TDA7293 Amplifier Design Blows Up Instantly

...driving my subwoofer...In the instant I turned on power, both chipamps literally exploded...
I have a few guesses:
Simplify mute:
Short R8, omit D1, omit C13
Short SW1
*one simple resistor makes the mute work quickly for favorable sequence

Simplify feedback:
Change R11 to 100k and R4 to 2.7k
Omit R5, C6, C4, C3
Short C5
Put 330u series to R4
*high gain makes a subwoofer amplifier with all-conditions stability

Boost the Bootstrap:
Change C7 to 100u
Omit slave's bootstrap cap
Jumper pin 12 of slave to pin 12 of master
Jumper pin 6 of slave to pin 6 of master
*force both chips bootstraps do the same thing at the same time so the outputs don't fight

Heatsink:
A heatsink is required before power up.
Thermal pads and sholder washers are required for voltage isolation.
*Chip amplifiers should actually be heatsinked before soldering them
** for test, power up without the slave chip in place

Other:
Fake chips explode because they're unstable at higher voltage and don't support modular parallel (not all 15 pins work), so there could be two unstable amplifiers in a death match fight. It makes an epic odor!
To prevent this, get the inexpensive TDA7293 parallel kit from ebay (those chips have all the pins working and connect right). TDA7293 Dual Parallel 170W BTL Mono Audio Power Amplifier AMP Board Assembled 699918445180 | eBay That has exactly the same chips as stocked by Mouser. I suggest to use the TDA7294S datasheet for power tolerances guidelines.

Other option:
Ebay boards for TDA8932's big brothers have omitted the input base stoppers, with harmonic distortion as the result (typical of Class D, attention to detail, except foobar the input); but, that problem isn't hearable from a subwoofer. Search Results: Some of those larger models can use the same transformer as your TDA7293 project. Same big bass dynamics, much less labor. If at higher volume, the efficiency is similar too; so, that also does need a heatsink.
 
Last edited:
Hi, I have the same problem. My TDA burns instantly. I have built the single IC version and I have the same problem. This circuit is as I always use with TDA7294, except by the connection of bootstrap capacitor. In the TDA7294 I use it between pins 14 and 6, while in TDA7293 I use it between 12 and 6.
Now I am thinking:
What about the working voltage of bootstrap capacitor in that configuration?
Should it be 2xVs?
The manufacturer does not mention that voltage.
I used TDA with +36V and -36V and a bootstrap capacitor of 50V. May be that's the mistake.
 
Last edited:
Hi everyone,

about 2 months ago I decided to take on the project of designing my own amplifier which is supposed to be driving my subwoofer in the future.

The circuit I chose to use is a dual TDA7293 parallel design.

I have attached my circuit diagram to this post. Please be aware, that I have already figured out and fixed the follow mistakes made:

- Capacitors C10 and C10P1 (1000u smoothing caps in negative supply rail) are polarized in the wrong direction, have turned them around on PCB
- Capacitor C3 polarized in wrong direction as well

I have also attached an image of my PCB layout.

The power supply is a 24-0-24V toroidal transformer with a rectifier board attached to it (which did not have any problems and has worked perfectly fine from the beginning). It has 2 63V 10.000uF caps per rail.

Now the problems started when I first turned on my amplifier board.

On the first try I had both TDA7293 chips soldered to the board. I have shorted the audio input signal to ground.

In the instant I turned on power, both chipamps literally exploded, meaning they blew off part of the black cover of the IC.

Ofter the past weeks I quadruple checked again for mistakes in my layout and then soldered on a new TDA chip (this time only the one in the master position, because I had heard of the possibility of instability in parallel mode).
But the same thing happened again, instantly after powering on, a flame shoots out the bottom of the IC and that's it...

My last hope are now the great people in this forum who might be able to help me figuring out, what the problem with my amplifier design is.

I'm looking forward to your responses and suggestions and please don't hesitate to ask for any more info required to assess the situation.


I think you have to check the feedback circuit. Please do one amp as explained in datasheet.
 
Hi, I have the same problem. My TDA burns instantly. I have built the single IC version and I have the same problem. This circuit is as I always use with TDA7294, except by the connection of bootstrap capacitor. In the TDA7294 I use it between pins 14 and 6, while in TDA7293 I use it between 12 and 6.
Now I am thinking: What about the working voltage of bootstrap capacitor in that configuration?
Should it be 2xVs? The manufacturer does not mention that voltage. I used TDA with +36V and -36V and a bootstrap capacitor of 50V. May be that's the mistake.
Bootstrap has only the voltage difference, less than 1 rail, inside the cap; however, double-check the polarity.

V- related explosions:
Screw terminals and other friction connectors, if used for power, cause a broken chip. Also, it is necessary to double-check with the ohmmeter between the chip's tab and heatsink--verify insulation works (no conductivity). At the power board, the diode/rectifier must not have screw terminals or other friction connectors. You need a soldering iron to secure these connections, for sure. Also, at the power board, put an LED on V- and an LED on V+, run at 8ma or less--these also provide clues as well as eventually drain the charge down to a couple of volts. Also, led's on = keep fingers away.

Overcurrent:
Initially check out the amp when an incandescent or halogen light bulb is in series with the mains plug. Have a clear bulb so that you can see the filament. An amp that is neither playing nor endeavoring to explode, results in a bulb filament almost too dim to see.

