LM1875 burned. Need suggestion for fix.

Hello friends.. This is my first post. ( This was not supposed to be it, but bad luck, i screwed up my entire project just minutes before i finish it:( )

The project was a 5.1 LM1875 home theater amp. And everything was working fine, and i was just going to close the cabin because it all went ok, and everything was perfect.
But my bad luck, i got careless because i was so excited. And burned whole 7 ic's.

My setup was build around a 5.1 tda2030 circuit board , which i obtained from a china made home theatre system, and i replaced the chips with original LM1875 from national, added extra heatsinks, boosted the supply with 18-0-18 torroids and 10000uf capacitors, 6amp diodes for bridge... etc. And everything was working perfectly. Connected it to my yamaha speakers and the music was like crystals.

I had 5 lm1875 for each channel, and 2×2050 for subwoofer which totalled 7 ic's. But i didint used subwoofer amplifier section because i have a powered yamaha sub. So i kept it for future, but i did check them, and it was working.

I had the circuit carefully analyzed, the gain was about 65 ( because there was no powerful pre amp, it got a single transister pre / channel, which gave a gain of 45 for 2030 ic, but about 65 for lm1857. So i kept the circuit as it is, and had no problems.

Then when i was about to finish the thing, i hooked up the board to wrong power line. My torroid have 40-0-40 lines and 18-0-18 lines, and a seperate 10.5-0-10.5 line for the digital ic's for volume/input/Bluetooth/usb ssections . ( the amp's power section is totally isolated from the digital section of the board, and hence it is perfectly working even though i fried the amp).

I hooked the amp's power section to 40-0-40 lines of torroid accidentally, and it fried all 7 ic's:gnasher: .

Now i replaced all 5 ic's... also removed 2×2050( removed because i am not going to use it any way, and i want it run lil bit cooler, so why add extra heat and draw some amperes for nothing).

This time i got duplicate lm1875 ( marked as d1875 , with no national logo) just because originals are not avilable in the market. Now when i power it, i get hum, like oscillations, may be 120hz or so and ic's getting hot even with out inputs.
The hum and heat problem happens even with out input.

So what might be the problem... ? I tried changing the power filter capacitors ( which had 25v ranges, but kept 65-80 v ones)
Also the resistors and ceramics are smd types, so didnt changed those.. only thing now i have to change is 22uf/25v capacitors used in 1875 feedback line ( the one which is connected to ground from leg 2, via resistor. ).

So what is your opinion, should i change those 22uf too?? Or what are the components that usually get damaged when high voltage is applied to 1875??

Now its only oscillations.. input have no effects at all. Nor touching the input legg with signal generator have any effects.. all 5 channels have same problem.

Please help me to fix this.. i really want that beauty back.
TDA2050/2030/LM1875 will die instantly if more than a total of 36 volts gets to them.
Replace all IC's and build a power supply that delivers around 32Volts. That will give a bit of room for error.
If you are using a dual rail, which is best for noise from the supply lines, ensure no more than +-16Volts. Regulation is not important.
A steep and expensive learning curve! Buy replacements from a proper supplier like Mouser, RS Components, Farnell etc.
Ya.. i know.. power supply does matters.. and i have keep it clean but unregulated. But this was an accident. Hooking up wrong wires.

But thats not the point. I want to know what is that oscilations. And which faulty component makes it.. Any way i got my 22ufs. Now in 50v range. I am gonna replace it and see again. Will report back if i get it fixed .


Paid Member
2009-02-11 12:58 am
First off, replace the fake LM1875s. The real ones are available directly from TI, as well as various distributors (Mouser, Digikey, Newark, Farnell, etc.)

It is possible that you fried some power supply caps as well. I'd removed the fake LM1875s and measure the supply voltage. Measure both DC and AC. With no load on the supply, you should have very little AC on the power supply to the chip amp. If you are seeing, say, >100 mV of AC voltage (ripple) one or more of the supply caps has probably been damaged by the over-voltage condition and needs to be replaced.
If, on the other hand, you're seeing a DC voltage below the ABS MAX spec listed in the LM1875 data sheet and low AC ripple on the supply lines, solder in some new, genuine LM1875 ICs.

Don't forget to measure the output DC offset before hooking your speakers to the amp for the first time. So without a speaker connected and the power on, measure the DC voltage on the output of the amp. This should be no higher than 10 mV.

Ya.. i found it.. its those stupid china copy ic's.. i doubted the poor passive components.. seems like the rejection ratio, and the gain of the original circuit with original ic's dont play well with the china ic's...
This should be a lesson for all.. it took me 2 days to figure it out.. never buy a chineese copy of any of these quality components..
Will soon fix things up. Orderd original ic's. And also got hold of 10000uf elna japan capacitors.... whewwww... excited again...