Article: Loudspeaker Crossover Inductors Part 2
In this second sweep of loudspeaker crossovers we look at some of the details involved in the construction of both DIY and semi-production quantities of Loudspeaker Inductors. Information is already available on the quality rankings and for completeness it ranges from Iron core, Ferrite core, Air core and Foil type Inductors. Apart from Foils, all use 'wire' as a significant part of the manufacturing process. So lets look first at the wire types available.
Loudspeaker Crossover Inductors.
Inductor Rods and Bobbins.
Winding Wires for Loudspeaker Inductors.
The most often used is copper and more rarely silver, due to its cost. The wires
come in different diameters, purity, strandings, insulation coatings, and coating operating temperatures.
The most common wire sizes for loudspeaker inductors range from around 0.5mm to 1.4mm diameter. This is the diameter of the actual copper content "not" the overall diameter of the wire which includes coatings. The wires are coated with very thin electrically insulating materials generally known as enamels. There are many of these coatings all with different performance characteristics. The coated wires are graded, G1 is a single coating, G2 is coated twice and exceptionally G3, three times. Thicker coatings provide for greater abrasion resistance and insulation but reduce the overall copper content in a layered coil.
Wire purity is usually advised as the number of 9's which are included in a wires description for purity. This is otherwise specified in percentage terms. Hence "4 nines" implies 99.99% purity, whilst 5 nines" would be 99.999%. The higher purity specifications are preferred for loudspeaker inductors based upon perceptions regarding the overall sound quality.
Self Bonding Wires.
By using self-bonding wires to form self-supporting coils these types of
coils can be used to manufacture Air-cored inductors of various shapes without a bobbin and so for production use
are more cost efficient by using less materials. However, this process requires reusable, collapsible coil formers
and requires some thought and jigging.
The winding wire, be it Grade G1 or G2, is further coated with a thin chemical layer which softens under heat and adheres to adjacent wire turns / layers. This tends to form a solid copper coil which is far less prone to coil turn vibrations. The vibration can otherwise be a source of signal noise and distortion in use.
To self-bond the coils and layers the wire needs to be heated around (100C - 120C). The techniques for this range from direct exposure using a precision hot air gun during winding, self heating with a large calibrated current or ovenised heating.
For accurate Inductance values three processes need to be consistent from
coil to coil, the wire tension, the number of turns and turns / winding layers. If you are winding a few DIY
coils then this might be overlooked somewhat providing you can keep the wire moderately tight whilst winding.
It's then relatively easy to wind more turns than needed and then by inductance measurement cut off short
lengths of wire until the target inductance value is reached. For volume quantities of inductors this cut and
measure system is wasteful in both materials and time, hence the consistency requirements as above.
There are tables which identify the maximum recommended tension that a chosen cable
diameter can sustain without change of properties or complete failure. The first failure mechanism is stretching. This
reduces the cable diameter and hence the electrical characteristics of the wire / inductor. The second and more
permanent is, it snaps.
It's not uncommon to use around 50% to 80% of the maximum safe cable loading. For normal loudspeaker inductors using 0.5mm to 1.25mm diameter copper wires this results in wire tensions of around 1.4Kg to 9Kg. For DIY use this is unhelpful without rather expensive specialised jigging.
There are a number of designs of wire tensioners ranging from simple mechanical to sophisticated electro / magnetic control. The mechanical versions, whilst fully functional, suffer from wire friction losses and are less able to provide high tension consistency due in part to physical differences between static and dynamic friction. Permanent magnet versions provide non-contact magnetic drag / back tension the wire spooler. In order to function correctly these devices require moderately constant wire velocity and pull through torque. These types of accessories are expensive and not really cost effective for the DIYer requiring just a few coils. It's much more efficient to purchase them, but hand winding a few is ok too.
Ferrite Rods, Cores and Bobbins.
There are a wide range of materials and shapes used in the manufacture of
Loudspeaker Inductors. The most popular being:- Air-Core Bobbins, (usually plastic mouldings),
Ferrite Rods and Bobbins
(ferromagnetics) and iron core inductors consisting of laminated steel sheets or iron power with a binder.(ferrimagnetics).
For the DIYer, air core bobbins are inexpensive and easy to use. They don't saturate and provide a very acceptable sound. Their one disadvantage is their restricted inductance range for an acceptably low Coil DCR. This can be overcome with thick wire but the coil dimensions become unreasonably large and therefore present difficulties with crossover layout.
Next up are Ferrite Rods. These are generally available in 12 and 19mm diameters and of length up to 50mm. Again easy to wind but more difficult to keep the wire on the rod. It needs end-caps. For a few coils you can usually make and glue end-caps out of thick card or thin sheet plywood. This problem is usually solved by using self-bonding wire. An advantage of Ferrites is their extended inductance range but with the disadvantage of saturation at high power levels.
Ferrite bobbins are available in a number of sizes and by volume can contain more ferrite than rods and so the larger varieties are harder to saturate than easily available rods, The bobbins are made from a central ferrite core with the ferrite flanges attached by adhesive.
