Global Free Energy Blog

I was going to use both of the plastic discs from version 3, but there's a 1 cm split near where the bearing was. What happened to that poor thing? My best guess is that when it hit 13500 rpm the glue gave away on one of the plastic discs, which cause the plastic disc to flutter in the wind, which caused it to crack. It's all strange because I did not see anything strange with the disc, except the odd noise.

I was going to use the other disc, but the bearing in version 4 is larger than version 3, so the hole was opened up, but I made it to large. What a shame because these discs are a pain to make since they're alway lopsided, so I have to painstakingly sand, view under high power magnifying glass, mark, sand, and repeat over and over till it's round. Two more round discs with the correct hole size were successfully made today.

Next, one of the discs was marked into 8 sections to place the magnets-- contact me for detailed instructions. That's also a long and tedious task. That's complete. Last, one of the magnets was glued to the disk with wood glue. You're probably wondering about the wood glue. Well, it appears to be the best glue that I have for the task. The epoxy did not work at all. Perhaps my fault. Maybe the epoxy mixture was not correct. This wood glue held up amazingly well for the "Tiny Orbo Replication" version 1 and 2. In version 1 & 2 the magnets were glued to the outer edge. So only one side of the magnet was glued on. Version 1 & 2 magnets must have been at least 5 times heavier than the version 3. Also, there's a lot of any air turbulence in version 1 and 2.  So it's amazing the wood glue held on to the magnets at high rpm.

All 24 of the magnets, sets of three, are glued together. So now there are 8 sets. That was supposed to be quick an easy, but turned out to be difficult. Regardless what I tied, the magnets kept shorting together. Liquid Paper and glue were tried, but I did not use that much glue. I was just about to get more elaborate so the magnets did not snap together, which is the probably cause of the short, until it just occurred to me that nearly all of the eddy currents will not be in the width direction. Rather, nearly all of the eddy currents will be around the circumference of the magnet. The Ni, Zn, Cu or whatever coating they used was sanded off on the outer part, so that should eliminate a lot of the eddy currents. Also these magnets are tiny. No concern, so I did not bother trying to insulate each magnet from each other, and just glued them together.

Tomorrow I should have the full day to work on version 4.

The metal coating on all 24 of the N50 disc magnets was sanded off this evening. Unlike the N52 square magnets, these N50 disc magnets have a zinc layer over a copper layer. Hmm, I wonder why they would place a copper coating. That goes to show that NdFeB magnets can have completely different performance. Copper has extremely low resistivity, and the eddy currents would have been high, thus preventing a significant amount of changing fields from external sources such as the toroid.

I was looking at the "Tiny Orbo Replication 3" and noticed a slight space between one of the plastic discs. I could not believe that over half of the N52 magnets were not even glued to one of the plastic discs, as it broke free from the glue. Most of the other magnets easily broke free by slightly pulling on the disc. A good tug broke the rest free. The other disc was also easy to break free except for a few magnets, but a few seconds with a razor blade freed all of the magnets. I can't believe it, this glue must be so cheap or I did a poor job at mixing both of the epoxy ingredients. Wood glue did a better job, as the "Tiny Orbo Replication" 1 & 2 held together fine, and the magnets are glued to only the outer edge of the disc. Maybe that Gorilla glue would work well.

Maybe that's why the "Tiny Orbo Replication 3" quickly went down hill toward the end of it's life since it was difficult to get it over 6000 rpm. Also, maybe that was the strange noise I was hearing toward the end of it's life, the plastic disc vibrating in the wind. It's amazing these magnets did not shoot off the disc. Actually the magnets were only moving less than 50 mph, so it's not that dangerous as long as you're wearing glasses or goggles.

The homemade razor blade bearing might not be such a good idea. It will probably work great for awhile, but it will probably quickly wear out since all of the wear is on a spot section of the blade. It would be a shame to waste the N52 magnets on that. Therefore the change in plan is to use my original ball bearing used in the "Tiny Orbo Replication" version 1 & 2, which is a steel ball bearing that just happens to be the right width for the new design. This old bearing is 4 mm wide. The N50 disc magnets are 1 mm thick x 3 = 3 mm, thus leaving 1 mm space for both plastic discs.

There is a design flaw in "Tiny Orbo Replication 3."  As far as I know, version 3 uses the same type of bearings used in computer hard disks, so I assumed that it was fine to apply pressure on the same axis that a magnetic head applies on a hard disk, axial load. Maybe they can take a little bit of pressure.

The following image shows axial and radial loads:

So "Tiny Orbo Replication 3" was placing an axial load on the poor bearing. According to WikiPedia, I need a Thrust bearing for version 3.

That's unfortunate since placing the toroids on the side of the disc allows there to be twice as many toroids. Also the toroids can be placed closer to the magnets. For now I'll test my homemade razor blade bearing for "Tiny Orbo Replication 4," unless someone can save me the time by telling me this will not work at high rpm.

