Adding impedance to headphones

[holland]

Quote:

Originally Posted by grndslm /img/forum/go_quote.gif I'm trying to clean up the muddy bass in the KSC75 (there are at least a thousand posts about KSC75 and adding impedance adapter, but none of them explained why they chose 75ohms)... plus I'm considering buying IM716 which also benefits from the P->S adapter as they are essentially Etys. The Supermini's extra 75ohm resistor suits my situation perfectly! Sure I'd prefer more resistance if possible, but now I don't have to add anything inbetween cables or have an adapter that can be lost. Perhaps the KSC75 would benefit from an additional 20ohms, but the tradeoffs for SQ prolly aren't even worth me thinking about this for another second.

Are they all using 75 ohms? No clue, I haven't searched. I would think some would have tried other values. Just piggy back on what works best if you don't want to experiment. I don't have a KSC75 so I can't comment more on that.

For im716, snip the pod out and put resistors in line. Again, experiment. I chose 68 as JSatch did. It sounded best to me amped. Of course, unamped it totally sounds like crap. Others chose 47, and others chose 100 IIRC. See how it varies?

 

[majkel]

Clutz, I guess you don't understand the idea of a resistor in series and mislead people with great effort empowered by lack of understanding. Let's try a different way: ideal voltage source (read: amp with no resistors) works like this for headphones: P(f) = Uamp*I(f). Current I(f) = U/Z(f), that's Ohm's law. On the opposite side, you can imagine a perfect current source (very high output impedance), then P(f) = I*I*Z(f)

When the amplifier is an ideal voltage source, it delivers more power to headphones when their impedance goes lower. If the amp is an ideal current source, then power dissipated in the headphones increases with the impedance. Using a resistor moves you a bit towards ideal current source approximation. If the resistor has a higher resistance than the headphones, more power is dissipated in the resistor. When headphone's impedance is higher than resistance, more power remains in the headphones. As long as there is no resistor, and the impedance is not flat all over the spectrum, you lose power delivery on the impedance bump (lower amplitude/volume here). If the resistor was of a high resistance, you would get amplitude rise instead, because the resistor has more influence on the current than the headphones. Putting a resistor with the resistance somewhwere in between makes power delivered closer to each other for two different frequency points.

 

[JaZZ]

Quote:

Originally Posted by majkel /img/forum/go_quote.gif Clutz, I guess you don't understand the idea of a resistor in series and mislead people with great effort empowered by lack of understanding. Let's try a different way: ideal voltage source (read: amp with no resistors) works like this for headphones: P(f) = Uamp*I(f). Current I(f) = U/Z(f), that's Ohm's law. On the opposite side, you can imagine a perfect current source (very high output impedance), then P(f) = I*I*Z(f) When the amplifier is an ideal voltage source, it delivers more power to headphones when their impedance goes lower. If the amp is an ideal current source, then power dissipated in the headphones increases with the impedance. Using a resistor moves you a bit towards ideal current source approximation. If the resistor has a higher resistance than the headphones, more power is dissipated in the resistor. When headphone's impedance is higher than resistance, more power remains in the headphones. As long as there is no resistor, and the impedance is not flat all over the spectrum, you lose power delivery on the impedance bump (lower amplitude/volume here). If the resistor was of a high resistance, you would get amplitude rise instead, because the resistor has more influence on the current than the headphones. Putting a resistor with the resistance somewhwere in between makes power delivered closer to each other for two different frequency points.

Clutz is right. The above theory doesn't make any sense. Any serial resistance will accentuate the impedance rise in a headphone's impedance curve, so with usual headphones it will create a bass emphasis, coupled with a bad transient response in this region.

I can't see how serial resistors could improve the KSC-55/-75's frequency response, unless you think it lacks bass (... at ~80 Hz).

Some headphones do indeed benefit from serial resistors (or a high output impedance, resp.) -- exactly because of the otherwise thin bass.

However, there's no established rule in the above (majkel) sense or in the form of the ER-4's 75-ohm adapter. I've tried various resistors with the latter and preferred something around 35 ohm (a variant called «ER-4X»). Note that the ER-4 explicitly doesn't belong to the «usual headphones» -- in that it uses balanced-armature drivers with completely different impedance response (no «bass hump» there) in comparison to conventional dynamic headphones.
.

 

[grndslm]

I heard it CLEANS the bass up as well. Obviously the amp will add more impact for sure, but I've read a few reports that these 75ohm impedance adapters do clean up the bass.... but maybe they're just crazy ??

