[ppl]

Tomo:
While most Headphones state the sensitivity similar to what Sony dose on my MDR-7506 and MDR-v6 as (106dB/Mw) this dose not imply 106 dB per Mw, rather it states 106dB@1Mw. Thus these Phones should produce 116dB@10Mw since as PRR states power is 10 log, while Voltage is 20 log. gives a more easy to understand spec. why thay do not spec Headphones like Loudspeakers are i can't say.

PRR:
No i have never put close to 1 amp into my can's in fact since My most powerfull portable amp can output just over 7 Volts RMS into a 50 ohm load i could not evean get close to 1 amp through the cans at pull output. Most of my listening is done at under 0.5 Volts and for the most part at about 100-200 Mv. this one would think should be well within the range that most opamps can Linearly drive. real world Listening test at these levels show quite different results. these phones operating in the 100-200 Mv range powered from an opamp able to output 80 Ma into a short circuit sound Thin and bright to me. maby that impedance rise you mentioned. If i power these from an Opamp Buffer combo with the Buffer able to output 160 Ma into a short circuit the sound is alot better More full and warm with more control over the bass, however the bass still is not ideal since it is slightly bloated. By using buffers able to output 1.8 amp of current (EL-2008) the Bass gets real tight and quick. the Midrange takes on a nice sence of ambience that was missing with both the opamp alone or the 160 Ma buffer (EL-2001) and this is at low levels at higher levels the bass on just the opamp compleatly falls apart with total loss of control of the Driver. a scope monitoring this did not reveil any clipping. Using the El-2001 the Bass at high levels got alot better but was on the verge of losing it. the midrange also gets muddy as like the bass was modulating the Mid's. going to the EL-2008 and at the same high level 0.5-1.0 Volt Everything is great with nice and tight bass as well as the bass not messing with the Mids. My old Koss pro4AA rated at 220 ohms sound better with a good buffer than with just an opamp, but i must addmit it takes only (1) El-2001.

Based upon my own listening test as well as measurements I have not found that just an opamp alone without some form of buffering to be sonicaly ideal. I look at headphone Amp design as it were a low powered Audio power amp so like any power amp should have an outout stage after the gain stage. now it is true that most opamps include an oiutput stage folling the voltage gain stage and thus an opamp could be considered a lowpowered monolithic Amp. so why follow this with an evean higher Current output stage. well like you stated the real open loop impedance of over 100 ohms vs a real open loop impedance of less than 10-20 ohms. But moreover the most critical reason for buffering is to Improve the Accuracy of the opamp and to remove load dependent tempature changes that can drmaticaly affect the Performance of the opamp. Opamp makers know this and that is why Buffers were made.

To Conclude! It may be true that in tomo's case by using a opamp able to output more than 100 ma and by using 600 Ohm phones most of the Nonlinearities may be removed but the tempature effects are not. In addition consider the back EMF of the phones and what this dose to the feedback and why a low real output impedance is required. But also in tomo case finding a buffer to Follow a 450 MHz. Opamp is quite a task unless you want to use the Buffer outside of the opamp's feedback loop. in that case a Low distortion Buffer is required. and one of the lowest distortion open-loop buffers i have found is the OPA-633/HA-5033.

 

[PRR]

> calculation to determine voltage requirements for your headphones.

How much voltage?

Depends on headphone power sensitivity, impedance, and desired listening level, of course.

On HeadRoom's site, they propose a level of 100 dB SPL as a guide, and publish the voltage needed to reach this level for the headphones they sell. 100 dB SPL is perhaps arbitrary: we know many people listen in the mid-80s, and others need more than 100 dB SPL. However, 100 is enough for most folks, and those who need more can do the math for their level.

Sadly, something in HeadRoom's webserver is busted: all the values are "0.0V". So I computed numbers from Rane's paper posted on HeadWise.
Quote:

Conclusion? If 100dB SPL is your goal, any headphone on this list (or similar) will reach that level with 3 Volts and 26 mA RMS (4.2V 40mA Peak). This is within reach of small chips working on 9V batteries.

