Headphone amplifier advice
TonyL
For various reasons I need:
1) Quad amps to be used, there's a requirement for other op-amps in the
units.
2) Single rail, around 12-15 V, I guess most modern op amps will do that
easily.
3) Not too current hungry, I want to power a number of these units from one
simple 12VDC transformer supply.
4) Lo-fi will do so long as distortion is not too gross. Intercom quality is
OK.
5) Reasonable cost.
I built a prototype using a LM324 that happened to be handy. Result was
gross crossover distortion when the output had to supply current to the
low-z earphones. I know these amps can be biased into class A operation but
I want reasonable power economy. Would be nice to just drop a pin
compatible device into my prototype, say a LM 387 ? Except the spec says
>600ohms load.
Advice/comments please ?
Arny Krueger
news:7badnanG9dsoxvfU...@bt.com
> I want to drive low-z (32 ohm) headphones using op amps.
> For various reasons I need:
> 1) Quad amps to be used, there's a requirement for other
> op-amps in the units.
That one is probably a non-starter, for reasons I'll detail below.
> 2) Single rail, around 12-15 V, I guess most modern op
> amps will do that easily.
Yes.
> 3) Not too current hungry, I want to power a number of
> these units from one simple 12VDC transformer supply.
The application defines the current drain.
> 4) Lo-fi will do so long as distortion is not too gross.
> Intercom quality is OK.
No need to compromise sound quality.
> 5) Reasonable cost.
> I built a prototype using a LM324 that happened to be
> handy. Result was gross crossover distortion when the
> output had to supply current to the low-z earphones.
Read the spec sheet and do the math. What sort of voltage will you need to
drive headphones? How much peak current does this net out to be?
> I know these amps can be biased into class A operation but
> I want reasonable power economy. Would be nice to just
> drop a pin compatible device into my prototype, say a LM
> 387 ? Except the spec says
>> 600ohms load.
> Advice/comments please ?
Quit pulling numbers out of your memory from days long ago, and do what
professional engineers do when they have a task like this: Go to several op
amp manufacturer sites like TI, National, Maxim, NJM and Philips, and see
what they recommend for the purpose.
There are even such things as op amps that are designed to drive low
impedance loads and work from single-ended supplies.
The odds that they will be quads is about zero, since headphone drivers are
typically used in pairs.
David Looser
news:WIOdnXSlstRW_vfU...@giganews.com...
> "TonyL" <TonyLacyR...@Nu-Ware.com> wrote in message
> news:7badnanG9dsoxvfU...@bt.com
>
>> I want to drive low-z (32 ohm) headphones using op amps.
>
>
> There are even such things as op amps that are designed to drive low
> impedance loads and work from single-ended supplies.
Indeed there are. Since 32ohm is now the standard for headphones there are
plenty of heaphone amps intended to drive 32 ohm h/phones on the market.
Mostly they are designed to work from single-ended supplies of around 3-12V
>
> 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.
David.
Phil Allison
"TonyL"
** Huh ??
The LM387 is a long obsolete, dual low noise pre-amp.
How about you try a TL074 - exact same pin out as the LM324 and no x-over
issues.
Will deliver about 20mA ( = 13mW peak ) into a 32ohm ear phone, which is
plenty loud.
..... Phil
TonyL
>>
>> 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. It drives low-z headphones fine. So, I'll try plugging in a
TL074 and see what happens before looking to a purpose designed hadphone
driver.
TonyL
>
> How about you try a TL074 - exact same pin out as the LM324 and no
> x-over issues.
>
> Will deliver about 20mA ( = 13mW peak ) into a 32ohm ear phone, which
> is plenty loud.
>
I'll give it a try. Thanks.
Woody
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 or
TDA2003, then put a resistor in series with it - something in the range
330-470R will usually suit. You will probably end up with a far more
acceptable result as the amp is specifically designed to drive audio
rather than just a general purpose amp.
--
Woody
harrogate three at ntlworld dot com
tony sayer
<TonyLacyR...@Nu-Ware.com> scribeth thus
Would one of those SSM -whatever -there -called series line drivers do
this quite well?..
--
Tony Sayer
Eiron
That's an impressive damping factor.
I wonder how the frequency response of headphones changes when
current-driven
rather than voltage-driven.
--
Eiron.
Arny Krueger
news:6t1ftcF...@mid.individual.net
> 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.
TonyL
Just as a FYI, this project isn't hifi. In fact, just the opposite. It is to
train students in simplex radio comms protocols. A modicum of noise and a
frequency response such as 300 Hz -3 kHz would be fine and simulate a radio
link even better. In fact, part of the spec is to be able to inject noise
onto the system :-)
With regard to the headphone audio amp issue, I may just do it in discrete
components. A complementary pair of gp transistors and a half dozen
resistors will probably cost about tuppence :-)
TonyL
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.
TonyL
Yes, I appreciate that a proper audio amp would do the job better. But I had
a spare amp section on the quad op-amp so I thought I would try to use that
before looking at other options.
tony sayer
<TonyLacyR...@Nu-Ware.com> scribeth thus
I'm sure a few contributors could spec you something in the 300-3400 Hz
area;))...
--
Tony Sayer
David Looser
news:xkp$X5D2w6...@bancom.co.uk...
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?
David.
TonyL
> Ah!, you should have said that..
>
> I'm sure a few contributors could spec you something in the 300-3400
> Hz area;))...
>
LOL !
TonyL
>
> 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.
>
Not in class AB mode ?
> 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?
Yes, I have the data sheet. With the headphones connected the resistor had
the effect of moving the crossover point up or down from the 0V level
depending on whether the resistor was connected to +Vcc or -Vcc. The value
of the resistor determined how far the crossover point moved. I guess the
idea is to set it so that the output stage goes into Class B mode at higher
signal levels so that the crossover distortion would be buried in the
signal. Unfortunately, it is so bad that it is still clearly audible at all
but very high levels. Moving the crossover point beyond the maximum output
voltage swing removed the distortion but left me with a class A output
stage, with associated high quiescent current. I'm willing to accept that
this problem may be due to trying to drive a low-z load with an
inappropriate device.
Eeyore
TonyL wrote:
> I want to drive low-z (32 ohm) headphones using op amps.
Then you're screwed if you need any volume. Most op-amps current limit at around
30mA =~ 20mA RMS. That's 13mW.
Graham
Eeyore
TonyL wrote:
> I built a prototype using a LM324 that happened to be handy. Result was
> gross crossover distortion
The LM324 is world reknowed for that. Why not use an AUDIO op-amp ?
Graham
Eeyore
Arny Krueger wrote:
> The odds that they will be quads is about zero, since headphone drivers are
> typically used in pairs.
Very true.
Graham
Eeyore
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
Eeyore
Arny Krueger wrote:
> "Eiron" <E1...@hotmail.com> 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
Eeyore
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
Eeyore
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
Eeyore
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
Eeyore
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
Unknown
It has a poorly biassed output stage - virtually pure class B.
d
David Looser
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.
TonyL
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.