Bessler Wheel Theory by Alex

[fletcher]

Hi Hans - I concur - I wrote this reply on a thread [BW.com] this morning about a 'runner' that didn't appear to be behaving as you might expect- btw Rainer professionally supports the product in industry.

edited for brevity ... Hi .. interesting design concept for sure - Rainer dropped in last night - he is having great difficulty accessing the board at the moment hence I'm posting on his behalf this morning after he rang me - I showed him your sim & got him to take a look at it as it seems a little peculiar to us - we tried a few things to test its integrity & got some interesting results.

If you anchor the pendulum bob to the background the system locks up & keels [this you would expect] - if you lock it down with a rod attached to the background the sim accelerates away as usual [this shouldn't happen] - Rainer went into the accuracy settings & increased the accuracy & it was still a runner, ... so he went deeper - he changed the way it calculates to Euler [a different calculation method from Kutta] & it keeled, regardless of accuracy settings i.e. number of frames per second - he then changed it back to the Kutta-Merson method of calculating [its original setting] & it keels this time.

His thoughts are that the problem lies in the way the gear forces are calculating thru - since the program is iterative then it's a discreet process, frame by frame - he thinks the forces calculating on the gears are out of sinc [so to speak] with the frames calculating [steps per min] - this would result [he thinks] in a fictitious force or torque on the gears thus turning the sim - apparently the Euler calculating method is another way for sims to do their totaling - he is suspicious of your initial 'runner' in that it keels in both methods after it is changed to Euler & then back again

He asked me to say that using WM as a tool is fine but it shouldn't be entirely trusted or relied upon [especially in this field] - he gave the analogy of would he trust to go flying with a brand new helicopter pilot - maybe, & maybe everything would be ok - he would have a lot more trust that things would go well with a pilot of 20 years experience.

We are not trying to rain on your parade [only give our opinions of the sims integrity & reliability] but suggest that you don't always trust the sim results - you are building the REAL DEAL so that should prove your principle beyond a doubt & we both wish you well with that.

Other things to look out for is use of polygons - the program can sometimes get confused & use the inside surface for reaction force calculations - try not to use the polygon tool.

Overlap error to big - the program misses a step in calculating giving erroneous results - reduce the iteration steps & watch for overlap tolerances.

Placement of pins etc not accurate enough - restraints break N.B. if you get a "restraint warning" do not trust the program after that point - means the program has left out some calculations it can not resolve - usually due to builder error.

Other ones that Rainer knows about - if you know what to look out for.

All in all, in the right hands, it is an excellent, accurate & reliable tool to assist analysis or design & planning of a build, but should not be taken as gospel - doesn't replace common sense & a good eye for detail & observation.

[hansvonlieven]

Thanks Fletcher, your comments are helpful.

Hans

[rlortie]

Hans,

Sorry that I misinterpreted the full reasoning for your simulation.

[When it is all said and done, there is no substitute for building.]

All I will say is that I agree and that I do not believe a simulation of  Alex's design would prove fruitful.

Ralph

[fletcher]

Hans .. may I suggest you delete your extra weight ball to get your sim rotating [so its not in the system] - then use the torque tool to add some rotational element [place it anywhere on the mech - give it a positive or negative value which dictates the direction the torque will occur - use say 100 Netwons or less - try making it active for say 2 seconds only [see graphic] - then, if the center of the sim is coloured select the coloured parts & go to window>appearance>patterns & select NO [this will make the parts transparent - then select view>system center of mass [CoM] - then run your sim.

Now you will see the entire systems CoM as it oscillates beneath & to the side of the Center of Rotation - this will show why it accelerates IMO.

[fletcher]

Can't seem to edit the above post - here's a closeup pic of the CoM - it should travel from side to side, swinging beneath the axle as the oscillations dampen down [assumng there is some air resistance [quick & dirty friction added] or pin friction etc.