[Ducker]

Quote:

Last 2 posters, this is a work on progress as I think of where the best spot for attenuation is, try other amps, and make more progress.  For instance, I need more attenuation for my system, so I have been loading up on the series resistor before the phones.  This brings the amp to a higher power point, too, so I am changing 2 things: output of the amp, and impedence the headphones are seeing.  However, the amp is still seeing the same impedence load. Changes in sound could be from running the amp at a watt instead of 1/4 watt, for instance.  Or from running the phones from a current source, where the current will be constant in the face of changing impedance.  This is neither right nor wrong, depends on the headphone, how it was developed/tested, and the listener. 
 
In a way, part of the different sound of different amps into a particula phone would have to do with the output impedance, not quality of amp.  It could be an advantage in having enough power to try the amp both ways, as a voltage source and as a current source. 
 
Since this is not a finished work (though I love what I heard this morn, running the HD800 from a current source tames the high end and brings out the bottom, as well as richens the mids) I don't want to be misleading readers into thinking this is a finished work with definitive answers.  Many forum guys make one experiment and issue dogmatic statements.  It takes a long time and a lot of testing and measuring to really figure out what is going on and black and white answers are often not available.  For that reason, int might be best to pm me and I'll let you know what I'm trying and how it works.  I might be able to give you some direction based on the headphones you are using.  No promises.......!  

Operakid,
Have you made any more progress?
I'd like to try this with a vintage amp.  Seems very strange to me that more people aren't trying this.
 
Thanks.

 

[Operakid]

Hi Ducker, I sent a PM so as not to go into great detail boring everyone.
 
In general matters about this, I have made a balanced cable for the 800s, using a proven cable we developed, which tests extremely flat, has no overshoot, and a good, flat impedance curve.  Most important, having separate grounds by virtue of being balanced, I can use it on any amp, whether it be dual mono, bridged, whatever.  
 
A quick listen yesterday (one hour of back and forth listening comparing stock cable to this) showed wider imaging, more natural highs, more powerful deep bass.
 
I can now try different speaker amps I have on hand tro see if their character into speakers is mirrored by their character into cans.  It will also be interesting to see if some are noisy with cans, or any other negative about using speaker amps with cans. 
 
As far as any other conclusions, further listening over the last week confirms to me that at least with the big tube amp I have been using I'd rather listen to this (admittedly expsive) set  of praemp  and speaker amp than the very well regarded headphone amps I have.  Certainly they are both huge compared to any headphone amp setup, but for folks with good electronics already, or who want to buy good electronics on the used market, this might be the ticket. 
 
Next I will try some of my favorite solid state amps, now that I have the balanced harness, and report those findings.  

 

[Ducker]

This is really interesting.
I'd like to buy a vintage amp and use it to drive headphones for my desktop setup.
 
Haven't seen the pm??
 
Thanks and please keep this thread going.

 

[Happy Camper]

The orthos got me trying speaker amps and I shake my head at the money spent on headphone amps.

 

[playitloud]

Quote:

Hi Ducker, I sent a PM so as not to go into great detail boring everyone.
 
In general timers, I have made a balanced cable for the 800s, using a proven cable we developed, which tests extremely flat, has no overshoot, and a good, flat impedance curve.  Most important, having separate grounds by virtue of being balanced, I can use it on any amp, whether it be dual mono, bridged, whatever.  
 
A quick listen yesterday (one hour of back and forth listening comparing stock cable to this) showed wider imaging, more natural highs, more powerful deep bass.
 
I can now try different speaker amps I have on hand tro see if their character into speakers is mirrored by their character into cans.  It will also be interesting to see if some are noisy with cans, or any other negative about using speaker amps with cans. 
 
As far as any other conclusions, further listening over the last week confirms to me that at least with the big tube amp I have been using I'd rather listen to this (admittedly expsive) set  of praemp  and speaker amp than the very well regarded headphone amps I have.  Certainly they are both huge compared to any headphone amp setup, but for folks with good electronics already, or who want to buy good electronics on the used market, this might be the ticket. 
 
Next I will try some of my favorite solid state amps, now that I have the balanced harness, and report those findings.  