There is a tendency for multiple layer windings to spread outwards and this has to be controlled. Large diameter wires with high wire tension can apply considerable sideways forces to the bobbin flanges. As ferrite is a non-flexible material this pressure can be sufficient to shatter the ferrite flanges. Being of moulded materials Air-cored bobbin flanges can flex but the bobbin flanges should be supported to provide consistency in turns per layer and therefore inductance value.
Iron-cored inductors are made from insulated soft iron laminations. Usually stamped and used in strip format, the laminations are coated and bonded together to form the iron core. The assembly is wrapped in insulating tape to prevent the corners shorting out the copper turns. Iron cores can usually be identified by their square or rectangular end section as ferrite rods are nearly always round..
Build an Inductor.
There are many designs of Loudspeaker Inductors based upon core varieties, bobbin shapes,
wire stranding and preferred target sound quality. Good inductors are expensive and better than good, are very expensive.
We look here at a few hints and tips for the DIYer to either make or source them.
(1) You need an Inductance meter. If you don't have one it's going to be tough, either buy one or buy finished Inductors, there's no way out.
(2) For small value inductors, say less than 0.4mH in value use an air cored bobbin, size 17mm by 40mm diameter
and thickish wire (0.8 to 1.0)mm wire should be fine. 0.8mm is slightly easier to handle. The resulting
sound quality is good. These smaller inductor values are usually used in Tweeter crossover sections and
are moderately small and so don't carry high currents.
For a few inductors the core formers can be cut from cylindrical wood dowels, (12 and 18)mm diameter will work. On the dowel ends glue flanges cut from either thick card or thin plywood.. The Dowel lengths can vary but between (20 to 30)mm is good. You can view the sizes known to work here, Bobbins. From a purist viewpoint the depth of coil "layers" should equal the coil "core" height / length, this saves a bit of wire and gives the most efficient coupling but......
(3) An accurate turns count is required. For small inductors you can afford to calibrate your coil bobbin first. This requires you to waste a bobbin, full of wire, but you end up with a calibrated design If you can use the same bobbin size for several inductor values you only need to do the calibration once.
Keep the wire tight and wind the bobbin to capacity. For consistency you need to keep the coils well packed, so no gaps and spaces. Try to maintain the same number of turns per layer. Remember, this is your calibration bobbin. Keep a careful count of the number of turns. Screw this up and it's coil in the bin. Measure its inductance, if it's less than you need, oops, use a larger / longer former.
Unwind the coil say 5 turns a time and measure the length of wire cut off, and the coil inductance. You need to keep the wire tight whilst doing this else it will unravel, changing the inductance value. Keep the data accurately listed. This list should look something like this.
TURNS REMOVED // WIRE LENGTH REMOVED // TOTAL WIRE REMOVED // L1mH // TOTAL TURNS//
Continue the unwinding and measurements process till the logical conclusion: no wire and no inductance. Plot a graph of the cumulative No. of turns against each Inductance measured. With a little more work you can also plot wire length against Inductance.
You now have a graph / graphs from which you can read off the turns OR wire length required for a particular inductance.
The same calibration procedures can be used for Ferrite bobbins where for the same value of inductance the wire length will be less.
(4) A single wire turn can make a large difference to the Inductance. With very few winds / turns the change is small but one turn on a fully wound coil can make up to 0.8% difference in inductance value. If you are looking for tight tolerances this shows the importance of accurately counting turns or wire length consistency. It is better to over-wind the turns required by say 10 turns and then trim the wire length, re-measuring the inductance as you go..
(5) There are formulae from which you can calculate the coil inductance from the mechanical details of your prescribed coil / bobbin geometry but it's not always reliable and you have no previous data with which to compare the calculated result so.....It is said that more consistent inductors can be wound knowing / using just the wire length, but (20 - 30) meters of loose wire usually has turns embedded into it which makes it knot easily. Removing the knots is difficult.
(6) If you are not using thermo-bonded wires, (you need the equipment to use it), then coil vibrations can usually be suppressed by using coil locking lacquer. This can be applied with a small brush painted over the layers as you wind them. This gets tricky if you have a bobbin in one hand and wire in the other. As this material is expensive, yacht varnish which does a similar job can be used. However, it takes longer to dry.
(7) Having made your coils you now need to mount them. Three easy methods
are:- (silicone, tie wraps, or non-ferrous bolts).
Silicone is good as the bobbins can be stuck, flange down, straight onto your board. It also absorbs vibration. Note however, Once set, it is very difficult to move or remove the coil so....
Tie wraps can be used in a variety of threadings to fix the coil in place where silicon may not be the first choice. Non-ferrous fittings such as M3, M4 or M5 brass or nylon bolts work well. If you can't find the required length then you can you can always make something out of threaded brass rod plus brass nuts/washers. If you used a wood dowel for a bobbin core then you need to drill a hole through it first..
(8) And ,lastly, if you have a small crossover board and a number of inductors which are short on real-estate then you should try to mount the coils each on different central axis. this will help to minimise any inter-coil pickup / interference.
(9) So there you have it. If it all sounds a bit too involved QTA Systems can provide either bobbins in various sizes or completed Inductors. There is a comprehensive selection of Loudspeaker Inductors here. Custom values are also available. All our inductors are wound to order, process time approximately 7 working days.