It's an issue I (perhaps obviously) feel strongly about. Many in the industry agree the whole impedance thing is a giant mess. So, even IEM use aside, I very much hope the trend continues towards output impedances < 2 ohms so designers are free to design better headphones and end users don't have to end up with "luck of the draw" pairings when connecting compromised sources to compromised headphones.
When posts, such as yours, claim or imply higher output impedances are somehow better I like to see the factual objective evidence supporting their claim or position. Otherwise, in my opinion, those promoting higher output impedances for questionable reasons are just making things worse.
When posts, such as yours, claim or imply higher output impedances are somehow better I like to see the factual objective evidence supporting their claim or position. Otherwise, in my opinion, those promoting higher output impedances for questionable reasons are just making things worse.
RocketScientist
I am in 100% agreement with what you are saying about the desirability of low Z output impedances. It's up to the manufacturers to sort out this mess due to that outdated IEC standard. The main problem at the moment is that very few headphone manufacturers even bother to state whether they meet that standard, or give the optimum source impedance for their headphones.
SandyK
Agreed. What's surprising is it's not done more. For many DIY'ers here it's not a big deal to add some resistance. But the majority of amp/dac buyers don't even own a soldering iron. So it would be good to have the option.
And the "offer very low output impedance" part is also critical. As many designs inherently have, or require, a higher output impedance for a variety of reasons. With such designs you often can't just bypass a series resistor to lower their output impedance and have them work correctly.
RocketScientest, it may just be that people prefer the sound of the +-7db plot in your graphic.
Sure, it measures like ****, but people might like it. Also, if you are talking IEM's, I would assume the coupling from the tranducer to the eardrum is very good on acount of the fact that the two form a sealed enclosure. A lot of material nowadays is badly recorded with over pumped base (some of that rap stuff is insane).
I think the real reason we have this mess is that back when h/phones first started to get popular and headphone sockets were provided by amplifier manufacturers, they just divided the input down froim the speaker output, or just inserted the dreaded 120 Ohm resistor in series with the socket - cheap and seemed to work ok.
I can see the engineering benefit of a low output impedance design (I'm using 2.7 Ohms on the pre-amp I am working on now), but maybe we need to provide a switchable damping resistor - e.g. 2 Ohms, 15 Ohms, 33 Ohms or 68 Ohms?
Sure, it measures like ****, but people might like it. Also, if you are talking IEM's, I would assume the coupling from the tranducer to the eardrum is very good on acount of the fact that the two form a sealed enclosure. A lot of material nowadays is badly recorded with over pumped base (some of that rap stuff is insane).
I think the real reason we have this mess is that back when h/phones first started to get popular and headphone sockets were provided by amplifier manufacturers, they just divided the input down froim the speaker output, or just inserted the dreaded 120 Ohm resistor in series with the socket - cheap and seemed to work ok.
I can see the engineering benefit of a low output impedance design (I'm using 2.7 Ohms on the pre-amp I am working on now), but maybe we need to provide a switchable damping resistor - e.g. 2 Ohms, 15 Ohms, 33 Ohms or 68 Ohms?
OK , I'm ready to do some simple tests. Add a series resistor and short it out to see the difference real time.
Based on what's been discussed about output impedance , I can try 120 ohms vs 2.2 ohms ( the present Zout). I have 32 ohms / 60 ohms/ 300 ohms hedphones.
Has anyone tried this already ?
Based on what's been discussed about output impedance , I can try 120 ohms vs 2.2 ohms ( the present Zout). I have 32 ohms / 60 ohms/ 300 ohms hedphones.
Has anyone tried this already ?
@Bonsai, I agree some might like it, but many don't--especially with IEMs as often have very "sharp" (narrow) impedance swings which then create narrow peaks or notches in the frequency response with a higher output impedance. If you've ever played with a parametric EQ (with variable Q) you know the effects of such peaks or notches can sound rather unnatural. I also agree a switched (or even internal jumper) option is a good solution. The user Peranders on here has designed a high quality headphone amp and he's included a jumper to have the output impedance be either 10 ohms or 120 ohms. The minimum of 10 ohms in this case is recommended by the chip manufacture for stability of the very high speed output stage:
Peranders-qrv09-headphone-amp
@ashock, In my DAC listening test, 20-some people downloaded high quality recordings made of the output of the DACs. In one case I used the SuperFi IEMs on the 6 ohm output impedance of the NuForce uDAC-2. These were "blind" files with generic names. The uDAC-2 recording with the SuperFi's was universally disliked. Everyone who voted on it, voted it the worst even when they didn't know what they were listening to. The trial results are here (it was a very informal listening test):
DAC Listening Challenge Results
I have also personally tried 120 ohms vs zero ohms in a blind ABX and the results, not surprisingly, are fairly obvious even with regular dynamic headphones. For example, with the Beyer DT770 Pro 80 ohm studio headphones, the bass is more boomy, exaggerated, with less deep bass extension with 120 ohms in series. With the zero ohm source, the Beyer's bass is much more tight, deep, and controlled.