The present plan for the "Tiny Orbo Replication 4" is to use the homemade razor blade bearing described in a previous post. The rod will be a 19 mil (5 mm) diameter drill bit. This bearing has razor blades on both ends of the drill bit. Each side there are 3 tiny razors to prevent the drill bit from moving. Two extremely thin plastic discs will be in the middle of the drill bit. Sandwiched between the two plastic discs on the outer rim will be 8 N50 NdFeB disc magnets, part number D1021,

D1021 are very tiny disc magnets, 4 mm diameter, 1 mm thick. That's almost the exact width of the MP1303P4AS amorphous core (not the plastic casing). Three of these guys will be stacked to form a magnet that's ~ 4 mm OD by 3 mm thick, for a total of 3 * 8 = 24 magnets. The zinc plating will be sanded off from all 24 magnets, which will be a tedious task since these guys are so tiny. The groups will be glued together. Do you think it's okay to glue them together without coating them? Hopefully the glue alone will prevent the magnets from touching each other, but there's a lot of force between these N50 magnets. I have room for one mistake since I have 25 of these.

The "Tiny Orbo Replication 4" will use the same MP1303P4AS toroids where the plastic casing is removed, thinly coated with epoxy, and with a *tightly* wound coil. The major difference is that the toroids will be placed outside the outer rim of the spinning disc so that it applies radial forces / vibrations.

I'm not sure what to do with the "Tiny Orbo Replication 3" now. It was made to be a self-runner, and it seems obvious to me that it would when all 16 toroids are on so as to balance it. ... That is, if the bearing did not have so much leeway / swing, but it does, and that pretty much kills it. How could it be balanced with such leeway?

To remove that epoxied bearing and replace it equates to a fast extermination. It would destroy the plastic discs. The bearing is a lesson learned. Although I *might* consider taming down the "Tiny Orbo Replication 3" by increasing the toroid - magnet separation distance by a significant amount to see if the bearings are completely shot, but just seems like a waste of time. The mistake was that I should have asked the bearing company how much play was in that bearing.

It seems there will be a  "Tiny Orbo Replication 4."  It should take a fraction of the time to build since most of the time building version 3 was wasted trial & error. So many errors. Also I now have access to nice drill press ... that does not wobble, LOL.  Version 4 will have -->

• Thicker magnets
• Dual bearings to prevent wobble. The rod will spin, instead of a single outer bearing.
• Slightly larger diameter disc.
• Thicker magnets. Each magnet should be just as thin, but stacked. The zinc coating will be sanded off. Each magnet will be glued together making sure each magnet does not touch so as to reduce eddy currents.

Something that keeps popping in my mind is an old experiment with an ridiculously thin drill bit (19 mils thick, 0.5 mm) spinning on two homemade bearings made of razor blades. It was amazing how low the friction was. The question is, will it handle high rpm. It probably will because the drill bit was so thin that it would hardly move, even at 20000 rpm. At 20000 rpm the outer edge would spin at 1 mph.  The razors were taken from a disposable bic shaver. The razors were removed from the shaver, and then they were cut into tiny pieces since one razor is to large. Actually, since the table split in two last night these tiny razors are on the carpet somewhere.

Is this the end of the "Tiny Orbo Replication 3?"  According to emails and blog comments, I think most people agree the tiny ceramic bearings are not happy. After all, it is true that is a lot of axial force on those little guys. Someone emailed me saying those bearings were not meant to have force in that axis, which is true. Yet then again, the design calls for 16 toroids, 8 on top and 8 underneath, which when balanced would eliminate such axial forces. I had no idea a few runs would destroy my ~ $12 tiny ceramic bearings!

We'll have to see if the bearings are shot, but I'm already thinking about the "Tiny Orbo Replication 4," which might have a stack of about 3 of the thin N52 magnets. It will still be 8 magnets, but each magnet will consist of 3 stack magnets where each magnet is electrically insulated to greatly reduce eddy currents. That's a bit premature. It would be nice to verify the power input with detailed scope analysis. My gut intuitive feeling is bothering me about the power input in that it seems the inductive losses (because I'm not capturing the energy back, yet) should at least be comparable to the resistance losses. Although, detailed scope analysis on the "Tiny Orbo Replication 2" clearly showed that nearly 100% of the inductance energy can be captured back.

The "Tiny Orbo Replication 2" captured back nearly all of the inductance energy by dumping it into a LED, as seen in an older YouTube video. This energy was verified on the scope.

I just posted a video of the monster,  "Tiny Orbo Replication 3" at YouTube,

http://www.youtube.com/watch?v=1aFxHTa9J6M

Tonight started great and ended in disaster. My homemade table that had a zillion things on it, including the "Tiny Orbo Replication 3" split in two. Everything is fine, all the meters, except the "Tiny Orbo Replication 3" wooden base broke. No big deal, as I was going to make a better one tomorrow, but it makes me wonder how much pressure was placed on the disc in order to break the wooden base.

