Blakbird

[GBC] The Akiyuki Project

Recommended Posts

I hadn't noticed you were Vancouver, Rebricker. This makes BrickCan a must for me! Looking forward to some awesome GBC action...

I'm especially excited about the train loader - that may be my favourite module of all. Keep up the amazing work!

Share this post


Link to post
Share on other sites

I have made about the same progress!

I finished the Lift Triggered by a stuck ball by using ldraw system(I tried to modify a little) and I guess that you also finish this one.

I just uploaded the file at my bricksafe folder for your reference.

800x450.jpg

Share this post


Link to post
Share on other sites

Instructions for Akiyuki's Cup to Cup module are now tested and complete and are available here along with a Brickstore parts list.

  • 65 pages
  • 1089 parts
640x360.jpg

Most of the credit for reverse engineering this module goes to djm who did the work, made full instructions, and presented them in this topic. There was discussion at the time that this was a simple module for Akiyuki and while that is true mechanically, having now built it I can say that there is some very complicated SNOT in here which is not at all obvious upon initial examination.

This is one of Akiyuki's earliest modules from when he was still learning his craft. Perhaps partially because of this, there are many different versions. The original version presented in his video is very crude aesthetically and does not have an output. A survey of his videos from various conventions and presentations reveal at least 3 later versions of the module with varying degrees of changes and improvements. Djm's version represents an amalgamation of these roughly at the mid-point of the design cycle. It has the nice looking ramp and input hopper and a small sensible output ramp. Later versions included a much longer (and less practical) output ramp. I built the model directly from djm's instructions which worked great, but there were a few things I wanted to change:

  • The bottom of the input hopper has a significant flat spot where balls can stall and not make it into the spiral lift. It is not unusual for 5 or 6 dead balls to accumulate here. I changed the slopes at the bottom of the hopper to more forcefully drive them into the lift. I can't tell whether or not Akiyuki did the same thing based on the views in the videos, but I think he did something similar.
  • The frame is weak torsionally because it has only a bottom rail. This makes the model fragile and a bit hard to move around. Later versions of Akiyuki's module added a top rail, so I added one as well. This makes the model very sturdy.
  • Later versions from Akiyuki included some transparent side windows on the input hopper to make the mechanism more visible. I added these.

The image below compares djm's version 1 with my version 2:

640x384.jpg

Visually, the changes don't look like much but it was quite a challenge to integrate them without screwing up any of the other mechanical or structural features. The windows, in particular, took me forever but I am pretty proud of my solution which uses train windows to allow the flags to protrude into the support column. Again, I can't be certain if Akiyuki did it the same way based on the videos.

Because of all the variations over time, there is no way to say that this an accurate rendition of Akiyuk's module or even a snapshot in time, but it captures what seem like the best features of all the versions (in my opinion).

Some of the timing is pretty tricky here. Because all of the 40 tooth gears driving the cups are in series, if you adjust one by even a single tooth then it messes up everything downstream. I ended up using the molding lines on the gears as a datum and painstakingly counting teeth before I got it right. It is also important that the spiral lift drop a ball exactly when there is a cup there to receive it. This operation also requires single tooth accuracy. Once you get it all set though, the module is extremely reliable. It never jams, and any dropped balls roll back down the ramp to recirculate back into the system. It can also handle large loads of balls in the hopper without issue.

For whatever reason, my wife is fascinated with this module more than the others. It has something to do with the sound the rhythm of the balls makes. It is quite entrancing.

Share this post


Link to post
Share on other sites

Kudos to Blakbird for the version 2 build instructions. He gave me the opportunity to preview the build instructions. I can vouch that his instructions are significantly cleaner and clearer then my original instructions. The bar is being set higher!

Regards,

David

Share this post


Link to post
Share on other sites

I hadn't noticed you were Vancouver, Rebricker. This makes BrickCan a must for me! Looking forward to some awesome GBC action...

I'm especially excited about the train loader - that may be my favourite module of all. Keep up the amazing work!

I have not confirmed a GBC layout just yet, but with registration just starting, that will be hashed out very quickly I'm sure! I would need to get a line on enough balls to run the show as they are a little more rare, and until I joined VLC (the local LUG) there wan't a real GBC interest. So we've got to find a reasonable way to acquire enough balls if the show is to have a GBC, as it generally works better if all balls come from one source, that way there is no "Mine and Yours" thing going on.

