Beyer T1

[Tyll Hertsens]

Looking only at the amp and the theoretical goal of very low output impedance & improving electrical damping factor adding an output resistor like this is madness, but by the perspective of what the headphones seem to want it makes the amp better.

Lemme try again:

Designers make sure the headphone have enough damping to work reasonably well with 120 Ohm output impedance.

If the headphones happen to be driven by an amp with a lower output impedance you will get better fidelity because the driver movement will more closely match the output signal of the amp.

If you put a Porche engine in a VW it will perform better even though the bug was designed for the smaller engine.

I think that gear should be compared at its best, and with such an easy guess at how to get there why shouldnt it?

A) Again, if the manufacturers don't tell you what the specified output impedance is there's no way to test at the designed spec.

2] Assuming we're talking about the goal of fidelity reproduction, all else being equal, the lower the output impedance the more fidelity it will have.

I'll reiterate, they use 120 Ohms because they know they'll be plugged into crap amps, not because they can make it sound better with 120 Ohm impedance in.

[Hopstretch]

I'll reiterate, they use 120 Ohms because they know they'll be plugged into crap amps, not because they can make it sound better with 120 Ohm impedance in.

Doesn't this start to fall away when we're talking about $1K+ headphones, Tyll? Who is going to drop that kind of cash on cans and then jack into their hawt 80s Aiwa integrated?

[cetoole]

Lemme try again:

Designers make sure the headphone have enough damping to work reasonably well with 120 Ohm output impedance.

If the headphones happen to be driven by an amp with a lower output impedance you will get better fidelity because the driver movement will more closely match the output signal of the amp.

Wouldn't you end up with a diaphragm which is overdamped? A low z amp provides the damping electrically, so if beyer is compensating for a lack of this by increasing diaphragm tension or additional external damping materials, and you suddenly add the electrical damping back in, it should push it too far the other way, instead of just getting better and better.

[Tyll Hertsens]

Doesn't this start to fall away when we're talking about $1K+ headphones, Tyll? Who is going to drop that kind of cash on cans and then jack into their hawt 80s Aiwa integrated?

Gotta agree that it's getting better, but I'd still bet people plug T1s and HD 800s more into pre-amp headphone jacks than dedicated headphone amps. Maybe I'm wrong; I'd love to see the numbers.

In any case, if more people are using dedicated amps the argument is all the stronger for low-Z amps.

Wouldn't you end up with a diaphragm which is overdamped?

I'll admit to being close to the edge of my knowledge here, so I'm not really sure, but I think they're two different things and lower output impedance will always drive the cans more accurately. I'll try to do some research on that.

[mypasswordis]

Doesn't this start to fall away when we're talking about $1K+ headphones, Tyll? Who is going to drop that kind of cash on cans and then jack into their hawt 80s Aiwa integrated?

Me, but replace T1 with K1000 and 80s Aiwa with 70s Yamaha.

My understanding is the lower the output impedance, the less effect there is on output voltage from the variation in impedance vs. frequency of the headphones. But there is such a thing as overdamping, and so there is an optimum damping factor range for each headphone. I bring up again the Pass discovery of the "current source" amplifier for efficient single driver speakers. Beyers in the past have been loose in the bass due to the bass reflex design, making producing the bass frequencies much more efficient so I don't see further undamping them helping the sound any.

[Tyll Hertsens]

Me, but replace T1 with K1000 and 80s Aiwa with 70s Yamaha.

That's the coolest part of your damn good post.

Current amplifiers have been around for a long time. I've often wondered why they aren't used in audio, and I'm not surprised the great Nelson Pass has played around with them. They were used in the scanning coil amplifiers when I used to work on scanning electron microscopes because they make nice linear sweeps because the electron deflection was directly related to the magnetic field stength which was directly related to the magnetic coil current. The problem is that a speaker is not quite like that because the inertia of the whole floppy mess is much more reactive then a nice lightweight electron beam. I'd have to look into it, but I would guess that the load reactance on a current drive amp would create the need for much higher voltage swings on the amp which makes it impractical except in tweaky applications.

