TDA2030... crack, poof, help!

[mdh]

I built a simple amplifier to drive a talkbox, using a tda2030 that I pulled out of a computer speaker. This afternoon, I managed to coax some blue-white smoke out of it by inadvertently shorting the output to ground. At least, that's what I think happened. I used is the single supply circuit from the data sheet. Here's the schematic:



I made a couple of substitutions due to availability of components in my junk box: I used a 2200 uF cap in place of the 2000 uF; 165k resistor in place of 150k; a 100k pot in place of the 22k pot; and 5 ohm resistor + 100 nF cap in place of the 1 ohm + 220 nF.  The speaker in the circuit was an 8 ohm 80 W compression driver.

Power supply was a wall wart with secondary rated 24 V at 1.25 A, into a bridge rectifier followed by a 2200 uF 35V cap. Supply voltage came out at about 34 V DC.

And now for the mea culpa... the short occurred when I stupidly unplugged the cable connecting the amplifier output to the talkbox while the amp was on. Crack, poof, acrid smell. I had them connected with a guitar cable, and I suspect that the plug shorted output to ground when I pulled it out.

So, this raises a few questions:

1. Why didn't the tda2030's output short circuit protection prevent this cataclysmic event?

2. What kind of cable should I use to connect the amp output and the speaker? Presumably, something that won't allow output and ground to short, right? XLR, perhaps? What is typical in this application?

3. Is there a simple way that I can build short circuit protection into the output of the amplifier to prevent this from happening in the future?

[Threefish]

Since no-one else has spoken up with any real help yet.....
- did you actually get anything out of it before you sent the poor little bugger to sleep with the fishies? You don't mention how the circuit itself worked.
I had a look at the data sheet and the absolute max V rating is +-18v. I would have thought that 34V would have started fireworks immediately, though a +-18V swing translates to 36V. This doesn't appear to be running on a bipolar supply though. I only see a +V connection. If it held itself together until the unplugging, this probably doesn't apply.
What's your grounding arrangement? I notice that the chip tab (heatsinked??) is connected to pin 3. If you are running it on a centre-tap bipolar supply, then you probably need to isolate the heatsink from everything else, or the chip from the heatsink, otherwise you'll get different voltages at "ground".
As an aside, I know it refers to it as a hifi component in the data sheet, but this looks like one of those chips used in car stereos (hence the mention of 14V and +-14V) that draws heaps of current, and will oscillate or motorboat in the absence of it. Is your power supply up to it?

I'm guessing that something you've done in the setup of this circuit has side-stepped the protection circuit built into the chip. I would hope that something as simple as unplugging it would not cause it to cook itself though.

[Wimpy]

Maybe teh supply voltage was near max and not quite enough heatsinkin did made the blue smoke come out, the datasheet says it is just current limiting 3,5 A and temperature limit - check that output cap too if it is okay

Maybe you could put a fuse in line with pwr supply

Best connectors are Neutrik Speakon, cannot think any more bulletproof than that but 6,3 mm jack is what guitar amps use so it should be ok, I have used two terminal connectors that they use for home/car speakers

[Paul Perry (Frostwave)]

I'm very suprised that a momentary output short would kill it (with the cap in series as well!) but, as Wimpy said, you are pretty close to the maximum supply rating voltage. Maybe disconnecting the load caused the supply voltage to rise enough to kill it. A sudden crack sounds more like a catastrophic breakdown, rather than an overload burnout.

[RDV]

I've blown the fuse/tripped breakers on many amps in the past using guitar cables rather than speaker cables. I think there's an impedance difference there.

Just a thought.

RDV

[Joecool85]

I'm pretty sure guitar cables can't carry more than a few watts safely...but I could be wrong.

[Nasse]

I can say guitar cables can carry more than few watts but you are right, not safely. I remember when I was a kid and we started a band the bassplayer had some "know how" how to use the cables.  His "know how"  was to use the most unreliable one for the speaker connection (the ones that crackled when moved or had some hum problems) and save the better one for instrument use. Never any exploded output transformers (just good luck) for us but I don´t wonder why we blew some EL34 tubes... Used to play Marshall 50 watt amp head at huge level, It was late 60´s model with small box and no master volume, loudest and cleanest 50 watts I have heard

[R.G.]

1. Why didn't the tda2030's output short circuit protection prevent this cataclysmic event?

I think that it was not a short circuit that got it. I suspect that it was ultrasonic oscillation. Guitar cord has higher capacitance than parallel cords, and that's enough to push some amps over the edge... especially if the amp does not have a series inductor on the output.

