[gyrodec]

It is often said that the HD600/650 really need a home amp with lots of voltage to get the best from them, and balanced is even better. Now, I was reading the HD600 specs and the sensitivity was stated two different ways: 96dB @ 1mW and 112 dB @ 1V rms (which would be only 3.333mA for the 300 ohm HD600s).

From these specs, it looks like a swing of +/- 2V would get you clean 112dB peaks, which seems more than loud enough. The 96dB, which seems like a good loud working volume for the non-peak parts, is only 0.54V.

So, it seems only a few volts +/- would be more than enough. I also measured the voltage across my HD600s at normal listening volume and it was arounf 150mV. So are the specs not really accurate, Did I screw up my ohms law calcs, or do HD600's only need a couple of volts to get really load?

 

[FallenAngel]

Has nothing to do with how many it needs to get "loud", all about driver control. You need enough voltage swing for the driver to behave as designed. Specs state minimums to get it moving and making noise.

 

[gyrodec]

FallenAngel

Surely the drive control comes for high damping factor / good instantaneous current delivery. As the voltage defines the loudness, as these are voltage driven amps, then you need the volts dictated by the sensitivity to get the desired loudness and no more. The control is another issue, but related to damping/current and not the voltage (assuming we don't go anywhere near clipping).

I don't see where I'm wrong, but if I am, please explain why (in detail is you have time - thanks).

 

[tomb]

Quote:

Originally Posted by gyrodec /img/forum/go_quote.gif FallenAngel Surely the drive control comes for high damping factor / good instantaneous current delivery. As the voltage defines the loudness, as these are voltage driven amps, then you need the volts dictated by the sensitivity to get the desired loudness and no more. The control is another issue, but related to damping/current and not the voltage (assuming we don't go anywhere near clipping). I don't see where I'm wrong, but if I am, please explain why (in detail is you have time - thanks).

First of all, peak-to-peak voltage does not equate to RMS voltage. More important, however, is that FallenAngel is correct. If you do the research, you'll find that sensitivity is most likely quoted at 1KHz. Senns have a bass impedance hump - rising to as much as 500-600 at 100Hz. That requires much more voltage swing at lower frequencies, among other things. dB peaks in typical music are also much higher than a sustained sound pressure level. All of that combines to mean that you need much more voltage swing capability to sound good, rather than to just make a sound. Without putting words in his mouth, I believe that's what Fallen Angel meant.

As for making measurements yourself, some meters are notorious at measuring VAC, period, much less RMS values. You may also find that the actual RMS voltage provided at the output of an amp is much lower than you would think, considering the power supply at the rails, etc.

Headphone sensitivities are somewhat useful to compare one against the other, but drawing independent conclusions from those specs are a bit risky, IMHO.

 

[amb]

Quote:

Originally Posted by tomb /img/forum/go_quote.gif Senns have a bass impedance hump - rising to as much as 500-600 at 100Hz. That requires much more voltage swing at lower frequencies, among other things.

A quibble about that statement. Yes, there is an impedance hump, but that hump is due to diaphragm resonance. You'll see the same sort of thing in speaker drivers too. Resonance, being what it is, means that the driver efficiency at and near those frequencies is much higher, so it doesn't need any more voltage there to reach the same SPL as other frequencies. In fact, if you were to look at the frequency response graph, which is measured using a constant voltage, but sweeping the frequency across the spectrum, you see an essentially flat response ("normal" response variations due to other factors notwithstanding, of course). If the higher impedance at the resonant frequencies required higher voltage, you'd be seeing a dip in the response there.

The rest of your points are correct. To achieve a "loud" SPL with sine waves doesn't require much voltage swing, but music is not a test tone. To reproduce clean, undistorted transient peaks require a lot more voltage. Every 6dB increment requires double the voltage, and 10dB is roughly twice the perceived loudness. So if you want 20dB of "headroom" above the steady state average for peaks, you'll need 10 times the voltage. Suddenly, we're talking about some pretty large swings.

Thanks to the way our ears work -- i.e. the logarithmic relationship between voltage and dB!

The Sennheiser HD580/600/650 series are not bad as far as sensitivity goes (especially for a 300 ohm headphone), but they can handle a lot of voltage and really do perform better when driven with an amp that's capable of supplying it.

 

[tomb]

By all means, quibble ... thanks for the insight into that phenomenon, amb.

 

[gyrodec]

I was thinking that 112 dB was about right for the top of those transient peaks and so 92 dB for the general music base level. If this is true, then +/- a couple of volts is still enough. BUT, perhaps 92/112 aren't in the ball-park. Does anybody have an idea about what their or normal listening levels / peak levels might be. Not questioning amb/tomb, just want to get my assumptions moved in line with reality.

 

[nikongod]

I listen quietly, and the HD-580/600 are a VERY weird headphone

Unlike a grado which is basically power based (get it more power, it sounds more good!) the hd580/650 are almost purely QUALITY based. The power levels in actual use are SMALL.

By quality based, I mean that the most important thing for them is the quality of the signal. I have put 1vRMS into the 580 as an experiment, and you can hear them clearly outside the room.... its LOUD.

Now, the idea that they can "work" on less than 1v is correct, but this does not take quality into account.

With a portable amp you have many things working against this goal for quality.
Limitations on PSU cap sizes, limitations on how far into class-a you can bias things, essentially being forced to use opamps, and a few others.

With a home amp, the designer/builder can avoid the problems associated with the compromises mentioned above to the benefit of the sound.

 

[jcx]

Quote:

Originally Posted by gyrodec /img/forum/go_quote.gif I was thinking that 112 dB was about right for the top of those transient peaks and so 92 dB for the general music base level. If this is true, then +/- a couple of volts is still enough. BUT, perhaps 92/112 aren't in the ball-park. Does anybody have an idea about what their or normal listening levels / peak levels might be. Not questioning amb/tomb, just want to get my assumptions moved in line with reality.

HeadWize - Article: Preventing Hearing Damage When Listening With Headphones (A HeadWize Headphone Guide)

read twice - for protecting your hearing and for live music peak SPL