Hi,
Now where is the fun in that?
We had such discussions years and years ago, my Evil Twin alter ego KYW even started a complete thread on a simple and direct "how can I test this" project (remained theoretical).
Here some links to previous discussions that focused on the practical side of things and where generally held in an open and non-contentious atmosphere, probably because of the absence of contributions from those intend on only defending their orthodox faith and keeping their minds closed to new ways of doing things...
Have a few links:
http://www.diyaudio.com/forums/multi-way/3889-current-driven-loudspeakers-tranconductance-amplifiers.html
http://www.diyaudio.com/forums/chip-amps/20108-gainclones-current-drive.html
http://www.diyaudio.com/forums/multi-way/39338-theoretical-design-current-feed-active-dipole-speaker.html
These also have quite a few sources on the subject.
For me it remains an old hat, going back to the early 80's... 😉
Now gentlemen, could you please continue the petty bickering? I find it most intriguing. (Lt.Cmndr. Data in Startreck TNG)
Ciao T
Now where is the fun in that?
We had such discussions years and years ago, my Evil Twin alter ego KYW even started a complete thread on a simple and direct "how can I test this" project (remained theoretical).
Here some links to previous discussions that focused on the practical side of things and where generally held in an open and non-contentious atmosphere, probably because of the absence of contributions from those intend on only defending their orthodox faith and keeping their minds closed to new ways of doing things...
Have a few links:
http://www.diyaudio.com/forums/multi-way/3889-current-driven-loudspeakers-tranconductance-amplifiers.html
http://www.diyaudio.com/forums/chip-amps/20108-gainclones-current-drive.html
http://www.diyaudio.com/forums/multi-way/39338-theoretical-design-current-feed-active-dipole-speaker.html
These also have quite a few sources on the subject.
For me it remains an old hat, going back to the early 80's... 😉
Now gentlemen, could you please continue the petty bickering? I find it most intriguing. (Lt.Cmndr. Data in Startreck TNG)
Ciao T
With such rings, it is indeed possible to reduce the inductance EMF and the detriments introduced by it, but not totally eliminate them. It is, however, not very common nor cost-effective, and does not address the other side of the problem, the motional EMF.Joachim Gerhard said:
In ribbon tweeters, the inductance is inherently negligible, and quite obviously it is this fact that makes the differences in favor of them, rather than the smallness of mass.
I wonder a little, however, how the heating could be solved mechanically, without undue size and cost.
Differential equations are discussed mostly only in the appendix section and for reference. Driver operation in both modes is treated with transfer functions, as they are a more practical way.john curl said:
ETM, your book arrived today. I have only glanced at it so far, having to do audio design first. However, I will get back here with feedback as soon as it is practical.
*raises hand* Question!!!
How much benefit can be had from using current drive on headphones? Is it the same as with conventional speakers?
- keantoken
How much benefit can be had from using current drive on headphones? Is it the same as with conventional speakers?
- keantoken
In electrodynamic headphones, the achieved benefits of current drive are ususally very minor compared with the improvement in loudspeaker operation. This is mainly because in the impedance of headphone transducers the relative proportion of the DC resistance is generally much higher than in speaker drivers, so the current components produced by the electromotive forces are left rather small even on voltage drive.
A greater problem is generally constituted by the unevenness of frequency reproduction and its dependence on the ear canal shape. As with loudspeakers, the frequency response of headphones also exhibits certain changes when moving to current-drive. These changes may, depending on the case, also result in undesirable impressions.
A greater problem is generally constituted by the unevenness of frequency reproduction and its dependence on the ear canal shape. As with loudspeakers, the frequency response of headphones also exhibits certain changes when moving to current-drive. These changes may, depending on the case, also result in undesirable impressions.
Hi ETM,
Now, if we examine this situation a little more closely, you will see that the average receiver drives headphones from the main power amp, using dropping resistors. These resistors are normally a couple hundred ohms or so, making the voltage amplifier appear more like a current source.
That should provide more of an answer to keantoken's question.
Hi keantoken,
In other words, most people are listening to their headphones with a source that appears to be more of a current source than a voltage source anyway. I would think that most normal dynamic headphones are therefore designed to be driven with a current source when it came to tailoring the response of those headphones.
