We were not discussing symmetric BL curves. We were discussing symmetric inductance.
BL curves are consistent with the force on the voice coil as a result of the current drive.
Le is the measure of stored energy within the system as a consequence of the current being driven into it.
They are not the same thing.
I am also discussing the energy dissipation as a consequence of eddy currents and flux dragging, however I believe that understanding is a ways away.
Small steps..
Jn
In circuit theory, we normally use terms as transfer impedance and transfer admittance and they both may be linear or nonlinear. The terms are used based on input and output variables. Admittance matrix are routinely used.
A nonlinear indefinite admittance matrix for modeling electronic devices - IEEE Journals & Magazine
You can do yourself with moderate effort. Take any fullrange driver with a rather cheaply built (=compromised) motor, closed box, align and/or EQ it to a resonable acoustic target. Then take an amp with enough voltage headroom that you can afford the voltage drop on a series resistor (30Ohms or so) to achieve high impedance drive and EQ it minutely to the exact same (+-0.2dB) FR target, incl. overall gain (DSP based EQ pretty much mandatory for this). Then compare the perceived sound and try measure two-tone IMD, say small 1kHz sine on top of a larger 100Hz'ish sine.
Sonically, I find typically two major effects of high-Z drive:
- MF/HF is cleaner, in general a more "see trough" translation.
- Bass is different, though not neccesarily better. The large signal behaviour around resonance is certainly different and it "detunes/overloads" with its own specific character, sort of.
The better the driver the less is the difference at MF/HF but the distinct bass behaviour, notably under large signal conditions, usually remains.
I mostly prefer some daming around resonance with most drivers, set up by specific feedback networks in the amp.
Last edited:
Actually the design shown in the MyRef threads have a current pump with a voltage feedback loop around an integrator, if you adjust the voltage vs current gain according to the driver needs, it provides pretty good results. I intend to give it a differential input and see if I can take advantage of some things learned from other composite amps to improve on it. This also gives and advantage such that it will benefit normal drivers and flatter impedance drivers as well. I am wondering to what extent I can shape the driver impedance through enclosure design.
It is the exact same thing.
Using a shorting ring to lower the inductance introduces magnetic drag and dissipates based on flux magnitude with velocity, and rate of change of flux. That energy loss is measureable as the Rs component of an Ls/Rs inductor model.
While that can equalize the equivalent inductance through the range of vc motion, it skews the dissipation with more occurring when the vc is closer to the bottom than the top.
Returning the a two sine example, if you are pushing the vc hard at 20 hz, with a 500hz second signal, the damping factor for the 500 hz component will be vc position sensitive.
Jn
Well, I have been entertaining the idea of a shorted voice coil former. Since there is no relative velocity between the voice coil and the former, there is almost no drag change. The tricky part is how to shoes the former so thermal expansion between different material does not destroy the driver.
Hasn't it been done before? Or are aluminum formers slit to break the secondary short?
That doesn't take care of the inductance slope though.
I would go composite non conductive, and I would make the magnetic circuit symmetrical so that the vc inductance is the same throughout it's excursion.
I recall a manu did this already, but I believe they were optimizing for power and spl. Their design has two voice coils, with two gaps of opposite polarities...the vc's were wired antiparallel so that both gap forces added, but the net NI on the magnetic circuit was zero. Another scheme would be to split the magnetic return path, half in front and half in back, so that the inductance is consistent on each side of the center. If your worried about the magnets, put copper shorting rings in the magnetic return path but not near the vc.
For sonic accuracy, optimization of different entities would be needed; you know, the normal engineering tradeoffs.
jn
Last edited:
hmm. That explanation is not floating my boat.
Aluminum expands at 25.5 ppm/C, copper at 16.6 ppm/C (both are alloy dependent of course, but most alloys are very close to those numbers.) The copper wire is bonded to the former, on the outer surface, or on both inner and outer. If there is a gap in the former, it can certainly absorb a radial expansion difference as one would think, but there is a problem.
The copper wire is bonded to the former everywhere along the circumference. When they expand relative, the bond CANNOT let go. The concept of using a slit in the former requires that the wire to former bond allow sliding. That is contrary to what is needed.
