A very simple wire tensioner

[Joe Gwinn]

One thing to emerge from the long thread (http://music-electronics-forum.com/t22650/) on the Tanac magnetic tensioner is that it seems that Tanac allows the hysteresis brake to do most of the work for fine wire, such as that used in pickups - there is no tension feedback from dancer to brake, and the constancy of tension comes mostly from the inherent properties of magnetic hysteresis.

This allows for a great simplification - all that's needed is a disk made of thin spring steel, some permanent magnets held nearby, with the distance adjusted to adjust the tension, a capstan, a spring-loaded wool felt pre-tensioner between wire spool and capstan (so the wire won't slip on the capstan, and a music-wire dancer (to smooth out the pulsations from winding on oblong bobbins).

My initial thought is a capstan with a thin spring steel disk attached, rotating on a pair of small ball bearings. Nearby is a mild steel plate with four or six little rare-earth magnets on it, held in place by an aluminum spacer with holes for the magnets, screwed to the mild steel plate. There would be some kind of mechanism to allow the magnet assembly to be moved closer and farther from the capstan assembly.

The wire groove in the capstan is made large enough to accommodate a small O-ring, which provides the friction to prevent slippage (patent 3,520,492, figure 3).

Similar mechanisms have been patented, for instance 2,724,065 (which seems to use Hysterloy, a form of Alnico 5 but with different heat treatment) and 2,519,882 (which probably uses hardened steel).

So, I'm thinking about possible approaches, the intent being a design simple enough for pickup makers to make their own.

[Mike Sulzer]

That was clever, finding patents relating to tensioning yarn, etc in the textile industry.


If the eddy currents are induced in the spring steel (am I correct about that?), then why mount the neos on a steel plate? Could you use three aluminum disks in a sandwich? Use short cylindrical magnets that fit into holes in the middle plate, which is as thick as they are long. Then a thinner plate on each side, with the whole assembly held together with three to six flat head screws passing through one outer plate and tapped into the other.

It might be fun to make the magnet assembly rotate with the spring steel fixed. Then you could measure the currents in the steel conveniently to verify how it works. I suppose you could also use the currents to measure the rotation rate.

[copperheadroads]

One thing to emerge from the long thread (http://music-electronics-forum.com/t22650/) on the Tanac magnetic tensioner is that it seems that Tanac allows the hysteresis brake to do most of the work for fine wire, such as that used in pickups - there is no tension feedback from dancer to brake, and the constancy of tension comes mostly from the inherent properties of magnetic hysteresis.

This allows for a great simplification - all that's needed is a disk made of thin spring steel, some permanent magnets held nearby, with the distance adjusted to adjust the tension, a capstan, a spring-loaded wool felt pre-tensioner between wire spool and capstan (so the wire won't slip on the capstan, and a music-wire dancer (to smooth out the pulsations from winding on oblong bobbins).

My initial thought is a capstan with a thin spring steel disk attached, rotating on a pair of small ball bearings. Nearby is a mild steel plate with four or six little rare-earth magnets on it, held in place by an aluminum spacer with holes for the magnets, screwed to the mild steel plate. There would be some kind of mechanism to allow the magnet assembly to be moved closer and farther from the capstan assembly.

The wire groove in the capstan is made large enough to accommodate a small O-ring, which provides the friction to prevent slippage (patent 3,520,492, figure 3).

Similar mechanisms have been patented, for instance 2,724,065 (which seems to use Hysterloy, a form of Alnico 5 but with different heat treatment) and 2,519,882 (which probably uses hardened steel).

So, I'm thinking about possible approaches, the intent being a design simple enough for pickup makers to make their own.

I'd Love to see this

[salvarsan]

Outstanding research.

Google Patents has these.

The first, MAGNETIC DRAG FOR CONTROL OF YARN TENSION, is a minimal, but explicit device.
It looks like it would be easy to craft by hand.

The other, CONSTANT TORQUE DEVICE, is somewhat underdocumented.

Researching tension control for yarn and fiber filaments was inspired, a reminder that
Gibson, after all, conscripted a modified yarn winding machine for making humbucking pickups.
Leesona, to hear them tell it, has never _not_ made fiber-handling machinery.

[David King]

Mike,

I think the point of the hysteresis drag device is that the drag is constant regardless of the rotational speed. That said I see no reason why the magnet holding plate shouldn't be made from a nonmagnetic material in the way you describe here.

It would be nice to measure that drag and perhaps a piezo torque sensor in the axel of the fixed disk would tell you exactly what you need to know electronically but i could also foresee a spring-loaded, mechanical torque sensor giving you the same info in real-time via a needle over a scale.

