TDA 7294/3 Debugging

[pompouspompei]

Hello ,


I just finished an amp with a tda2050 , and now im moving forward with a tda 7293. Couldnt really get it to work , and i am at my wits end.


I am building the circuit on a solderable breadboard(General purpose PCB ? the name varies i gues.)



I just have a multimeter to debug. Dont really have access to an oscilloscope now.




My power supply : a 24-0-24 5A transformer buffered by 2 x 10000uF 63V caps and 2x 10uF 63V caps.


My speaker : a 80W(?) 8ohm speaker ( its around 8-9 inches in dia).


I have burnt (shorted ) 3 TDA7293s till date using the default schematic as per the datasheet with a few changes :
https://www.st.com/resource/en/datasheet/tda7293.pdf



1. a 220uF aluminium cap with a 2.7k and a 56k in the feedback network
2. a 4.7uF aluminium cap with a 0.1uF ceramic in parallel to it in the non inverting terminal.
3.a 47uF bootstrap cap.
4. 1000 uf Bypass caps along with other bypass caps like 2.2uf and 0.1uF (ceramic)

3. I have used a heatsink salvaged from a computer PSU and is isolated from the tab of the IC.
4.Mute and stdby pulled to the Vcc via the resistor and capacitor as given in the reference . ( i also tried just using the resistors once)



My input :
1. Line input from my phones jack (used the left channel) . applied a sinusoidal signal of 0.5 - 1.2 V RMS and 400Hz freq.



My outputs :
1. none.


2. A static hum from my speaker which appeared to be proportional to the input voltage regardless of the input signal ( music /sine). Turns off after a couple of seconds after i turn off the power input. Voltage measured was varying from 3 - 10 Vrms with a constant sinusoidal input of ~1V and a gain of approx 22. Didnot run it for more than a few seconds



3. added a 47uF cap to to the output , ended up blowing the chip ( no output from the speaker- maybe a high pass filter was formed?)


4.Measured it with no load , voltage varied from 3 - 8 Vrms , chip burnt after a couple of seconds.


Can anyone help me out? I am at my wits end now.


PS: The chips might be chinese .


If any information is required , feel free to ask.


Thanks!

 

[Mooly]

I would suggest using 'bulb tester' in the primary of the transformer to limit current in the event of a fault. It will save destroying parts.

As a normally functioning amp like this draws minimal current I would suggest no more than a 40 watt bulb, less if you have one.

Layout is critical with high gain chips like these, particularly how you return such things as the feedback return, the input ground and the speaker back to the zero volt line. Make sure the speaker return can not influence the other ground returns. In other words run the speaker directly back to the power supply.

The application notes are not always fully developed circuits. It might be worth adding a Zobel network to the output. Again return this to the main PSU ground.

Always check that the speaker output is at zero volts before attempting to connect a speaker.

 

[pompouspompei]

Hello,


Although I am using 4A fuses on either rails , I will try out the bulb too.


Regarding the layout , ive optimized it as much as I can ( smaller loops and whatnots) . I guess , at this point , the only way I can improve further is go for a proper PCB with proper layout and ground planes.


I used a Zobel network , but the amp still got shot. the fuses havnt blown once.

Always check that the speaker output is at zero volts before attempting to connect a speaker.

Will do that.



Btw , As ive mentioned before , i tested the Amp without a speaker at the output , but it still blew in a few seconds. I dont think that this is normal behaviour right? Is it possible for you to deduce something from this? Runaway oscillations due to poor layout?


Thank you!

 

[Mooly]

4A fuses would allow the chip to be instantly destroyed in the event of a fault. Use the bulb which puts a limit on current draw. The more current the circuit draws, the more bulb lights and the more it lights the higher its resistance becomes.

Blowing the chip with no load almost certainly points to instability. Definitely add the Zobel network to the output before trying again. A 10 ohm and series 0.1uF cap would be suitable.

 

[Mark Whitney]

Can you post some photos?

 

[pompouspompei]

I use an 8ohm and 0.1uf . i will try again under no load with a zobel and get back to you. Usually the Vcc and the output gets shorted. So i guess more or less its instability



I will get back to you as soon as i test it.


Thanks!

 

[pompouspompei]

Sure! What would you like? The actual ckt or the diagram?

 

[Mark Whitney]

Photo's of whatever you have built. It doesn't take much to blow these TDA chips. If you want to try something different the LM3886 is far more robust.

 

[nigelwright7557]

I found the tda7294 is very touchy about the feedback resistor loop.
It has to be very short or chip oscillates.
Do not run without a heat sink.
Do not alter gain of tda7294 it needs to be a minimum of 22 or it becomes unstable.

Might be worth uploading your schematic in case something is wrong/missing.

+/-35VDC is max power supply volts.

 

[pompouspompei]

Here is an image of the board



TDA7293 - Album on Imgur


The following is the ckt which im using now more or less. Any idea if adding a cap across the feedback resistor to make a low pass filter would help?

Im using a TDA7293 on 33V rails(24-0-24 ,5A) with a gain of 21.7


Ill resolder a new chip and get back to you guys


Thanks for all the help!

 

[NanoFarad]

For tda7294 i suggest 30db min gain.
You should use a good quality input cap, 47uf min as bootstrap & 100uf as feedback. Probably you'll get much better low end response if you use a bass booster circuit at input(my experience),,,& DON'T use fuse at supply rails!
Good luck.