If 100 dB SPL is not enough, you probably want to look for 110 dB SPL. Triple the numbers above (3^2=10, near enuff). 10 volts and 80 mA may be needed.

-PRR

 

[ppl]

PRR:
Thanks for redoing rain's list to include Recent cans. Also i would like to point out that these are RMS levels and peak levels considering a 10 dB dynamic range of most music and Hope-fully more than that on Audiophile recordings, then the values you listed for power, will have to be 10 times larger. Also consider that the current limit on most Chips limit rather quickly so evean peaks will be limited.... Just a thought.

 

[PRR]

> PRR: Thanks for redoing rain's list to include Recent cans.

Well, but I didn't add anything to Rane's list. I did shorten some model numbers to fit.... if by accident they now look like new-models, it should be understood that they come from Rane's list. I can't find info on modern cans.

I am amazed at how many model numbers on that 20 year old list are still available. True, there may be refinements between a 1984 HD-600 and today's HD-600, but I suspect the gross numbers for impedance and sensitivity are about that same today.

> these are RMS levels and peak levels considering a 10 dB dynamic range of most music .... the values you listed for power, will have to be 10 times larger.

I understood HeadRoom's words to mean: "average" listeners should be sure the voltage needed to drive a phone to 100dB SPL is less than the maximum voltage available from the jack.

Allowing a little peak clipping for best battery utilization, we could say 10 dB headroom. Meaning HeadRoom thinks that 90dB SPL average will satisfy most buyers.

I agree: 90 isn't soft and won't disappoint most listeners. More might be nice, but in limited-voltage duty such as portable battery players, we can't go overboard.

However, most of HeadRoom's "Volts for 100dB SPL" listings are blank, and I could not find where they say what the typical output voltage is for different mass-market devices.

I computed some numbers for the headphones. And the most common portable player is the 2-AA cell WalkMan-type, which must make 0.5 to 1.0 volts available at its headphone jack.

The 32-ohm Sony phones probably work fine (loud) on 2-AA WalkMen. They need 0.1V to 0.3V to make 100dB SPL. The AKG 240 and its ilk are known to be unsatisfying on 2-AA WalkMen, and they need 3 volts to make 100dB SPL. So HeadRoom's 100dB SPL guide is probably a good starting place.

In live recording work I often need 110dB SPL (5 volts into AKG240 is not quite enough). In a quiet office, 75dB SPL average 85 peak might be ample for many listeners. So it is only a benchmark, not a golden ruler.

> Also consider that the current limit on most Chips limit rather quickly so evean peaks will be limited....

Current limiting on small chips is normally FAR faster than any audio transient. The only way to design is a chip that can make full current with ease. And many chips will get nasty at currents less than the limit.

Of course things like R-load or CCS-loaded SE amps just won't make more than a certain amount of current. (Tomo's thoughts grew out of a thread about such an amp.)

Power chips may also have heat limiting. They may be slower. You can sometimes cheat them for an audio peak. But when they limit the amp goes wild or dead for many milliseconds. So we don't want to go there for good listening.

-PRR

 

[ppl]

PRR:
Thanks for the clarification on Headrooms avg. 100dB Level, We seem to be in agreement on the Current limit issue and i would like to add that I have found that All IC output stages are sensitive to Heat and will limit sooner at higher temp's. the Burr Brown BUF-634 is a clasic example of this and is so stated on the data sheet, moreover I also agree that most IC output stages start to get real nasty at currents way below their max Current ratings. This may be yet another reason that high current Output stages sound better to me than low Current ones.

Distortion also starts to get really high at currents more than 1/4 the Rated output Current. I measured the BUF-634 as a stand alone Buffer and the THD + Noise went from a nice 0.01% @ 20 Ma of output current @ 5 Volts RMS output to just under 1.0% at the rated 250 Ma of output Current. note>> the load impedance had to be reduced to get 250 Ma Flowing through the Load resistor. Evean at 50 Ma of output current & 5 volts RMS output the THD+ Noise was 0.6% also note that what could be considered exsessive heatsinking was needed to get these numbers (200 Sq. In.) for the To-220 package. I wouls suspect that similar results would also happen with other Buffers as well. Also i would suspect that Opamps would be worse behaved under High load conditions relitive to there rated output, However the Opamps would have the Bennifit of feedback so actual distortion numbers might be lower due the feedback reduction.