Edited March 8, 2010 by Tyll Hertsens

[mypasswordis]

Heh, I use the pigtail and plug it into the speaker outs so I avoid the series resistor off the the 1/4" headphone jack.

Here's the article on current source amps: http://www.passdiy.com/pdf/cs-amps-speakers.pdf

The reason it's a "current source" amplifier is it's really just a normal (high-powered) voltage source amplifier with a 47 ohm resistor in series at the output, which is the Thevenin equivalent of a real current source with a resistor in parallel. The reason it works is because speakers have impedances that are usually much smaller than 47 ohms so it seems like a current source, though it obviously won't work on headphones. Though many people have said they liked the K1000 with the F1, which is a current source (transconductance) amp. Well, for speakers anyway.

[jpelg]

I'm still not convinced this all matters to a 600-ohm headphone driver.

Can someone show me some calculations that take driver impedance into consideration?

[Tyll Hertsens]

I'm still not convinced this all matters to a 600-ohm headphone driver.

Can someone show me some calculations?

No need, you're right, it doesn't matter much with a high impedance driver.

The hub bub about matching the 120 Ohm thing isn't important,

and it's especially not important when you just over power the problem with a good amp.

The current drive thing is really cool though ... been reading some AES papers.

[Tyll Hertsens]

FWIW this was a pretty okay article about headphones and impedance.

[NightWoundsTime]

No need, you're right, it doesn't matter much with a high impedance driver.

I may be pulling one of these, , but the differences and benefits are not negligible in any system I've heard the T1 with so far. Maybe this cable that Ari put together for me has some magic mil-spec pixie dust that's doing more than just some ohmage .

Edited March 9, 2010 by NightWoundsTime

[jpelg]

Is there any measurement data to show if the T1's have a significant variance from their nominally-rated impedance across the FR range?

[mypasswordis]

I may be pulling one of these, , but the differences and benefits are not negligible in any system I've heard the T1 with so far. Maybe this cable that Ari put together for me has some magic mil-spec pixie dust that's doing more than just some ohmage .

It should make a difference. Assuming an SS amplifier with 0.2 ohm zo, damping factor of 600/0.2 = 3000. Damping factor of 600/120.2 = 5.

Is there any measurement data to show if the T1's have a significant variance from their nominally-rated impedance across the FR range?

Yes but the impedance curve seems to be outside of the Headroom graph.

[Tyll Hertsens]

I may be pulling one of these, , but the differences and benefits are not negligible in any system I've heard the T1 with so far. Maybe this cable that Ari put together for me has some magic mil-spec pixie dust that's doing more than just some ohmage .

Nah, I'm just being my normal, old school, objective self; placebo effect be damned,

the humming bean is an exquisite sensor, we may not really know what's in our head and what's not, but we hear stuff.

Is there any measurement data to show if the T1's have a significant variance from their nominally-rated impedance across the FR range?

Gotta get some files from another computer ..... .... here you go:

http://putts.smugmug.com/HeadRoom/On-Line-Content/t1-impedance/805803381_RNtX3-O.jpg[ATTACH=CONFIG]2785[/ATTACH]

[Tyll Hertsens]

Yes but the impedance curve seems to be outside of the Headroom graph.

Yeah, I've asked about that before; I can't make it hit the priority list high enough though. C'est la vie.

[jpelg]

Gotta get some files from another computer .....here you go

Thanks, T. It appears from that graph that it never drops below 600-ohms.

Doesn't that indicate the amp's output impedance would have even less of an effect?

[aerius]

Interesting...the sharp peak in the impedance graph at ~8kHz matches up with a similar peak in the frequency response graph at about the same frequency. I'd suspect this is a resonance of some sort though I'd need to see a cumulative spectral decay plot to be certain.

[Tyll Hertsens]

Doesn't that indicate the amp's output impedance would have even less of an effect?