I eventually plan to publish more on this topic including some better run blind listening trials that include others actually listening to headphones (vs recordings of the output).
Peranders-qrv09-headphone-amp
@ashock, In my DAC listening test, 20-some people downloaded high quality recordings made of the output of the DACs. In one case I used the SuperFi IEMs on the 6 ohm output impedance of the NuForce uDAC-2. These were "blind" files with generic names. The uDAC-2 recording with the SuperFi's was universally disliked. Everyone who voted on it, voted it the worst even when they didn't know what they were listening to. The trial results are here (it was a very informal listening test):
DAC Listening Challenge Results
I have also personally tried 120 ohms vs zero ohms in a blind ABX and the results, not surprisingly, are fairly obvious even with regular dynamic headphones. For example, with the Beyer DT770 Pro 80 ohm studio headphones, the bass is more boomy, exaggerated, with less deep bass extension with 120 ohms in series. With the zero ohm source, the Beyer's bass is much more tight, deep, and controlled.
I eventually plan to publish more on this topic including some better run blind listening trials that include others actually listening to headphones (vs recordings of the output).
RocketScientist, I am in the process of building a "Panda" amp, a Class A design from china that was discussed on RockGrotto. Many members who have K701s commented that it was the best sound they had ever heard from the K701's, particularly with an output impedance of around 100-120 ohms. I think someone mentioned before that the 701's were designed around the IEC standard, so this makes sense. I have some K601's that I suspect will get the same benefit from the power that the Panda can provide, and I have wired 100 ohms in series with the output on their recommendation. I am also building an adapter cable with 100 ohms inline that I can use to compare the SQ of some of the other amps I am building, using the K601's to see which way they sound better..
MrSlim, I agree when headphones are designed for 120 ohms (or some other value significantly greater than zero). I know Sennheiser, Grado, Ultimate Ears, Westone, and others design mainly for zero ohms.
The K701's are known for having relatively accurate flat bass performance. If they were designed for 120 ohm outputs, then they'll be overdamped on a zero ohm source and the bass will likely sound thin. Even if they're designed for zero ohms, some might prefer the "warmer" bass you'd get with 120 ohms in series.
With a headphone like the Beyerdynamic DT 770-80's, with their significantly bumped up low frequency response, the added upper bass boost (and loss of control) from a higher output impedance generally makes for worse sound.
So it's really a mess and that won't change unless the industry moves to zero ohms as a defacto standard. In the meantime, resistance options are the best solution--especially for those of us with many different headphones.
The K701's are known for having relatively accurate flat bass performance. If they were designed for 120 ohm outputs, then they'll be overdamped on a zero ohm source and the bass will likely sound thin. Even if they're designed for zero ohms, some might prefer the "warmer" bass you'd get with 120 ohms in series.
With a headphone like the Beyerdynamic DT 770-80's, with their significantly bumped up low frequency response, the added upper bass boost (and loss of control) from a higher output impedance generally makes for worse sound.
So it's really a mess and that won't change unless the industry moves to zero ohms as a defacto standard. In the meantime, resistance options are the best solution--especially for those of us with many different headphones.
Thanks for the link to the SolderDude paper. Much of that is similar to the impedance article on my blog. He did a better job than I did in some areas, but I also have to disagree on a few points:
He says 120 ohms has been adopted by "most headphone manufactures" and, while that may have been true 10 or 20 years ago, I don't think it's true today. It's certainly not true if you change the wording to"most headphones sold". I think the majority of headphones sold today are designed for zero ohms (or some compromise much less than 120 ohms). The iPod changed everything.
Experts in the field who have met with lots of manufactures like Tyll Hertsens, founder of HeadRoom, claim most headphones are now designed for low impedance sources. The 120 ohm standard didn't account for the massive growth of high quality portable audio and was meant more as a professional (i.e. studio) standard. That's why companies like Grado ignore it and many others are doing the same.
I don't agree the sole reason to lower output impedance is for battery powered gear as he implies. It makes vastly more sense as a standard for all headphones and all sources. Zero ohms eliminates the massive compatibility mess we have now. And it also allows headphone designers to count on electrical damping from the source which lets them design significantly better headphones. This is something speaker designers have had the luxury of doing for 40 years. We should give headphone designers the same ability instead of being handicapped by a 120 ohm (or, in reality, totally unknown) source impedance.