Do not be afraid of boring everyone. There is quite a lot of information of low impedance headphones (Hifiman etc) on speaker amplifiers. There is NOT a lot of information on high impedance headphones on speaker amplifiers. Many headfi-ers do not try is because of the impedance mismatch. I happen to be very interested in this topic because I only have a speaker amplifier (see my profile). I am planning to use the Sennheiser HD 800 on my speaker amplifier as well. However, I am still waiting on the cable that allows me to hook up the phones to the 8 ohm speaker terminals. Any information on how to maximize the results of the Sennheiser HD800 on the speaker amplifier are much appreciated. Of course, if would be very glad if you can send me a pm with detailed information. However, I do think that there are more headfi'ers struggling with this issue than you might think. So , why not share it in this thread. Maybe you can change the title into: high impedance headphones and speaker amplifiers. I think that covers the topic better, and addresses something that has not been addressed before in detail. 

 

[Ducker]

Quote:

The orthos got me trying speaker amps and I shake my head at the money spent on headphone amps.

This is what I want to know.
Nothing wrong with buying headphone amps.  I've owned a couple and probably will buy more.  But I'd like to know how speaker amps and headphone amps compare in performance and cost.
I'd like to know what my options are.

 

[Operakid]

OK, I won't be afraid of boring everyone.  
 
I like to work out details and finalize experiments before stating things as facts.  30 years in the audio industry has resulted in my getting fed up with quick opinions of "gurus and experts" based on quick experiments they have done, incomplete auditions, and their seat of the pants, elevator physics opinions, with no physics or engineering backgrounds.  I learned and practice the "scientific method" in my physics and engineering education, part of which is to change one and only one thing at a time.  It takes a long time to do this, try combinations, measure, test, watch waveforms.   So I'm leery of spouting off here before I have checked a wider variety of amps.
 
This much I know: my favorite tube amp is bettering my headphone amps into dynamic headphones and bettering my Stax amps into my Stax cans, with deathly silent background, big dynamics, and very, very refined sound.  Far less between me and the performer.
 
More to come as I run more tests.  Unfortunately, my day job of being an audio manufacturer, coming out with more products, etc. trumps my work with headphones.  It started as another way to listen to my products, to evaluate my work, in addition to my usual suspects in speakers for testing.  I could quit now, being satisfied with my headphone setups, which indeed show me more in certain respects than any speaker I've ever heard.  Except, my curiosity has been piqued about how things have always been done and alternative ways of doing them with headphones.  So I'll keep experimenting, obtaining different transformers (for Stax), making different voltage dividers for dynamic cans and seeing where it leads.  

 

[playitloud]

Quote:

OK, I won't be afraid of boring everyone.  
 
I like to work out details and finalize experiments before stating things as facts.  30 years in the audio industry has resulted in my getting fed up with quick opinions of "gurus and experts" based on quick experiments they have done, incomplete auditions, and their seat of the pants, elevator physics opinions, with no physics or engineering backgrounds.  I learned and practice the "scientific method" in my physics and engineering education, part of which is to change one and only one thing at a time.  It takes a long time to do this, try combinations, measure, test, watch waveforms.   So I'm leery of spouting off here before I have checked a wider variety of amps.
 
This much I know: my favorite tube amp is bettering my headphone amps into dynamic headphones and bettering my Stax amps into my Stax cans, with deathly silent background, big dynamics, and very, very refined sound.  Far less between me and the performer.
 
More to come as I run more tests.  Unfortunately, my day job of being an audio manufacturer, coming out with more products, etc. trumps my work with headphones.  It started as another way to listen to my products, to evaluate my work, in addition to my usual suspects in speakers for testing.  I could quit now, being satisfied with my headphone setups, which indeed show me more in certain respects than any speaker I've ever heard.  Except, my curiosity has been piqued about how things have always been done and alternative ways of doing them with headphones.  So I'll keep experimenting, obtaining different transformers (for Stax), making different voltage dividers for dynamic cans and seeing where it leads.  

 
I most certainly agree with 'the scientific method' as I am a scientist as well (not in physics though). I my case I will use an 8 watt 300B amplifier and drive the sennheiser HD800 from the 8 ohm speakers taps.  I am aware of using resistors in parrallel to match impedance. However, what value should I use. If we take the nominal impedance of the sennheiser at 300ohm, I can use a 10 ohm/10 watt, a 15ohm/10watt, or a 20ohm/10watt non inducutive resistor. The 10 ohm resistor would result in an impedance of 9,6 ohm, the 15 ohm resistor in a 14,28 ohm impedance and the 20 ohm resistor in a 18,75 ohm impedance.  On the basis of impedance matching you could argue for the 10ohm resistor. However, that also means that more power runs through the resistor and  there is less power left for the headphone. Personally, I like to have as much power as possible running though the headphone, without suffering from an impedance mismatch between headphone and speaker amplifier.
 
