doug72

Probably got one of the two drive differentials the wrong way round !

Thanks for the comments, there were several times when I almost gave up but it finally came together on Xmas Eve. The long drive axle is 28L long and still had a slight tendency to twist - but by using a friction pin on the bucket wheel outboard axle bearing it keeps all the gear teeth in contact. I was determined to get it working without the use of ball control load/unload gates. Final step is to place a white board so only the ball circuit can be seen with a dummy drive cord as per the YouTube video. There might be a Mark 3 version next year as I received a Mindstorms 31313 set as an Xmas gift.

Thanks for your comment, the biggest challange was getting the ball timings right. Designed as a stand alone GBC module, don't think it would meet the speed criteria for intergrating into a GBC layout but some elements might scale up. i.e the steeper unit and the 2 arm ball lift maybe using BWE buckets !!

Mark 2: Now running OK with four balls in the circuit. After initial Mark 2 re-build still had a few problems which now finally sorted out. Main issue was getting balls to follow the circuit without falling off !! Bucket wheel has a friction pin at the outer end to ensure gear train doesn’t have any backlash. Top ball run rebuilt and a cage made to ensure balls drop onto the bucket wheel which now has two rings of curved gear quadrants. Next lower ball run rebuilt to also ensure the balls load into gondoliers correctly. Speed now increased and running smoothly. See new video of this in action.

Try removing the s from https. if that doesn't work: I find the best way is to upload to UT and then open your YouTube video and then copy the url from the address bar and paste that into your post.

I use a Canon PowerShot AS2300 HD compact camera for photos and for videos at 640 x 480 or 1280 x 720HD usually hand held.

Had thoughts about a chain drive but I don’t have any except the old style large Lego links which use the old toothed gears and I would have to completely rebuild steeper unit to be able use a chain drive. Think a chain drive would have same issues unless kept tight using a jockey wheel. The new drive is working OK.

Mark 2: Progress and new problems solved. The new steeper unit is working very smoothly and transferring balls onto the top ball run OK. The Bucket wheel however has proved awkward to align consistently so that balls drop into each bucket. This time the balls are loaded from the side. The long drive shaft (22L) from the 4:1 reduction output on the steeper unit had two universal couplings plus two knob gears for right angle drive to the bucket wheel which resulted in a lot of torsion and gear backlash causing bucket wheel to intermittently get out of sync. Replaced the drive shaft with an 22L axle built up using 11 x 2L axle connectors and 9 x 2L axles which is much stiffer and more resistant to torsion.
 Both universal couplings eliminated and in order to keep drive parallel two 24T gears used and the right anlgle drive knob gears replaced by two 12T double bevel gears, result a big improvement in bucket positioning but gear backlash still a slight problem. Solution: 
 The axle at the outboard side of the bucket wheel was supported in its bearing using a tan frictionless axle/pin.
 This was replaced by a blue axle/pin with friction and this ensures the drive system is always under constant tension resisting any torsion or gearing backlash. Buckets now all pause for loading balls in same place. Bottom ball run to build next.

Mark 2: Progress.

 Learning from the Mark 1 version which had some problems with synchronising the whole system, the drive input from the L motor is now at the two arm ball lift and hard coupled to the bucket wheel vi a reduction gearing.
 The final bucket wheel drive is via two knob gears without any need for a steeper unit. The L motor dives the 24/16T differential using a worm gear engaging with the 24T gear.
 Speed is adjusted using a PF speed controller. Output from 24/16T differential drives the two arm ball lift directly and the bucket wheel via a 2 stage 4:1 reduction gear box. Four revolutions of the two arm lift rotates bucket wheel one revolution in eight steps. The other output from the differential is held stationary until the striker pins on the rotating arms engage with the knob gear. Then the two 24T gears mounted on friction pins start to slowly rotate and also rotate the knob gear until it release the rotating arms again back up to for full speed until the next cycle. 
This gives a long enough dwell time for the ball transfers to take place at both gondoliers. 
At each revolution of the ams there are two periods at full speed and two at slow speed. Everything now much smoother in operation without any jerkiness. Next stage build the two ball runs and the ball gondoliers.

Found a German seller who has plenty of white visors BUT cost for 12 pce = £3 - Shipping to UK £7.00 !!! Now started work on a Mark 2 Version having identified failings on the Mark 1 build.

