Magnetic Tensioner (heads up for the Brits)

[David King]

I was wondering about this disc some more and it occurs to me that the ideal material would be a medium-hard magnetic, non-conductive ferrite material if such a thing existed. I wonder if an old hard drive would qualify...

[ken]

You mean like the media disk itself out of the drive? Cool idea...

ken

[Jimbosticks]

I couldn't resist the temptation to come back just for quick look. Mr. Gwinn and others keep throwing around patents, saying this is a hysteresis device, but neither the original Tanak patent (US4526329) or the one which followed 6 1/2 months later (US4570874), make any mention of "hysteresis" or "eddy currents". They also don't say the disk gets warm, or is made of "tempered steel".

" The braking torque generating means provided applies the brake torque to the main tension pulley 12 by means of a permanent magnet 38 and opposing magnetizable disk 39, such as an iron plate."

An iron plate is a "soft magnetic". Iron doesn't have a lot of retentivity. That's why millions of grade school students make electromagnets for school projects using soft iron nails. You apply the current, and the magnetic field builds. Turn off the current, and most of the field is gone. I've found that units that sit for a long time develop some very weak irregularly placed "poles", on the disks, and that these spots cause the wire feed to become "lumpy". Demagnetizing the disk seems to help with the "lumpiness".

Still, I'm not a physicist or engineer, and don't pretend to know the actual goings on with this unit. It just seems intuitive to me that the closer the magnet is to the disk / plate, the more "drag" you get. And letting the unit sit idle at a high tension setting will eventually impress some weak "poles" on the disk. That is why I have suggested all along that the disk is or SHOULD Be made of soft iron.

I DO think the guy who patented it probably knows more about how it works than the rest of us. And Tanac isn't talking. Someone here asked me what Tanac told me about how the unit worked, and what could be causing our problems. The answer is, neither the USA office, or the Japanese office had the courtesy to reply to numerous emails. (I think I contacted the European office too.) And that is not a hallmark of a company that makes a quality product or wants repeat business.

Anyway, MY big concern is how to keep the six units we use working consistently (on the days they are run). I just did our quarterly tension tests and one of the units had an increase of 10 - 20 grams since the last inspection. All across the adjustment range. In this case however, demagnetizing the disk did not fix the problem, and the threads didn't appear to be worn badly enough to tip the magnet into the plate. So if the smart people here have any other ideas what could be going wrong with these terribly expensive units ($1000 USA), I'm all ears. I have two to three disassembled units which we took out of service for the same issues.. At one point, they went back to Tanac, and we were charged a heap of money (~$300 ea.) for a "calibration" - which fixed nothing. If someone here wants to offer something for them, I will talk to the boss about it. I haven't the authority to sell them on my own.

I believe in paying my way. People here have asked repeatedly for pictures of the inside of the units. Here's some! You will notice the plastic threads in the faceplate, and the threaded brass "slug" (sits behind the knob), which runs in those plastic threads. The face of the magnet has 8 distinct poles arrayed radially around the magnet (45 deg apart). The ferrous disk, which can be seen inside the unit is glued to a plastic wheel. The friction brake at the top of the disk is released when the tension arm pulls down to the horizontal position. I regret that I failed to get a picture of the iron filings sticking to the disk in clumps.

[Mike Sulzer]

I couldn't resist the temptation to come back just for quick look.

If you read my comments earlier, you see that I agree with you partly, that the patent not only does not mention hysteresis, but appears to be describing something else altogether. However, it cannot work as described. Introducing drag requires dissipating energy, and the described mechanism involving attractive and repulsive force does not do it. Since a magnetic disk is essential, the only explanation that looks reasonable is that it is hysteresis damping. This does require a material that can retain magnetism, but the disk clearly can retain at least some magnetism according to your observations of it.


So what is going on? Could the properties of the disk change so that it retains "lumpy" magnetism? This seems unlikely, but who knows.

Have you checked carefully to see that the magnets are OK? Maybe the lumpy magnetized disk could result from problems with the magnets.

[David King]

Jumbosticks,
Thanks for dropping back in and adding internal photos to the discussion!

If it is a soft iron disc (and it's very hard to tell from the photos what either side of the mechanism are made of) then I can believe that the retentivity of the iron could change over time. Anyone who has had an old fuel pump relay go bad on a hot day knows how this can happen...

Looking at the threads in the plastic faceplate, it's hard to imagine how they could wear out unless that brass was very roughly finished. They are indeed unlikely to exist anywhere else.

The magnetic disk looks like ceramic, I don't see the 8 alternating segmentations in the photos but undoubtedly they are alternating NSNSNSNS poles.

[Steve Conner]

Thanks for the photos, Jimbosticks!