Anyhow, I took some video footage. Not sure it's worth putting on youtube. The vast change in rpm was documented on video. It could be 11000 rpm one moment and then quickly decrease. I seems to be the bearings. Maybe it's failing. Toward the end, prior to the table snapping, it was so bad that I couldn't even get it to go over 6000 rpm. It kept rapidly speeding up, then rapidly slowing down, repeatedly over and over.

I analyzed some of the "Tiny Orbo Replication 3" on the oscilloscope. It's confirmed that my power figures in previous posts are correct in that the inductance losses (since the inductance energy is not being captured back) are nothing in comparison to the wire resistance losses. IOW, nearly all of the losses is due to electrical wire resistance. These tightly wound cores are very unusual. It has *less than* 0.1 uH, yet ~ 1 amp (not that much for an "Orbo Replication") is enough to turn it into a monster. It just blows me away that regardless how close the magnets were placed near the toroids that it could easily overcome the cogging effect. In fact, the "Tiny Orbo Replication 3" loved the magnets as close as possible. If I built another "Tiny Orbo Replication," the magnets would be ~ 3 times as thick. Actually it I would layer three of these magnets with an electrical insulator between each magnet to nearly eliminate eddy currents. That's another great thing about this motor is that there's nearly no eddy currents.

I think the pressurized (tightly wound) cores could be a major discovery for the "Orbo Replications."  Nearly 100% of the losses is due to electrical wire resistance. The inductance is almost nonexistent. So far I have not used the "Tiny Orbo Replication 3" circuit. I quickly put together the original "Tiny Orbo Replication 1" circuit, which is basically the hall effect switch connected to a MOSFET. So it's a simple RL circuit. During the coil pulse, it takes ~ 12 micro seconds for the current to rise! That's how low the inductance is, very low. From there it's all resistive. The voltage is 0.27 volts, and the current was close to 1 amp. I did not measure the duty cycle, but looked close to 20 degrees on the scope. That comes to 0.27 V * 1 A * 20 degrees / 360 degrees * 8 magnets = 0.12 watts, which nearly 100% is joule heating (wasted heat)!

... Baffling!  Absolutely amazing!!  Sean, you did it!!!

Wow, I just temporarily taped the rewound toroid down, moved the magnets as close as possible to the toroid, and started the motor. With the hall effect chip bouncing all over the place as it's also not mounted, the "Tiny Orbo Replication 3" with just one toroid *rapidly* accelerated to over 13000 rpm. This thing was a beast! Since it accelerated so fast, it was difficult to say what the maximum rpm was before an usual event occurred. As I was trying to hold down the entire setup to keep it from vibrating off like a rocket, I could feel the power, and hear the noise.

Although I did not analyzed this on scope, yet, most of the losses are probably from joule heating, but I could be wrong. The joule heating losses were at most 75 mW. If I was a gambling man, I'd guesstimate the total output in heat was over a half of watt, but who knows until it's analyzed on the oscilloscope. It's very possible most of the losses went into inductance, but as discovered from the "Tiny Orbo Replication 2" nearly all of such inductance energy can be captured back.

Anyhow, something unusual occurred when it hit ~ 13500 rpm. It was like it hit a brick wall, and began to rapidly decelerate. Maybe the disc began touching the epoxied toroid. I have no idea, but it did not sound like that. I immediately touched the MOSFET, then the 2N2222, then the hall effect chip. They were cool. Very strange. This kind of stuff bothers me, just like the my 1st "Semi Orbo replication," which showed a rapid drop in temperature far below ambient temperature, but then leveled of back to normal temperature, all caught on video. Still to this day that has never been replicated to that degree. Only the occasional mild temperature drop below ambient has been replicated.

A youtube video will be posted as soon as everything is mounted and running smoothly. And eventually I'll have to bore the tears out of everyone by posting a youtube video of the long and outdrawn efficiency measurements.

One of the rewound toroids is complete. For this coil more time was spent on winding. This time I pulled harder on all of the turns, all 3 layers. This gave a bit of an unexpected surprise. The first time, the entire coil was ~ 2 uH, but the rewound coil measures at 0.0 uH, resolution of 0.1 uH! This is crazy! Hmm, not sure how many turns there are, but apparently an air coil this size would have less than 0.1 uH. Anyhow, it appears the pressure from the turns significantly increases the cores longitudinal dipole alignment, which gives it a squarer BH-loop. It is well known that pressure in the correct axis can give cores a squarer BH-loop.

Some information that I forgot to blog about is the weight of thin N52 NdFeB magnets. Each magnet weighed in at 0.16 grams. All 8 weighed 1.31 grams, which comes to 0.164 grams per magnet.

Just now the entire disc, which contains two 9 mil plastic discs, 8 N52 magnets, metal rod, tiny ceramic bearing, and some glue, weighs 2.29 grams.

Not that it matters much, but the rewound MP1303P4AS toroid weighs 3.31 grams. An empty toroid with only the epoxy weighs 1.24 grams. A stock untouched MP1303P4AS toroid with the factory plastic casing weighs 1.65 grams.