Woody wrote:

A mistake that many people make with lo-Z headphones is to run an amp
straight into them and then wonder why there is so much residual noise.

The easiest method is to use a small power amp such as a LM386

Oh FFS ! Grow up !

Graham

Arny Krueger wrote:

"Eiron" wrote

That's an impressive damping factor.
I wonder how the frequency response of headphones changes
when current-driven rather than voltage-driven.

The impedance curves of headphones are documented on the web, and are
typically pretty non-uniform.

Probably why so many people like dedicated headphone amps, and drive them
from low impedance, low voltage sources.

Studios tend to use standard power amps.

Graham

David Looser wrote:

A complementary pair of transistors, as suggested, would have a bandwidth
plenty wide enough for video! But the linearity would leave rather a lot to
be desired, more crossover distortion than the LM324 probably. But I wonder
if Tony L has read the data sheet for the LM324?, it clearly states that
with an AC coupled load (I assume he has capacitively coupled his
headphones?) a resistor should be connected from the output pin to ground
"to avoid crossover distortion", has he done that?

The LM324 is a piece of crap in the first place ! It has an intentional near
dead-band.

Graham

TonyL wrote:

I'm willing to accept that this problem may be due to trying to drive a low-z
load with an
inappropriate device.

Designed about 40 years ago too. Well, maybe 35. Why are you obsessing over
obsolescent devices ?

More than 35 in fact.

"Since their introduction by National more than 25 years ago, the LM324 and
LM358 quad and dual op amps have been top performers" (but not for audio).
March 10, 1998
http://www.national.com/news/item/0,1735,24,00.html


Graham

TonyL wrote:

David Looser wrote:

The odds that they will be quads is about zero, since headphone
drivers are typically used in pairs.

Agreed. I know of no quad op-amps that will do the job. I think Tony
should accept that he will need a separate h/phone amp chip.

All comments noted. Thanks.

The *only* problem with the LM324 in this application is the crossover
distortion.

A consequence of its internal design.

Graham

TonyL wrote:

Don't know. But if I can't use the spare section on the existing quad op-amp
to drive the headphone then I may as well go for a purpose designed heaphone
amp as suggested by others in here, or add a bipolar complementary pair.

Yes, buffer the op-amp with a complementary pair. Don't forget to bias them on
either and set the current with emitter resistors. You can even take the
feedback from the junction of the emitter resistors but be advised to include a
load 'isolating' resistor of say 47 ohms.

Graham

On Tue, 13 Jan 2009 08:22:26 +0000, Eeyore
wrote:



TonyL wrote:

David Looser wrote:

The odds that they will be quads is about zero, since headphone
drivers are typically used in pairs.

Agreed. I know of no quad op-amps that will do the job. I think Tony
should accept that he will need a separate h/phone amp chip.

All comments noted. Thanks.

The *only* problem with the LM324 in this application is the crossover
distortion.

A consequence of its internal design.

Graham

It has a poorly biassed output stage - virtually pure class B.

d

Don Pearce wrote in message news:496c541a.346449453@localhost...
On Tue, 13 Jan 2009 08:22:26 +0000, Eeyore
wrote:


TonyL wrote:


The *only* problem with the LM324 in this application is the crossover
distortion.

A consequence of its internal design.

Graham

It has a poorly biassed output stage - virtually pure class B.


Whether the design is "poor" or not depends on what application you have in
mind. THe LM324 was one of the first op-amp designs to work from a single 5V
rail, making it a convenient way of adding analogue functions to a logic
board. The input common-mode range includes the -ve supply rail making it a
good choice for comparators etc., and the output can drive TTL directly.

And the LM324 *can* work quite well as an audio amp (though possibly not a
headphone driver) as long as a suitable resistor is connected between the
output pin and the -ve supply rail.

I entirely agree that there are far better audio op-amps around these days,
but the LM324 still has it's uses, and it's dirt cheap!

David.

Eeyore wrote:
TonyL wrote:

Don't know. But if I can't use the spare section on the existing
quad op-amp to drive the headphone then I may as well go for a
purpose designed heaphone amp as suggested by others in here, or add
a bipolar complementary pair.

Yes, buffer the op-amp with a complementary pair. Don't forget to
bias them on either and set the current with emitter resistors. You
can even take the feedback from the junction of the emitter resistors
but be advised to include a load 'isolating' resistor of say 47 ohms.


Thanks,

Probably the simplest solution for this application, a few pence worth of
discrete devices. Your tips regarding class AB biasing+feedback+load
isolator resistors duly noted.