[GBC] The Akiyuki Project (2010 - 2020)

[doug72]

On 3/5/2019 at 6:47 PM, Doug72 said:

Thanks - that was quick
Cost is £6.12 UK
I have all the parts except the LEGO Gear Wheel Z12 with Hole / Crossaxle (24014) which are the robot hands

Catch & Spin GBC.

Now purchased the instructions and built this module after sourcing the extra parts required.
Had to use colours I had available.
Didn’t have any eyes so made them from 1L round white plates with the stud part blacked using a marker pen.

Robots are slightly out of phase with each other. (to be corrected).
The yellow robot is working OK

Tested and module working but have an issue with the Red robot that  is noticeably slower and performs an additional opening & closing of the hands before reaching the position when the roller drops down. Hands then re-open & to pick up ball OK - then robot rotates to open and drop ball down discharge chute.
Robot sometimes fails to rotate and drop into the loading position

All gears checked and freely turning. Correct friction free pins used.

Anyone got ideas how to fix this ?

See video.

 

 

Edited April 12, 2019 by Doug72

[doug72]

3 hours ago, Doug72 said:

Anyone got ideas how to fix this ?

Problem solved.

The half Z28 turntable for the red robot was 180 degs out of phase with the 24T gear below that has the short crank arm.

Now tested and Red robot working at same speed as the Yellow robot.

 

 

Edited April 12, 2019 by Doug72

[Juroen]

I finally ordered the last pneumatic hoses, so I can finish the ball cleaner and start on the compressor.

[Lego_GBC_NL]

22 minutes ago, Juroen said:

I finally ordered the last pneumatic hoses, so I can finish the ball cleaner and start on the compressor.

That's good to hear! Was getting a bit worried as it was so quiet lately around the Ball Cleaner builds. Hope you can finish it soon and all works well 

[Juroen]

5 hours ago, Lego_GBC_NL said:

That's good to hear! Was getting a bit worried as it was so quiet lately around the Ball Cleaner builds. Hope you can finish it soon and all works well 

The webshop where I bought my pneumatic hose was down for almost 3 months, so I had to wait for them to reopen. As I wanted all hoses to be the same type. I thought that 5 meters would suffice in the first place. But that is not nearly enough. So I ordered an extra 5 meters.

[Ankoku]

Great to see the cleaner coming along. It is such a big and difficult project to do.

I rather like the twin colour Catch & Spin!

I have started to build the Akiyuki Cup-to-Cup Type 1 module.

As ever, it seems that the instructions do not match the Akiyuki original and thus don't seem to work as well. Has anyone tried to work out what is going on in the original design?

This is nothing like what is in the instructions.

Also, using 2x Technic, Liftarm 1 x 3 Thin is smoother and more secure than Technic, Axle and Pin Connector Perpendicular 3L with Center Pin Hole.

[9v system]

2 hours ago, Ankoku said:

Great to see the cleaner coming along. It is such a big and difficult project to do.

I rather like the twin colour Catch & Spin!

I have started to build the Akiyuki Cup-to-Cup Type 1 module.

As ever, it seems that the instructions do not match the Akiyuki original and thus don't seem to work as well. Has anyone tried to work out what is going on in the original design?

This is nothing like what is in the instructions.

Also, using 2x Technic, Liftarm 1 x 3 Thin is smoother and more secure than Technic, Axle and Pin Connector Perpendicular 3L with Center Pin Hole.

i think @djm made a modification on his version, so maybe ask him

[Ankoku]

The instructions version seems to be finicky and wants to jam.

I will start changing it back to the original. Already noticed a few differences in the images that will take some of the variance away. e.g. I prefer things in fixed positions rather than where you need to adjust it.

[djm]

8 hours ago, 9v system said:

i think @djm made a modification on his version, so maybe ask him

Way back on page 29 of this thread.

Akiyuki's original model used one mechanism, which was what was originally replicated when we reverse engineered the module. The top of the spiral lift changed in a subsequent Akiyuki iteration. I came up with an approximation of it subsequently, which is what my "delta" to the model recreates.