I think I understand what's going on. If the ceramic magnet really is magnetized NSNSNS and so on (this is a quite common arrangement in floppy disk drive motors and could be available off the shelf) then some of the poles could put a stronger field into the iron disc than others. Either because the magnet tilts as mentioned above, bringing some of the poles closer to the iron disc than others, or because of some aging process within the magnet.

If the machine is left in this state for a while without turning the disc, then the lumpiness could set in, as the disc sits in an asymmetrical field and takes on an asymmetrical magnetisation.

If it were me I would tell the operators to back the tension knob off to minimum any time the unit isn't in use. Of course the risk is that the extra twiddling will wear out those cheap-ass plastic threads. :-(

[Joe Gwinn]

I couldn't resist the temptation to come back just for quick look. Mr. Gwinn and others keep throwing around patents, saying this is a hysteresis device, but neither the original Tanak patent (US4526329) or the one which followed 6 1/2 months later (US4570874), make any mention of "hysteresis" or "eddy currents". They also don't say the disk gets warm, or is made of "tempered steel".

It's true that Tanaka does not say that he is using a hysteresis brake design, but he is, as discussed elsewhere in the thread. I put a lot in posting #27 et seq.

" The braking torque generating means provided applies the brake torque to the main tension pulley 12 by means of a permanent magnet 38 and opposing magnetizable disk 39, such as an iron plate."

An iron plate is a "soft magnetic". Iron doesn't have a lot of retentivity. That's why millions of grade school students make electromagnets for school projects using soft iron nails. You apply the current, and the magnetic field builds. Turn off the current, and most of the field is gone. I've found that units that sit for a long time develop some very weak irregularly placed "poles", on the disks, and that these spots cause the wire feed to become "lumpy". Demagnetizing the disk seems to help with the "lumpiness".

The key word (now underlined in the quote) is magnetizable. Soft iron is not magnetizable in this sense.

Tanaka's description of attraction and repulsion is a clumsy statement of how hysteresis brakes work. I bet that Tanaka speaks no English, and that the Japanese patent is far clearer.

Still, I'm not a physicist or engineer, and don't pretend to know the actual goings on with this unit. It just seems intuitive to me that the closer the magnet is to the disk / plate, the more "drag" you get. And letting the unit sit idle at a high tension setting will eventually impress some weak "poles" on the disk. That is why I have suggested all along that the disk is or SHOULD Be made of soft iron.

The first two statements are true, but the conclusion (to use soft iron) does not follow in that it undermines the hysteresis mechanism that generates the needed drag.

I DO think the guy who patented it probably knows more about how it works than the rest of us. And Tanac isn't talking. Someone here asked me what Tanac told me about how the unit worked, and what could be causing our problems. The answer is, neither the USA office, or the Japanese office had the courtesy to reply to numerous emails. (I think I contacted the European office too.) And that is not a hallmark of a company that makes a quality product or wants repeat business.

Hmm. Japanese society is very polite. Consider the attitude and tone of voice of your earlier postings in this thread.

And Tanaka does know how the unit works, and documented his theory of operation in his patent. If one is telling him that the patent is wrong, it may require some proof and persuasion.

Anyway, MY big concern is how to keep the six units we use working consistently (on the days they are run). I just did our quarterly tension tests and one of the units had an increase of 10 - 20 grams since the last inspection. All across the adjustment range. In this case however, demagnetizing the disk did not fix the problem, and the threads didn't appear to be worn badly enough to tip the magnet into the plate. So if the smart people here have any other ideas what could be going wrong with these terribly expensive units ($1000 USA), I'm all ears. I have two to three disassembled units which we took out of service for the same issues.. At one point, they went back to Tanac, and we were charged a heap of money (~$300 ea.) for a "calibration" - which fixed nothing. If someone here wants to offer something for them, I will talk to the boss about it. I haven't the authority to sell them on my own.

Even when the $300 recal failed, Tanac was still silent?

I'll offer $50 for your worst MT-100 unit.

I believe in paying my way. People here have asked repeatedly for pictures of the inside of the units. Here's some! You will notice the plastic threads in the faceplate, and the threaded brass "slug" (sits behind the knob), which runs in those plastic threads. The face of the magnet has 8 distinct poles arrayed radially around the magnet (45 deg apart). The ferrous disk, which can be seen inside the unit is glued to a plastic wheel. The friction brake at the top of the disk is released when the tension arm pulls down to the horizontal position. I regret that I failed to get a picture of the iron filings sticking to the disk in clumps.

Very interesting. Brass on plastic ought to work OK. Is there any reason one cannot replace the plastic faceplate?

In their sales literature, Tanac states that the magnet is ferrite.

By the way, where are you geographically, what are you winding, and what wire sizes and tension?

[Jimbosticks]

Thanks for the photos, Jimbosticks!