He said output impedance has "very little to no" influence on the damping factor of "high ohmic" headphones and that's misleading. What he says is true only if the output impedance is < Headphone Z/10 (i.e. with 250 ohm headphones the output impedance could be up to 25 ohms, but not the standard 120 ohms). That's often not the case. My 80 ohm Beyer's are a good example of a plainly audible difference in damping factor. And there are 250 headphones that also change audibly with less damping from a 120 ohm source.
But the rest of it is good info. I think this is a highly misunderstood topic (less so here on DIYaudio but trying reading the impedance posts over at Head-Fi!).
He says 120 ohms has been adopted by "most headphone manufactures" and, while that may have been true 10 or 20 years ago, I don't think it's true today. It's certainly not true if you change the wording to"most headphones sold". I think the majority of headphones sold today are designed for zero ohms (or some compromise much less than 120 ohms). The iPod changed everything.
Experts in the field who have met with lots of manufactures like Tyll Hertsens, founder of HeadRoom, claim most headphones are now designed for low impedance sources. The 120 ohm standard didn't account for the massive growth of high quality portable audio and was meant more as a professional (i.e. studio) standard. That's why companies like Grado ignore it and many others are doing the same.
I don't agree the sole reason to lower output impedance is for battery powered gear as he implies. It makes vastly more sense as a standard for all headphones and all sources. Zero ohms eliminates the massive compatibility mess we have now. And it also allows headphone designers to count on electrical damping from the source which lets them design significantly better headphones. This is something speaker designers have had the luxury of doing for 40 years. We should give headphone designers the same ability instead of being handicapped by a 120 ohm (or, in reality, totally unknown) source impedance.
He said output impedance has "very little to no" influence on the damping factor of "high ohmic" headphones and that's misleading. What he says is true only if the output impedance is < Headphone Z/10 (i.e. with 250 ohm headphones the output impedance could be up to 25 ohms, but not the standard 120 ohms). That's often not the case. My 80 ohm Beyer's are a good example of a plainly audible difference in damping factor. And there are 250 headphones that also change audibly with less damping from a 120 ohm source.
But the rest of it is good info. I think this is a highly misunderstood topic (less so here on DIYaudio but trying reading the impedance posts over at Head-Fi!).
Last edited:
@ Javin and gcl,
Zero output impedance is always best, but for commercially available HP amplifiers, they would have constant failures from shorted cable, HP drivers and connection to leaf blowers!
Here's the real deal:
1. Any output impedance/resistance will dissipate some of the amplifier power. Example:
Amp output voltage swing through 120 Ohms into a 120 ohm headphone (let's assume that is DC Resistance), then 1/2 of the power amp voltage swing is dropped in the HP amp 120 ohm series resistor. Additionally, that resistance DOES effect the frequency response of the complex impedance of the headphone.
The reason for output resistors as mentioned is to protect the HP output amplifier device from a short circuit - which causes smoke! Yes, if the smoke escapes, the device no longer works
I designed a headphone system called Hear Back for Studio, and live applications.
I did quite a bit of research in the area of HP amplifiers and headphones used in recording studios. The problems I was faced with was output power vs. HP impedance.
For example the popular AKG K-240 600 Ohm headphones will handle a watt of power. Ohm said P = sq rt of PxR or 0.2W x600ohm = about 11 volts...hmmm, won't happen using +/- 15 volt rails and an output series current limiting resistor, because we still have to swing 22 volts at the headphone driver!
I selected a OPA551FA +/- 30 V rail amplifier. A great amp with thermal protection and flag to indicate a fault!
Now with a small current limit series output resistor, I could realize over 200mW into 600 Ohms using higher than normal voltage rails, I believe I selected +/- 18 volts.
NOTE: at a lower impedance current is the concern, at a higher impedance, voltage is the dominate concern. Thank you Mr Georg Simon Ohm!
I hope this helps my fellow DIY users.
I think you are both basically saying the same thing...just from different points of view
Zero output impedance is always best, but for commercially available HP amplifiers, they would have constant failures from shorted cable, HP drivers and connection to leaf blowers!
Here's the real deal:
1. Any output impedance/resistance will dissipate some of the amplifier power. Example:
Amp output voltage swing through 120 Ohms into a 120 ohm headphone (let's assume that is DC Resistance), then 1/2 of the power amp voltage swing is dropped in the HP amp 120 ohm series resistor. Additionally, that resistance DOES effect the frequency response of the complex impedance of the headphone.
