DC offset in LM12 circuit

[cm961]

Hey guys,

I've just started a new project using an amplifier that I bought off ebay. I gutted it and am replacing the amp boards and adding filters to do a bi-amp kinda thing for my home made studio monitors. Anyways, the heatsinks already have 4 LM12s on them, so I plan on using them. I know sound quality is going to be harsh (opinions?) but it'll be a lot better than the amp I'm using now, and also the crossovers in my speakers will no longer be needed and they are terrible. So overall it should be a lot better. Now that my life story is finished, here's my question.

I'm new to power op amps, and my circuits (the basic one from the LM12 data sheet) have a large DC offset. The only thing I'm doing different is the gain, R2 = 30k instead of 3.3k. Is it true that power op amps provide a large DC offset when they have a high gain? If so, then I can lower the gain, but will need a stage in front of it to bring it up, perhaps with a NE5532?

Any ideas greatly appreciated!

Pete
www.peterbinsted.com

[ThingyNess]

You can add a large-value electrolytic capacitor in the bottom of the feedback network (between the last resistor which would normally be connected to ground and ground) to cause the amplifier's gain to roll off at very low frequencies, and go all the way down to unity at DC.

That way, instead of multiplying the input offset voltage by the closed-loop gain, you only get the raw input offset voltage appearing on the output.

Doing this is pretty much standard on any DC-coupled voltage-feedback audio amplifier I've seen, but it's an easy thing to overlook (especially if you're just going from the datasheet for a part like an LM12 which isn't designed specifically for audio)

The LM12 is spec'd for an input offset voltage of only 15mV or so, so if you roll off the closed-loop gain to 1 as noted above, you should be fine (assuming you're using a non-inverting configuration.)

Look at the datasheet for the LM3886, another power op-amp, but this time designed with audio in mind.

http://www.national.com/ds/LM/LM3886.pdf

The capacitor I'm talking about is the electrolytic labeled "Ci" in the typical application on the first page. Ideally you want this cap to be as large in value as possible, to move the rolloff point down as low in frequency as you can to minimize phase and frequency response variations in the audio band. If you can fit it on your board/layout, 100-220uF is a good place to start. Bigger is better, if you can fit it in there.

[halojoy]

What I can see LM12 does not have so high DC-offsets.
It has also very low bias currents.

Power OPs use to have a lot higher offset.

So I am not sure LM12 is the problem.
must be something in the circuit.

/halo has only read the datasheet

[jackinnj]

the LM12 has gotten to be a very expensive chip. Used to be $9 or so, now they are about $30.

[cm961]

Thanks for the responses guys. I'm going to throw in that cap tonight and see what happens.

I didn't have to pay for my 4 LM12s, the amplifier I ripped apart already has them all wired and mounted to the heatsinks, so all I have to do is hook up a few caps and resistors (and an inductor) and voila, a 4 channel amp.

Pete