So the real question is, to what extent can I deviate from the 8 ohm impedance of the speaker amplifier without suffering from impedance mismatch. 
Second question, how can you indentify an impedance mismatch. What kind of 'distortion' do you hear when there is a mismatch in impedance.
Third and last question, I am planning to use a Mills non inductive resitor in parallel. Is that OK, or can it be improved upon?
 
Operakid, I appreciate you effort in this, and value you expert answer.

 

[Operakid]

What a complex subject impedance matching can be.  Given the reasonably low frequencies we are at in audio we don't suffer from partial wave wave reflections and standing waves in the cables since wavelengths are so long.  But, we still should strive for a goal impedance that satisfies the needs of the source and load devices.  
 
The resistor in parallel with the outputs of the amps is for allowing us to use an amp comfortable with a speaker load into a much higher impedance load (headphones).  Since most amps are very happy with an 8 ohm load, something around 8 ohms would let the amp see its desired load and prevent any oscillations due to the high load impedance (cans).  For efficient headphones like the HD800 5 to 10 watt resistors should be safe.
 
As far as throwing away power with the parallel and load resistors, this is a fact of life.  However, the power waste is very minimal with efficient headphones.  for example, I am using the same volume setting that I use with my 100 db efficient speakers,  That means that at 85db I am using between 1/16 and 1/32 watt total average steady state power out of my amp.  Instant impactful peaks might be using a watt.  That's one of the beauties of using a speaker amp: even a 5 watt SET has lots of headroom into efficient headphones.     
 
Now we have a reality that the speaker amp will likely have more gain than would be practical for use with high efficiency headphones such as the HD800.  To be in a useful volume range with the preamp and to prevent blowing the headphones we can load the + line going to the headphones with a series resistor.  I started with the standard load (though many headphone jacks don't totally follow this load) of 120 ohms (I believed this is the value, no need to be exact).  For the HD800, being so efficient, a 1 to 2 watt resistor should be safe.  
 
Be aware that bridged amps and dual mono amps may destruct themselves hooking up a common ground set of headphones with a resistor network.  Common ground would be a wire set  that terminates in a headphone plug.  The plus and minus headphone grounds are connected at this plug.  There may be other amp designs, other than bridged and mono, that would destruct also.  A good check would be to check the resistance between the 2 negative speaker connectors of your amp without anything plugged into the amp.  If they are shorted together (very low resistance) then you are safe with the common ground load of the headphones.  If not, you better have a balanced cable set where each channel of the headphone has a separate ground.  
 
When using common ground headphone cables with an amplifier that has common ground outputs I run ground to the headphone connector ground from only one channel of the resistor network.    This brings ground together in the jack anyway, but doing it this way prevents any parallel resistors in the network (i.e., resistors going from + to -) from introducing quite a bit more crosstalk between left and right channels..    .  
 
Caution!  Disclaimer!  
 
Hooking up the described setup could blow your amp and/or headphones if a wrong connection is made.
 
Hooking up the described setup could blow your amp and/or headphones if the amplifier does not like its ground terminals to be connected.  Check with your amplifier manufacturer to be safe.
 
If you are not sure of what you are doing do not do it.  I'm not suggesting that anyone hook up their speaker amps to drive the headphones.  I am relating what is working for me.  
 
I also am not suggesting that folks should not buy headphone amps.  I don't need to be an enemy of those makers.  In my case, I already had perfectionist amps and preamps, some costing $40k each.  In that situation it only made sense to try to utilize them.  And it is no criticism of a $2k ore $5k headpone amp that a $20k or $30k or $60k amp and preamp set might outperform it.    

 

[playitloud]

Operakid, thank you for your answer. So you recommend to match the impedance as closely as possible. Hence, a 10ohm/10watt resistor in parallel is prefered to a 15ohm/10 watt resistor. Correct? Just out of interest, what is the difference in sound between a well matched impedance and a not well matched impedance between headphone and amplifier. Again, thank you fo your effort.