 

On 10/3/2016 at 11:24 PM, djm said:

As per my immediately previous post, I've been getting some GBC modules ready for my LUG's Brick Show. One of the modules I pulled out of storage was Akiyuki's Cup to Cup. One of the challenges I found when running this module is getting the ball to reliably load from the spiral lift into the first cup. The build instructions I wrote reflected the original design of Akiyuki's i.e. using a Technic #4 angled connector. Sometimes the balls load adequately into the first cup but at other times the load fails and the ball drops back to the bottom. I found the less level the table, the less reliable the loading mechanism was.

I recalled from when I first wrote the build instructions, there were actually two different designs for the loading mechanism. I had chosen the one from the original design as it was much easier to see how it was built in the images which were available. The images available for the second loading mechanism were too unclear for me to be able to discern how it was built.

Fast-forward to now. In readying the Cup to Cup for the show, I became a bit frustrated at the unreliability of the loading mechanism. So I resolved to revisit the images of the second loading mechanism and see if I could work out how it might be assembled. Suffice to say, I'm making this post to identify that I have managed to come up with something akin to said second mechanism. I have tested the assembly and I am now getting 100% successful loading of the balls into the first cup.

The solution requires some fairly minor changes to the rear wall of the elevator, after which it will look something like this:

I have produced an LDraw and a PDF file which contain just the rear elevator wall element of the module, should you wish to incorporate this loading mechanism into your build. The parts which are different in the second loading mechanism are coloured red in the LDraw and PDF files, to highlight the changes. Obviously the expectation is that they would be built in Dark Bluish Grey and Black, as appropriate. It should be noted that Blakbird improved the slopes used in v2 of the build (see page 14). The LDraw and PDF files are still using the original slope from v1.

Regards,

David

Regards,

David

[Ankoku]

Found a couple more shots of it in various YouTube vids.

 

Looks as though it is something like this:

Weirdly, Stud.IO on has 7 lengths of Ribbed Hose, the length needed for this module is not one of them. There are about 30 different lengths in total.

[Ankoku]

For anyone with the 2 plate Ball Factory, this is how Akiyuki deals with splitting the plates for transport.

[HRU_Bricks]

On 4/15/2019 at 1:35 AM, Ankoku said:

Found a couple more shots of it in various YouTube vids.

 

Looks as though it is something like this:

Weirdly, Stud.IO on has 7 lengths of Ribbed Hose, the length needed for this module is not one of them. There are about 30 different lengths in total.

Those flexible hoses,  such a pain to model in LDD and Studio

On 4/13/2019 at 11:38 PM, Juroen said:

I finally ordered the last pneumatic hoses, so I can finish the ball cleaner and start on the compressor.

Looks really good,  have you got it running? 

 

 

[Ankoku]

Yeah, I tried bending it in Studio.IO and it instantly looked like a Slinky gone wrong. So I hit undo.

The tools for hoses could do with some work.

[Ankoku]

I just realised that the axle that goes through the centre of the spiral lift is probably a single 16L, rather than 2 separate axles which float around. I am going to leave it as is, as I want to cap off the top with another tile.

That said, I do suspect it is a 16L Axle in the original.

[Juroen]

On 4/16/2019 at 5:16 PM, HRU_Bricks said:

Looks really good,  have you got it running? 

Yes, all motors are running fine on the EC3-brick. Last thing to do is build a compressor. 

[Ankoku]

I am starting to feel that this bit:

Sits one plate higher than it does in the instructions. I believe that makes the height difference between the pin points of both tubes the same. Thus the upper tube would now hug as close to the rotator as the lower one.

It also mean the ball drops into the cup without the need for such a strong angle on the cup arm.

Still experimenting, but those are my initial feelings.

[Ankoku]

Okay, so here is the height adjustment in action.

This is with zero effort on syncing or cup angle, so it could certainly be better. It was my first attempt with the adjusted height and original guiders.

Here is one notch faster on the LEGO 8878 Power Functions Rechargeable Battery:

Not sure which vid it occurs in, but you see the spiral lift slow a little due to the balls getting stuck slightly.
This does bug me, but it occurs on even the latest version of Akiyuki's Cup-to-Cup. I will look into it further, but may be a design flaw. Never noticed it on the Ball Factory.

 

Anyway, the outcome of the change was better than expected.

The way the balls get lost behind it really bugs me.

And yes, this is the smallest amount of testing possible, but it feels like I am going in the right direction.