If the eddy currents are induced in the spring steel (am I correct about that?), then why mount the neos on a steel plate? Could you use three aluminum disks in a sandwich? Use short cylindrical magnets that fit into holes in the middle plate, which is as thick as they are long. Then a thinner plate on each side, with the whole assembly held together with three to six flat head screws passing through one outer plate and tapped into the other.

It might be fun to make the magnet assembly rotate with the spring steel fixed. Then you could measure the currents in the steel conveniently to verify how it works. I suppose you could also use the currents to measure the rotation rate.

[David King]

This might seem extra complicated for our purposes but I see one of the patents uses electromagnets to vary the drag by varying the current applied rather than by mechanically adjusting the distance between the magnet disc and the spring steel disk.

[salvarsan]

It would be nice to measure that drag and perhaps a piezo torque sensor in the axel of the fixed disk would tell you exactly what you need to know electronically but i could also foresee a spring-loaded, mechanical torque sensor giving you the same info in real-time via a needle over a scale.

I had pondered over measuring tension using a parallel beam load sensor.

American vendors have priced themselves out of the market so it is easier to gut the sensor from a cheap 100 gram digital scale in the $5-10 range.

Who sells piezo torque sensors?

[Joe Gwinn]

That was clever, finding patents relating to tensioning yarn, etc in the textile industry.

Thanks, but I must confess that it's a well-traveled path. The textile industry is orders of magnitude larger than the wire-winding industry, and started far earlier. The first industrial revolution was largely about the mechanization of textile production. The wire-winders adapted textile machinery to wire.

If the eddy currents are induced in the spring steel (am I correct about that?), then why mount the neos on a steel plate? Could you use three aluminum disks in a sandwich? Use short cylindrical magnets that fit into holes in the middle plate, which is as thick as they are long. Then a thinner plate on each side, with the whole assembly held together with three to six flat head screws passing through one outer plate and tapped into the other.

There are eddy currents in the spring steel disk to be sure, but that's not the main effect, as the permeability is low and the resistivity is high. The drag arises mostly from magnetic hysteresis.

It would be difficult to use the eddy currents to measure speed, as the voltages are very low, and hard to extract from the metal. I would be tempted to instead put a row of holes in the spring steel disk, and use an optical sensor to mark their periodic passage.

The purpose of the mild steel plate is to close the magnetic circuit, thus strengthening the magnetic field at the spring steel disk. One could also have a second stationary mild steel plate on the other side of the spring steel disk to further increase the field, but at the expense of somewhat reducing the adjustability of that field.

It might be fun to make the magnet assembly rotate with the spring steel fixed. Then you could measure the currents in the steel conveniently to verify how it works. I suppose you could also use the currents to measure the rotation rate.

One can certainly do this, but the main reason to spin the spring steel disk versus the magnet assembly is that the rotational inertia of the disk is far less than that of the assembly. Given that the wire speed pulsates when winding bobbins, the lower the rotational inertia of disk plus capstan, the smoother the winding process.

List of moments of inertia - Wikipedia, the free encyclopedia

For every action, and equal and opposite reaction: The magnet assembly will experience the same magnitude of torque as the capstan, but directed in the opposite rotational direction. If one mounts the magnet assembly on an axle, one can measure the torque with a simple arm, spring and scale assembly. Some kind of mechanical damping (an airpot for instance) would be needed to low-pass filter the pulsations from winding oblong objects, or it may prove difficult to read the scale for all the bouncing around.

[Mike Sulzer]

The drag arises mostly from magnetic hysteresis.

Right you are. I am getting my effects confused,

The purpose of the mild steel plate is to close the magnetic circuit, thus strengthening the magnetic field at the spring steel disk.

I think the effect of the steel plate would no be big enough to make it worthwhile for that purpose. The neos are in that circuit as well, and their permeability is only about 2% greater than a vacuum, and so they do not direct the flux induced in the plate to the spring steel.

If the field were judged inadequate, one could instead use slightly larger magnets.

[Joe Gwinn]

I think the effect of the steel plate would no be big enough to make it worthwhile for that purpose. The neos are in that circuit as well, and their permeability is only about 2% greater than a vacuum, and so they do not direct the flux induced in the plate to the spring steel.

If the field were judged inadequate, one could instead use slightly larger magnets.

I think the mild steel plate roughly doubles the field at the spring steel disk (by shorting out the field in the back half-space), but haven't modeled it in FEMM. I do note that the patented hysteresis drags generally use mild steel plates. One can certainly use a larger magnet instead. Having a big field out the back may be a nuisance as well - tends to collects tramp iron.