Not many people have a dedicated headphone amplifier of good quality. Of those, they may not even know their headphone output is driven directly. Most pro and semi-pro audio equipment do use a direct out, as do CD players and cassette players. Often the headphone driver turns out to be an NJM4556 op amp, probably going into current limit on louder passages. So to answer your question yourself, an easy experiment might be to listen to your headphones while plugged into your receiver, then try a purpose built headphone amplifier. There are some projects here on the site, and they are available commercially if you can get a store to allow you to try a headphone amp out at home.
I know what my observations are, and the existence of the headphone amplifier product should give you a hint as to how things shake out in the practical world. But I would recommend you do try this out for yourself. Avoid trying the less expensive headphone amplifiers. The models I have seen employ a very rudimentary amplifier, something you wouldn't listen to at home. Units using the NJM4556 op amp are probably only good at low levels, unless you have a pair of AKG headphones that have a higher impedance. I have a pair that I changed to use the European DIN connector, eliminating the common between channels. Still, a single NJM4556 isn't what I would consider as a good amplifier.
-Chris
To put this another way, since headphones represent a low power application, there is less interaction between the source and the driver.
Now, if we examine this situation a little more closely, you will see that the average receiver drives headphones from the main power amp, using dropping resistors. These resistors are normally a couple hundred ohms or so, making the voltage amplifier appear more like a current source.
That should provide more of an answer to keantoken's question.
Hi keantoken,
In other words, most people are listening to their headphones with a source that appears to be more of a current source than a voltage source anyway. I would think that most normal dynamic headphones are therefore designed to be driven with a current source when it came to tailoring the response of those headphones.
Not many people have a dedicated headphone amplifier of good quality. Of those, they may not even know their headphone output is driven directly. Most pro and semi-pro audio equipment do use a direct out, as do CD players and cassette players. Often the headphone driver turns out to be an NJM4556 op amp, probably going into current limit on louder passages. So to answer your question yourself, an easy experiment might be to listen to your headphones while plugged into your receiver, then try a purpose built headphone amplifier. There are some projects here on the site, and they are available commercially if you can get a store to allow you to try a headphone amp out at home.
I know what my observations are, and the existence of the headphone amplifier product should give you a hint as to how things shake out in the practical world. But I would recommend you do try this out for yourself. Avoid trying the less expensive headphone amplifiers. The models I have seen employ a very rudimentary amplifier, something you wouldn't listen to at home. Units using the NJM4556 op amp are probably only good at low levels, unless you have a pair of AKG headphones that have a higher impedance. I have a pair that I changed to use the European DIN connector, eliminating the common between channels. Still, a single NJM4556 isn't what I would consider as a good amplifier.
-Chris
ETM,
I just read your previous reply (#157) to a post I made. A couple things strike me taking this into account with other things you have posted.
Firstly, this is indeed in the commercial area, and this is an ad for your book. Okay, cool. I brought up my concerns and they've been aired enough. I look at things from a more physical viewpoint, how stuff actually works and interacts with other things. At this point I believe you are being overly defensive to the point where you are not reading the points I have posted. Either that or there is a language barrier of sorts that we may be reacting to. I freely admit that I'm treating this thread as I would any other, and you should probably be given more latitude with your claims as a result.
Secondly, I'm not sure whether you are purposely misinterpreting what I've said for public reasons, or whether you truly do not understand. It doesn't matter either way really, the result is what matters. My previous recent comments are not "absurd commentary", "Rubbish" or "totally irrelevant verbiage". They are in fact on topic and deal with the effects on a dynamic speaker, or driver. Also, your comment that "Output impedance is a property of amplifiers" has been taken erroneously. In fact, when speaking of transferring energy to the air, the concept of impedance mismatching between the electromechanical device and the air is an important one. It helps describe the losses that occur in that transfer of energy. The fact that you made a point of dismissing this either points to your lack of consideration for concepts you don't normally use (which does not make those concepts any less valid, by the way), or an effort on your part to discredit another viewpoint no matter how valid or not it may be.
