Hi!
I have recently bought 2x TDA2050 (plus some caps, and a transformer).
I wanted to connect vs+ and vs- to their corresponding pins on the IC, but the ic just blows up after connecting the psu to it.
Im using a 18-0-18 volt 5amp transformer.
I put a bridge rectifier between 18v outputs of the transformer, after that a 4700uf 35v cap between every rail and ground (central tap).
I was getting around 25.5v after rectification so i used extra diodes to lower the voltage a bit and now im getting around 24v after rectifying.
Just for extra security i put a 22k resistor between the cap - side and its corresponding rail, so i was getting around 17,5v bc the caps weren't fully charging.
After that i connected Vs+ to pin 5 and Vs- to pin 3. When i plugged it in a lot of smoke came out of the chip just before it cracked.
Am i doing something wrong? (I know the chips are pretty fake bc they're not in production for couple years but idk).
Any idea?
Thanks guys!
In my run for my first DIY amp i have 2 fried TDA2050 and a fried cellphone :c
I have recently bought 2x TDA2050 (plus some caps, and a transformer).
I wanted to connect vs+ and vs- to their corresponding pins on the IC, but the ic just blows up after connecting the psu to it.
Im using a 18-0-18 volt 5amp transformer.
I put a bridge rectifier between 18v outputs of the transformer, after that a 4700uf 35v cap between every rail and ground (central tap).
I was getting around 25.5v after rectification so i used extra diodes to lower the voltage a bit and now im getting around 24v after rectifying.
Just for extra security i put a 22k resistor between the cap - side and its corresponding rail, so i was getting around 17,5v bc the caps weren't fully charging.
After that i connected Vs+ to pin 5 and Vs- to pin 3. When i plugged it in a lot of smoke came out of the chip just before it cracked.
Am i doing something wrong? (I know the chips are pretty fake bc they're not in production for couple years but idk).
Any idea?
Thanks guys!
In my run for my first DIY amp i have 2 fried TDA2050 and a fried cellphone :c
It is not really clear how you managed to reduce some 24V to 17.5V with a 22KOhm resistor? Or, how your cellphone got fried from all this?
Anyway, a "real" TDA2050 should stand +/-25V so the voltages you mention should not cause the smoke. Unless, your admittedly not genuine TDA2050 is in reality a re-marked TDA2030 that stands only +/-18V (max.).
If you have a variable power supply, it is always clever to start with a lower supply voltage such that you know if the amplifier at least worked with a lower supply voltage. If you only have this fixed voltage supply, you have to take the chance but if it goes wrong, it is difficult afterward to say what happened. Other overload situations can cause the chip to burn such as a DC error at the output, excessive loading or self-oscillation.
You may try with LM1875 in the future as it is very similar to TDA2050. It can be bought at a very reasonable price.
For a class AB amplifier, it is often clever to do the very first test as an idle mode test (preferably at a lower supply voltage) without any input signal and a rather high impedance load at the output. If the amplifier has a DC error at the output, the flaw can be found without a huge current though a low-impedance load.
Only when all measurements indicate a healthy construction, you put a real low-impedance speaker at the output and later an input signal.
Last edited:
Sadly i dont have a regulable psu nor oscilloscope :c
I reduced the voltage just stacking a bunch of diodes in series, and the resistors were just to prevent the caps from charging too much (So they filter just a bit).
I though i was the one making some error. I'll buy a LM chip next time.
/// My fried phone was because of MY error, this will sound pretty stupid. I was without internet but had a screenshot of a tda2050 schematic, i tried to figure out which pin number was which pin in the IC and connected my input to pin 4. Just fried it :C
Thanks for the answer!
I reduced the voltage just stacking a bunch of diodes in series, and the resistors were just to prevent the caps from charging too much (So they filter just a bit).
I though i was the one making some error. I'll buy a LM chip next time.
/// My fried phone was because of MY error, this will sound pretty stupid. I was without internet but had a screenshot of a tda2050 schematic, i tried to figure out which pin number was which pin in the IC and connected my input to pin 4. Just fried it :C
Thanks for the answer!
An externally hosted image should be here but it was not working when we last tested it.
That's a pretty basic schematic. I forget the resistors at the caps but that will do.
When i turned on the PSU the poor TDA just smoked and cracked. Tbh it didnt have a good smell
Nigel, thanks for the tip! I didnt know zener diodes could be used for voltage regulation (Im pretty new to electronics sadly).
Thanks to everyone
I believe this chip "MUST" be on a heatsink before full power is applied. The no-signal idle heat is several Watts, and the small chip can only handle a Watt+ without a heatsink.
We almost always connect the several resistors around an amplifier before we apply power, so it "knows" to settle its DC voltages to happy values (about zero on all non-power pins). I do not know what happens if you don't.
