LM4780 as a sub amp... odd driver behaviour

[rabbitz]

Put a LM4780 in parallel into a sub and the woofer has an odd behaviour I haven't struck before.

As the volume is turned up on the sub amp, the woofer moves from the central voice coil position to a new position about 5mm away. Reverse the phase and it does the same thing but in the opposite direction.

The LM4780 has a high DC offset of 135mV which I think could be the culprit.

First question is, has anybody experienced this and what was the cause?

Second question is, how do you get the DC offset down to a more sensible level on chipamps or is the the luck of the draw with chips?

[pinkmouse]

What schema are you using?

[rabbitz]

This one.

http://www.diyaudio.com/forums/showt...057#post761057

It's a kit from Peter.... PCB is so nice, I didn't want to put solder on it.

http://www.specializedkits.com/lm4780.shtml

[pinkmouse]

Ok, do you get this offset with the input shorted to ground?

[rabbitz]

Offset only checked with nothing on input.

[pinkmouse]

Short the input and measure again then please.

[rabbitz]

Too late..... after much swearing, the plate amp went back in and burned some midnight oil and changed it to a stereo amp to replace a P3A in the workshop.... ah, that's better. No funny driver business and nice sound.

The offsets checked out OK in stereo mode so the parallel must have been adding the offsets or some other relationship such as (OS1 + OS2)^todays temp / days to the full moon.

When you talk about shorting the input, that's using a resistor or such?

I've never experienced the moving driver before and I've done heaps of amps and speakers. One of the great mysteries.

Cheers & thanks.

[pinkmouse]

Originally posted by rabbitz
When you talk about shorting the input, that's using a resistor or such?

No, just get a plug and short the centre pin to the screen. It just makes sure there is absolutely no input, as some GCs can get very strange if the input is left floating.

Glad you got sorted anyway!

[rabbitz]

Thank you.