The only proper use of the slit is to allow differential expansion as the unit is being cured at elevated temperature. The former will expand more than the copper, the gap will close a bit, then the epoxy (polyimide) will harden and lock the parts together. When the unit is cooled, there will be built up shear forces between the copper and aluminum, but that would be considered in the design.
You could try using a wrap insulation on the copper wire which allows it to slide within the wrap, and that would zero the shear forces. However, you now rely on the insulation strength against both hoop forces and BLI forces. That won't work either.
Any attempt to mitigate differential expansion by layer design is also of little use. If the copper is on the outside, it would require an operational temperature rise that was 25/16 times that of the aluminum to get the same expansion total. Achieving that level of thermal control in the gap is not possible and requires heat flow in only one direction.
That is why I recommended a composite structure. E glass has an expansion coefficient very close to copper. A former made with carbon oriented axially along the pole tip axis and E glass circumfrentially (sp) would match the expansion rate of the copper wire. As an added bonus, it will retain strength down to 1.8 Kelvin, just in case your room gets very cold.
As to the slit, I believe it can only be relied upon to break the shorted secondary of the voice coil.
jn
I thought it was well known that there is a paper insulation between the copper wire and former, but maybe not.
For the reduction in Le to work, the former cannot have a slit. I am not familiar with e glass, but the stiffness to density ratio would also be a limiting factor if wide band response is a goal.
For the reduction in Le to work, the former cannot have a slit. I am not familiar with e glass, but the stiffness to density ratio would also be a limiting factor if wide band response is a goal.
Last edited:
It doesn't matter. There still has to be a solid bond that supports shear forces caused by TCE differentials. I don't believe a sheet of paper insulation can support the forces and still allow slip.
Again, engineering tradeoffs.
Fiberglass cloth can be made with E glass and S glass. PC boards use one of them, E I think. We used S glass because boron tended to stop radiation so would heat up in our applications. Because it takes a kilowatt of fridge power to remove 1 watt of heat from 4.5K, we tend to be finicky with materials.
jn
Agreed. I'm not saying paper cannot be there, but that I can't see it supporting circumferential slip between the aluminum and copper while withstanding axial forces in use. It has to be a total locked system.
True. But I also believe you'd have to have aluminum on both sides of the voice coil to shield all the iron from the field.
Jn
I suggested this before:
and you responded with:
Seems to me that you are contradicting yourself.
This has always been my understanding.
As far as the expansion at different rates, no glue is perfect and even epoxy has some flexibility so it seems to me that this differential expansion would be small enough that the bond could absorb it.
No, I am not.
Think about where the flux is going with an extended pole piece. Say for example, the vc has extended entirely out of the gap. Now, all the turns of the coil are trying to magnetize the pole piece. But the tip of the extended pole piece is floating in the air. How does the flux get from the extended tip back into the complete magnetic circuit? The air path has a high reluctance.
When the coil is in the fully opposite location, buried deep into the gap and beyond, the reluctance path is the entire magnetic circuit.
What I said was to have half the magnetic circuit on each side of the gap, so no matter where the coil is, there is a complete iron return loop.
jn
jn
I get you now and I agree. What I should say is that extending the pole piece, while not perfectly symmetric, is more symmetric than it would be otherwise. Agreed?
There are drivers done as you suggested. They have not made a big splash in the marketplace however. Perhaps it is not such a big deal. I will ask my designer friend at a big driver manufacturer what he thinks. I'll let you know what he says (if he responds, as you say, it might be confidential.)
I get you now and I agree. What I should say is that extending the pole piece, while not perfectly symmetric, is more symmetric than it would be otherwise. Agreed?
There are drivers done as you suggested. They have not made a big splash in the marketplace however. Perhaps it is not such a big deal. I will ask my designer friend at a big driver manufacturer what he thinks. I'll let you know what he says (if he responds, as you say, it might be confidential.)
Yes, total agreement.
It may just be a solution in search of a problem.
I have no problem with confidentiality. I also have no issue with anything I've posted being passed to or even used by industry. If I were worried about it, I wouldn't be posting it.
Having a different perspective and experience, some of my experience may be of use to somebody, as opposed to being lost in the dustbin of history.. If your friend would like to discuss inductance test issues, he is welcome to ask. As are you.
jn
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.