 

[Operakid]

The resistor between + and - is not going to be a "one size fits all" proposition.  The "ideal" resistor would be the one that puts your amp in the sweet spot of what impedance the output section likes best.  For one amp it might be 4 ohms, for another 6 ohms, for another 8 ohms, and maybe 16 ohms for some.  It depends on the design of the output transformer if you use a tube amp with transformers.  For an OTL amp I would think it would be higher.  Given my OTLs have about 2 ohm output impedance, which is pretty high, I will try a 20 ohm resistor across outputs on those.  Finally, beyond the "ideal load" for a given amp, it will involve selecting the tone you wish to hear.  It's like loading for a phono cartridge.  Guys who think there is one correct, ideal load are wrong, they are merely picking what works, in a sonic sense, for them in their system.
 
With my transformer type tube amp I have a 10 ohm as it was easily available.  I understand its transformers are optimized for 4 or 5 ohms so maybe that would sound better, although I would eat more power.  To be truthful, the setup sounds so good with the 10 ohm resistors I have not strayed from there.....I almost don't care if it sounds better, the results are beyond what I ever expected out of headphones as it is.  I have other amps to try, that's more important to me than optimizing any one particular amplifier when the results are already this good.  I could experiment forever if I allowed my self to, bur in reality I have achieved my original goal, which was to have other sets of transducers to be more effective in evaluating my own products.  The side benefit is that I now have additional setups that I love listening to for pleasure.  

 

[playitloud]

Quote:

The resistor between + and - is not going to be a "one size fits all" proposition.  The "ideal" resistor would be the one that puts your amp in the sweet spot of what impedance the output section likes best.  For one amp it might be 4 ohms, for another 6 ohms, for another 8 ohms, and maybe 16 ohms for some.  It depends on the design of the output transformer if you use a tube amp with transformers.  For an OTL amp I would think it would be higher.  Given my OTLs have about 2 ohm output impedance, which is pretty high, I will try a 20 ohm resistor across outputs on those.  Finally, beyond the "ideal load" for a given amp, it will involve selecting the tone you wish to hear.  It's like loading for a phono cartridge.  Guys who think there is one correct, ideal load are wrong, they are merely picking what works, in a sonic sense, for them in their system.
 
With my transformer type tube amp I have a 10 ohm as it was easily available.  I understand its transformers are optimized for 4 or 5 ohms so maybe that would sound better, although I would eat more power.  To be truthful, the setup sounds so good with the 10 ohm resistors I have not strayed from there.....I almost don't care if it sounds better, the results are beyond what I ever expected out of headphones as it is.  I have other amps to try, that's more important to me than optimizing any one particular amplifier when the results are already this good.  I could experiment forever if I allowed my self to, bur in reality I have achieved my original goal, which was to have other sets of transducers to be more effective in evaluating my own products.  The side benefit is that I now have additional setups that I love listening to for pleasure.  

OK, I will try different values of resistors to see if and in what way it makes a difference. Thanks for your reponse, and good luck with your experiments

 

[calibro]

Good evening from Italy (sorry form my american english...).
 
I use a similar setup (speaker amp on cans) and I don't use any resistor.
 
I read this discussion and I have some questions:
 
1)If I understood the resistor in series should avoid self-oscillating in the amplifier. It's correct?
 
2)A low feedback amplifier is more "robust" about self-oscillating. It's correct? How I could see the self-oscillanting in an amplifier?
 
3)The output impedance of the amplifier should be a fraction of the impedance of cans. This to avoid variations on the frequency response of the system. It. correct?
 
4) A flat impedance can (for example an ortho can) should not lead to variations on the frequency response, even with an high output impedance amplifier. It' correct?
 
5)I'm very happy with my setup (see signature) but: how could I improve the signal to noise ratio??
 
Thank you in advance for yours contributions.
 
In Italian forums the topic was discussed ad follow:
 
http://www.ilgazeboaudiofilo.com/t2938-partitore-resistivo-per-cuffie
 

 

[Operakid]

Non e' necessario scusarmi per il tuo inglese, va bene. I tuoi punti sono chiari.
 
No need to apologize for your English!  It is clear.  
 
