What is/are output impedance for?

Discussion in 'Sound Science' started by billsonchang007, Dec 15, 2012.
First
 
Back
1 2
4 5
Next
 
Last
  1. stv014
    To return to the topic, I show some loopback measurements with a (cheap and rather old, but that is not relevant here) 56 Ω headphone driven from a 100 Ω source. Unfortunately, the drivers are not very reactive, so the difference is not as large as it could be with some other models - HD598 would be more interesting, for example - but it is there. Here is what happens to the frequency response:
    imp_fr.png
    That is not too much, but it is definitely enough for a positive ABX test. Now the distortion, with the loopback shown in yellow, and the microphone input (at the same loudness) in red. Again, the effect is not major, but at the resonant frequencies (~100 Hz and ~350 Hz) it is within 15 dB of the acoustic distortion, or about 5.6 times less):
    imp_thd_2.png
    With 56 Ω resistors instead, it would stay below -96 dB for most of the frequency range.
     
    ultrabike likes this.
  2. jaddie
    Quote:
    I've clearly opened the can of worms here.  
     
    I've never stated that response variations are related to impedance variations per se, except when the driving impedance is too high, a condition that is unlikely given the statistical trend that shows that the typical models with larger impedance variations are higher impedance where it has no impact. 
     
    Perhaps I'm at fault here for not making the point clear, though I've tried now several times.  I've mentioned something that while true, doesn't matter.  
     
    Look, it doesn't matter.  I've found something that is true, and may be hard for you to accept, but I've said now several times, it doesn't matter.  What does matter is that a headphone amplifier has a low source impedance.  And if it does, these variations don't matter.  
     
    I'm pretty sure we agree, and now we're just arm wrestling.  I do appreciate the smily, but when you post stuff like that in forums, people are going to take it a statement that has validity. We're dealing with folks that are forming opinions.  Somebody's going to take that out of context, and ignore the indication that you're being sarcastic.  
     
    And, finally, I never said high impedance is worse than low impedance.  It isn't.  
     
  3. xnor
    You talk about larger impedance variations, but it has been pointed out before that all that matters is the ratio between min/max impedance.
     
    Example:
    Zout = 30
    HP1: Zmin = 30, Zmax = 60 -> +2.5 dB at the resonant frequency
     
    Zout = 250
    HP2: Zmin = 250, Zmax = 500 -> +2.5 dB at the resonant frequency
     
    In both cases Zmin/Zmax = 0.5, but HP1 has only a +30 ohm and HP2 a +250 ohm peak in the impedance curve. With the same damping factor, the results are the same.
     
    With a fixed Zout of 30 ohms HP2 could have a Zmax of 10000 ohms and still would not even reach +1 dB at the resonant frequency. It's the ratios that matter.
     
  4. mikeaj
    Many Grados, Denons, cheap headphones, Ultrasones to some extent, all the planar magnetics in particular, etc. have lower impedance that's mostly flat. Correlation? It's the ratios that matter; that said, I think the only point that was being made was that many lower-impedance headphones have relatively resistive impedance. Anyhow, it's worth noting that many lower-impedance IEMs do in fact exhibit extreme impedance variations, particularly the multi-driver models that actually have passive crossovers.


    So hopefully to summarize in a way that doesn't lose all the important details...


    The output impedance is a property of an amplifier's electrical output. Like other amplifier parameters, it depends on the design of the circuit. When using headphones (speakers), the amplifier actually has to drive the series combination of the headphones and its own output impedance.

    In general, impedance is a function of frequency and may be different at different frequencies in the audio frequency range.

    If the amplifier has relatively high output impedance compared to the headphone's impedance (thus lower electrical damping), and if the headphone has large variations in impedance values over the audio frequency range (thus higher susceptibility to damping issues), the sound of the system may be changed. Lower damping means overall less control of the amplifier over the headphones; in terms of the sound, you can get unintended frequency response variations and a little higher distortion.

    If you use an amplifier with low output impedance relative to any headphone you use, then these effects become negligible, so you don't need to worry about them. Generally, if the nominal quoted impedance value of the headphones is several times greater than the amp output impedance, there should be no problems. There are vastly diminishing returns on higher ratios, so more is not worse but not really any better in practice. That said, some people subjectively prefer lower levels of electrical damping than the common by-the-books assessment, for listening with certain headphones. Depends on the headphones and the person.


    Impulse response is the resulting output (response) of the system when you input an impulse. Step response is the output of the system when you input a step. An impulse is just a very brief (and high) spike. A step looks like a step on a staircase: it suddenly jumps up to a higher level and stays there. The impulse and step response behaviors of a relatively well-behaved system—like audio hardware, which is mostly designed to be that way—gives us information about how the response might be for other kinds of signals too. For a perfect linear time-invariant system, the impulse response alone is enough to tell us completely about the response for any input. For real-world systems, we can look at deviations from the ideals.
     
  5. stv014
    Quote:
     
    With full size dynamic headphones, the most obvious audible effect of high output impedance is typically more bass, which is indeed a popular effect.
     
  6. BillsonChang007
    Thank you guys for all the great explaination(thought my brain are now fully loaded)! Christmas is/are around the corner :D ! MERRY CHRISTMAS EVERYONE :)
     
  7. sinquito
    Oh man, I'm not going to approve the next semester's electrical measurements let alone 5th semester electromagnetic theory :confused:
     
  8. BillsonChang007
    Please allow me to ask one more question :)

    I have an iPod Nano 3rd & iPhone Original. I can't find the output impedance of these two. So how can I measure it?
     
  9. jaddie
    Not precisely what you asked for, but will put you in range:
    http://www.head-fi.org/t/583426/ipod-output-impedance
     
  10. stv014
    Quote:
     
    While someone could very well have already measured those devices, to test it yourself, you basically need to generate a test signal, and measure the output voltage of the device with and without a known load, then the output impedance can be calculated with a simple formula. The exact method depends on what you have for the voltage measurement and load. Without a DMM and resistors, it is even possible to measure the voltage drop with a PC sound card, a headphone of known impedance, and a splitter.
     
  11. jaddie
    Quote:
    That's on the right track, if a little incomplete.  
     
    Hang on a second...Let me have a shot...I'll just type "measuring impedance" into the search field...
     
    Gimme one...more...There!  I've just Googled that for you!
     
    http://www.learnabout-electronics.org/ac_theory/impedance73.php
     
    See the bottom figure.
     
  12. BillsonChang007
    Looks like I don't have the right equipment for the test. A build in pc soundcard: YES. A headphone of known impedance: YES. Other than these two, NO :xf_eek:

    I have had installed these... http://audio.rightmark.org/products/rmaa.shtml but no idea what it does XD
     
  13. jaddie
    Well, you're a good part of the way there.  But, if I may, why the need? Isn't the published data close enough?  You're not designing a headphone amp, are you?
     
  14. xnor
    Just get a splitter and connect only the line-in. Set the level so that it doesn't clip the input. Play a sine wave and record it, note the recorded level (Vnoload).
    Now also connect the headphone (with impedance Rload at the frequency you're testing) to the splitter and repeat the above process, but don't change the level. Note the recorded level (Vload).
     
    calculate: Zout = (Rload * (Vnoload - Vload)) / Vload
     
  15. BillsonChang007

    I'm having my holiday & I'm bored + I don't play games a lot~ & feeling like improving my knowledge in audio equipment :)

    As for designing a headphone amp...nope...I'm still young~
     
First
 
Back
1 2
4 5
Next
 
Last

Share This Page