I wanted to find out more about how you are looking at your current drive concept to possible add to what I already know, not to necessarily pick a fight. Since any input from me is disruptive, I'll left this alone. It would be nice if you'd read some of the comments I've made again, from a different direction as they are not designed as an attack. BTW, you showed a similar response to Joachim Gerhard when he mentioned "If you put a correctly designed copper ring or cap into the voicecoil gap it does not matter any more if the amp has current or voltage drive.". Joachim is quite right about this, and I tend to use the Peerless drivers that employ this very concept. Guess why?
Have fun ETM, Chris
I just read your previous reply (#157) to a post I made. A couple things strike me taking this into account with other things you have posted.
Firstly, this is indeed in the commercial area, and this is an ad for your book. Okay, cool. I brought up my concerns and they've been aired enough. I look at things from a more physical viewpoint, how stuff actually works and interacts with other things. At this point I believe you are being overly defensive to the point where you are not reading the points I have posted. Either that or there is a language barrier of sorts that we may be reacting to. I freely admit that I'm treating this thread as I would any other, and you should probably be given more latitude with your claims as a result.
Secondly, I'm not sure whether you are purposely misinterpreting what I've said for public reasons, or whether you truly do not understand. It doesn't matter either way really, the result is what matters. My previous recent comments are not "absurd commentary", "Rubbish" or "totally irrelevant verbiage". They are in fact on topic and deal with the effects on a dynamic speaker, or driver. Also, your comment that "Output impedance is a property of amplifiers" has been taken erroneously. In fact, when speaking of transferring energy to the air, the concept of impedance mismatching between the electromechanical device and the air is an important one. It helps describe the losses that occur in that transfer of energy. The fact that you made a point of dismissing this either points to your lack of consideration for concepts you don't normally use (which does not make those concepts any less valid, by the way), or an effort on your part to discredit another viewpoint no matter how valid or not it may be.
I wanted to find out more about how you are looking at your current drive concept to possible add to what I already know, not to necessarily pick a fight. Since any input from me is disruptive, I'll left this alone. It would be nice if you'd read some of the comments I've made again, from a different direction as they are not designed as an attack. BTW, you showed a similar response to Joachim Gerhard when he mentioned "If you put a correctly designed copper ring or cap into the voicecoil gap it does not matter any more if the amp has current or voltage drive.". Joachim is quite right about this, and I tend to use the Peerless drivers that employ this very concept. Guess why?
Have fun ETM, Chris
Thanks Anatech for your advice. I don't know if I will need to buy an amp, since I will (hopefully) soon have my own design working, and I can try all manner of output impedances I want.
I didn't realize it but I already answered part of my question. My signal generator has 50 and 600 ohm switchable output and I was testing the impedance of my headphones, and the impedance was very flat with both impedances.
- keantoken
I didn't realize it but I already answered part of my question. My signal generator has 50 and 600 ohm switchable output and I was testing the impedance of my headphones, and the impedance was very flat with both impedances.
- keantoken
Hi keantoken,
As for testing using a resistance, try to install a resistor at least 10 X the rated impedance of your headphones. If you can find the resonant peak, then go for 10 X that value. Music would be the appropriate source for your testing, something off a CD perhaps?
-Chris
That's even better than building a kit! Perfect, and I hope your design exceeds your expectations. 😉
As for testing using a resistance, try to install a resistor at least 10 X the rated impedance of your headphones. If you can find the resonant peak, then go for 10 X that value. Music would be the appropriate source for your testing, something off a CD perhaps?
-Chris
Well my headphones are Portapros, known for their very generous bass, so I don't want to put my ears out. To avoid that I can probably make a passive RC filter at the front of the amp.
However mathematically, a simple resistor couldn't match the impedance of an amp truly set up for current drive. A resistor large enough for this would require a voltage drive with a high voltage swing, and to approach an ideal current source would require infinite voltage swing and infinite resistance. My plan is to convert my amp between voltage a current drive. Then I can add series or parallel resistors to explore in between.
Also, correct me if I'm wrong, but doesn't moving to current drive invert certain phase characteristics? If so, then a parallel resistor with current drive might still differ from a series resistor with voltage drive.
