A moving coil speaker is essentially a linear electric motor albeit an extremely simple one. When the coil moves an electromotiv force, also known as a voltage, is generated in it. The reason is an electric field is setup due to moving inside a magnetic field. The electric field is such that its direction is along the coil's wire. Every turn experiences this force, and this implies, the effect is additive for the number of turns in the coil. The back electromotive force, EMF, is directly proportional to the speed of the coil and the magnetic field strength in the gap. The EMF alone does not determine the amount of current drawn into the coil, but the amount of force required to drive the speaker cone against the surrounding air does: this is actually speaker action which creates a longitudinal wave we all know constitute sound. Had the wire ohmic resistance been zero, the current drawn would only depend on the force required to interact with the air, but resistance, being part of the wire, and hence must act in series, limits the visible voltage applied to coil according to Ohm's Law. Proper mathematical analysis is required to determine whether the series resistance contributes to speaker distortion. If its effect is proportional to current, my intuition tells me, it should not add anything to distortion.
I'm not sure either. It is a similar situation when you take a driver, with the terminals open. Then give it a push. Clearly, there will be a voltage generated at the terminals, but there will be no current as the circuit is open.Hmmm.... Good point. But where does the back-EMF go then? I don't buy Dave's claim that it becomes zero – or if it does I don't understand how and would like to know – but I also doubt it goes into the infinite output impedance of the ideal current source and results in infinite voltage across the speaker. Maybe we should recommend that those who use current output amps install a spark gap across the speaker. Clearly there's a market for cryogenically treated, rhodium plated spark gaps now...
Tom
Where does the energy go? Or is there no energy dissipated? In this case the speaker eventually comes to a rest because of the mechanical losses.
If you would have shorted the terminals, it comes to a rest quicker because more energy is dissipated as current - that's your damping.
Remember how we used to transport those little analog meters? With shorted terminals to damp movement.
In fact, this story also comes around to the lack of damping with current drive, only the mech losses give damping. That is also why you see that resonance peak in the freq response - it is not damped.
Jan
The magnetic field strength varies, depending on the position of the voice coil.
See the various articles about current driven loudspeakers.
It isn't. The resistance varies with voice coil temperature, which varies with dissipation. It can even double at high volumes.
There is also the self-inductance of the voice coil. It depends on all the iron around it. As a result, it changes with voice coil position and it has hysteresis effects.
My previous post (number 121) is about speakers driven normally with a voltage amplifier as is 'almost' universally done.
The beauty of voltage driven speakers is they are simple and economical to produce, and yet, produce sound reproduction that is a very good representation of the original. Although other archaic methods may exist, the expenses involved and the added extra technical complications, do not justify them, and this applies to most people in the field of audio.
If I understand well, current driven speakers need a variable magnetic field depending on the coil position which mandates the need for a coil position sensor. This introduces magnetic hysteresis in the core.... I do not need to go further. Too many disadvantages for only a little or no improvement.
This is enough for anyone attentive to expenditure like manufacturers to reject it for something much simpler and economical.
The beauty of voltage driven speakers is they are simple and economical to produce, and yet, produce sound reproduction that is a very good representation of the original. Although other archaic methods may exist, the expenses involved and the added extra technical complications, do not justify them, and this applies to most people in the field of audio.
If I understand well, current driven speakers need a variable magnetic field depending on the coil position which mandates the need for a coil position sensor. This introduces magnetic hysteresis in the core.... I do not need to go further. Too many disadvantages for only a little or no improvement.
This is enough for anyone attentive to expenditure like manufacturers to reject it for something much simpler and economical.
I only build my loudspeakers only with fullrange drivers and with no crossovers. They are all dsp corrected as you have like this the biggest liberty in the design of the speaker, without taking care of thiele small parameters and you can tame peaky resonances at any frequency. This you can do today with a simple smartfone and equalizer software for free.
Most drivers out there, over 95% are conventional drivers with no copper rings. All these loudspeakers profit from less distortion.