 

[SonicDawg]

Just wondering about an OT question: if you want to maintain the same impedance after balancing a pair of headphones, do you simply add half of the original resistance to each channel? (e.g.: 60ohm adding 30ohm to each)

 

[majkel]

I base my thoughts on my hearing only and I'm explaining how it works. I'm completely not interrested what people think if it makes sense or not. For me does make sense with HD595 on some music genres because it gives me the sound I like more than without resistors. I know even more people preferring that "resistor" sound, dissregarding the source and amp they use. It is true, and I mentioned it before, that on a impedance bump you will get higher amplitude. Ety's behave in the opposite way - there is a dent in the bass range of the impedance, so adding resistor adapter lowers the bass giving impressions of emphasized highs and details. Given this example we can state that headphones don't necessarily need high damping factor like speakers, especially that headphone's transducers are very light compared to loudspeakers, so there is no such need to attenuate their inertial movement electrically, and resistors will not deteriorate the pulse response as they would for loudspeakers.

 

[Clutz]

I've talked with who I know, knows what they're talking about- because your math seems correct to me, but there seemed like an underlying flaw in your logic. The friend agreed that you were mostly correct- and your mathematical explanation is fully correct, however I was correct in stating that you were wrong at some level.

Well, I'll just quote him anonymously
Quote:

He seem to make the erroneous assumption that constant _power_ delivery with respect to frequency is desirable. It isn't. Most headphones produce the designed frequency response (hopefully close to being flat, with some tweaks to compensate for ear cup and ear cavity acoustics) when driven by a constant _voltage_ across the frequency band, even if the impedance isn't flat. Adding a resistor in series with the transducer forms a voltage divider with the transducer. Since the transducer has non-linear impedance, then the voltage divider's attenuation ratio varies with frequency, causing deviations in the frequency response. Yes, adding resistance like this has a dramatic effect, but in most cases it's not a good thing.

So you can enjoy the sound of placing resistors inline with your headphones, and that's great - but don't act like it's universally accepted as a good idea, because it isn't.

I stand by my original position that there is a flaw in your logic.

 

[Clutz]

Quote:

Originally Posted by SonicDawg /img/forum/go_quote.gif Just wondering about an OT question: if you want to maintain the same impedance after balancing a pair of headphones, do you simply add half of the original resistance to each channel? (e.g.: 60ohm adding 30ohm to each)

No, you leave them alone.

The impedance of each of the sides of the headphones is the nominal impedance. The left channel and right channel are not run in series with each other- the only thing they have in common for an unbalanced pair of headphones is the ground- but you wouldn't measure the impedance of the headphones from left channel to right channel - you would measure it from left to ground and right to ground.

 

[mdarnton]

I'd be wary of making assumptions about what to do based on rules.

I have a P>S adapter, and have tried it on my various phones/amp combinations. The results aren't predictable at all from anything I know, and rarely the result is better, but in most the effect is definitely negative. For instance, it cleans up my Edition 9s nicely on my laptop (a pairing that doesn't make sense on any level whatsoever), but with the Ed9/Opera combo the adapter makes everything mushy. The bass is affected, yes, but the hit on everything else is undesirable.

I just tried the ESC75/Opera combination with and without the adapter, and to me the adapter ruins everything just as it does with the Ed9/Opera pairing--yes, the bass lightens up, but the treble disappears, and vocals move back, crispness is gone--they sound like the phones that came with my walkman.

As Clutz' anonymous source says, it's always a distinct possibility that the people who design these things might actually know what they're doing, I guess.

 

[grndslm]

Maybe your Opera is broken...

 

[majkel]

Clutz, your anonymous friend is not necessarily right in saying that you always need to have constant voltage drive for headphones. This is the rule for loudspeakers, not headphones. Here some facts:
1) 120 ohm resistors were recommended for some types of headphones in the past.
2) No tube amplifier has "zero" output impedance, especially OTL. These means a tube acts like having a resistor inside (look at their limited transconductance values) That's why HD595's bass increases on tube amps.
3) See the schematic in the datasheet of TPA6120A2 headphone op-amp. They suggest connecting 10 ohm resistors in series. Constant voltage drive idea is spoiled here whenever headphone's impedance isn't flat.
4) See some service manuals for hifi components. My CD player's headphone output is buffed with 150 ohm resistors in each channel. Holly crap!

5) I have never found a statement or recommendation regarding amplifier's output impedance for any headphones. Constant voltage drive (providing high damping factor) seems natural but as long as headphones with a flat frequency response don't exist, it is allowed to modify it by any means, for instance thru their uneven impedance. It's the most puristic solution, no sound processing and additional circuits needed, and logic has nothing to do with it but hearing.

 

[JaZZ]

Quote:

He seem to make the erroneous assumption that constant _power_ delivery with respect to frequency is desirable. It isn't. Most headphones produce the designed frequency response (hopefully close to being flat, with some tweaks to compensate for ear cup and ear cavity acoustics) when driven by a constant _voltage_ across the frequency band, even if the impedance isn't flat. Adding a resistor in series with the transducer forms a voltage divider with the transducer. Since the transducer has non-linear impedance, then the voltage divider's attenuation ratio varies with frequency, causing deviations in the frequency response. Yes, adding resistance like this has a dramatic effect, but in most cases it's not a good thing.