The neo magnets would have alternating polarity, N S N S ..., in effect replicating the field of the multipole ferrite magnet used by Tanaka, which has all poles on one face of the magnet. If one can easily buy such magnets in small quantities, that would work too.

[Mike Sulzer]

I think the mild steel plate roughly doubles the field at the spring steel disk (by shorting out the field in the back half-space), but haven't modeled it in FEMM. I do note that the patented hysteresis drags generally use mild steel plates. One can certainly use a larger magnet instead. Having a big field out the back may be a nuisance as well - tends to collects tramp iron.

The neo magnets would have alternating polarity, N S N S ..., in effect replicating the field of the multipole ferrite magnet used by Tanaka, which has all poles on one face of the magnet. If one can easily buy such magnets in small quantities, that would work too.

I believe doubling is the correct upper limit, one that would apply when the two dimensional circuit analogy is valid. This would require that the flux resulting from the magnetization induced in the plate be directed through the magnets into the spring steel, in effect just increasing the flux present from the magnets. In practice, the field induced in the plate points in the correct direction (towards the spring steel) only very near the magnets, and much of the flux would be expected to exit from the plate and return to it on paths that do not take it through the spots illuminated directly by the magnets. (The magnets, having low permeability, do not direct the flux.) In effect, the properly directed magnetization illuminating the spots on the the spring steel is further away than the magnets, and it represents a relatively small amount of magnetization since the flux paths change direction in the plate..

[Joe Gwinn]

All true, but what point are you trying to make?

It also occurs to me that the mild steel plate could be fixed, with the magnet assembly (aluminum plate with six neo pill magnets mounted in a ring) moving towards or away from the spinning spring steel disk.

[Mike Sulzer]

All true, but what point are you trying to make?

Well, there are two, one specific:

1. If you are expecting that factor of two doubling in the field, you will be disappointed. It is not going to happen.

And one general:

2. Noting the relationship between E = IR and F = (phi)(Rm) does not establish that the second relationship applies to the magnetic circuit in question. J = (sigma)E and B = (mu)H do apply, at least when we have linearity. They both require solving a differential equation.

The second of these limitations applies to this case: (Magnetic circuit - Wikipedia, the free encyclopedia)

Limitations of the analogy

When using the analogy between magnetic circuits and electric circuits, the limitations of this analogy must be kept in mind. Electric and magnetic circuits are only superficially similar because of the similarity between Hopkinson's law and Ohm's law. Magnetic circuits have significant differences, which must be taken into account in their construction:

• Electric currents represent the flow of particles (electrons) and carry power, which is dissipated as heat in resistances. Magnetic fields don't represent the "flow" of anything, and no power is dissipated in reluctances.
• The current in typical electric circuits is confined to the circuit, with very little "leakage". In typical magnetic circuits not all of the magnetic field is confined to the magnetic circuit; there is significant "leakage flux" in the space outside the magnetic cores, which must be taken into account but is difficult to calculate.
• Most importantly, magnetic circuits are nonlinear; the reluctance in a magnetic circuit is not constant, as resistance is, but varies depending on the magnetic field. At high magnetic fluxes the ferromagnetic materials used for the cores of magnetic circuits saturate, limiting the magnetic flux, so above this level the reluctance increases rapidly. The reluctance also increases at low fluxes. In addition, ferromagnetic materials suffer from hysteresis so the flux in them depends not just on the instantaneous MMF but also on the history of MMF. After the source of the magnetic flux is turned off, remanent magnetism is left in ferromagnetic circuits, creating a flux with no MMF.

The case under discussion here has a lot of leakage flux since the magnets cannot direct it where we would want it.

[Joe Gwinn]

Again, all true, but I think we are over-thinking this. I'm just going to try it.

[Mike Sulzer]

Again, all true, but I think we are over-thinking this. I'm just going to try it.

Good idea. With a 3/16" by 1/32" neo button measured 3/16" away, I get a 60% increase using a tele bridge bottom plate. That is closer to a factor of two than I expected. Your prediction was better, although this is very dependent on the geometry. (The same plate makes only about a 5% difference on the pickup.) The key to getting close to the factor of two is a very thin magnet. If it is about as thick as the plate and has a significantly larger diameter than its thickness, then the flux from the magnetization induced in the plate and the flux from the magnet are nearly the same and spread in approximately the same pattern; then you can approach the factor of two. This fact can be a guide in determining whether the plate is worthwhile depending on the magnets you use.