I am using a simpler network he shows in your link, his has more versatility for manipulating the attenuation for the headphone level.  Mine simply uses the first parallel resistor and one series resistor.  He has 2 parallel resistors and 2 series resistors.  
 
In his example he lets the amplifier see a higher impedance than I do, he has higher impedance across the amplifier leads.  Still, he will be preventing oscillation, and he will be operating the amplifier nearer its intended use, just not as near as I do.  Of course, this is not related to his circuit, this is simply due to his resistor value choice.  My guess is that he wishes to waste less amplifier power.  For me, all speaker amplifiers have the power to waste, and I chose to let me amplifier see a more normal 10 ohm load and waste a little more power.  I am sure I am still using maximum 1 watt peaks,  I do not think this is a big matter.  In any case, it is easy to change out his resistor values for lower load impedance. 
 
So, you can follow his circuit, or you can use my simpler 2 resistor model, which works great for me.  With my circuit, I am forced to have the headphones be fed by a very high impedance in order to attain the desired attenuation.  To the headphones this looks to be even more of a current source driver.   I do not see the negative to this, and listening tests show no negatives.  So I will continue to use it.  Perhaps I will sometime try his circuit to see if there is any sonic advantage.  One can experiment forever, I have such good results that I will stop where I am for now.  
 
One note is that I hear no undesirable tonal shifts using my setup with the HD800.  The only changes are desired changes: a nicer treble (but nobody should be fooled, the HD800 will always have generous treble) and greater quantity mid and upper bass.  Dynamics are better than I have heard them, bass more powerful, deeper, more impactful, so I do not think the HD800 dislike seeing the high source impedance.           

 

[Operakid]

Quote:

Good evening from Italy (sorry form my american english...).
 
I use a similar setup (speaker amp on cans) and I don't use any resistor.
 
I read this discussion and I have some questions:
 
1)If I understood the resistor in series should avoid self-oscillating in the amplifier. It's correct?
 
2)A low feedback amplifier is more "robust" about self-oscillating. It's correct? How I could see the self-oscillanting in an amplifier?
 
3)The output impedance of the amplifier should be a fraction of the impedance of cans. This to avoid variations on the frequency response of the system. It. correct?
 
4) A flat impedance can (for example an ortho can) should not lead to variations on the frequency response, even with an high output impedance amplifier. It' correct?
 
5)I'm very happy with my setup (see signature) but: how could I improve the signal to noise ratio??
 
Thank you in advance for yours contributions.
 
In Italian forums the topic was discussed ad follow:
 
http://www.ilgazeboaudiofilo.com/t2938-partitore-resistivo-per-cuffie
 

 
To specifically address your quesrtions:
 
1 It is the resistor in parallel to the amplifier + and - that prevents oscillation or bad behavior on the amplifier's part.
 
2.  I do not know about different amplifier designs and their resistance to oscillation.  It can be hard to tell if an amplifier is oscillating at a high frequency.  If it is at a disastrously high level you can tell when the ampilfier melts down.
 
3.  This is a complex subject that is rarely addressed in high end audio.  Current solid state amplifiers are voltage sources.  Some speakers react better to a voltage source, some better with a current source.  I am sure the same can be said of headphones.  Part of it depends on what amplifier the designer worked with when sound-shaping the headphones.  And part will be just plain old luck.....trial and error in listening.
 
4.  That would appear to be correct.  A dead flat impedance should act the same with a current or voltage source, I believe.
 
5.  If you are using no means of attenuation between the speaker amp and headphones then you must be at a very low volume setting on your preamp and this will lead to a poor signal to noise ratio.   The resistors will turn some power into heat, causing you to increase the volume setting, this will take you further away from the noise floor of the components and you will hear less of that noise (for a greater signal to noise ratio).  You will also be protecting your headphones from damage.  And, if you are not using the parallel resistor across your amplifier terminals your amplifier will be happier driving a lower impedance load if you add that resistor.  
 
If you are using a headphone jack, when you make the voltage divider you should only use one ground wire whether that be for left or right channel.  Grounds (negative lead to headphone) are summed at the jack and this compromises stereo separation.  Possibly this is why he uses larger parallel resistors.  If you use both ground leads out of the voltage divider and combine them at a headphone jack, which sums the grounds, the + leads of left and right channels will be connected with a resistance equal to the summed parallel resistors....which is a pretty severe level of loss of separation.