The above statement is entirely correct.

Quote:

Originally Posted by majkel /img/forum/go_quote.gif Clutz, your anonymous friend is not necessarily right in saying that you always need to have constant voltage drive for headphones. This is the rule for loudspeakers, not headphones. Here some facts: 1) 120 ohm resistors were recommended for some types of headphones in the past. 2) No tube amplifier has "zero" output impedance, especially OTL. These means a tube acts like having a resistor inside (look at their limited transconductance values) That's why HD595's bass increases on tube amps. 3) See the schematic in the datasheet of TPA6120A2 headphone op-amp. They suggest connecting 10 ohm resistors in series. Constant voltage drive idea is spoiled here whenever headphone's impedance isn't flat. 4) See some service manuals for hifi components. My CD player's headphone output is buffed with 150 ohm resistors in each channel. Holly crap! 5) I have never found a statement or recommendation regarding amplifier's output impedance for any headphones. Constant voltage drive (providing high damping factor) seems natural but as long as headphones with a flat frequency response don't exist, it is allowed to modify it by any means, for instance thru their uneven impedance. It's the most puristic solution, no sound processing and additional circuits needed, and logic has nothing to do with it but hearing.

Well, of course there's nothing wrong with experimenting with serial resistors. It's just that in most cases you get bad results if your goal is a flat acoustic frequency response. Because -- as you know -- in first place the inevitable bass resonance gets accentuated by doing so.

Of course not all amplifiers offer near zero ohm output impedance; (OTL) tube amps have to deal with up to 100 ohms and more. It's a logical consequence of the design, but in fact an unwanted one, which makes them unsuitable for some headphones. It's illogical to deduce a virtue from this necessary evil. The same applies to buffer resistors serving as short-circuit protection. BTW, the high (buffer) resistor values of CDP's headphone outputs are the main reason for their bad reputation.
.

 

[majkel]

Are we talking about the results of resistors in series, or a philosophical and logical aspect of this? Whatever works, is logic and makes sense. Neither KSC75 nor HD595 are hi-end expensive headphones intended to be extremely flat and accurate. Same thing with the amplifiers in the same price range as HD 595's. They are constructions far from being perfect and the idea of direct voltage drive is a kind of a missed shot here because it's only achievable in SS. Why forbid people using tube amplifiers just because of the idea of voltage drive and tell them it isn't logic because the headphones' impedance isn't flat? This really makes no sense and has nothing to do with logic. It is a kind of intolerance and forcing the idea of supremacy of logic all over listening pleasure IMHO. I would never use resistors for expensive headphones because I trust and believe these devices are intended to work their best as they are. But sorry, not this price range. I would also never use any resistors for any Grados, AKG K501-701, Senns HD 580-650, because it deteriorates their sound. Sonically proven and undisputable. Also, any other headphones with a flat impedance all over the spectrum won't benefit from additional resistors.

 

[JaZZ]

Quote:

Originally Posted by majkel /img/forum/go_quote.gif Are we talking about the results of resistors in series, or a philosophical and logical aspect of this? Whatever works, is logic and makes sense...

Don't get me wrong -- I'm not generally opposed to (experimenting with) serial resistors. My concern was that your esthetically pleasing formula (you remember?) made it look like there was some kind of technical advantage from this configuration. After all you don't seem to stick to this idea anymore.

(In fact, if you want to get frequency-neutral current flow from the amp, go as high with the resistor values as you're ready to sacrifice efficiency. But then again, what is frequency-neutral current flow worth in the context of music reproduction? You don't listen to amplifiers, but to headphones.)
.

 

[majkel]

Quote:

Originally Posted by JaZZ /img/forum/go_quote.gif Don't get me wrong -- I'm not generally opposed to (experimenting with) serial resistors. My concern was that your esthetically pleasing formula (you remember?) made it look like there was some kind of technical advantage from this configuration. After all you don't seem to stick to this idea anymore. .

It could have looked like you said but it would obviously make no sense if I stated something like "you always have to make power matching". Sometimes it's convincing for people when there is a mathematically explainable reason for doing something, not only perceptual gain. That's why I'll still think that matched power dissipation is a good starting point if somebody wants to change something by very cheap and often working means. In another thread I found info that both Ety's ER4P and ER4S contain resistors in series, 22 and 100 ohm respectively which clearly means that headphone manufacturers don't care at all for the damping factor that amps could provide. The same is with Ety's siblings AltecLansing im716 and their switch, being nothing more than switched passive RC filter. Linkwitz went event further, designing passive RLC filter to compliment Shure e2c IEMs.