[Ankoku]

I can replicate a pause and strain in motor with no balls. Interestingly, it seems to be roughly X amount of revolutions. It seem quite rhythmic. So probably every 15-20 would-be ball exchanges, it will pause slightly. No idea why.

[Ankoku]

Also works without issue on top speed.

[rskamen]

On 4/12/2019 at 5:52 AM, Doug72 said:

Problem solved.

The half Z28 turntable for the red robot was 180 degs out of phase with the 24T gear below that has the short crank arm.

Now tested and Red robot working at same speed as the Yellow robot.

 

 

@Doug72 While awaiting parts for the Train layout, I started building the Catch & Spin Module. Glad you were able to figure out your timing and speed issues. I always get nervous I won't get it done properly. I am following the instructions carefully and reading the appendix at the end of the pdf file. I am about halfway finished. May need to supplement with a few different colored pieces as well. I will keep you posted. Let me know if you have any specific do's or don'ts, or what to watch out for. Hopefully will get this finished before Easter and my northern family comes down for a weeks visit to the warm Florida sunshine. Thanks

[Ankoku]

Looks as though he no longer has easy access to the latest version of Cup-to-Cup Type 1. Here is a better quality image, with a favourable angle from the version you see in the 2012 video: "LEGO Great Ball Contraption (GBC) Layout 2012.9". I guess that is when he last took photos of it.

I am curious as to his thinking in regards to moving the motor position which you see from 2013 onwards. Was it for better access to the motor in case you needed to change it?

Did it relate to the pauses I am seeing, which may be to do with the mechanism which rocks the gears back and forth.

Interestingly, the ball guide which stops it progressing further round had the slope at the same level as a the tile. I have the slope one plate up.

Question for @djm: In your version, you use the Slope 45 2 x 8, instead of 2x Slope 33 3 x 2 & 2x Slope 18 4 x 2. Do you think the 45° degree slope plus area of flat surface works better than having the ramp right up to the tube?

In the 2013 videos and after, he is using 33° slopes, but prior he is using 45°.

 

 

[djm]

12 hours ago, Ankoku said:

Question for @djm: In your version, you use the Slope 45 2 x 8, instead of 2x Slope 33 3 x 2 & 2x Slope 18 4 x 2. Do you think the 45° degree slope plus area of flat surface works better than having the ramp right up to the tube?

I can't recall whether there was a conscious reason for making that choice or not. It could just as easily have been me making a mistaken identification of which parts were used in the model. I always considered my reverse-engineering of Akiyuki's GBC modules to be an "interpretation" rather than a replication. Perhaps experiment and see whether there is a difference in behaviour.

David

[doug72]

19 hours ago, rskamen said:

Let me know if you have any specific do's or don'ts, or what to watch out for.

Found you do have to check & double check each stage before moving on.
I had to rebuild both robots twice before I was sure they were built correct.
Difficult at times to determine exactly how parts connected.

Also had to rob older MOCs for some parts to co-ordinate colours etc !!!

Even with an engineering background - still hard to figure out how it works !!!

Made a small MOD as did not like how the ball input was only attached on one side, plus 

Added support legs to keep it stable. see image below.

Edited April 19, 2019 by Doug72

[Ankoku]

It was only connected on one side? That would bug me too.

For me, with the Akiyuki modules, I build them as the instructions say and if they work like in the videos, then I am happy to leave it at that. If they don't, that is when I start trying to work out where the differences are and fix them. The Ball Factory was my first ever GBC anything and I went in believing it to be correct. It was only watching it in operation that made me start looking into things.

Obviously, I do make some changes. Catch & Release got a lot of mods as I didn't like the issues I could see with it, even in his videos. I am not entirely puritan! :P

My main issue with Cup-to-Cup Type 1 at the moment, is the gear assembly beneath. It is just floating there, not well secured, which means if there is any resistance, it moves around and can act like a spring once the resistance is surpassed.

I have been hand cranking it and there is resistance is the cup-to-cup part, so I will loosen everything up there before making any changes.

[doug72]

50 minutes ago, Ankoku said:

It was only connected on one side? That would bug me too.

Correction:- My mistake
Seems I missed out page 84 steps 1 to 3 and only put the 1 x 10 studded beam before moving onto page 85. 
I however came up with almost the same solution but additional legs also help keep the input box from drooping especially when moving the module from its storage shelf.

Edited April 19, 2019 by Doug72