee]

The board with the heatsink in this pic is the actual board we're talking about: https://www.facebook.com/photo.php?fbid=527547070662147&l=74cc8efd9d

 

[Rattanee]

I'm suspecting the amps might be heating up due to something not right in the circuit. After all they're not more then hand-warm when they start farting out white-noise. Still worth a shot so we'll see.

 

[davenn]

I'm suspecting the amps might be heating up due to something not right in the circuit.

that was going to be one of my next suggestions

it is 10W per channel 20W total. total dissapation is likely to be in excess of 40W per chip
its going to need a serious heatsink


D

 

[Rattanee]

We'll see tomorrow. If all else fails I'll build both stages separately with an exact copy of the reference pcb design and start gluing them together on vero or somesuch.

 

[Rattanee]

I think I may have the problem down. Amplifier ground is disconnected from the PC's jack ground. However the bridged amp works different to the stereo amp, and would need that ground to maintain a clear signal. I'm going to try separating the ground on input 2 of the bridged amp, and linking that single one to the pc ground.

Also, how would I go about filtering the output of the PC switcher enough so that it could be used for powering the amp? I don't think putting a huge cap in there would solve the noise issues.

 

[Rattanee]

Okay I have no idea what's going on once more... I pulled out the first PCB I made with this amp, and it works like a charm even though it has all the grounds tied together. (though that is a stereo only amp, no second bridged amp on it) It also has a comparatively miniscule heatsink and doesn't heat up near as much... hrrmm....

 

[BobK]

Of course, all the grounds NEED to be connected together!

Bob

 

[Rattanee]

Of course, all the grounds NEED to be connected together!

Bob

Yeah, but I was getting a lot of noise, and someone in another thread suggested removing ground from the jack going to the PC. I have a feeling my problems might be stemming from trying to use both a bridged and a normally configured amp. Right now I'm making a board to build the two amps separately and test them as such, see what happens. Hopefully they'll work fine. If so, I'll have to figure out how to combine them without them interfering with each other.

 

[duke37]

The amps feed a Zobel network which will provide a heavy load on the amp at high frequencies. Does the amp get hot with no external load?

You have to stabilise the amp (no white noise) with appropriate capacitors near the amp. If you have a scope, check for high frequency oscillation.