I think I understand what's going on. If the ceramic magnet really is magnetized NSNSNS and so on (this is a quite common arrangement in floppy disk drive motors and could be available off the shelf) then some of the poles could put a stronger field into the iron disc than others. Either because the magnet tilts as mentioned above, bringing some of the poles closer to the iron disc than others, or because of some aging process within the magnet.

If the machine is left in this state for a while without turning the disc, then the lumpiness could set in, as the disc sits in an asymmetrical field and takes on an asymmetrical magnetisation.

If it were me I would tell the operators to back the tension knob off to minimum any time the unit isn't in use. Of course the risk is that the extra twiddling will wear out those cheap-ass plastic threads. :-(

Excellent suggestion, and one I had wanted to do. However, although I am responsible for the maintenance of the units, I don't have any authority on the production floor. So I don't know if any of the workers will listen to me.

Incidentally, for those who wondered, there are no visible physical manifestations of the poles on the ceramic magnet. However the manufacturer magnetized the ceramic with a NSNSNSNS radial polarization. On the threads.. something is wearing, because with all the daily adjustment our operators give the machine, they get very loose in a couple years. Since I am not seeing any brass particles inside the boxes, I don't think it is the brass which is wearing.

[Jimbosticks]

......

Hmm. Japanese society is very polite. Consider the attitude and tone of voice of your earlier postings in this thread. And Tanaka does know how the unit works, and documented his theory of operation in his patent. If one is telling him that the patent is wrong, it may require some proof and persuasion.

Even when the $300 recall failed, Tanac was still silent?

I'll offer $50 for your worst MT-100 unit.

Very interesting. Brass on plastic ought to work OK. Is there any reason one cannot replace the plastic faceplate?
......
By the way, where are you geographically, what are you winding, and what wire sizes and tension?

Joe,

I can generally tolerate it when people carefully explain to me why I am wrong about something. However, I get excitable when I think someone is suggesting I am not being truthful about my observations. I got that impression from some of your postings, and perhaps I "lashed out". Sorry about that. I'm not normally a hot head.

I do usually begin my conversations respectfully. And that is how I began with Tanac. I never referred to the patent when I contacted them, because I wasn't even aware of it until you pointed it out. But after the $300 / unit calibrations failed to correct the problem, Tanac USA stonewalled. They wouldn't put me in touch with their engineers, or offer any more assistance. (Tanac Japan never made any replies.) At that point I got a little confrontational with the home office. I know the Japanese are a smart, industrious people who are generally very polite. However, I purchased a LEMON Toyota a few years back, and found that some are completely unwilling to admit when they make a mistake.

I have replaced the plastic faceplate / window on a couple units. (It is expensive.) It has generally helped with the gradual climb in tension. However the threads on this latest machine weren't very loose. I thought perhaps the machine had been left on the highest setting for too long, and the disk had become heavily magnetized. But the iron filings didn't reveal a big problem. Just a few small, randomly placed spots where the filings clung to the disk. I eliminated that with the demagnetizer, but the tension is still offset about 8 grams from the readings I made three months ago. The operator had been complaining about breaking wires.

We're in Central NY, USA. Small company. We manufacture custom inverters and drivers for LCD backlighting. Everything is done by hand. No automation. The workers who wind our transformers for those inverters use wire gauges from 38 to 44. I'm not sure the boss would take $50 for a broken unit, but I will ask. You're in Boston?

[Mike Sulzer]

Strong ferrite magnets have a preferred axis of magnetization. (How Ferrite Magnets are made) They can be magnetized either positive or negative along that axis. The axis of magnetization for this device should be along the axis of the disk. Then the eight different regions can be magnetized with opposite polarities. I wonder if the magnetization deteriorates near the boundaries between these regions?

An interesting alternate to the ferrite disk would be to drill holes in a non-magnetic, non conducting disk and insert a number of small neodymium rod magnets in order to get a similar pattern of field and the same strength. The field would be extremely difficult to alter over time.

[Joe Gwinn]

I can generally tolerate it when people carefully explain to me why I am wrong about something. However, I get excitable when I think someone is suggesting I am not being truthful about my observations. I got that impression from some of your postings, and perhaps I "lashed out". Sorry about that. I'm not normally a hot head.

It's true that I didn't believe some of the claims about the mechanism, but I would submit that it's an overreaction to conclude that because one is questioned that the questioner believes that a lie was told. People are sincere but wrong all the time, especially about technical things, and nobody is exempt.

I do usually begin my conversations respectfully. And that is how I began with Tanac. I never referred to the patent when I contacted them, because I wasn't even aware of it until you pointed it out. But after the $300 / unit calibrations failed to correct the problem, Tanac USA stonewalled. They wouldn't put me in touch with their engineers, or offer any more assistance. (Tanac Japan never made any replies.) At that point I got a little confrontational with the home office. I know the Japanese are a smart, industrious people who are generally very polite. However, I purchased a LEMON Toyota a few years back, and found that some are completely unwilling to admit when they make a mistake.

