Max "Safe" DC Offset

[morfic]

Tangent suggests 20mV as a safe threshold for DC offset.

Is this an accurate amount, too conservative, too loose?

And are the effects indeed cooked headphones, or does the shift due to the constantly applied voltage "just" bias the headphone?

Thanks,

Daniel

[AudioFreak]

It very much depends on the type of driver you are feeding but excessive dc offset will indeed cook a driver if it's severe enough or for long enough.

[nigelwright7557]

I really cant see 22mV frying headphones !

A lot depends on what you are driving, speakers can be from 1watt to many hundreds of watts. Poking a few millivolts of DC into a big speaker just wont harm it.

On the other hand a volt of DC offset can dissipate quite a bit of heat in the MOSFET of the rail the DC is offset to.

A classic sign of too big an offset is one transistor of the pair gets hot and the other doesnt.

[morfic]

I am referring to opamps, not discrete circuits, although in the end i see no difference in the effect on the headphones.

While building my cmoy based on tangent's docs i noticed just how much DC offset a bipolar opamp with large input bias current can create in a circuit not designed for it.
I didn't see the 20mV as a "right above that it'll fry" value

It just made me wonder, in my cmoy i have control over the resistors involved in the amp and i can later just for fun change their values to control the LT1364's high DC offset. But what about the people opamp rolling in amps under the assumption it's all good?

I put the LT1364 in a hybrid amp and while the LME49720 in it shows ,1mV offset, the LT1364 still almost scores 30mV. (It's btw 300mV in my cmoy using the stock resistor values, 100mV with resistor values i happened to have available).
I am impressed how well National controlled the input bias current, since i saw no other (popular) bipolar opamp reach nA range, bipolar opamps seem to be typically more in uA range.

So i was wondering, how much past the 20mV is safe, how much heat is a constant voltage in mV range causing on the driver, or is it all just affecting the dynamic range with the shifted "zero point" (offset + signal reaching the max the headphones could deal with sooner?)

I can't possibly listen to the headphones as loud as they can manage to dish out, so is the DC offset still of concern?
This might actually spill over in my decisions of buying used headphones down the road if damage is indeed a possibility.

Thanks,

Daniel

[Geek]

Opamps?

I have a bunch of generic oldies, TL08x, TL07x and NE553x and all the ones I tested were maybe at largest, 5mV

Maybe I just lucked out?

Cheers!

[peranders]

.... times the gain

[Geek]

Originally posted by peranders
.... times the gain

Which were all followers... gotcha!

[kipman725]

opamps aren't designed for driving headphones directly thats why there are DC offset problems. Looking at the datasheets of most opamps they specify a minimum load of 600Ohm's or greater. Companies make dedicated headphone amplifier chips that are not much more complex to use and will work much better, I don't understand why people don't just use those.

[AndrewT]

Originally posted by morfic
the input bias current, since i saw no other (popular) bipolar opamp reach nA range, bipolar opamps seem to be typically more in uA range

look at the datasheet for input offset current.
Some opamps have offset current cancellation and these generally don't respond to tweaking of the input pin impedances.

[morfic]

Thanks for all the good info.

I would like to come back to the actual question of what is safe and what really happens to headphones once you go past that though.

thanks,

Daniel