http://www.youtube.com/watch?v=Uw1BgD72vaA

Building instructions now available!

I just built this yesterday and I must say that it is absolutely perfect. It worked perfectly from the beginning, and it much more user-friendly than the Tilted Twister, and it works better and faster than it, too. I would recommend building this for anybody who has a LEGO Mindstorms 2.0 set.

Anyone want to try?

Is there any difference to the to competetists here?
viewtopic.php?p=269510#p269510
(please scroll down)

The two robots in the thread you linked to are known as tilted twisters. Reading about this mindcuber, it looks like it is designed to be an improved version, and it is by a different designer. I have built tilted twister, and it works alright.......if you don't mind the cube falling out.

Cool. I will try out in holidays.
My tilted twister did work, however... the cube sometimes fell out and sometimes was not solved properly. I hope this one will be as good as it seems to be!
Thanks for sharing.

No, "Tilted Twister" is a name that Hans Andersson gave to his LEGO NXT Rubik's Cube solver because of the nature in which it manipulates the cube with a tilted structure.

My MindCuber is a completely different mechanical and software design and is not "tilted" in any way

Thanks for sharing, I do hope many others enjoy building and using MindCuber

I hope you find MindCuber more reliable if you try it.

I have demonstrated it at several events since last I first published the video and it is always very reliable for me. In one test at home while I was trying to make it as reliable as possible, it solved one cube 80 times in a row without any mistakes on a single battery charge. Of course your "mileage may vary"

I'd love to hear from anybody who tries it.

The mileage is fine, but what's really great about this is the speed vs. battery charge curve. It appears to be a decaying exponential function. As the battery gets lower, the speed decreases slightly, but when it reaches about 65%, the speed doesn't change AT ALL. But it's much faster when leaving it plugged in with the rechargeable battery. That's what I do.

Thanks. It sounds like yours is pretty reliable then too?

You're right that leaving the charger plugged in does keep it running at maximum speed although when demonstrating at shows I prefer to run it from the batteries to emphasise that it is a self-contained autonomous robot.

And it would be great to hear if anyone is inspired to adapt the MindCuber software to solve a Void Cube

http://www.youtube.com/watch?v=9QL5DZBo8WQ

My robotics team and I got this running today and we were using a Dayan Zhanchi. it was somewhat reliable but didn't live up to my full expectations. my question is does it work better with a Rubik's brand cube? Perhaps we are just experiencing the usual technical difficulties that you get when building complex machines from small plastic building toys. Either way this works 10x better than Tilted Twister.

Thanks for building MindCuber.

You are right that building a robot out of small plastic pieces held together by friction is not exactly "precision mechanics" but in my experience it can be very reliable with a smooth cube and the right lighting conditions (not too bright).

What is unreliable?

1) the scan?

2) the mechanical solve?
if so, does the cube sometimes not tilt correctly? or does it sometimes jam when trying to turn a face? both? or something else?

3) or something else?

Have you tried aligning the turntable and calibrating the color sensor using the white face each time the program is run as described in the "How to..." document?

Did you include the stickers in the turntable to help the cube slide back into position?

There is also a possibility that the scan arm has been constructed incorrectly but is close enough for the scan to work sometimes. For example, several people have connected the horizontal beam across the scan arm in the wrong holes or missed the black pegs that stop the scan motor as it turns towards the cube to calibrate its position when the program is first run. It might be worth double checking that the scan arm is constructed exactly as shown in the build instructions.

I did most of my testing with a Rubik's branded cube as I assumed (perhaps incorrectly!) that this is the cube that most people who tried MindCuber would have. However, as long as one face is white and the others are distinct colors the scan ought to work. The scan is unlikely to work reliably if one of the faces is black. Some people have also found that the scan is more reliable with cubes that have a black plastic body than a white one. I have not tried cubes with colored bodies but I suspect that might cause a problem with the scan. Some people have discovered that cubes with stickers tend to scan more reliably than those with plastic tiles on the faces.

I hope some of this helps but if not, please describe in more detail what is unreliable and perhaps show me a video of the problem I will try to help.

Thanks for your complimentary remark about it working better than TiltedTwister. I have not built Hans Anderrson's design myself as I didn't want to be influenced by it but I have seen a few videos and I think it is a really impressive creation!

Thanks again.

The tilting mechanism is the problem, however my coach mentioned that he couldn't find some orange plates that were supposed to go on the cube table. These must be the stickers. We'll try the stickers, and even without them it will work pretty consistently.

thanks for your help!

Yes, the stickers are intended to help the cube slide back without catching on the middle of the turntable.

I used stickers in the instructions as this keeps the design 100% LEGO (the stickers are included in the standard NXT set)! However, an alternative, which lasts longer than the stickers, is to cut a piece of paper as a square to fit in the turntable with some "tabs" of paper left sticking out at each corner that can tuck under the corners of the turntable to hold the paper in place.

I use the paper approach when demonstrating MindCuber at exhibitions as it typically solves the cube about 120 times a day and the stickers tend to peel of with that amount of use!

I hope it works consistently for you whichever approach you take. Please let me know if this helps.

Thanks again for building MindCuber.

Your MindCuber looks great!

Early on 2009 i brought a NTX 1.0 to make an old version of the TiltedTwister:

http://tiltedtwister.com/tiltedtwister.html

Since then i never had courage to disassembly my NTX in order to build another projects,

i had to adjust some little details such as making a power supply to the brick instead of using batterys what helped a lot and other minor things...

Looking to your's MindCuber make me want to buy another NTX, this time a 2.0 to make a MindCuber now.

Thanks.

If you do decide to buy NXT 2.0 to make MindCuber I'd love to hear about it!

Very interesting, this thread has inspired me to start a Minimalistic Mindcuber Competition.