Share this post


Link to post
Share on other sites
On 11/28/2015 at 2:49 AM, djm said:

I'd estimate that this is about 20% of the build. The two other "triggers" are fairly similar to the first which is visible in the diagram. The other two are primarily variations on a theme.The trigger I have built is about 80% done - it needs the upper surface and horizontal rotating "finger".Once the triggers are done, next will be linking them to the "control panel". There is one link out of the six required showing in the picture above. This will be followed by some framework and brickwork. The floor and the superstructure at this stage look fairly straight forward.

Since we were discussing multiple versions of the Cup to Cup module, I've also noticed that there are multiple versions of the ball shooter. Specifically, Akiyuki seems to have changed the back panel significantly in the newer video. Which do you intend to use?

Original:

640x289.jpg

Newer:

640x500.jpg

Share this post


Link to post
Share on other sites

Since we were discussing multiple versions of the Cup to Cup module, I've also noticed that there are multiple versions of the ball shooter. Specifically, Akiyuki seems to have changed the back panel significantly in the newer video. Which do you intend to use?

Hmmm, I don't know. A bridge I have yet to cross.

At the moment, I've encountered a 2 LDU "challenge" when digitally building the trigger. I'm likely to resort to building physically (it will never catch on) to see if the digital build is correct and whether or not it matters.

David

Share this post


Link to post
Share on other sites

Hmmm, I don't know. A bridge I have yet to cross.

I assume Akiyuki changed it for a reason, so I'd recommend the newer version.

At the moment, I've encountered a 2 LDU "challenge" when digitally building the trigger. I'm likely to resort to building physically (it will never catch on) to see if the digital build is correct and whether or not it matters.

I had a similar problem with the Bucket Wheel Tower. There is a 2 LDU interference on one of the ramp slopes. In the physical build, it is not a problem and actually makes the ramp line up better.

Share this post


Link to post
Share on other sites

I assume Akiyuki changed it for a reason, so I'd recommend the newer version.

Indeed, I've found when reverse engineering his GBC's I'll start off saying "why is that done like that" and then later will come the "Ahhh, that's cunning" moment.

I had a similar problem with the Bucket Wheel Tower. There is a 2 LDU interference on one of the ramp slopes. In the physical build, it is not a problem and actually makes the ramp line up better.

I've done the physical build of the trigger and can confirm that it isn't a problem. Indeed, it helped me to understand what another part is likely there for when I thought it appeared redundant. .... Ahhhh, that's cunning.

David

Edited by djm

Share this post


Link to post
Share on other sites

The latest news:

I finished building the cycloidal drive today and it works perfectly. There were a few portions of the instructions which had a poor building order and a couple of areas that were physically impossible to put together. I'll make corrections and updates to the file and then release those instructions this week.

I'll be building tilted rotors next. I'll let you know how it goes.

The last two parts for my Cup to Cup v2 arrived (the window flags) and to my chagrin I realized that my clever train window solution didn't work. The flags interfere with the windows. I had to switch to 1x2x3 panels but they don't exist in DBG so I had to use transparent black. It works fine and looks OK, but now the instructions and parts list are wrong. I'll make the updates soon. If you are planing to build this model, don't buy the train windows.

After that my next build is the fork to fork for which I have instructions ready. I also have a draft for spiral staircase which Rebricker has reviewed for me, so that will be next. He is also doing some work on the fork (different from fork to fork) file. And finally, dim is still working on the basket shooter, after which I'll take over on the instructions. So we'll be busy for quite some time yet!

Share this post


Link to post
Share on other sites

The latest news:

I finished building the cycloidal drive today and it works perfectly. There were a few portions of the instructions which had a poor building order and a couple of areas that were physically impossible to put together. I'll make corrections and updates to the file and then release those instructions this week.

I'm so sorry!! I forgot to change these poor building order but anyway you do that. Thanks^^

Edited by jesuskyr

Share this post


Link to post
Share on other sites

Many thanks Blakbird, Rebricker and others for this incredible effort. I finalized the Ball Factory and am in the final stage of completing the Invisible Lift. Next on my list will be the Cycloidal Drive or the Fork to Fork (both are fascinating).

Share this post


Link to post
Share on other sites

Instructions for Akiyuki's Cycloidal Drive module are now tested and complete and are available here along with a Brickstore parts list.