Yupper. That article I pointed to makes that point. It is more complicated though, because it's really the phase angle (amount that voltage leads or lags current) that makes the headphone a difficult load to drive. Simple sine wave sweeps is one thing, but instantaneous changes in random signals relative to the current nature of the signal can set up very complex relationships that deliver a whole different set of problems. It's why a lot of analysis is done with impulse responses, MLS sequences, and transient measurements. It is friggen complicated; I don't really know shit about it, but I read quite a bit lately. The funny thing is you'll still get camps of differing opinion publishing AES papers. Who the fuck knows WTF to believe sometimes.

Interesting...the sharp peak in the impedance graph at ~8kHz matches up with a similar peak in the frequency response graph at about the same frequency. I'd suspect this is a resonance of some sort though I'd need to see a cumulative spectral decay plot to be certain.

From what I've been reading today, the first hump (around 100Hz) is the mechanical resonance of the driver --- the pendulum effect, so to speak, of the diaphragm. It's mildly inductive before the peak, at the peak it's purely resistive, and then is capacitive after the peak. Tradition says that the characteristic impedance of the cans is the floor after the first peak, so the T1 indeed shows 600 Ohms. Then gently rising slope starting at 1 kHz is the ever increasing electrical inductance of the voice coil itself.

My guess is the little glitch you see around 8kHz has to do with acoustic frequency shaping. I'm coming to the conclusion that you want a low response between 1-2kHz to keep things from sounding harsh, and then you want a bump somewhere between 5-8kHz to bring out air and imaging. But that's kind of a guess on my part; none the less the glitch is either an acoustic tweak, or some kind of mechanical resonance in the headphone.

[Dreadhead]

I disagree that 120 is a negligible number compared to 600ohms. I've looked at this in the past (even getting it wrong once ) so using the same basic technique as here (Best Amp to stick between DAC1-USB and HD-650 - Page 6 - Head-Fi: Covering Headphones, Earphones and Portable Audio) I did the numbers for the T1 and a 120ohm output impedance should lower the bass curve by around 4 dB which is a bit I would say and would certainly clear up some of the bass mismatch between the T1 and the HD800.

Edit: If I could get the FR and Z chart data I could even make up a spreadsheet that plots the modified FR vs output impedance if you'd like.

Edited March 9, 2010 by Dreadhead

[Tyll Hertsens]

Edit: If I could get the FR and Z chart data I could even make up a spreadsheet that plots the modified FR vs output impedance if you'd like.

Our policy is the we don't ever distribute the actual data from our measurements, but I do feel the same desire to see some more meaningful information extracted from our data and available on the web. It'll just take time.

[nikongod]

I still want to see a FR graph at 120 ohms that can be plotted against the same beyer cans driven from ~0 ohms.

We can argue theory of what amps and headphones do when run a certain way until we hate eachother, but can a FR graph be argued? I guess it could, but maybe it shows that there is no change.

[morphsci]

Our policy is the we don't ever distribute the actual data from our measurements, ....

Hmm, maybe it is time for a non-commercial entity to get out there and do some measurements then. Putting more varied expertise on this problem could be valuable.

[Tyll Hertsens]

Hmm, maybe it is time for a non-commercial entity to get out there and do some measurements then. Putting more varied expertise on this problem could be valuable.

Indeed. But how to finance it? Would you pay a membership? How much?

[Dreadhead]

Our policy is the we don't ever distribute the actual data from our measurements, but I do feel the same desire to see some more meaningful information extracted from our data and available on the web. It'll just take time.

Tyll, thanks for letting me know. I figured that that would be the answer. Oh well

I still want to see a FR graph at 120 ohms that can be plotted against the same beyer cans driven from ~0 ohms.

We can argue theory of what amps and headphones do when run a certain way until we hate eachother, but can a FR graph be argued? I guess it could, but maybe it shows that there is no change.

Sorry buddy, I'm a bit too loaded up to pull the numbers down by hand with a ruler but I think this would be good too.

[The Monkey]

Indeed. But how to finance it? Would you pay a membership? How much?

I wonder if a hearing society would consider a research grant for such a thing.