I tried Esa Merilainens simple resistor method. He told me himself that you have some benefits in using a resistor of 4 to 8 ohms for this kind of drivers of 4 to 8 ohms and you can hear some progress in distortion. I tested this with now three fullrange drivers with success. I do not lose too much amplifier power and you get for free a better hifi-like sound with simple fullrange drivers. I did not try and know that resistances can have this effect before and omitted resistances.
I always thought to omit some resistances but it showed that less damping for little fullrange drivers up to 13cm makes the bass better.
All this you can simply hear with an A / B loudspeaker comparison with the balance knob of your amplifier and some mono recording. One modified speaker versus unmodified.
Listening test fullrange driver with simple resistor method - approximation to current drive:
https://www.diyaudio.com/community/threads/new-15-full-range-fane.308652/post-7076903
https://www.diyaudio.com/community/...w-distortion-with-a-2-way.334757/post-6711184
Reducing distortion from the loudspeaker diaphragm:
https://www.diyaudio.com/community/threads/full-range-pics.87174/post-6711011
Most drivers out there, over 95% are conventional drivers with no copper rings. All these loudspeakers profit from less distortion.
I tried Esa Merilainens simple resistor method. He told me himself that you have some benefits in using a resistor of 4 to 8 ohms for this kind of drivers of 4 to 8 ohms and you can hear some progress in distortion. I tested this with now three fullrange drivers with success. I do not lose too much amplifier power and you get for free a better hifi-like sound with simple fullrange drivers. I did not try and know that resistances can have this effect before and omitted resistances.
I always thought to omit some resistances but it showed that less damping for little fullrange drivers up to 13cm makes the bass better.
All this you can simply hear with an A / B loudspeaker comparison with the balance knob of your amplifier and some mono recording. One modified speaker versus unmodified.
Listening test fullrange driver with simple resistor method - approximation to current drive:
https://www.diyaudio.com/community/threads/new-15-full-range-fane.308652/post-7076903
https://www.diyaudio.com/community/...w-distortion-with-a-2-way.334757/post-6711184
Reducing distortion from the loudspeaker diaphragm:
https://www.diyaudio.com/community/threads/full-range-pics.87174/post-6711011
^if there is need to have the electronic damping fully effective on a fullrange driver main resonance, or low shelving, just put bypass inductor parallel to the series resistor. This could also compensate for bafflestep if there is need to.
If there is any cone breakup then parallel notch (in series) would reduce distortion amplified by the breakup. Handling cone resonance peak with DSP will drop only the original signal to that cone breakup but distortion products "generated in the driver" are still amplified by the breakup. Increase impedance in the circuit on around the breakup and also distortion is EQ:d.
Anyway, passive filter can have two function, impedance manipulation and filtering. If filtering is done actively then there is more freedom to do impedance manipulation.
If there is any cone breakup then parallel notch (in series) would reduce distortion amplified by the breakup. Handling cone resonance peak with DSP will drop only the original signal to that cone breakup but distortion products "generated in the driver" are still amplified by the breakup. Increase impedance in the circuit on around the breakup and also distortion is EQ:d.
Anyway, passive filter can have two function, impedance manipulation and filtering. If filtering is done actively then there is more freedom to do impedance manipulation.
Last edited:
Well, there's the air (spring) also if you use a closed box. Which I think one should. Perhaps especially in the CC case?
//
No, they don't need that, and neither do voltage-driven loudspeakers, but the magnetic field is never perfectly constant in a practical loudspeaker.
That's again an undesired effect that all loudspeakers have. Its effect on the sound is worse with voltage than with current drive.
It not novel and has been around for 70+ years.
A usable mixed impedance drive (low output impedance near resonance and increasing above) is cheap and easy to implement... at least with conventional linear amplifiers. If it (or even "pure" current drive) results in better performance then it is a viable option, at least for manufacturers who are aware of the idea.
As I said, the technique can be quite beneficial for lesser quality drivers or some specific types of drivers like AMTs.
That's not a proper A/B as the resistor also changes the frequency response which quickly is the dominant factor why it sounds different. To judge the true effect of high(er) source impedance the frequency response must be dialed in to be exactly (<0.1dB) the same. The same if obviously true for any kind of A/B measurents.
You shall changed only the variable of interest, and that is the different damping, not the frequency response.
You can measure it - but you can also hear it with a loudspeaker which will work both with voltage and current drive.Sure, i see that
My next step is to build a low distortion current amp. I could need some help with this on how to adjust the real world output impedance of such an amp.
When done, i am able to do some measurements myself...
Just modify only one mono amplifier and leave the other one in voltage drive. Adjust both loudspeakers to the same loudness. You can record both speakers in A / B comparison with a Zoom H1 or better.
Put the recording online and everyone can hear it with a good headphones.
This listening test is good to learn how distortion sounds to the ear.
There are many people lacking listening experience who would spontanously say that a distorted speaker sounds better because it sounds louder!
A loudspeaker with very little distortion sounds less spectacular.
If you need a proof for listening to the effect of distortion reduction this is true. It can be done with a EQ.
But if you want to hear the effect without doing a measurement then A / B listening tests help to find out a difference as the ear is easy to betray. For small fullrange drivers more bass and highs are welcome.
Not all tests have to be an a scientific level - and you can criticize the arrangement. Nearly nobody does that on the highest level. For more secure listening tests you would need a panel of listeners and do blind testing, better on both sides - listener (hiding the speakers) and the person who changes the sources.
As I proposed it - changing immediately between two sources without the need to rewire loudspeakers helps in doing a listening test for anyone interested to come to know how it sounds to him. And you can make anyone participate by recording and using good quality headphones.
That.
Open terminals= no energy dissipation at the voice coil.
The difference is important, same as between "behaving like a Bass drum" giving a defined frequency/musical note to "as dead as a sandbag"
I am going to sum up the messages that have arrived me:
- benefit for current drive strongly depends on the drivers that were used,
the cheaper the driver - the higher the benefit
- benefit of current drive depends on the output impedance of the current amp,
the higher the output impedance - the higher the benefit
- current drive can get problematic around fs, more critical at higher output impedance
- the benefit of current drive (if any) shows up in lower intermodulation (and that is very welcome to me!)
- to really compare current drive with voltage drive individual equalizing is a must
- the interaction between amplifier and speaker is hardly understood, even specialists dont agree about current drive
- the current amps to build are all based on Esas ideas how to do it, none of them is really low THD (when tested with resistive load)
- my question for help on building such an amp (detail: output impedance) seemes to be unheared...
- my question to the musical timing when the current is in phase with the controlling voltage (and the amps
output voltage is therefore phaseshifted) is still not answered...
- benefit for current drive strongly depends on the drivers that were used,
the cheaper the driver - the higher the benefit
- benefit of current drive depends on the output impedance of the current amp,
the higher the output impedance - the higher the benefit
- current drive can get problematic around fs, more critical at higher output impedance
- the benefit of current drive (if any) shows up in lower intermodulation (and that is very welcome to me!)
- to really compare current drive with voltage drive individual equalizing is a must
- the interaction between amplifier and speaker is hardly understood, even specialists dont agree about current drive
- the current amps to build are all based on Esas ideas how to do it, none of them is really low THD (when tested with resistive load)
- my question for help on building such an amp (detail: output impedance) seemes to be unheared...
- my question to the musical timing when the current is in phase with the controlling voltage (and the amps
output voltage is therefore phaseshifted) is still not answered...
MarcelvdG said:
Oh, please, come on! Perfection? Did I ever mention perfection and that component qualities or characteristics are perfect, whatever is meant by the latter. However, it is extremely desirable that a speaker magnet assembly provides a reliable magnetic field, and one requirement of that field, is constancy of intensity. Another such requirement, but in the case of resistors, is their material must have a 'linear' characteristic. Yes, for real materials one may get other coefficients for the polynomial representing voltage against current, but the most prominent, the coefficient of the linear term, must be far bigger than the rest. Pedantically, it would not be linear, but for real applications, it can satisfy most situations.