  • 133 pages
  • 2081 parts
640x480.jpg

Most of the credit for reverse engineering this module goes to jesuskyr who did the work making the LDraw file based on Akiyuki's video and creating most of the instruction steps. Before I knew he had made a file, I was working on it myself. I started with the internal and external rings. I thought it would just be a simple matter of copy and rotate, but nothing is ever that simple. The external ring has 10 lobes but they are not all the same; there are 2 types of construction. Likewise, the internal ring has 9 lobes with two types of construction. On top of that, there are also the radial supports which are not uniform. When I finished, I had this:

400x400.jpg

Making a 10 and 9 sided figure which is mostly empty in the middle out of only elements of only integral length is not easy, but Akiyuki makes it look easy. Even though it looked just right on the computer screen and the angles all came out perfectly, I was worried that I had missed something and wanted to confirm it would work so I made a quick animation. With an external gear of 10 teeth and an internal of 9 teeth, the overall ratio is 9:1 (reversed) so I put those rates into POV-Ray and came out with this:

anim1.gif

It looks perfect! (Incidentally, the inner wheel has 120 degree symmetry so I only animated the first 3 revolutions and then looped the frames. The effect is invisible.) It was at this point that I found jesuskyr's file and was happy to see that it agreed almost completely with my own, so from that point forward I switched to his file.

This module worked out really well. There were only a few small tweaks I had to make to the build. Most importantly, there were a couple of things that couldn't be assembled and therefore required some pin changes and build order adjustment. I also found that the 3rd stage of the output ramp had a slope that was too shallow so balls would not roll down reliably. I adjusted its exit down 1 stud and then adjusted the attach points appropriately. That worked great. I also found that the feeder ramp (the little ramp that feeds the wheel) was installed too tightly and would fall apart after a while. Again, I found a way to move things one stud and remove the problem. After that, it works flawlessly. I don't think it has dropped a single ball. It does take a considerable amount of power and uses the capability of that L motor. This particular model does bad things if you run it backwards, so it incorporates a ratchet which prevents that from happening. If you try to run it backwards, the clutch gears just slip.

There is minimal timing to set here. The only thing to adjust is the indexer to make sure a ball is released close to when the wheel needs it, but there is a wide acceptable tolerance band. You'll see a small turntable used as a gear in this model. Why? The ring gear of the turntable drives the wheel, and therefore is geared based on 56 teeth. The indexer needs to be able to release exactly one ball per cycle. Because the outside of the small turntable has 28 teeth (half of 56), this is possible. It would be very difficult to achieve this gearing any other way and could require many stages.

The cycloid ring is not the only great part of this module. Don't neglect the 7 stage stair lift at the input hopper which is really fun to watch. Each stage is built differently and becomes increasingly complex. You can see what I mean in the instructions.

Share this post


Link to post
Share on other sites

Progress update on the basket shooter;

640x480.jpg

This is a rendering of the build-to-date in LDD. I used LDD to POV-Ray Converter to turn the LDD file into a more aesthetically pleasing image than the screen capture capabilities available within LDD.

I'd estimate I'm somewhere in the range of 40%-50% of the way through the reverse engineering. I think (hope!) I'm over the main hump, which I believe is the triggers. The triggers have a substantial degree of commonality but just enough differences that made the reverse engineering a bit painful. However, I'm reasonably confident that they are pretty close to Akiyuki's design.

Chug, chug.

David

Share this post


Link to post
Share on other sites

Progress update on the basket shooter;

Great Work!! I'm looking forward to your basket shooter.

Edited by jesuskyr

Share this post


Link to post
Share on other sites

I've updated the instructions and parts list files for the Cup to Cup module to call out panels instead of train windows to allow the window pass-through. I've built this version and can confirm it works.

I've also now completed physical constructions of Tilted Rotors and Fork to Fork. I found lots of issues with the instructions for Tilted Rotors, so it will be a while before I can get the changes incorporated, however the problems have all been solved and the module works well. Fork to Fork works flawlessly and just needs some build order improvements.

Share this post


Link to post
Share on other sites

Instructions for Akiyuki's Fork to Fork module are now tested and complete and are available here along with a Brickstore parts list.

  • 129 pages
  • 1743 parts
640x360.jpg

Most of the credit for reverse engineering this module goes to jesuskyr who did the work making the LDraw file based on Akiyuki's video and creating most of the instruction steps. This module went together very smoothly even with the first draft of the instructions. I had only a couple of observations:

  • The model was much too slow and obviously did not support the 1 ball/s spec. After checking Akiyuki's video, I could see that his was much faster. A closer look at the video led me to believe that two gears on the same axis were NOT on the same axle. After fixing this, the speed went up by 3x and the clutch gear started working, so I am convinced it is right.
  • Jesuskyr's interpretation uses 5 more stages on the snake slide than Akiyuki's. This means that the output is too low to pass balls to a mating module. If you want to connect it to other modules, it is a simple matter to remove the latter stages. When used alone the lower output is nice.