Authentic die-shrink, functional clone, counterfeit, label-swap, stability? The answer could be a 'safer start' schematic, and some notes:
The attached schematic is for the initial effort of getting the amp to work. The mute led's are optional OR can be a Zener for more effectiveness. In this example, there must NOT be any cap bigger than 270uF on the amp board (The larger caps go on the power board). The ringing lowest-bidder 100n caps that come with kits, should be replaced with ordinary ceramic caps (disc preferred due to lowest chances of ringing). For the feedback resistor, the resistor pins must contact chip pins, no vias, no traces. For the feedback-shunt resistor, that resistor must take a straight path between chip pin and coupling cap, no vias, no traces. Those two resistors fit under-board easily, and do use the values from this schematic (attached). The popular, tidy row resistor placement is decorous but anti-functional, so that needs corrected first.
 

Attachments

  • TDA7294safestart.gif
    TDA7294safestart.gif
    32.5 KB · Views: 635
Last edited:
Hi Marco,

The symptom you described has only two likely sources: A low impedance path between SPK+ and one of the power rails, or full-power oscillation. There are enough clearance oopsy-doodles in your board layout to cause the former. And enough long, convoluted, loping traces with ill-advised vias to cause the latter.

At the very least, the clearance issues should be addressed. Start with the ones marked with an orange squiggly in the pic below. Use an X-acto or utility knife, a 10x or 15x magnifier, and your ohmmeter to verify.

Then consider cutting and replacing some of the traces marked with dashed purple, with a direct connection using an insulated jumper of 'hook up' wire. It may be possible to get it working without re-laying out the whole board.

If a new layout IS under consideration, I'd strongly recommend a few not-so-minor changes. Move the output terminals to between the chip-amps, and remove all the Mute/Standby circuitry OUT from in between them. Since such circuits have no performance, stability, or high-current considerations, they should never be allowed to increase the spacing between components that DO.

All grounds are NOT created equal! Try to avoid 'daisy chaining', where a row of supply bypass capacitors have an input, output, or feedback ground picked off either somewhere along the way, or from one end or the other. Also, R6/C15 needs to go directly and with a short path to the high frequency supply bypass caps.

As much fun and satisfaction as it can be to design your own board, it can also be far more frustrating and time consuming than the value of the hard-won lessons learned. I really think it would be worth the effort to attempt the board design in ST's spec PDF.

Good luck and all the best,
Rick
 

Attachments

  • marco1601; TDA7293 parallel amp layout-bits.jpg
    marco1601; TDA7293 parallel amp layout-bits.jpg
    821.5 KB · Views: 556
Hi everyone,

about 2 months ago I decided to take on the project of designing my own amplifier which is supposed to be driving my subwoofer in the future.

The circuit I chose to use is a dual TDA7293 parallel design.

I have attached my circuit diagram to this post. Please be aware, that I have already figured out and fixed the follow mistakes made:

- Capacitors C10 and C10P1 (1000u smoothing caps in negative supply rail) are polarized in the wrong direction, have turned them around on PCB
- Capacitor C3 polarized in wrong direction as well

I have also attached an image of my PCB layout.

The power supply is a 24-0-24V toroidal transformer with a rectifier board attached to it (which did not have any problems and has worked perfectly fine from the beginning). It has 2 63V 10.000uF caps per rail.

Now the problems started when I first turned on my amplifier board.

On the first try I had both TDA7293 chips soldered to the board. I have shorted the audio input signal to ground.

In the instant I turned on power, both chipamps literally exploded, meaning they blew off part of the black cover of the IC.

Ofter the past weeks I quadruple checked again for mistakes in my layout and then soldered on a new TDA chip (this time only the one in the master position, because I had heard of the possibility of instability in parallel mode).
But the same thing happened again, instantly after powering on, a flame shoots out the bottom of the IC and that's it...

My last hope are now the great people in this forum who might be able to help me figuring out, what the problem with my amplifier design is.

I'm looking forward to your responses and suggestions and please don't hesitate to ask for any more info required to assess the situation.

TDA7293 blowing up is related to MUTE and STBY pins.
Remove the switch and connect the circuit directly to +VCC and your problem will be solved, if there are no errors.
The IC DOES NOT like the STBY pin left OPEN or connected in any other manner.

The TDA7293 works perfectly in PARALLEL bridge to deliver around 320W/8E while not needing any current sharing resistors on its outputs, just as mentioned in the datasheet.

bimbla.
 
TDA7293 blowing up is related to MUTE and STBY pins.
Remove the switch and connect the circuit directly to +VCC and your problem will be solved, if there are no errors. The IC DOES NOT like the STBY pin left OPEN or connected in any other manner.
Datasheet application of standby works perfectly. However, the datasheet application for the mute circuit is an awkward thing that I avoid. The datasheet treats the mute as large signal, which is wrong because it IS a small signal input. And, a bit of wrong at the input causes decreased stability.

I honestly think that the majority of explosions involve sprawling layout to hinder stability (if the gain is too low) and screw terminals glitching the v- while the amp is running. However, in such conditions, an awkward mute circuit making stability slightly worse, could hasten explosion in a very small way.

Anyhow, if you've got a single-layer board, screw terminals and 20-ish gain (mostly bad case), there's nothing you can do to either the mute or standby to prevent explosion.

However, if you've got a compact dual-layer board, securely soldered connections and 30-ish gain (mostly good case), then you might consider streamlining that mute circuit (being mindful that it is small signal).
 
Last edited: