If B changes because of the voice coil current even with NO hysteresis it would still be a nonlinear effect. The fact that there is minor loop hysteresis is not a requirement for this nonlinearity. B changes with i and hence there is a Bl(i) contribution regardless of any hysteresis.
I guess we disagree on this one, but as I said, we agree on the result and its solution, just not the details.
Then please explain the cause of the distortion? Why would you deny something that is so clearly there in the measurements?
David S.
I never removed the copper. I assume that it's air under the raised section. It's doubtful that they would have machined the pole-piece into this shape under it nor filled it with copper.
Dave
If the extended copper helps reduce flux modulation outside the gap, even possibly in the absence of a square-ended pole, it ought to affect BL(x) to some degree. The flux lines will still be fairly strong just above the gap in the area near to the top plate surface. Flux modulation may still occur in this area as well, but probably with less effect. With a midrange unit such as the 13m with a small surface area, it's probably of more importance.
Dave
Then I guess I'm being dense today. To say that B is non constant is to say that the DC operating point is driven somewhat by the AC signal. This may be true but doesn't require or force nonlinearity. If all the materials in the structure were hysteresis and nonlinearity free, then the opperating point could be somewhat soft and we might see a level drop from the theoretical output, but no distortion because of it.
Finally, I can imagine a system with nonlinearity but free from hysteresis, but why would we conjecture this to be the case when we know that hysteresis is present and inherent in the magnetic material?
David S.
Some material/threads you guys may or may not be interested in:
ATC created their s.l.m.m technology to counteract hysteresis:
Extremely high performance monitors
further discussions here:
http://www.diyaudio.com/forums/feastrex/70709-exciting-new-line-fullrange-drivers-feastrex-2.html
ATC's S.L.M.M white paper
http://www.atcloudspeakers.co.uk/download.php
ATC created their s.l.m.m technology to counteract hysteresis:
Extremely high performance monitors
further discussions here:
http://www.diyaudio.com/forums/feastrex/70709-exciting-new-line-fullrange-drivers-feastrex-2.html
ATC's S.L.M.M white paper
http://www.atcloudspeakers.co.uk/download.php
Last edited:
In the world where cost is no big deal, then the approach is to raise the magnet portion until the gap area is very saturated. When the pole tips are very saturated, the magnet system looks more like a constant H (what the constant H current driven field coil hifi drivers claim to exploit).
The opposing H (Amp turns in the voice coil) do not push the magnets operating point up and down as much and mostly alter how the saturation is taking place.
The copper shorted turn produces an internal current IF the flux field moves through it. That transformer coupling produces an internal voltage and because it is one thick copper short, that current can be high. That current produces an opposing magnetic which opposes the motion of the flux. As this is a transformer (follows the other rules for them) this effect only is useful above a hundred Hz or so.
Keep in mind, THD is the sum of the distortion products, not a ranking of audibility by harmonic number and level against a pure tone background or weighted by frequency or account for masking like how we hear. It is the sum of all the drivers interacting nonlinearities.
It is pretty clear at least in one area, that when you use a cone driver and horn load it, it is the motor that is responsible for the majority of the distortion.
Essentially all of the problems loudspeakers themselves have as “parts” are problems that get larger faster than the desired signal when you increase the input level so one of the easiest ways to get very low distortion is to have the potential for a very great maximum output being part of the design.
As they used to say “headroom is your friend” and also a heavily saturated magnetic gap costs more and making loudspeaker drivers is largely about the cost.
Best,
Tom Danley
Danley Sound Labs
The opposing H (Amp turns in the voice coil) do not push the magnets operating point up and down as much and mostly alter how the saturation is taking place.
The copper shorted turn produces an internal current IF the flux field moves through it. That transformer coupling produces an internal voltage and because it is one thick copper short, that current can be high. That current produces an opposing magnetic which opposes the motion of the flux. As this is a transformer (follows the other rules for them) this effect only is useful above a hundred Hz or so.
Keep in mind, THD is the sum of the distortion products, not a ranking of audibility by harmonic number and level against a pure tone background or weighted by frequency or account for masking like how we hear. It is the sum of all the drivers interacting nonlinearities.
It is pretty clear at least in one area, that when you use a cone driver and horn load it, it is the motor that is responsible for the majority of the distortion.
Essentially all of the problems loudspeakers themselves have as “parts” are problems that get larger faster than the desired signal when you increase the input level so one of the easiest ways to get very low distortion is to have the potential for a very great maximum output being part of the design.
