[xnor]

The term you seem to be looking for is counter or back electromotive force (EMF).
 
Sure, in a dynamic headphone with a clearly visible resonant frequency in the audio range the back EMF will be at its maximum there, but it will also be generated at other frequencies.
 
I guess where impermanente is coming from is the damping factor (DF), which is just the ratio of impedances. The term implies that a high DF provides more damping (in other words energy absorption) for the back EMF.

 

[impermanente]

Ok I see what you mean now, I apologize for my analogies, maybe are a little bit misleading. But when I say that damping works at all frequencies I mean in the transient phase, sharp changes of signal etc.
 
So for example for the LCD 2 in sequence:
 
1. I apply a signal to the speaker.
2. The diaphragm moves.
3. I stop the signal (or abrupt change).
4. In an ideal world the diaphragm does not have mass and it is not linked to anything so it would stop or it would change direction instantly! Unfortunately in the real world the diaphragm does have a mass (the ortho ones are very small comparing to dynamic though) and it is connected to the frame so it doesn't stop instantly. Without any "brake" it would tend to vibrate without stopping but then the air and other internal energy dissipations would eventually stop it.
5. If from the point the signal stops, the diaphragm is connected to an amp according to the CDR (critical damping resistance) principle then it will get to the resting position quicker because of the electric damping effect.
 
- The fact that the diaphragm does not stop instantly is sure, especially for large excursions at low frequency. It is enough looking at waterfall plots, impulse responses and using common sense, to be sure about it.
- The fact that when no signal is applied to the speaker the diaphragm is damped if connected to an external circuit (in CDR mode) it is also a true thing (unless I am completely missing something here). This will occur in whatever situation and for any speaker (including orthos), if I close the speaker in a short circuit, its oscillation will be damped (more or less depending on the speed of the diaphragm), this is a basic principle of electromagnetism, if I try to move the diaphragm it will be stiffer than the same speaker connected in an open circuit. The same type of damping as for a sphere in a viscous liquid.
 
Dear Steve I hope I explained myself a little bit better, does this makes sense for you?

 

[Steve Eddy]

Quote:

Ok I see what you mean now, I apologize for my analogies, maybe are a little bit misleading. But when I say that damping works at all frequencies I mean in the transient phase, sharp changes of signal etc.
 
So for example for the LCD 2 in sequence:
 
1. I apply a signal to the speaker.
2. The diaphragm moves.
3. I stop the signal (or abrupt change).
4. In an ideal world the diaphragm does not have mass and it is not linked to anything so it would stop or it would change direction instantly! Unfortunately in the real world the diaphragm does have a mass (the ortho ones are very small comparing to dynamic though) and it is connected to the frame so it doesn't stop instantly. Without any "brake" it would tend to vibrate without stopping but then the air and other internal energy dissipations would eventually stop it.
5. If from the point the signal stops, the diaphragm is connected to an amp according to the CDR (critical damping resistance) principle then it will get to the resting position quicker because of the electric damping effect.
 
- The fact that the diaphragm does not stop instantly is sure, especially for large excursions at low frequency. It is enough looking at waterfall plots, impulse responses and using common sense, to be sure about it.
- The fact that when no signal is applied to the speaker the diaphragm is damped if connected to an external circuit (in CDR mode) it is also a true thing (unless I am completely missing something here). This will occur in whatever situation and for any speaker (including orthos), if I close the speaker in a short circuit, its oscillation will be damped (more or less depending on the speed of the diaphragm), this is a basic principle of electromagnetism, if I try to move the diaphragm it will be stiffer than the same speaker connected in an open circuit. The same type of damping as for a sphere in a viscous liquid.
 
Dear Steve I hope I explained myself a little bit better, does this makes sense for you?

I see what you're saying, but damping still comes down to resonance. If there's a resonance in the audio band that's underdamped, then exciting that resonance will cause overshoot and ringing. However if the resonance is critically damped, overdamped, or there is no resonance in the audio range, then the driver will handle transients without ringing and without any need for damping.
 
se

 

[xnor]

Looking at impulse responses and CSD plots, some orthos show ringing and their impedance is nearly completely flat.
 
There also seem to be problems with diaphragm flexing causing deep nulls or narrow peaks.

 

[Steve Eddy]

Quote:

Looking at impulse responses and CSD plots, some orthos show ringing and their impedance is nearly completely flat.
 
There also seem to be problems with diaphragm flexing causing deep nulls or narrow peaks.

 
Just as you would get with something like a traditional loudspeaker driver. There will be such things as cone resonances as the cone starts breaking up at higher frequencies. But those aren't something you can really control by way of electrical damping. Need some sort of mechanical damping for those. But then that brings up other issues.
 
se

 

[impermanente]

Quote:

I see what you're saying, but damping still comes down to resonance. If there's a resonance in the audio band that's underdamped, then exciting that resonance will cause overshoot and ringing. However if the resonance is critically damped, overdamped, or there is no resonance in the audio range, then the driver will handle transients without ringing and without any need for damping.  
se

 
thanks Steve,
 
So you are basically saying that since the resonance of the LCD2 is outside the audible spectrum (it is basically an harmonic oscillator so it must have a resonance frequency) it is useless damping it, one could not hear any ringing (unless one can hear ultrasounds).
 
Also reading your previous reply to xnor you are also saying that the reason of transient problems in the CSD plots is caused by other resonances of the diphragm in itself especially at higher frequencies (it has not an infinite rigid structure), that obviously cannot be controlled by electric damping.
 
But what about if one electrically damps the inaudible resonance at high frequency, which could cause the breaking up of the diaphragm and possible audible transient problems in the audible spectrum?

 

[stv014]

Quote:

But what about if one electrically damps the inaudible resonance at high frequency

 
If there is any such resonance that can be electrically damped, shouldn't it show up on an impedance vs. frequency measurement with a sufficient bandwidth ?