The reason for output resistors as mentioned is to protect the HP output amplifier device from a short circuit - which causes smoke! Yes, if the smoke escapes, the device no longer works
I designed a headphone system called Hear Back for Studio, and live applications.
I did quite a bit of research in the area of HP amplifiers and headphones used in recording studios. The problems I was faced with was output power vs. HP impedance.
For example the popular AKG K-240 600 Ohm headphones will handle a watt of power. Ohm said P = sq rt of PxR or 0.2W x600ohm = about 11 volts...hmmm, won't happen using +/- 15 volt rails and an output series current limiting resistor, because we still have to swing 22 volts at the headphone driver!
I selected a OPA551FA +/- 30 V rail amplifier. A great amp with thermal protection and flag to indicate a fault!
Now with a small current limit series output resistor, I could realize over 200mW into 600 Ohms using higher than normal voltage rails, I believe I selected +/- 18 volts.
NOTE: at a lower impedance current is the concern, at a higher impedance, voltage is the dominate concern. Thank you Mr Georg Simon Ohm!
I hope this helps my fellow DIY users.
I think this is very misleading. The vast majority of IC devices used to drive headphones, at least these days, have internal thermal and current limit protection. A few have constraints over maximum currents and/or voltages, but those are often well outside how they're being used in a typical headphone amp.
It's also very simple, and cheap, to add short circuit protection to discrete output stages as is done in 99% of audio power amps. These circuits can be designed with a threshold that's unlikely to ever be activated in normal use and hence don't compromise the performance.
And even for IC's that don't have sufficient internal protection, there are other ways to protect the IC. You can monitor the peak current, for example, and disable the power supply or use a power supply that has inherent current limiting.
It's entirely possible to design an ultra low distortion, ultra quiet, very fast, high current, near zero output impedance, and even relatively inexpensive amp that will survive any short circuit you can throw at it. So, in my opinion, there's little excuse for tossing a big resistor in the output as short circuit protection with all the sonic compromises that involves.
The National LME49600 Reference Design is an example of just such an amp. It will deliver 250 mA of current (or 500 mA if you parallel two of them) at levels of distortion that are so low they defy measurement. You can buy $99 complete audiophile headphone amps that use these and similar devices.
EDIT: It's also worth pointing out if you want DC output protection (to avoid turning your $1000 Audeze LCD-2's into junk if your amp, or even something upstream of it, should malfunction) that circuitry can also include over current protection to protect the amplifier as well as the headphones.
Last edited:
Hey Rocket Man,
This is the second reply today I've seen today where you make the statement - "this is very misleading"...just sayin'
The LME49600 looks like a pretty nice solution...as for output protection in power amplifiers I assume you're speaking of some sort of crowbar system and current monitor that prevents the amplifier from doing more than it can. While that has been done for years those are not always space or economically feasible. I assume you haven't put designs out there that require competitive market analysis and general mass distribution.
There are always trade offs where cost is crucial IF you intend on serving a world market. It's either esoteric gear and priced accordingly, or it reaches the mass market by being very good quality and competitive.
Given the 1K price of $4.5 each for the LME49600, let's assume a bridged HP amp using two of those, a pot/knob, op amp front-end, power supply, I/O connectors, PCB, chassis, not to mention protection circuitry, tech support, User's guide, etc., I doubt you can find one like that for $99.00, unless all the other parts are sub- standard quality. The $99 "audiophile units" sound more like a $99 urban myth, if they cover all your claims for the $1000 Audeze LCD-2's you speak of!
For the record, I said nothing about doing this...quite the opposite!
Pretty much, yeah
Before modern solid state amps, the output impedance of tube amps was, as I understand it, rather variable (pardon the multi-tapped transformer pun) much like headphone outputs today. Some used clever feedback schemes yielding relatively low output impedances, and some did not.
I'm not a vintage tube amp expert but I have worked with some "old time" speaker designers. When some tube amplifier output impedances were often in the same ball park as the dc resistance of the voice coils that changed the damping and system Q. That, in turn, compromised the overall design of some speakers--just like headphones are sometimes compromised today.
Once solid state amps were widely powerful, reliable, and with uniformly low output impedance, the first really good "acoustic suspension" sealed speakers like the original AR's, Large Advents, etc. were developed. They had deeper and better bass than any of their tube powered predecessors could manage from a similar box size. It was a big milestone in "hi-fi". We just have to get the headphone world, after 40 years of impedance confusion, to catch up.