Tanac sells a zillion of these tensioners (and winding machines). If they generally treated their customers this way, they would not be selling zillions for long. Car dealers are quite a different matter.

The Japanese don't always speak directly as well, especially about problems. Anyway, my guess is that you became excitable a tad too early, and Tanac simply shut down. There is a big culture difference here.

I have replaced the plastic faceplate / window on a couple of units. (It is expensive.) It has generally helped with the gradual climb in tension. However the threads on this latest machine weren't very loose. I thought perhaps the machine had been left on the highest setting for too long, and the disk had become heavily magnetized. But the iron filings didn't reveal a big problem. Just a few small, randomly placed spots where the filings clung to the disk. I eliminated that with the demagnetizer, but the tension is still offset about 8 grams from the readings I made three months ago. The operator had been complaining about breaking wires.

So, somehow, the setting creeps, to higher tension levels. From what to what? Roughly how much wire has passed through the tensioner during this period? At what winding speed in rpm?

We're in Central NY, USA. Small company. We manufacture custom inverters and drivers for LCD backlighting. Everything is done by hand. No automation. The workers who wind our transformers for those inverters use wire gauges from 38 to 44.

Ahh. Lots and lots of ferrite-core transformers running somewhere between 20 KHz and a few MHz. I assume that the bobbins are squareish. What's the most extreme bobbin length to width ratio? This ratio sets the level of pulsation in wire payout speed.

I'm not sure the boss would take $50 for a broken unit, but I will ask. You're in Boston?

Tell him that the hope is that I'll figure the problem out. Broken units don't generally command large prices, and I want your worst unit.

I am in the Boston suburbs.

[Mike Sulzer]

You can get a "Chinese copy" new for $100.00. Aliexpress.com : Buy Magnetic tensioner for TANAC CNC coil winding machine from Reliable coil winding tensioner suppliers on KIM CHEN INDUSTRY INTL LIMITED

[David King]

What these need is a feedback loop to keep tension steady over time and then an impulse demagnetizer to zap the disk every time the unit gets turned off.
Perhaps the Chinese copies have fixed the frailties in the originals (that would probably be a first).

[Joe Gwinn]

Lumpiness cause and cure

As for lumpiness, this being caused by formation of permanent poles on the magnetizable disk, I recalled that this had been discussed in one of the various patents. I finally got around to do the research.

It turns out to have been discussed in Tanaka's original patent, 4,526,329, specifically starting in column 4 line 41 and ending in column 5 line 47. In column 5 line 20, the resulting "ripple" is mentioned - this is the same thing as lumpiness.

Also discussed is how to prevent lumpiness by procedural means. The key observation is that while going from high tension (small d) to low tension (large d) can cause lumps, the converse (go from low tension to high) does not. This is implicitly utilized later in the patent, specifically column 7 line 63, where the description of how to use the tensioner starts. In this section, mention is made (column 8 line 17 or so) of hand demagnetization, which isn't really explained. I think what is meant is that one rotates the capstan with one's finger while simultaneously performing the demagnetization maneuver, slowly varying the tension setting with the other hand.

[Joe Gwinn]

Tanaka MT-300 Magnet Tension Unit - photos and some analysis



In the photos, the background grid is inches by inches. The pictures are about 2 MBytes each.

The used Tanaka MT-300 Magnet Tension Unit arrived a few weeks ago, and I've been studying it. This unit is well-used and dirty, but almost everything seems to be there. The snell (pink ceramic curl) is broken, and the anti-slip O-ring in the capstan has become hard and is fragmenting.

The case is made of high impact polystyrene, with no wall thinner than 3 mm, which is adequate.

The hysteresis disc is made of some kind of spring steel, tempered to a brown color, which will make it harder and more magnetic than blue temper, but more brittle.

The ferrite magnet is 8mm thick by 45mm in diameter, and has 8 poles in the face towards the hysteresis disk.

There is no mechanical feedback from the dancer arm to the hysteresis brake mechanism, so the mechanism of patent 4,526,329 is not fully implemented. Patent 4,570,874 is the closer description. This newer mechanism is far simpler than the original design. One assumes that the simplification was found adequate.

If there is any feedback, it's via the big black box at the bottom, which houses a pneumatic cylinder that can dynamically alter the back tension.

The MT-100 photos provided by Jimbosticks in posting #48 of the present thread show that his MT-100 has the same construction, presumably with lighter springs and a lighter dancer.

In short, the magnetic tensioner part is operated open-loop, and the back-tension adjustment controls only the spring force on the dancer arm.