The Great Ball Contraption Wiki

Forum - Akiyuki train

From The Great Ball Contraption Wiki
Overview > Great Ball Contraption > Customs > Akiyuki train Write a reply

A bit out of date, but here's what I have posted about my version of Akiyuki's GBC train:

Posted by ALittleSlow (administrator) on 17 February 2015 at 12:16.

do you know of any instructions for the mechanical parts that load and unload this train?

Posted by on 7 October 2015 at 06:25.

Yes. Josh Coalson built the Akiyuky Mechanical Train for the BayLUG area at Maker Faire 2014 and we ran it with my GBC modules for two full days. The train car had some problems, and we recently found a BayLUG member who started building GBC modules after Bricks By the Bay 2015 last August, and that member had an idea about how to fix the problem we saw at Maker Faire.

I contacted Josh and he sent me LDD files for the train car, loading station, and unloading station. I've sort of got them working and will take them to the BayLUG meeting this Saturday, October 17 and hopefully work out the bugs with Josh and the new GBC builders.

I don't think Josh would have a problem with me sharing his LDD files, but the polite thing to do is to check with Josh first to be sure. I'll post again early next week.

I just found out today about this site from an email from Tom Atkinson.

Posted by Walt White on 15 October 2015 at 23:33.

I finally ran my loader and unloader satifactorally this past weekend at Greenfield Brick Expo. It was the first time I didn't feel like I had to watch it constantly. I think two factors made the biggest difference. 1) I used disposable AA lithium batteries. 2) I replaced the springs in the reversing mechanism with rubber bands.

The lithium batteries gave me consistent power and momentum, eliminating two of the factors involved in tuning the performance. The rubber bands reduced the torque required to run the reversing mechanism. This put less stress on just about everything.

Posted by ALittleSlow (administrator) on 16 October 2015 at 00:18.

The "upload file" utility (part of the Brickmedia site) only allows certain file types, so I can't upload the .LXF files. They are tiny files and could easily be attached to an email if you want to send me an email address. Or to some other site. I tried the BayLUG discussion site, and it refused the .LXF file as well. I'll try some other sites, but wanted to let you and anyone else contact me directly if you want them soon. Walt

Posted by Walt White on 18 October 2015 at 19:21.

Digging into ancient history, I just uploaded five .LXF files to Brickshelf!

When I log out of Brickshelf and enter the URL the site says the folder is not yet public. When (or if) it does go public, I'll post the description of them. The URL is: if you want to check yourself


Posted by Walt White on 18 October 2015 at 19:33.

Brickshelf is still moderating the LDD files...........

Questions/comments for ALittleSlow: I like your idea of replacing the shock absorber spring with a rubber band setup. I had two problems with the loading station that need fixing before I would take it to a show:

The first problem was that I never got the drive ring mechanism working. It could be just my fumble-fingers, but I wonder if it's really needed. All it seems to do is stop the ball loading process half a second before the train car is reversed so no balls would fall on the track. I ran it briefly at the BayLUG meeting without the drive ring and just ran the loading process the whole time the train was in the station, and didn't have any balls fall on the track. Have you found it necessary?

The second problem was that the train car behaved fine in the loading station while I built and tested it, and then when I tested it again a week later it slammed into the hard stops at the end of track and dislodged the ball tray balanced on top of the train car. I moved the ramp that puts the car in neutral back two studs, and it still slammed. I tried adding pieces of pneumatic tubing near the hard stops, but that didn't seem to help. I'll keep moving the ramp back, but wondered if you have seen the same problem that the system behaves differently on different days.


Posted by Walt White on 19 October 2015 at 20:13.

I never used that particular loading station. At first I used a modified Philo shovel, but recently modified Rafe Donahue's conveyor by adding a trap door and an overflow. It does sound like you don't need the driving ring. I would stress test it by tossing 40 balls into the hopper at a time. If you don't get any leaks from that, you should be fine.

Yes, consistency has been a two-year battle. Some of the issues I've had:

  • Different speed based on battery charge. Fixed this with the lithium batteries, as mentioned earlier.
  • Drive wheels shifting position on the axles. I've heard that Akiyuky used modified LEGO elements to shim his wheels to get them to stay in place on the axles. I found a purist solution to that last year. I just finished a new MLCAD model and instruction with my current (and final!) train design. Will post soon.
  • Dirt on the wheels. Since there are only two drive wheels, and not four like a normal LEGO train, wheel grip is a significant problem. I clean them periodically with a pencil eraser.
  • bad wheels. I discovered that not all the train wheels are actually round. Or maybe the axle holes are off center. In any case I've built a train car that visibly bounced its way along the track, when it moved at all. The red ones tended to be worse than the black ones. Check the roundness of your wheels and put the most true ones on your drive axles.
  • differing approach slope. Depending on my setup, sometimes the train approached downhill, sometimes level, sometimes uphill, sometimes around a corner. Obviously that affected the momentum coming into the "station". That's not been a problem with clean wheels and lithium batteries as long as the final piece of track is level.
  • shift axle drift. I think the bushings on the 12L axle shift a bit when they're hit with the spring. I filled mine with something like 9 half-bushings, and that seemed to fix it. Combined with replacing the spring with a rubber band, this is not a problem at all now.

I spend a year refining my ball tray and tray support as well, in part because of this problem. The tray has a "skirt" all the way around the bottom, and the outer corners are supported. You'll see it in the instructions. It probably could still be improved, but not by me.

I would be careful about moving the "neutralizing" ramp back. I have mine as far foward as possible, even using a half-stud offset for this purpose. Eventually you'll drift the other direction and end up stopped in neutral before it engages the gear at the station.

I've tried a lot of things to soften the blow at the end. So far I've had most success with a plain solid stop. I tried rubber stops, but the train would bounce back inconsistently. If the tray and engine contact your hard stop simultaneously maybe that will keep the two together? I think mine work that way, although the stops are not monolithic.

Hope that helps.

Posted by ALittleSlow (administrator) on 20 October 2015 at 03:01.

My train car instructions are now available at

Posted by ALittleSlow (administrator) on 20 October 2015 at 03:16.

Posted by Walt White on 18 October 2015 at 19:33.

»Digging into ancient history, I just uploaded five .LXF files to Brickshelf!

When I log out of Brickshelf and enter the URL the site says the folder is not yet public. When (or if) it does go public, I'll post the description of them. The URL is: if you want to check yourself


If you get impatient waiting for the folder to go public, you could always post deep links - i.e. something ending in .lxf.


Posted by Captainowie on 20 October 2015 at 08:32.

Have no experience with this module, but I still love your detailed breakdown of issues. Working on a different kind of train module, some of these points are highly relevant. (Really can't skimp on a pair of drive wheels?)

Posted by Torso on 20 October 2015 at 10:05.
Pages: [01] 02 03

To protect the wiki against automated spamming, we kindly ask you to solve the simple sum below and enter the answer in the box (more info):