Like us on Facebook and Follow us on Twitter


OS:Dreieck Magnet Motor

Lasted edited by Andrew Munsey, updated on June 15, 2016 at 1:09 am.

  • 2 errors has been found on this page. Administrator will correct this soon.
  • This page has been imported from the old peswiki website. This message will be removed once updated.
Image:Dreieck magnet motor magnetism 300.gif

by Congress:Founder:Sterling D. Allan

Pure Energy Systems News

January 21, 2010

I was contacted by Dietmar Hohl from Austria yesterday regarding the design below that they are open sourcing.

No prototypes have been built yet. You are welcome to attempt that if you are inclined. As a next step for them, it looks like they are hoping to get someone to calculate the model in Maxwell 3D. Their 2D computer modeling predicts three times the "push" as any "pull" opposing it.

Feedback provided on Jan. 26 explains the There was an error working with the wiki: Code[1] in this model.


''Date: January 20, 2010 8:04 AM

Subject: New Permanent Magnetmotor-concept!''

Hallo Allan,

Thanks for your nice reply. How are you?

For you and our friends in the States a new OS-concept for an permanent-Magnetmotor with a calculated factor of 1:3 under

A/ German:





B/ English:

1/ PMM Concept Rev. C under

2/ Setup of "Project: Dreieck_A_N40" under

3/ Data/Results of FEEM-simulation "Project: Dreieck_A_N40" under

4/ Results (chart) of FEEM-simulation "Project: Dreieck_A_N40" under

5/ The zipped FEEM simulation file for "Project: Dreieck_A_N40":

5.1/ dreiecke_A_N40.FEM (The FEEM-file)

5.2/ PMM_xy-F_linar.lua (The script for simulation in 0,1mm steps)

5.3/ dreiecke_A_N40_dxf.dxf (The DXF-File for import in a other simulator-program)


Allan we need your help in calculating the model in MAXWELL 3D.

Have a nice day and thanks!

Dietmar Hohl -) Linz/AUSTRIA


Image:Dreieck magnet motor rotor-stator dimensions bf16.gif
Image:Dreieck magnet motor magnet specs bf32.gif

In the News

Image:Dreieck simulation error 95x95.jpg
There was an error working with the wiki: Code[1] > OS:Dreieck Magnet Motor > OS:Dreieck Magnet Motor - "Hohl's and many other's mistake is to take a limited 'snapshot' of a complete cycle. This limited zone or region of force, integrated (averaged) over the incomplete cycle path, then shows an apparent 'gain' or non-zero result!" 3D simulations suggest net zero gain. (PESWiki Jan. 26, 2010)
Image:Dreieck magnet motor 95x95.jpg
Latest / There was an error working with the wiki: Code[1] > OS:Dreieck Magnet Motor - No prototypes have been built yet for this design developed in Austria, which they are open sourcing. Their 2D computer modeling predicts a three-fold gain. They're hoping for help with 3D modeling. (PESWiki Jan. 21, 2010)


Simulation Error

On January 26, 2010 6:31 AM, Stan wrote:

Hohl's and many other's mistake, myself included in the early days is to take a limited "snapshot" of a complete cycle. This limited zone or region of force, integrated (averaged) over the incomplete cycle path, then shows an apparent "gain" or non-zero result!

So the neophyte gets all excited about discovering over unity! When in actual fact it is a false conclusion based on an incorrect integration of force. More correctly, the distance over which the linear motion and force are taken, is not a "complete cycle" and therefore can be non zero due to this error.

Hohl's models, appear to be using FEMM, a free 2D magnetics simulation software. I doubt he set up the model correctly, but that cannot be seen in his screen shots. Despite that distinct possibility (because the Help in FEMM leads most to not setting up models to accurately render open field magnet systems properly), the primary error Hohl makes is to take a smattering of force values of his armature, only while it is between the ends of the stators in his linear model. (linear motion as opposed to a circular motion)

With ANY and EVERY linear motion of two or more arrays of permanent magnets, a complete cycle is only when you go out beyond the ends of the stationary magnets, far enough on both sides, to come to near zero force on the "armature". Then you must integrate the force-displacement values for this entire path length, from zero force on one side through the array, and to zero force on the other side when the path is linear motion!