The model is incredibly reliable: I haven't had a single problem with it. Like the Cycloidal Drive, it has a ratchet to prevent you from running it backwards. Running it backwards will not make anything terrible happen, but the "ferris wheel" will turn the wrong way. This module can support a large load of balls in the input hopper without jamming.

There is really no timing to adjust here. The forks are virtually self aligning based on the fact that knob wheels are used which have only 4 teeth. It is really obvious how to orient them. The relative timing of the ferris wheel doesn't matter. It will still deliver 1 ball per cycle.

Share this post


Link to post
Share on other sites
  • The model was much too slow and obviously did not support the 1 ball/s spec. After checking Akiyuki's video, I could see that his was much faster. A closer look at the video led me to believe that two gears on the same axis were NOT on the same axle. After fixing this, the speed went up by 3x and the clutch gear started working, so I am convinced it is right.

Thanks for your kind explanation^^ and endeavor for this project!!

Please let me know when you want to reverse engineering the other Akiyuki's GBC.

Edited by jesuskyr

Share this post


Link to post
Share on other sites

Thanks for your kind explanation^^ and endeavor for this project!!

Please let me know when you want to reverse engineering the other Akiyuki's GBC.

You can always check the first post to see which models are still needed. If you are looking for another one to work on, I'd recommend either the Catch and Release or the Pneumatic Module, but feel free to choose whatever interests you.

Share this post


Link to post
Share on other sites

I'd like to (try) to make an LDraw file for the marble run.

No idea when I'll be finished, but I'd like to use LDcad's flexible part system a little more :wink:

Just one little thing: it looks like he used a lot of those rigid hoses. Would it be better to make 2 long hoses in the LDraw file, or really try to mimic his setup of a lot of smaller hoses?

EDIT: if anyone's working on the elevator; it looks like the exit of the balls is a little different in the marble-run video. There's a Bionicle tooth that turns to push the balls away and there's a very precise groove to direct the balls into the next module (the marble run in the video).

Share this post


Link to post
Share on other sites

@Blakbird after looking through the instructions i see why the mechanics behind the lift wheel looked odd to me. I've never seen that tiny worm gear before. Any only available in 2 sets from last year! WOW! Good eye! I managed to build that part with a regular worm gear when I was Rebricking it, but I prefer the accuracy! Great job to jesuskyr and to you as well Blakbird beautiful instructions as always!

Share this post


Link to post
Share on other sites

Just one little thing: it looks like he used a lot of those rigid hoses. Would it be better to make 2 long hoses in the LDraw file, or really try to mimic his setup of a lot of smaller hoses?

If the idea is to eventually make instructions, then it would be best to use individual lengths so they can be added in steps. Furthermore, it would be ideal to actually use lengths which exist so that it can really be built. It is a big project!

@Blakbird after looking through the instructions i see why the mechanics behind the lift wheel looked odd to me. I've never seen that tiny worm gear before. Any only available in 2 sets from last year! WOW! Good eye! I managed to build that part with a regular worm gear when I was Rebricking it, but I prefer the accuracy! Great job to jesuskyr and to you as well Blakbird beautiful instructions as always!

That worm gear was introduced last year apparently as an intended replacement for the existing version but in a 3L length with a fixed attachment method instead of sliding. However LEGO pulled the design from that set (newer versions of the set contain the old worm gear). Apparently there was something inadequate about the design. No one knows if it will ever come back, so it could turn out to be a very rare part since the only way to get one was in the early versions of those 2 sets. Oddly, in both cases the worm gear was extra and only used in the B-model so it has never been used in a set. Works great in the GBC module though!

By the way, for those who haven't noticed I've updated the montage in the first post with all of the modules which are currently complete or near complete.

800x450.jpg

Share this post


Link to post
Share on other sites

Create an account or sign in to comment

You need to be a member in order to leave a comment

Create an account

Sign up for a new account in our community. It's easy!

Register a new account

Sign in

Already have an account? Sign in here.

Sign In Now

  • Recently Browsing   0 members

    No registered users viewing this page.