Headphone impedance and performance

[Guest]

Archived from groups: rec.audio.high-end (More info?)

While I was thinking about circuits to drive headphones, it occurred
to me that I have never in my life seen a graph of headphone impedance
v. frequency. As far as I know headphones have an impedance, and
that's it. As we all know, loudspeakers are designed to be driven
with a source impedance as close as possible to zero ohms. This is
because of wild impedance variations wrt frequency that are typical of
all speakers. My question is: Does this same issue crop up with
headphones? i.e. will there be a gross difference in frequency
response if headphones are driven by a high source impedance?

The sensitivity of speakers is usually specified as SPL when driven by
2.83 volts. With headphones, it's almost always SPL when driven by 1
milliwatt. Because headphones vary in impedance from about 16 ohms to
600, there will be a 6 to 1 variation in voltage required to produce
that 1 milliwatt. Any voltage divider that has to drive the entire
range of impedances will be quite a compromise in performance if the
headphone impedance varies significantly.

Anyone out there that can throw some light on this situation?

Norm Strong

 

[Guest]

Archived from groups: rec.audio.high-end (More info?)

normanstrong wrote:
>
> My question is: Does this same issue crop up with
> headphones? i.e. will there be a gross difference in frequency
> response if headphones are driven by a high source impedance?
>

Yes. Just like any other speaker.

I have seen impedance vs frequency charts for headphones. My
generalized observations:

- multi-driver headphones have significant impedance variations
- single driver headphones may exhibit one or more impedance peaks in
the 2K to 8K Hz region (most likely on purpose)

As long as you keep the output impedance of the driving amp and any
associated divider network at least 10-20 less than the nominal rated
headphone impedance, you should not have significant response
variations show up beyond what the designer intended.

A simple in-line variable resistor can be used to reduce amp output if
the headphone impedance curve is relatively flat.

> The sensitivity of speakers is usually specified as SPL when driven by
> 2.83 volts. With headphones, it's almost always SPL when driven by 1
> milliwatt. Because headphones vary in impedance from about 16 ohms to
> 600, there will be a 6 to 1 variation in voltage required to produce
> that 1 milliwatt. Any voltage divider that has to drive the entire
> range of impedances will be quite a compromise in performance if the
> headphone impedance varies significantly.
>

A T-type resistive divider is recommended as the best approach to get
attenuation without undo response variations. When designing the
T-type resistive divider, the headphone impedance needs to be taken
into account as part of the design - there is no universal divider
here. If you need to accommodate headphones with different
impedances, the best recourse is a switched network design.

The real problem arises when connecting a pair of headphones, one high
impedance and one low impedance to the output of an amp, the sound
levels can vary significantly. This requires a bit more patience to
design the appropriate divider network.

> Anyone out there that can throw some light on this situation?
>

Hope the above helps.

Best regards,

Terry