So I constructed a FEMM model according to Hohl's magnet dimensions and spacing, but I used fewer numbers of magnets in the array, simply to save space and solve time. (because you have to go out past the ends of the stator, 5-10 times further than the stator array is long)

Image:Dreieck simulation errorjpg.jpg

See the following animation of the simulation:

: Click here (1.6 Mb gif)

The simulation takes the armature out -100 mm from center, and moves it 2 mm per step for 101 steps. So it ends up at +100 mm. Force values are taken on the armature at each step. However the animation frames are skipping to every 4th step to keep the animation file size reasonable.

Even this large excursion leading into the stator array, and trailing out of the array is not enough distance to get to zero force. However I moved both the stator array and armature array so that the center of mass of each array is at point 0,0 in the FEMM geometry grid. This means equidistant values out past the ends have roughly the same "weight" regards their forces with respect to a force from this center of mass.

Now see the following chart, plotting the force at each step, and the integral of those force-distance steps: (the average is the integral if each step is equal displacement)

: HOHL-LINEAR-01.pdf (12 kb pdf)

Notice that the integral (average) of the force over this complete motion cycle, is practically zero! It is not exactly zero because there are simulation and math errors, which are always present. However the error margin is probably two to 10 times higher than this actual number, therefore we can consider the result to be essentially zero net or average force through the entire cycle!

The error Hohl and many made, is to take forces on the armature, only within the stator array region, and assume this means when you close the array into a circle, it will make a non zero result. It will not!

The armature is strongly attracted to the outer ends of the stator array, even when it is between those stator ends! You are "seeing" this attraction to the ends effect when you do not perform a complete linear motion cycle out to near zero force on both sides of the stator!

To further illustrate this rudimentary physics/math principle, I also have 3D magnetics simulation software and performed a 3D simulation of a circular array of one stator ring and one rotor ring of his magnet geometry.

See the following:

: HOHL-01.pdf (108 kb pdf)

This pdf, shows the result of one complete cycle in a rotary or circular form, using 3D magnetics simulation. I set the magnets at a cg radius of 51 mm, and then did a circular array of 20 magnets on each of the rotor and stator assemblies. That means one cycle comprises 18 degrees of rotor motion.

I did the 3D simulation with 36 steps of 1/2 degree each.

Again the result is zero average torque for a complete cycle. The torque wave form is not symmetrical, but the area under the curve, is identical above zero and below zero. (average with uniform step size, or integral reveals the area under the curve)

Please be very careful about speculations, simulations and conclusions about them with regards to "complete cycles" of motion and force or torque integrals of that motion cycle!

This is a VERY common mistake, even very learned persons can and have made. It would be very unfortunate for persons to read about some false conclusion and waste money or time trying to build it, when such a fundamental mistake is at the heart of the matter!

Now I must also add, that just because 99.9% of these kinds of structures are in fact a net zero result, this does not mean all examples are unity or no net output! There are ways to coax a small non zero result from certain kinds of magnetic structures!

But the details, and level of precision you need to find those, is far above what most of you are doing with tools such as FEMM!!! Do not get me wrong, it's a fantastic tool. And a very powerful one. But with seemingly minor missteps in setting up the model conditions, you can get huge simulation errors, on top of this "cycle" issue!

I've tested FEMM against the real world on the bench many times in many ways. And found what works and what does not with how you set up the model parameters.

I plan to generate some tutorials and post them to my website on this, but anyone who has a serious inquiry and an existing model they would like to have independently cross checked can go to my site and email me.

You can see a write up of "sim vs reality" here:


QD Mechanic


Dietmar Hohl


email: []