As they used to say “headroom is your friend” and also a heavily saturated magnetic gap costs more and making loudspeaker drivers is largely about the cost.
Best,
Tom Danley
Danley Sound Labs
When cost is no object you can go for Electromagnets on the bass end and a group of high power neodymium magnets on the midrange drivers:
Grande Utopia speakers technologies: EM drivers and multiferrite magnet
Grande Utopia speakers technologies: EM drivers and multiferrite magnet
So despite of compensational measures (shorting rings) the key
parameters affecting flux modulation in the magnetic circuit itself are
- coercitivity of the magnetic material ?
- saturation of the pole pieces ?
Using less coercitive material i should have more saturated pole
pieces to minimize flux modulation and approach constant H conditions ?
Best is to have both ?
Can it be put that simple ?
parameters affecting flux modulation in the magnetic circuit itself are
- coercitivity of the magnetic material ?
- saturation of the pole pieces ?
Using less coercitive material i should have more saturated pole
pieces to minimize flux modulation and approach constant H conditions ?
Best is to have both ?
Can it be put that simple ?
Sorry but I have always been lead to believe that this is primarily an issue with the magnetic material itself.
I know that magnetic steel has a saturation curve and is nonlinear, but steel was always present in Alnico structures and now in neodymium structures. But it was the conversion to Ferrite magnet structures in the late 70s that saw the rise in distortion and forced the alternative designs (for those that cared about woofer distortion).
Saturation of the steel parts may be a potential solution but it is a very expensive solution. It requires you to design a structure where the realized B field is, by definition, less than it could be. That is you use magnets so large (or core poles and top plates so thin) that you push heavy saturation somewhere and lose B strength. Its like designing a 16,000 Gauss structure but only getting 10,000 Gauss.
The alternative is a cheap aluminum ring at the base of the core pole as Mac and JBL and others use (and also compare that to the Grand Utopia scheme of a 20-pounds-of-copper field coil, plus large power supply).
I know that magnetic steel has a saturation curve and is nonlinear, but steel was always present in Alnico structures and now in neodymium structures. But it was the conversion to Ferrite magnet structures in the late 70s that saw the rise in distortion and forced the alternative designs (for those that cared about woofer distortion).
Saturation of the steel parts may be a potential solution but it is a very expensive solution. It requires you to design a structure where the realized B field is, by definition, less than it could be. That is you use magnets so large (or core poles and top plates so thin) that you push heavy saturation somewhere and lose B strength. Its like designing a 16,000 Gauss structure but only getting 10,000 Gauss.
The alternative is a cheap aluminum ring at the base of the core pole as Mac and JBL and others use (and also compare that to the Grand Utopia scheme of a 20-pounds-of-copper field coil, plus large power supply).
Interesting material but they again imply the underlying issue is about steel nonlinearity (HYSTERESIS in bold green letters) which I don't believe to be the case. From what I can gather they are using pressed powdered iron for the rings surrounding the coil. Such material has similar magnetic properties but since it is composed of many electrically insulated grains there can be no large eddy currents.
This is a similar technique to the ferrite cores used for antennas.
David S.
Here is a Seas driver that uses Neodymium magnets.
E0037-08S W15CH001
I have heard a Proac speaker with doped paper cones using the same Seas motor system.
Beautiful midrange, one of the best small passive speakers I have heard.
E0037-08S W15CH001
I have heard a Proac speaker with doped paper cones using the same Seas motor system.
Beautiful midrange, one of the best small passive speakers I have heard.
yes, the extra magnet increase BL, Rmec, decrease all 3 Q...But one friend of mine said whatever you change on the magnetic field you obligatory unbalance it thus distorsion obligatory increase...
Maybe we are not talking about the same thing. I hope this makes sense. Pardon my calligraphy.
Attachments
Casull
Your understanding is the same as mine. Between H1 and H3 there can be a small hysteresis loop, but the major effect is what you show. It has to be remebered that the flux changes act instantaneously and so will distort the waveform not just change the gain.
Hysteresis in the steel is completely insignificant.
Your understanding is the same as mine. Between H1 and H3 there can be a small hysteresis loop, but the major effect is what you show. It has to be remebered that the flux changes act instantaneously and so will distort the waveform not just change the gain.
Hysteresis in the steel is completely insignificant.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.