I think I need to buy some 1k wirewound rheostats.
- keantoken
However mathematically, a simple resistor couldn't match the impedance of an amp truly set up for current drive. A resistor large enough for this would require a voltage drive with a high voltage swing, and to approach an ideal current source would require infinite voltage swing and infinite resistance. My plan is to convert my amp between voltage a current drive. Then I can add series or parallel resistors to explore in between.
Also, correct me if I'm wrong, but doesn't moving to current drive invert certain phase characteristics? If so, then a parallel resistor with current drive might still differ from a series resistor with voltage drive.
I think I need to buy some 1k wirewound rheostats.
- keantoken
Hi keantoken,
If the headphones are hot on the bass, connecting them to a voltage source type amp would increase the damping effect (to over-damped in speaker jargon). This may reduce the subjective amount of bass you hear. It should be tight though! 🙂 It really depends on how your headphones are designed. In no situation that I can think of would you get even more bass to the point where it might damage your hearing - unless you crank up the volume!
You do need higher voltage swings. Since you are dealing with headphones here, this shouldn't pose any real problems. Just use your amplifier and 2~3 watt resistors (in case you like the song).
Red herring. Don't worry about that at all. It is a non-issue.
-Chris
Okay, so don't turn the volume up.
If the headphones are hot on the bass, connecting them to a voltage source type amp would increase the damping effect (to over-damped in speaker jargon). This may reduce the subjective amount of bass you hear. It should be tight though! 🙂 It really depends on how your headphones are designed. In no situation that I can think of would you get even more bass to the point where it might damage your hearing - unless you crank up the volume!
No, perfect current drive should have infinite impedance. Now back to our regularly scheduled reality. To approximate the actual effects of current drive, it is accepted that you only need to increase the source impedance to about 100 X the transducers impedance. To be really safe, make that 100 X the impedance of the resonant peak. Keep in mind that the magnitude of the resonant peak depends more on the mechanical "Q" than anything else. Mechanical Q for a raw woofer is normally well above 1, try more like 3 or so. It's been years since I have measured the T-S parameters of any drivers, so my memory may be off there. I have seen resonant peaks of 60 ohms though. My test jig (in a box and everything - with an AC meter too!) uses a 1K0 resistor to make the audio oscillator "look" like a current source to the speaker under test. It also has a precision 10R00 resistor for calibration purposes. I used it a lot, and the numbers always made sense to the point of a finished speaker had close to the projected numbers. Not too shabby for the late 70's and early 80's!
You do need higher voltage swings. Since you are dealing with headphones here, this shouldn't pose any real problems. Just use your amplifier and 2~3 watt resistors (in case you like the song).
My first guess would be common collector output for starters, and no voltage feedback from there - so it has to be inherently linear to start. It will be running open loop pretty much.
Nope. No connection between phase and the type of drive at all. Besides, what would you use as a reference for phase, and would it matter since both channels will still have the same phase in relation to each other?
Those are equivalents. Check up on Norton and Thevenin equivalents. Here's a quick link you can use.
Red herring. Don't worry about that at all. It is a non-issue.
Why? Use fixed resistors.
-Chris
No need for common collector, my amp is class A. I can just use a current feedback resistor, as shown on the cover of Esa's book. Yes, series/parallel resistors can only approximate current/voltage drive; using OLG and feedback of either type does a much greater job (I probably won't hear the difference, but hell). Here it is:
http://www.diyaudio.com/forums/headphones/159815-rush-cascode-headphone-amp-jlh-output-stage-3.html
I think I could try this out right now, with the output jack on my soundcard. I'm weary of that though since the output distorts at higher output voltages into any load.
- keantoken
http://www.diyaudio.com/forums/headphones/159815-rush-cascode-headphone-amp-jlh-output-stage-3.html
I think I could try this out right now, with the output jack on my soundcard. I'm weary of that though since the output distorts at higher output voltages into any load.
