No, it is possible to calculate the impedance (and phase) vs. frequency of a load from the voltage falling on it when driven by a source with any known non-zero output impedance. One of the utilities from the last link in my signature does exactly that. Actually, an output impedance comparable to that of the load makes the measurement more accurate than if it was close to zero (in that case, the current should be measured instead, but ADCs that expect a voltage input are obviously more common).

Measuring it with a high output impedance allows the cans to be as undamped as possible and lets the phase swing more easily for a strong response when measured. Since all the cans are measured similarly, it allows for better relative evaluation.

So does the mechanical vibration affect the impedance results? And is that a fair assessment on the headphone when they should be driven by a low impedance source (isn't that what headphones are designed for)? The electrical damping would change the mechanical response. Tyll, i love your site and your measurements. It is a great resource to the headphone community! Thanks There are still a bunch of missing pages in your headphone measurement explaination and im anxious for those to get posted Cheers

If the driver is reasonably linear, it can be modeled as a network of resistors, capacitors, and inductors, which do not change significantly with different sources. For example, the impedance peak at the primary resonance frequency can be simulated with a parallel RLC circuit. In any case, even with a higher source impedance, the driver itself as a load remains the same. However, mechanical factors like whether the headphones are actually worn do have an effect on the measured impedance curve. Edit: here is a comparison of the same headphone measured with damping factors of about 23 and 2.3. This might not be perfectly accurate, since even a small error in the assumed output impedance can skew the results in the high damping factor case, but it does show there is no significant difference in the resonance peak with increased source impedance:

How much of a damping ratio would provide the best sonic results? Some say 8-10. Some people with tube amps run at about 5 or thereabouts. One person I ran into told me at least 20. Personally, I think I have noticed a difference between a 8 and a 10 damping factor. Wouldn't too low of a damping factor distort both the high and low FR? Basically have mid-frequency bloat? Instead would this depend on the specific amp? Or maybe how linear the impedance curve is on a set of headphones? If this has already been discussed, please provide me with a URL.

The difference is insignificant (note: do not compare the left vs. right graphs, but the blue vs. red traces on the same graph instead), and is mostly the result of measurement inaccuracy. The analysis utility needs to know the output impedance exactly for correct results, and when it is low, even 0.1 ohm error will visibly skew the graphs (especially with the zoomed in Y scale I used). Additionally, the slight (~1 degree at 23 kHz) phase difference in the high frequency range is probably because the sound card output that was used for testing has some serial inductance, and since that is not taken into account by the analysis (which assumes pure resistive input and output impedances), the low output impedance case is skewed towards measuring a less inductive load. In any case, the reactance in that range is mostly from voice coil inductance, rather than back EMF. I could do the test again with 1000+ ohms output impedance, but I would expect it to match the 110.7 ohms measurement fairly accurately, probably better than what the above graphs show, since the source inductance becomes insignificant relative to the resistance.

LO has high output impedance. How come? What difference does it make if output impedance of the line out was significantly different like 20ohms vs 400ohms?

Very informative thread, but I have a question that I couldn't find the answer to here: Is there a primer on head-fi that provides instructions (for an amateur like myself) for measuring the output impedance of a headphone amp, presumably using a multimeter? Is it possible to measure the impedance from the output jack without opening up the amp?

Yes. You just need a (possibly DIY) cable that makes it possible to measure the output voltage, and at the same time connect a load (preferably a resistor) to the output. Use any software of your choice to generate a sine wave at a frequency where the AC voltage mode of the multimeter works reasonably well (it does not even need to be perfectly accurate, since it is only the ratio of the voltages that will matter), and measure the voltage with and without the load under otherwise identical conditions. It is easier to get accurate results if the resistance of the load is not very different (by orders of magnitude) from the output impedance, but it obviously should not be so low as to risk damaging the amplifier. Something like 33 to 100 ohms should be OK at <= 1 Vrms unloaded voltage. Once you have measured the voltages and know the load resistance, you can use the following formula: Zout = Zload * ((V_unloaded / V_loaded) - 1) A somewhat more advanced topic is measuring the output impedance as a function of frequency, for example with capacitor coupled outputs. This can be done with a sound card loopback; I have a program that can generate impedance vs. frequency graphs, but a simple RMAA frequency response test is enough to show if there is significant output reactance (e.g. capacitor coupling causes a low frequency roll-off, from which the capacitance can be calculated). With simple multimeter-only testing, one could measure at a low and mid-range frequency (e.g. 50 and 400 Hz) to see if there is any serial capacitance.

Have new pair of 4 driver earphones which need to be treble boosted on all sources except when attached to speaker + amp (B&W MM1 in this case) does this tend to indicate portables are too weak to drive earphones?

One of the things I don't understand is how peak sound pressure levels relates to general volume settings. I have an audio technica ad700 (98db/mw, 32 ohm) connected to an amp that can do 1.3v, 90mA, 1 ohm output impedance. That comes out to 115db peak, but I find that I'm frequently using the low end of the analog pot(it's an Alps) + about -15db of digital volume control to get a reasonable volume. Now, what would it be like if I switched to a q701(105db/v), with a peak of 107db? And I assume if I got a k612 (101db/v) and had a peak of 103db/v, that would just be unusable? Also considering an sennheiser hd 380 (110db/v), which would have a peak of 112db. That sounds like it would be pretty close to the audio technicas, but idk if it works out like that.

Hi, i have an old receiver from sony which has an input of 50k ohms and input sensitivity at 500mv my question is if i can use this input for good results with headphones like the Hd650 or Hd 700. I use the receiver with my Hd 598's and works perfectly...

the input specs only tell you that you can plug in an ipod with a line out cable(which is 500mv), and things like that.

Calyx coffee is a great example of the roll off from output capacitor. It has a capacitor to block DC offset and this same capacitor reacts with the load outputting a high pass filter response. Their headphone out has capacitor output.