I am not clear how you smoked the power chip, with those tiny 1N4148 diodes "dropping". The chip can pass many Amperes. The 1N4148 is rated for 100mA (0.1 Ampere). The power chip's idle current (with normal bias) may be just-about the 1N4148's maximum rating.
We almost always connect the several resistors around an amplifier before we apply power, so it "knows" to settle its DC voltages to happy values (about zero on all non-power pins). I do not know what happens if you don't.
I am not clear how you smoked the power chip, with those tiny 1N4148 diodes "dropping". The chip can pass many Amperes. The 1N4148 is rated for 100mA (0.1 Ampere). The power chip's idle current (with normal bias) may be just-about the 1N4148's maximum rating.
Oh sorry PRR, i forgot to put the real diodes in there, that was the default model of the schematic software sorry!.
But it is pretty strange, it literally blowed up as soon as i connected it, literally. It was instantaneous. Next time i'll put some resistors and use a zener diode to regulate the supply voltage c:
I'll test with the LM chip next time :_
But it is pretty strange, it literally blowed up as soon as i connected it, literally. It was instantaneous. Next time i'll put some resistors and use a zener diode to regulate the supply voltage c:
I'll test with the LM chip next time :_
The first one was used in this schematic http://diyaudioprojects.com/Chip/DIY-TDA2050-Hi-Fi-Chip-Amplifier/TDA2050-Typical-Application-Circuit-Schematic.png
It also started smoking as soon as i connected the power
It also started smoking as soon as i connected the power
Somewhat unrelated, but I had a strange experience with the TDA2040-based amp that I refurbed (re-cap and resistors change; from standard carb to low-cost yageo MF, with a slight change in NFB; from 56k/2k7/22uF combo to 47k/2k2/47uF - which shouldn't be problematic at all). The amp worked "just fine" before the refurb; but after the reworks it started oscillating wildly.
I test-powered it with my simple 9V analog brick and the amp module got hot quick; I measured slightly over 3V AC of HF oscillation - but nevertheless it played back audio when I plugged in some music source ... I guess I somehow made a Class-D TDA2040 amp, ho-hum.
Nothing that I tried could stop this oscillation - except the last measure that I took; I plugged a 1nF ceramic between +/- inputs of the TDA chip and that stopped the oscillation - finally I settled for a 220pF similar to what the appnotes for LM3886 prescribes.
AFAIK this "shouldn't work" for TDA chips; but in this case, it solved the problem.
Perhaps you could try something similar if your TDA's show self-destructive behavior.
I test-powered it with my simple 9V analog brick and the amp module got hot quick; I measured slightly over 3V AC of HF oscillation - but nevertheless it played back audio when I plugged in some music source ... I guess I somehow made a Class-D TDA2040 amp, ho-hum.
Nothing that I tried could stop this oscillation - except the last measure that I took; I plugged a 1nF ceramic between +/- inputs of the TDA chip and that stopped the oscillation - finally I settled for a 220pF similar to what the appnotes for LM3886 prescribes.
AFAIK this "shouldn't work" for TDA chips; but in this case, it solved the problem.
Perhaps you could try something similar if your TDA's show self-destructive behavior.
I had a similar problem with a tda7294.
Massive oscillation on the output and got hot quickly.
I too put a capacitor across the +/- inputs to solve the problem.
I read the datasheet again to find that you musnt have a gain of less than 22 or the tda7294 becomes unstable.
I had reduced the gain because I had a valve front end and this is what caused the problem.
The circuit I had here as it was before had a gain of 21,74 and seemed "ok"; the new values I put in were supposed to be in the safe territory with gain of 22,36. These TDAs surely were temperamental.
And I put "fair quality" components in this particular circuit; Yageo resistors; CrapXon GL low ESR and Samwha RD lytics; AVX mlcc ceramics and Kemet polys. Which should be a fair step up from what this amp came with. The TDA's were ST, vintage retro originals. What I did though was to replace a crappy disc ceramic in the Zobel with a same-value poly - but this really shouldn't under no condition make the amp oscillate as it did.
And I put "fair quality" components in this particular circuit; Yageo resistors; CrapXon GL low ESR and Samwha RD lytics; AVX mlcc ceramics and Kemet polys. Which should be a fair step up from what this amp came with. The TDA's were ST, vintage retro originals. What I did though was to replace a crappy disc ceramic in the Zobel with a same-value poly - but this really shouldn't under no condition make the amp oscillate as it did.
That's a pretty basic schematic. I forget the resistors at the caps but that will do.
What do you mean by 2x 1N4148? 2 in series? That lowers the voltage by 2 Volt, the diodes can't take the current of the amp though. 2 parallel does not work either because of the tolerance, one will switch sooner than the other, leaving the current to one diode. That means, while diodes will drop the voltage, that's not a working solution in your application.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.