- keantoken
Hi keantoken,
Well, do whatever works. I have to admit that I concentrate on designing and building amplifiers that reflect the current state of the art. So jumping in to design something that runs in a current source mode requires that I fight my normal inclinations. I can see I wouldn't do a good job with this without a huge amount of effort. Even still, it's too different from the paths normally traveled.
Go for it and let us know how it turned out.
-Chris
Well, do whatever works. I have to admit that I concentrate on designing and building amplifiers that reflect the current state of the art. So jumping in to design something that runs in a current source mode requires that I fight my normal inclinations. I can see I wouldn't do a good job with this without a huge amount of effort. Even still, it's too different from the paths normally traveled.
Go for it and let us know how it turned out.
-Chris
Too many pages to say I've reviewed the entire thread yet:
But my basic impression, "This is rediculous." Current mode is
no more valid than voltage mode is no more valid than power
mode (impedance matched) when true load impedance can't
be defined as a simple unchanging number. A Zobel network
isn't smart enough to get you entirely out of this frying pan.
Only a servo that knows (or can accurately predict) the
physical movement of the transducer can hope to do so.
Merely changing to current mode doesn't get you there.
If it did, then Class-D amplifier running open loop through
choke would be ideal for (insert esoteric audio application).
I deal with headphone amps like this professionally all day.
They all without exception have feedback loop for voltage
mode, and sabotaging that feedback definitely won't help.
By "professionally", I mean I'm test monkey at the end of
the production line. I don't design, I just measure. And on
rare occasion, troubleshoot and fix production errors.
But my basic impression, "This is rediculous." Current mode is
no more valid than voltage mode is no more valid than power
mode (impedance matched) when true load impedance can't
be defined as a simple unchanging number. A Zobel network
isn't smart enough to get you entirely out of this frying pan.
Only a servo that knows (or can accurately predict) the
physical movement of the transducer can hope to do so.
Merely changing to current mode doesn't get you there.
If it did, then Class-D amplifier running open loop through
choke would be ideal for (insert esoteric audio application).
I deal with headphone amps like this professionally all day.
They all without exception have feedback loop for voltage
mode, and sabotaging that feedback definitely won't help.
By "professionally", I mean I'm test monkey at the end of
the production line. I don't design, I just measure. And on
rare occasion, troubleshoot and fix production errors.
Last edited:
Hi,
NOW THAT is ridiculous.
Lookind at the basic function of an electrodynamic speaker it is clear that current is one of the two main controlling variables (the other is the magnetic field in the gap). So clearly, current mode is absolutely valid, the other two modes you mention are suboptimal from an academic viewpoint.
In practical terms anything that gives the results desired is valid of course.
So if you desire power compression and high levels of harmonic distortion as well as electrical damping of the resonance of the driver voltage drive is the valid solution.
if you desire freedom from compression and low distortion while taking care of the drivers resonance mechanically current drive is valid.
Any shunt network in parallel with the driver will disable current drive at the frequencies at which it works.
Hardly. The influence of the voice coil inductance can be equalised quite easily and the fundamental resonance can be damped by mechanical means. A servo is not needed, but can be employed gainfully if desired.
You may wish to read the thread after all, you save us going over ground already covered including EE101 stuff like the way how coild in magnet fields behave.
Ciao T
NOW THAT is ridiculous.
Lookind at the basic function of an electrodynamic speaker it is clear that current is one of the two main controlling variables (the other is the magnetic field in the gap). So clearly, current mode is absolutely valid, the other two modes you mention are suboptimal from an academic viewpoint.
In practical terms anything that gives the results desired is valid of course.
So if you desire power compression and high levels of harmonic distortion as well as electrical damping of the resonance of the driver voltage drive is the valid solution.
if you desire freedom from compression and low distortion while taking care of the drivers resonance mechanically current drive is valid.
Any shunt network in parallel with the driver will disable current drive at the frequencies at which it works.
Hardly. The influence of the voice coil inductance can be equalised quite easily and the fundamental resonance can be damped by mechanical means. A servo is not needed, but can be employed gainfully if desired.
You may wish to read the thread after all, you save us going over ground already covered including EE101 stuff like the way how coild in magnet fields behave.
Ciao T
- Status
- Not open for further replies.