2. What kind of cable should I use to connect the amp output and the speaker? Presumably, something that won't allow output and ground to short, right? XLR, perhaps?

Cheapest is 18 Gauge lamp cord. There are others, but whatever you use, you want either low capacitance, or the series inductance to isolate the cable.

What is typical in this application?

18g lamp cord.

3. Is there a simple way that I can build short circuit protection into the output of the amplifier to prevent this from happening in the future?

No and yes. No because it's probably not a short that got it, and yes, use the series inductor you'd make by winding one or two layers of 18G magnet wire over the body of a 10 ohm 1W or 2W resistor.

[mdh]

Wow, thanks for all of the responses. The chip does have a heatsink, and I believe it's sufficient; it got warm when cranked through a cheap 12 ohm 12 W (or so) speaker, but definitely not too hot. I played it quite effectively through that speaker and the talkbox before the unplugging catastrophe. It is a single (not bipolar) power supply. The wall wart being used to power the thing is actually the one that belonged to the speakers that the chip came out of (though that was a stereo application, and I didn't look to see if the original circuit used a voltage divider or something; it was unregulated, though, I know that), which is why I wasn't worried about the voltage.

I get the sense that you guys think this wouldn't have happened if I was using a 1/4-1/4 speaker cable instead of an instrument cable? Pardon my ignorance, but is that what is typically used to connect a head to a cab?

RG, thanks for the series inductor recommendation; is this just a (not very low) low pass filter to prevent the ultrasonic oscillation?

Again, many thanks for the responses. This is a great forum.

[RDV]

I get the sense that you guys think this wouldn't have happened if I was using a 1/4-1/4 speaker cable instead of an instrument cable? Pardon my ignorance, but is that what is typically used to connect a head to a cab?

a 1/4-1/4 speaker cable

RDV

[Paul Perry (Frostwave)]

I had momentarily considered high frequency oscillations..... that's the reason I hate to do anything with chip amps as a class..... but i must say, anything that is so unstable that the cap of the output cable can destroy it, isn't properly designed and/or laid out anyway.
I'm still not clear at all whether the failure occured because the output was momentarily shorted, or was permanently opened, or whether it was a voltage issue (though, if the transformer used was the one from the original gear, I guess not).
Also, did the amp work at all before it blew up??? That's the $1.50 question (price of a new TDA2030 from Mouser, or $15 if you want NTE printed on it).

[mdh]

I had momentarily considered high frequency oscillations..... that's the reason I hate to do anything with chip amps as a class..... but i must say, anything that is so unstable that the cap of the output cable can destroy it, isn't properly designed and/or laid out anyway.

Fair enough. I looked at the datasheet, and the only layout considerations in there read: The layout shown in Fig. 16 should be adopted by the designers. If different layouts are used, the ground points of input 1 and input 2 must be well decoupled from the ground return of the output in which a high current flows.

How should this decoupling be accomplished? Could anyone recommend threads or other resources that discuss layout considerations for chip amp circuits?

Also, did the amp work at all before it blew up???

Yes, it worked very well before the unplugging incident.

[Paul Perry (Frostwave)]

I guess, just run the ground points and the speaker ground via separate (fat) wires to the power supply. I must confess, I never persisted with these single chip amps. But there are literally millions out there, in every radio, TV etc.
The other thing (and I'm not being a smartarse here, at least not deliberately) is, just never disconnect the speaker!! There are a lot of things that only work properly when  loaded... think about it, there's a feedback loop, and suddenly there is no load & it is a suprise (at all frequencies!!) and BANG you have a rf oscillator.

[RDV]

Could anyone recommend threads or other resources that discuss layout considerations for chip amp circuits?

Just put chipamp in the search here or go to www.diyaudio.com.

Have Fun!

RDV

[Sir H C]

With the +-18V as the "absolute maximum" you can expect smoke if you are up there.  You should usually derate from absolute max to what you use by about 33% reduction.  In general the 36V number would suggest that the process limits are set there, and with a load you can have momentary swings beyond the rails when you short out.  I had this fun problem with a customer who tried running a part at the absolute max and not the recommended maximums.  They had a 10% fall out, and so we had to retest, take the yield loss and still tell them to redesign for the lower voltages.  Running near the absolutes is never recommended.

[mdh]

Good point, Sir H C. I will probably redesign, possibly going with the bipolar schematic and a +-9 or +-12 supply. I have a 25.2 V ct transformer sitting around, and a pair of 7809/7909 or 7x12's as well. Although, on top of that, I think I'm going to make the talkbox all-in-one, so that I don't risk unplugging the driver while the amp is on.

Thanks again for all the tips, everyone!