Lm4780 problem, 3.5V on output when power on...

[jaste]

Hi!
Just finished my stereo gainclone using two lm4780 chips each running in parallell mode. Before powerup everything semms nice and there is no short circuit and the powersuply works fine delivering 36.7 volts unloaded.
But when I turn the power on I read 0,1 Volts across one channel (out gnd --- out), a bit high but no problem? The other channel reads 3.7 Volts and the chip gets really hot!
Does anybody have an idea of where I sholud start looking? Will try to post some pictures later on...

 

[jaste]

Pics

Here's some pictures of the setup, i forgot to mention that it's the setup from audiosector running so schematics can be found at www.audiosector.com

Link to pictures: http://www.flickr.com/photos/23725950@N08/

 

[Leolabs]

check and recheck the component values.

 

[ttan98]

my experience, mine was working fine ie o/p=26mV then after using for a little while o/p goes to 1v, why,

fault: the i/p wiring from the pot to i/p was broken.

I suggest

1. check the input is tie low ie low resistance to gnd.
2. check the i/p to gnd resistor is properly soldered in.

cheers.

you are using the chipamp as parallel pair, make sure the gains on both amps are very close to each other, otherwise you have problems down the track.

cheers.

btw, 0.1 o/p is too high on the other channel as well.

 

[jaste]

Thanks for the replys!

But;

"1. check the input is tie low ie low resistance to gnd."

The ristance to gnd from input can't be lower then the i/p to gnd resistor, right?

 

[jackinnj]

Measure the DC Component with the inputs connected to ground -- if there is no "offset" the problem is with your source.

 

[AndrewT]

Hi,
if you want to check the amp for output offset, then you must eliminate any possibility of input offset confusing the measurement.
Jack is correct.
Fit a shorting plug to the input when checking for output offset.

 

[jackinnj]

One other thing. I don't know if Peter has "optimized" this design for parallel, but I think that it's a bit "hot" with gain of 33. The folks at National comb over values of Rf and Rg for optimal performance and use a gain of something like 20 or 21.

 

[jaste]

Thank you all again for your replys!

I just measured output offset when shorting the input and voila 23 mV offset on both channels, sweet!

Now I just have to figure out what´s wrong with the input section, any hints?

 

[jaste]

Wait a second, I havent tried it with a source yet.
All my measurments have been without any source. To me it sounds strange that the chip would get very hot when the power is on but there's no input but maybe that's the way it is and my amp is actually working just fine? If it's so I shouldnt have my amp on when there's no source?

 

[jackinnj]

Wait a second, I havent tried it with a source yet.
All my measurments have been without any source. To me it sounds strange that the chip would get very hot when the power is on but there's no input but maybe that's the way it is and my amp is actually working just fine? If it's so I shouldnt have my amp on when there's no source?

I'm gonna bet that without any input source your LM4780 is oscillating. It would be helpful if you had a scope. You might be picking up some RFI, or oscillating of its own volition.

FWIW, I have been able to get the LM4780 to oscillate as high as 1.5 MegaHertz -- and it gets very unhappy.

 

[Peter Daniel]

LM4780 chips produce more offset than LM3875s

It is highly recommended to always have the source connected when amp is on, if no source is connected it's best to short the input to ground.

With source connected, the DC offset us usually much lower too, as the input impedance is in parallel with source's output impedance.