This may be the reason that most folks seem to like Discreet transistor's as an output stage rather than in Monolithic form. I my-self also like these better. I also believe that current limit Hinders Performance on all aspects as well as on sound quality, however Current limit is a nessisary evil as every time the Phone plug is inserted or removed the Output is shorted and a designer can't expect the User to stop playing music prior to Pluging in or unplugging the Phones.

 

[zorin]

Can somebody wake up from the long hibernation ?
What voltage would an amplifier need to supply to drive 600 Ohm Beyerdynamic T1 ? The telling feature of these headphones is a wide impedance spread between their nominal impedance  of 600 Ohm and a peak impedance of around 1250 Ohm of the bass end, centered on 100 Hz.
 

 
So an amplifier would need to supply enough voltage to deal with the 1300Ohm peak. What would this voltage need to be ?
+13V to 15V / - 13V to 15V for  a peak voltage of 26 to 30 Volts ?

 

[jcx]

the frequency response is usually measured with low Z Vsource - the amount of current at any frequency is determined by the Headphone Z at that frequency, so the power "absorbed" by the headphone varies verses frequency with constant V drive - but to no consequence for SPL vs frequency which is controlled by V alone if the amp output Z is low compared to the headphone
 
http://www.innerfidelity.com/images/BeyerdynamicT1.pdf
 
with the T1 spec of 600 Ohm nominal Z and 102 dB re mW sensitivity/efficiency you need ~ +/-1.1 Vpk for 1 mWrms, giving 102 dB SPL (they state a 500 Hz ref frequency)
 
for 120 dB SPL you would need ~+/-8.7 V from your amplifier output, with ~  +/-15 mA where the Z is close to 600 Ohm, but only 7 mA would be needed at the 100 hz Zpeak of the T1
 
many monolithic op amp IC can manage that I, V when powered from more than +/-12 V supply, although some may show a little more distortion at 15 mA
 
I'd look for those rated with at least 2x the output current or parallel a couple "A47" or O2 style
 
the O2 output op amp has highish output current drive ability so just one JRC/NJM4556 would be happy with 600 Ohm or higher load  but the  dual "9V" battery may not be enough supply V if you really want >115 dB SPL clipping free
 
some discrete SS amps, or buffers for op amps used in headphone circuits can't swing very close to their supply rails

 

[zorin]

Thanks for replying,....I am in a conversation with an amplifier manufacturer, he makes a very good transportable 'class A' hybrid amp but this amp's driving voltage is 5V, so its headphones optimal top driving range is 300 Ohm. - http://www.analog2p.com/product/tur06.html - By my reckoning this voltage is not big enough to deal with T1's peak impedance of 1250 Ohm [as measured by en.goldenears.net] / 1400 Ohm [as measured by innerfidelity.com] centered at 100Hz. He told me he can modify the amp in order to raise the driving voltage to 12V. I worry if this voltage is enough. By my estimate, as I do not have enough knowledge to do the calculations, a voltage of 13 to 15 Volt would be needed. If you say 12 V is enough I'll tell him to go ahead and make an amp for me. Since this amp would have to drive well both 42 Ohm W3000ANV which needs big current and 600 Ohm T1 which needs big voltage I think of suggesting to him doubling the number of op-amps in a 'cascade' mode. ALO Audio did this with their RX-Mk3B portable amplifier. - http://www.aloaudio.com/rx-mk3-b-plus - I just do not know enough details about Rx-Mk3B if those 'doubled' op-amps are of the same kind, to deal with either increased current or voltage requirements, or if they are different ones to deal, each with either current or with voltage requirements. What do you think about doubling op-amps to deal with higher voltages ? In a matter related to this, I saw a bit of an information that 'the Continental', a portable amp made [discontinued since 2013] by  ALO Audio to specifically drive Bayerdynamic T1, has an "output voltage" of 26V. This number seems to be too high to be +26V / -26V.