Replicating the Perendev motor

Clickbank Promo Tools
A movie clip made by Robert Calloway at

Or download this Movie.(right click: save target as...)

2006-02-23 A flat Perendev device experiment

Clickbank Promo Tools
1 Oct.2004
Click for
click to enlarge click to enlarge click to enlarge
15 Oct.2004
Click for
click to enlarge click to enlarge click to enlarge
click to enlarge click to enlarge click to enlarge click to enlarge click to enlarge
15 May 2007
Click for
click to enlarge click to enlarge click to enlarge
click to enlarge click to enlarge click to enlarge

Here is an explanation I received by email, about the use of pot magnets in the Perendev Device:
The Perendev has movement, because the magnets on the rotor are sealed with steel. This technique makes the magnet stronger where it isn't sealed: top and bottom, but makes the magnetic fields (flux) closer. They isolate the magnetic field around the magnets.
The steel will not get attracted to the stators because the magnets on the stators are probably also sealed but the magnetic force between top of magnets (not sealed) on the rotor and those of the stators is the strongest and that is why the simulation on this site doesn't work.
This way of isolating magnets is very known: look at pot magnets! The steel is a better guider of magnetism than the air and that is why the magnetic flux chooses to travel through the steel and you are cutting the field short!
There is contact between the steel (=external shell) and the magnets. Steel gets magnetized (attracted), it is a way of isolating that is frequently used with pot magnets. Every magnet dealer sells them!
If the Perendev motor works, then it has to be with that kind of isolation;
otherwise the repelling power is equal to the attracting power on the sides of the magnets!

Trying to understand the Perendev motor

Calculation number of magnets: by Tero Ranta
On top of the middle white disk there is a metallic flange that has 6 holes 60 degrees apart (there is a threaded rod in one of the holes). Now you can draw lines that go through the 6 holes (black lines). Then project the black lines downwards the same heighth (yellow) as the metallic flange and in the same direction as the shaft (red line). The projected lines are blue and the angle between them is 60 degrees. By comparing the blue lines with the outer perimeter of the disk (green) you can count that there are 5 magnet attachment nylon screws per 60° segment. Thus you have 30 magnets on the disks.
Click here to see a video of the device at
by Jason Owens.
I have been studying the Perendev motor lately to see just how they have the offsets set up on the main wheel. I was also looking at some diagrams Dan made of the offsets and I finally figured out exactly how they did it. I took a clip from the flash video and the video of the working model and I could see just how the offsets were made by doing a cheap line test.

Danís idea came pretty close but from my observations, the distance between the offsets is one magnet width as opposed to half the width like in the diagram. What I did next was make a simple animation that I could study to see just how they accomplished the off balanced effect in the motor:
I attached the animation to this e-mail so you can look at it but I can see clearly here how the balancing effect is happening. From this idea, I figured out a way to modify my original tri-phase designs so that these three-phase offsets can be applied. I simply segmented the stator magnets up into three sections (to represent the offsets of the three wheels on the Perendev) and I got this final layout:
  Click here if you want to see a replication of this simulation, together with a torque calculation and results.
This here is also an animation that I attached to the e-mail for you to look at. It basically has the same three offsets as the Perendev motor and beautifully accomplishes the unbalanced effect. As far as I can see, the magnets on two of the stator sets are always helping the third one enter a line up, then the phase shifts and two different stator sets help a third one line up again; and itís all seamless too! I know that this diagram looks like an earlier model that I made but this one appears to actually work just from looking at the simulation animation.

It appears that the motor will cog when the lines of flux are at a maximum density. Besides tracing out the magnetic lines have you done any calculations on the forces?
You can do this by differentiating the energy in the magnetic cicuit.


Hi Georg,
Unfortunately, I really donít have any means to test the amount of force that the magnets have on each other. I have Maxwell 3D magnet analysis software and I heard that can be used for force analysis but I really donít have any background in the theory and the math necessary to use it. The only calculations (if you want to call them that) that Iíve done was to organize the geometry of the rotors so that they are offset, but I really didnít do anything more high tech than eyeing the picture to see if it looked ok. How do you calculate the amount of force acting on each magnet exactly? Knowing this would take much of the guesswork out of my designs. I really have no idea if it will work until I can bench test it. I did manage to build a rough prototype (which I will be posting picture of soon) but it didnít work mainly because I didnít cut the pieces out exact enough for them to line up like they needed to, but I also noticed the cogging problem you mentioned. Iím still confident that the tri-phase concept works; I just need to design a configuration that will work with it. What do you think about the whole design concept in general? The main idea is to always have two magnets helping to push one magnet past the sticky spot. Iím beginning to wonder if I even need to angle the magnets on the wheel for this to work. But the angles, spacing, size of the wheel, and number of magnets are the variables I have to work with. However, it would be invaluable if I could find a way to simulate the forces that all the magnets exert on each other. If I can do that on the computer, then I wouldnít need to do so many bench tests. What do you think

Butch LaFonte sent the following post about the Golden Ratio to this page.
Hi Butch,
I played around a bit with the cad system last night, and here is what I came up with:
For all practical purposes you probably donít need accuracy greater than 1/1000 as it gets difficult to measure and even things like surface finish and slight eccentricity are starting to become significant factors, the transition of the ratio through Phi happens with an offset just over .508 inches.
Here is the work up:
To the nearest 1/1000 ..
Using magnets 1.000 inches in diameter with a center offset distance of .508 inches
The Crescent moon shape is 0.4852 square inches
the Cats eye shape is 0.3002 square inches
For a ratio of 1.61626
Going for another decimal place we see the transition happening like this:

Extending the offset distance to .5081 gives
Crescent moon shape is 0.4853 square inches
Cats eye shape is 0.3001 square inches
For a ratio of 1.61713

Phi = 1.6180339887499...

Extending the offset distance to .5082 gives us
Crescent moon shape is 0.4854 square inches
Cats eye shape is 0.2999 square inches
For a ratio of 1.61887

I am sure you wouldnít remember me from years ago, but I was one of the people who joined your group about 4 years ago. I told you at that time I anticipated getting some good software and I would do some 3D work for you when I got it. Things didnít work out quite as I planned and I didnít get the software till a few months ago.
But I now have a copy of SolidWorks 2004 and it is all I had hoped for and more!
If you are interested in sharing your insights I would love to know where you are going with the Phi relationship. As I think about it, this would be very close to the region where you would get maximum attraction or repulsion in a plane 90 degrees to the magnet faces, and you may be onto something quite important here.

Clickbank Guide & Tools

Area calculation for GOLDEN RATIO research

Diameter circle: