Gap width vs winding thickness

[Vikash]

I'm curious about the air gap width vs the winding width (thickness of all wire layers summed). You can get more B if the gap is narrower, but more windings (hence L) if gap is wider. I think I remember Dan mention that it's better to have more force coming from B?

Anyway, if I have a large gap does it make sense to wrap shorter windings but with more layers (thicker) which redudes the effective gap-air distance and subjectively increases the resulting force. The goal being increased efficiency.

[Otherwise]

I think it's true that more B from narrower gap is better. If you wind too much, the Qes could go higher as a result of increased resistance of the coil. So, the lower the resistance/inductance is, the higher the current trough the coil will be. Some drivers are sold in 4 and 8 ohm coil versions and it often seems to be that the 4 ohm has a lower Qes.

I don't really know too much about electricity etc but "someone" once told me the above

[sreten]

You can get more windings by making the gap longer.

Making the gap wider is a bad idea.

Filling the gap with windings does not make it appear thinner.

The gap needs to be small enough to allow the poles to saturate.

Once its thin enough theres no point making it thinner, conversely
if its too wide you need more magnetic force, i.e. a bigger magnet
to create the force required for the gap.

Making the gap longer also requires more magnet to provide
the extra flux needed.

For efficiency you need an optimum gap width and length to
suit the fitted magnet.

Your issue then is excursion capability, how much of the coil is
in the gap, the lower the excursion capability the higher the
efficiency and the lower Qe thus Qts is.

sreten.

[Svante]

This is an interesting question. As I have understood it, the volume of air in the gap multiplied by B is proportional to the energy stored in the magnet. This is given that the pole pieces can carry the magnetic flux to the gap and there is no stray field.
This gap, we should do our very best to fill up with copper. We would use a thin wire and many windings if we want a high impedance, and vice versa, but the gep should be filler with copper conducting the signal current.
Now, given the above, is it better with few layers of wire? This would make the gap narrower, and given a certain magnet energy, the height of the gap could be made longer. If we use, say 50% of the voice coil's length to be within the gap, and leave the rest as overhang, the Xmax value would increase for the VC with few layers. Also, the surface/volume ratio would increase, so heat dissipation would improve.

Are there any reasons at all to make the voice coil have more than one layer? Mechanical precision?

[mwmkravchenko]

Hello Vikash!

You have the correct idea on how things work. Ideally you would like to have many layers of wire in a gap to give a good B factor. The trade of is that many layers will need a small diameter wire and therefore push up the R of the coil. The higher R value will raise the Qes and limit the high frequency limit because of the higher inductance. More wire can also mean more mass and then less efficiency. There are happy mediums. You can have twice the number of turns by choosing wires with 1/2 the circular mills. The only problems is that your R goes through the roof. I think that I remember you playing around with FEM and that is a good place to start fooling around with the possible combinations keeping in mind your final goals of Qts, Qes R and so on. You can also do rough and ready calculations with a wire table and calculating the length that you will end up using. It will prove or disprove an idea very quickly based on diameter, resistance and how much will fit into your gap.

I'm going through the same joy in trying to come up with an optimum voice coil for our Parthenon. The shop may be closed but the mind still works!!!!

MArk

[sreten]

Originally posted by Svante
This is an interesting question. As I have understood it, the volume of air in the gap multiplied by B is proportional to the energy stored in the magnet.
Are there any reasons at all to make the voice coil have more than one layer? Mechanical precision?

When I was a young man the gap flux density was
given in Gauss and the total flux given in Maxwells.

I've always thought its the total gap CSA that counts,
having a wide gap means you need more magnetic force
to create the flux.

Filling the gap with copper is important as it minimises
the resistance of the coil, so how wide the gap is I
suppose is determined by coil resistance.

I don't know why but multilayer voice coils appear to
have higher power handling than single layer, perhaps
the extra layers help each other not to start coming
apart, when a single layer starts coming apart you
get a hotspot and failure is consequently inevitable.

sreten.

[Vikash]

A bit more background...

I've resurrected my DIY transducer project. One of the main problems was in manufacturing a usable former/voice coil. The two issues before were a) the gap size (1.2mm) was IMO too small for an experimenting project, and b) the former wasn't the right size.

So I've opened the gap by 1mm and had an aluminium former made to fit correctly into the gap (pic in first post). The drop in sensitivity is relatively large even though I've now overcome the VC scraping issue by using a better former.

Now, my mistake was in doing both things simaltaneously (manufacture a former and open the gap at the same time) as it would seem that having a suitable former would have been enough.

So anyway, I have a large gap to work with. This is a given. I think, my bottom line curiousity is based on the observation that flux travels poorly through air and by putting more copper in the gap, the distance is reduced (before the coil actes upon it). So the underlying question was whether in doing it would increase efficiency.

Excursion is a non-issue for this project.

Filling the gap with windings does not make it appear thinner.

Yes I see that the flux must travel between poles to complete the circuit and perhaps irrespective of how much opposing force is in it's path and/or where the flux lines are cut. Though I'm still trying to make clear sense of all these thoughts

The current winding is with 0.125mm wire and is approximately the same height as the top plate. It measures 7.6 Ohms already with only a single winding. If I change the winding to a third of its height with 3 layers, will it have an effect on efficiency? I also have thicker wire up to 0.265mm to play with. What's the best approach to increasing effeciency?

[CeramicMan]

Efficiency could be affected by the aluminium former. As the conductor moves through a change in the magnetic field, eddy currents form inside it and it thus dissipates energy (reduces Qms).

Usually no-one should care about the B quoted for a speaker, but rather it's more interesting to know what the BL product is. If you want a stronger force from the magnet motor, then increase the total amount of current travelling within the magnetic field. You can do that by either increasing the amplifier voltage or by using the limited amount of space in air-gap more effectively. The latter method reduces Qes.

If the voice-coil wires were superconductive, then the force would be huge even at low voltage. But, since you have to make do with copper, if you use thin wires a high proportion of space taken up will be for the enamel coating, and if you use thick wires there will be more unused space. Another option is to use extremely thin ribbon wire.

CM

[Ron E]

I would think that the spreadsheet I attached in the earlier DIY driver thread should help with most of these questions.

I tried adding a magnet design section using some equations and data from magnet suppliers on the internet, but I was getting gap flux values like 4T for what seemed like typical ceramic magnet configurations - either I am doing something wrong, or gaps are thoroughly saturated. I haven't yet compared to an FEM simulation....

I would suggest playing with FEMM for the magnet system design.

[sreten]

Surely the aluminium former will form a shorted turn in the gap ?

sreten.