Great, look forward to see how you use them. Got my white ones from a Lego seller in France. Doug

Having proved that the “Mock Perpetual Motion “ machine can function, I have found it can be temperamental at times and get out of sync. Driving the main bucket wheel via a steeper unit and another steeper unit for the 2 arm ball lift can be jerky and if too runs too fast it upsets the timings for ball transfers. Now working on an improved Mark 2 version. Propose that the L motor will drive the two arm ball lift directly via a steeper unit using a 24:1 reduction and a PF speed controller. 
The bucket wheel will be hard linked to the 2 arm ball lift via gearing & shafting to rotate bucket wheel using a 4:1 reduction gearing. i.e. as two arm lift rotates for each half rev. it will rotate the bucket wheel 1 step. (1/8 rev ) The steeper unit will slow down the arms as they pass the ball loading and unloading points. It should then be easier to adjust the system to keep in everything in sync. A dummy rope drive and pulleys to be installed to simulate the YouTube video with all the Lego drive mechanisms hidden behind a white panel. See mock up of proposed drive system:- The L motor drive input via reduction gears is at the Red axle connector. Red beam = 2 arm ball lift. Yellow beam = 8 step bucket wheel. Dummy drive shown. Yellow 15L beam shows where the screen would be to hide the drive mechanism.

Like most things today they are not designed to be taken apart as componant parts are snapped together. Usually if dis-assembled the tabs get damaged and fail to engage when re-building. Its a fact of life nowadays. Only thing I suggest is to try a drop of thin oll on the axle shaft where it enters the motor housing and let cappillary action draw the oil along the axle shaft into the motor housing.

Thanks for the comments, it has at times been a very frustrating but challanging build and almost gave up. Its as close as possible to replicate the so call PM machine. In my version it replicates the action in that the ball rolls along the top ball run, drops into the bucket which then advances one step (1/8th rev) Getting the motor speed correct is critical to get ball tranfers to work OK in sync. Surprised that no one has yet asked what the purpose of the white disc with 8 holes and two small wheels on the two arm axle is for !! Have an idea for a Mark 2 version that will hide all the mechanisms behind a white screen and only use one steeper unit. PS still hunting for errant orange balls in my study !!

You ar so right, spent ages trying to get things to working smoothly as intended. Eight buckets now in use. Solution was to increase the rotation speed of the two arm ball lift and as predicted a lot of tweaking and adjustment to speed was required. Now faster running at 3rd step on a speed PF controller.
 Additional bracing added plus measure to ensure balls do not skip out of the buckets. A 2:1 step up gear (24/12) was required to run the two arm stepper unit but this resulted in arms rotating the wrong way, so 1:1 gearing (16/16) was required at the output of the other steeper unit to correct the rotation. The postioning of the two arm lift is achieved by removing the black 12T double bevel gear, move arm to correct position and replace the gear. Photos showing extra gearing required & new video.

Thanks, your observation is correct, at the moment only 4 buckets are loading. I think I have a solution which hopefully will not alter the ball run timings and also not require a ball control gate to load the two arm ball lift. Every time you improve things its been a case of 2 steps forward, 1 step back.

At last after much trial and tribulations have successfully achieved the gaol of replicating the “Perpetual Motion” device which must have had a power input hidden somewhere. My version all driven by one L motor and two stepper units, without any control gates, its all down to timing and fine tuning of various sections along the ball runs. Spent a lot of time chasing lost orange balls around my study !! PS: now to scale it up and use BWE buckets 6145856 ! Heres the video:-

Lego have built a large version of this, 230500 bricks, weight 731 Kgs, Time to build 1660 hours. See Lego facebook page. Note weight shown in video is wrong due to an extra zero !!

After more thought have continued with this build. Gondoliers for transfering balls built and an unloading sytem devised. Top gondoler is shown in the tipping postion. Next build the ball run back to the bucket wheel. Biggest challange will be loading balls into each gondoller and into each bucket on the bucket wheel at the correct moment. Some form of control gate will be needed, activated by trips as they rotate. Update Now built the return ball run and amazingly the balls roll along and drop into each bucket without any control gate !!!! Needs a little bit more tweeking.

See Efferman custom parts recent posting.

Perhaps the Efferman micro motor & mount would be the smallest.

Trying to get the two arm ball lift has proved impossible to get it to function as desired in sync. with bucket wheel as it rotates in 8 steps. After many attempts to progress this build have hit many dead ends and failure. At the moment I can only conclude it is not going to be successful in replicating the "so called" perpetual motion machine. Along the way some usefull mechanisms & ideas were found for possible future GBC builds. Please note I know perpetual motion is impossible, I was attempting to make a powered version, as the original YouTube video must have had a hidden drive.

Loading & unloading steeper unit for ball transfer now built. The rotating arms when they contact with knob gear slow down to transfer balls, then speeds up to next step. Ball loading and unloading gates to now figure out along with ball runs. Note: the black 36T gear is only to hold the rotating arms and is not connected to any other gear.

Update: L motor is strong enough to run both stepper units just need to determine the right gear ratio for the rotating loading arms. Has added bonus that the arm rotation slows down as it engages with the knob gear and then speeds up when it dis-engages to give time for ball transfer for loading and discharging.

Yes, can't double skip now, the 4 lobe knob gear prevents it. Aslo it gives a soft start when bucket wheel starts to rotate to next step. Have tidied up the structure for knob gear drive. Now building a similar unit for the two arm ball lift which only requires two steps per rotation. Hopefully will be able to drive that from the L motor as well.