 

[jcx]

I would think that's true
 
high frequency resonance likely involves cone breakup, multiple modes of diaphragm motion - electrical damping only works effectively for net/averaged voice coil or conductor motion so is best only for single mode motions - especially the bass mass spring resonance
 
for a voice coil I'd guess axi-symmetric modes with the coil a velocity node could also be damped electrically
 
for planars the serpentine track and multiple stripe mag field makes higher order diaphragm vibration mode interaction with electrical damping very complex
 
 
 
it is possible to make a amp with negative output resistance over the audio range - but it would be unstable without a load, or even a load outside of its expected load impedance range

 

[mogulmaster]

Hey guys, I've searched all over the place and can't really get a definitive answer. 

I have 10 ohm e09k paired with 35 ohm he-400. definitely far from 1/8th optimal ratio - it sounds pretty amazing to my ears, but I want the least amount of distortion possible because I have sensitive ears to fatigue. 
 
Should I be fine? Or look at different amps

 

[xnor]

With planar magnetic headphones the 1/8th rule of thumb doesn't really matter. From the output impedance side it should be fine.

 

[Sko0byDoo]

Very interesting topic...learn quite a lot.
 
Question, will it be wise to match the orthodynamics 'phones' impedance to transfer the max power from the amp, if no damping needed?
 
I'm thinking of building a diy tube amp that pumps out 1.5W/ch.  A current thought is to get 32-ohm OPTs that will be great for the HD800, per 1/8 rule.  Will these OPTs be ok for the 50-ohm LCD-3 (with damping ratio of only 1.5)?  
 
Thanks for any insight! 

 

[xnor]

Any increase in output impedance will reduce efficiency, that is the voltage across the load will decrease and so will the power.
 
1V into 0 ohm output impedance + 50 ohms load = 20 mW
1V into 50 ohm output impedance + 50 ohms load = 5 mW

 

[nikongod]

  Very interesting topic...learn quite a lot.
 
Question, will it be wise to match the orthodynamics 'phones' impedance to transfer the max power from the amp, if no damping needed?
 
I'm thinking of building a diy tube amp that pumps out 1.5W/ch.  A current thought is to get 32-ohm OPTs that will be great for the HD800, per 1/8 rule.  Will these OPTs be ok for the 50-ohm LCD-3 (with damping ratio of only 1.5)?  
 
Thanks for any insight! 

 
I would be concerned that a transformer designed for a 32ohm load would ring with a 300ohm load. 
Why not get a transformer with multiple secondaries, or at least a multi-tapped secondary, that is better suited to the wide range of impedances that headphones represent? A transformer set up like this will make getting good gain structure much simpler.
 
On that note, as you said, electrical damping factor is not important with orthos. Assuming the amp with a 32ohm transformer can swing adequate voltage, and does not ring with a 50ohm load, the orthos will be fine with it.

 

[Sko0byDoo]

  Any increase in output impedance will reduce efficiency, that is the voltage across the load will decrease and so will the power.
 
1V into 0 ohm output impedance + 50 ohms load = 20 mW
1V into 50 ohm output impedance + 50 ohms load = 5 mW

 
I'm still chewing on this thought...not quite sure if I get it.  I thought the main function of the OPT is to provide either voltage or current swing for a given output power per Zload?  The farther the reflected Zload (via the OPT) is away from the tube plate impedance, the lower the power delivered?  
 

   
I would be concerned that a transformer designed for a 32ohm load would ring with a 300ohm load. 
Why not get a transformer with multiple secondaries, or at least a multi-tapped secondary, that is better suited to the wide range of impedances that headphones represent? A transformer set up like this will make getting good gain structure much simpler.
 
On that note, as you said, electrical damping factor is not important with orthos. Assuming the amp with a 32ohm transformer can swing adequate voltage, and does not ring with a 50ohm load, the orthos will be fine with it.

 
Really?  32-ohm will be under-damped for a 300-ohm load?  Dang, it's true that the HD800 is a tough dog to drive.  
 
Multiple secondaries would be ideal to do a trial-and-error.  But a 32-ohm multi-tapped, with a good, broad freq. response OPT is hard to find.  Jack at Electra-print offers psss, but only with single output.

 

[nikongod]

  Really?  32-ohm will be under-damped for a 300-ohm load?  Dang, it's true that the HD800 is a tough dog to drive.  
 
Multiple secondaries would be ideal to do a trial-and-error.  But a 32-ohm multi-tapped, with a good, broad freq. response OPT is hard to find.  Jack at Electra-print offers psss, but only with single output.

 
The HD800 will be overdamped (I personally like the HD800 from a highish to really high output impedance source. Like 100ohms Zo and up) the transformer will be underdamped. 
 
Not all 32ohm transformers ring when presented with a 300ohm load, and you can always wire a resistor in parallel with the headphones to dampen the transformer, but I would be cautious. Weird world we live in. 
The other advantage of a transformer with multiple secondary options is that the gain structure of the amp/system will be easier to get under control. IME amps with just one secondary are too often compromises of too much gain for low impedance headphones or not enough gain for high. Most people put up with too much gain for low impedances even though they could do without a gain stage, and save a bit of distortion that way. 
 
If you have a bunch of money for the project:
Sowter makes multi-tapped SE transformers for headphone loads. 
Lundahl makes multi-secondary single ended and push-pull transformers that can be configured into headphoneish loads. 
 
If you don't have much money (me):
Edcor makes some new "double autoformers" that look intriguing for headphones, although I dont think that they will play nice with high voltage DC or DC current - they do look cool for parafeed tube or transformer coupled SS though. Edcor will make you just about anything if you ask nicely and put up a design fee