I have done several commercial products that had all the usual economic trade offs. The LME49600 is very close in price to the OPA551 you mentioned using in your commercial design. And this is DIY audio where people are not usually looking to save every last penny.Hey Rocket Man,
This is the second reply today I've seen today where you make the statement - "this is very misleading"...just sayin'
The LME49600 looks like a pretty nice solution...as for output protection in power amplifiers I assume you're speaking of some sort of crowbar system and current monitor that prevents the amplifier from doing more than it can. While that has been done for years those are not always space or economically feasible. I assume you haven't put designs out there that require competitive market analysis and general mass distribution.
There are always trade offs where cost is crucial IF you intend on serving a world market. It's either esoteric gear and priced accordingly, or it reaches the mass market by being very good quality and competitive.
Given the 1K price of $4.5 each for the LME49600, let's assume a bridged HP amp using two of those, a pot/knob, op amp front-end, power supply, I/O connectors, PCB, chassis, not to mention protection circuitry, tech support, User's guide, etc., I doubt you can find one like that for $99.00, unless all the other parts are sub- standard quality. The $99 "audiophile units" sound more like a $99 urban myth, if they cover all your claims for the $1000 Audeze LCD-2's you speak of!
The $99 amp is very real and even included three expensive chips for a USB DAC! The original ("mark I") version of this amp used a pair of LME49600's (one for each channel) and sold for $99. The manufacture has since switched to a single TI TPA6120 which is also short circuit protected. Here's the info including a good photo of the rather nice looking PCB:
HA Info NG98 MKII
You can buy this amp/dac, delivered for free, for $99 here:
NG98-MKII-USB Headphone Amplifer/DAC
If you'd prefer just an amp, complete with Li-Ion battery for portable use, using the LME49600 try this:
Little Dot Mark I Headphone Amp
Those don't have an output relay for DC protection, but this $55 amp does:
HiFi Headphone Amplifier
There's no need to bridge the LME49600 for any sane application. That was discussed in the LME49600 thread I started. So 99% of the world only needs 2 of them.
I'm not saying the above Chinese examples are the highest quality designs on the planet, but you could remove all the expensive DAC hardware (PCM2704, DIR9001, AD1852), substitute DC protection, and lower the price to probably $89 or even less. So it obviously can be done for $99.
I have an NG98 MKII on the way for extensive testing and blind listening. I'm looking forward to seeing just how much you get for $99 delivered to your door ready to plug in.
EDIT: And for those who might argue none of the above examples qualify as "audiophile", it's worth noting the Chinese Little Dot MK III earned a spot as a prestigious Class B component in Stereophile's annual Recommended Components list.
Last edited:
Stereophile's publisher has a new site that deals with head-fi exclusively; it's called "Inner Fidelity". It's basically a one-man show by writer Tyll Hertsens. Whether it is intended to remain online only or will actually turn into a news stand mag is unknown, but it's online now in any case.
Anyway, there you find the comparison of three models of the Beyer Dynamic DT880 which comes in three impedance versions (32, 250, 600). A lot of the issues raised so far in this topic are addressed to a certain degree.
The review can be found at:
A Comparison of Beyerdynamic DT 880 32 ohm, DT 880 250 ohm, and DT 880 600 ohm Headphones | InnerFidelity
The site itself which has some other interesting articles beyond simple headphone and phone amp reviews is here:
Home Page | InnerFidelity
(Currently the Beyer Dynamic DT880 review is prominent on the first page, but that's likely to change. So, right now (15Apr2011) you could alternately just go to the main page and find the article mentioned above. Your choice).
Anyway, there you find the comparison of three models of the Beyer Dynamic DT880 which comes in three impedance versions (32, 250, 600). A lot of the issues raised so far in this topic are addressed to a certain degree.
The review can be found at:
A Comparison of Beyerdynamic DT 880 32 ohm, DT 880 250 ohm, and DT 880 600 ohm Headphones | InnerFidelity
The site itself which has some other interesting articles beyond simple headphone and phone amp reviews is here:
Home Page | InnerFidelity
(Currently the Beyer Dynamic DT880 review is prominent on the first page, but that's likely to change. So, right now (15Apr2011) you could alternately just go to the main page and find the article mentioned above. Your choice).
Last edited:
Thanks, I didn't know the site went live. Last I knew it was still in "preview" mode and I thought scheduled for May. But, AFAIK, Tyll and I agree on most issues regarding impedance. He's a good guy and I think the site is a great idea--especially his plan to conduct lots of objective measurements. He has an Audio Precision analyzer and can do a lot. For those who don't know, Tyll founded HeadRoom and ran it up until last year I believe.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.