[EV3] Omnidirectional Robot Concept and Discussion

[TheMindGarage]

That connection is basically assuming that sqrt(8) = 3 (it isn't :P). I haven't tried that connection myself - it might be close enough to fudge it, but you'd risk breaking or bending something. If you're going to do that connection type, the best ax to use between the two 135-degree joiners (in order from best to worst):

• 16 studs (error of 0.03)
  
• 6 studs (error of 0.07)
  
• 9 studs (error of 0.10)
  

After that, your one is the next best.

[Jim]

Basically we're doing Pythagoras here......I kind of overlooked that.

The margin is minimal, but I'm all in favor of "no fudging" so I can build a valid model in LDD.

Thanks for the input guys. Still struggling with the design of the corners.

[TheMindGarage]

I've realised that it's possible to make a valid LDD if the diagonal axle is shorter than the gap, but not if it's longer. You just need to build the frame without the diagonal axle, then put the axle in. Best axle lengths are 6 studs or 3 studs.

[BusterHaus]

Let me know what you think of this. LDR file.

[Josephiah]

I'm growing very fond of this new part by the way. It almost feels like cheating when you're using this part, it's so convenient.

Much of the studless building process feels to me like a chess game where you're constantly having to think "right, in three moves' time I'm going to need a pinhole exactly HERE, and in THIS orientation". This part allows you to place that pinhole exactly where you like with less disruption to the rest of your structure, so it's a great little convenient shortcut!

Fascinating project, by the way - will be following this with interest.

[Jim]

Let me know what you think of this. LDR file.

Amen brother! I love it. That's exactly the kind of construction is was looking for. I will definitely see how this works out. Thanks!

(color scheme might be a challenge, but I can face that challenge)

This part allows you to place that pinhole exactly where you like with less disruption to the rest of your structure, so it's a great little convenient shortcut!

That's indeed why I like it. It only takes up a single stud, can be placed in multiple directions. So versatile.

Fascinating project, by the way - will be following this with interest.

Cool! Progressing slowly, but this will probably be my Magnum Opus, so no need to rush

[Jim]

CONCEPT V4

Okay, it's been a while since I have posted an update. I have been very busy with all kinds of other (non-LEGO) projects, so my building time is limited. 
I was focused on designing the wheels and wheel covers etc, but I realised I needed to focus on the base first. Panels and stuff will come later.
My goal of making it (more or less) modular needs a simple approach. 

I have been inspired by the base of the Bucket Wheel Excavator. This made me rethink my first approach.
After some versions, I hereby present you with version 4 of the base.

DIGITAL PROTOTYPE

BASE V4

The actual build of the inner base.

WHEELS

Four wheels will be attached to the base. 

BASE V4 VERSION 1

The wheels can be attached as shown in the digital design.

BASE V4 VERSION 2

However, this design might be sturdier. Both designs (v1 and v2) need to be reinforced to carry the weight of the upper section.
I will continue to make the other three wheel sections so I can test the proof of concept.

Maybe I will rotate the wheel sections 180 degrees, so that the wheel is closer to the base, to improve the weight distribution.

Again, I have noticed that there's no uniform Red color, like yellow. I tried to make a picture, but it's not as noticeable as in real life. 

Hopefully I will post some pictures with all four wheels attached next week.

[Myers Lego Technic]

First, great Renders!

Interesting idea! I look forward to more progress in this field.

[JJ2]

Nice progress, you will need to place some liftarms on the inside of the build to keep the wheel sections from separating from the main chassis

Probably on the 5L part of the frames where they join

[Jim]

7 hours ago, JJ2 said:

Nice progress, you will need to place some liftarms on the inside of the build to keep the wheel sections from separating from the main chassis

True. After shooting the pictures, I did some bracing to om the wheel section, but I need bracing on the base as well. 

I will continue to work on the other three wheel sections and after that I will think about the bracing. I like it to be modular, so I will aim for removable braces, using dog bone frames and 5x7 frames.

[The_Icestorm]

Nice setup, everything looks nice and compact, as well as sturdy. Can't wait to see the finished thing!

One question: how much clearance is there between the ground and the wheel-frame?

[DrJB]

One thought comes to mind. Depending on how much load you want the robot to carry, do you think one single motor per wheel will be sufficient? The 'weakest' part is typically the axle from the motor (and thus the wheel) and there is so much load it can carry, but not more.

[Jim]

12 minutes ago, DrJB said:

One thought comes to mind. Depending on how much load you want the robot to carry, do you think one single motor per wheel will be sufficient? The 'weakest' part is typically the axle from the motor (and thus the wheel) and there is so much load it can carry, but not more.

To be honest....I'm not sure. I can use a motor on the other side of the wheel as well. But I seriously have no clue as to how this will perform. That's one of the reasons I am going for a modular approach. If one design doesn't work, I like to be able to replace the wheel units fairly easy. Using two motors per units will need a multiplexer, or simply two EV3 units. I am not sure if I can attach 8 motors to a single brick (2 per port). I highly doubt it. These are the things I need to find out. And that's also part of the fun.

There is a 12t bevel gear attached to the motor by the way. So it is geared down a lot already. It won't be very rapid.

Yesterday I have finished the other wheel units, so I will be able to attach them soon. 

Maybe I will use the single motor setup to start programming the robot. After all it will need quite some math to navigate properly. I'd like to do a PID controller to let the robot use a compass etc. That will be a challenge too.

[Kristof]

Although this boxy modular approach looks great and perhaps is very beneficial for prototyping, it seems it restricts you from making realy sturdy base WITH the wheels attached. There will always be that connection point which will either need extensive amount of bracing to be somewhat sufficiently reinforced, or will be prone to bending. I think at least the base should be built as a single structure, using as much of the favorable structural geometry (inspired by real life structures, e.g. gridges...) as possible to let it withstand the load in the center, while not becomming dense and heavy pack of parts that will spring and bend anyway :)

The best way to prevent some connection place from bending is to implement such braces that would be 'drawn' or squeezed if such bend occured. That usually lead to some diagonal braces, though I am sure you know that, possibly better than I do :)

[Jim]

4 minutes ago, krisandkris12 said:

That usually lead to some diagonal braces, though I am sure you know that, possibly better than I do :)

Nope, I don't 

This project is meant to learn a thing or two, so I will see along the way. You are absolutely right about the base needing to be sturdy. But I still like to try a modular approach, albeit with a lot or bracing. We'll see.

Thanks for input though, it's appreciated!

[Kristof]

4 hours ago, Jim said:

But I still like to try a modular approach, albeit with a lot or bracing. We'll see.

Thanks for input though, it's appreciated!

I wish I could help in more specific way :D My knowledge is rather theoretical, beyond that I am quite poor technic builder.

What I found though is that technic liftarms are considerably stiffer against bend when the pinholes are facing in the 'bend plane' compared to them being perpendicular to it. It works sort of like an I beam. For your build, that would mean having some liftarms with holes aligned vertically. That is, however, often tricky to achieve if most of the build has this horizontal fashion :) Can be that the extra weakness introduced by some intricate joinery in order to get some paralel liftarms in ther wouldn't in fat overweight the positive effect of such bracing :D

[Jim]

On 2017-5-6 at 0:40 PM, krisandkris12 said:

Although this boxy modular approach looks great and perhaps is very beneficial for prototyping, it seems it restricts you from making realy sturdy base WITH the wheels attached. There will always be that connection point which will either need extensive amount of bracing to be somewhat sufficiently reinforced, or will be prone to bending. I think at least the base should be built as a single structure, using as much of the favorable structural geometry (inspired by real life structures, e.g. gridges...) as possible to let it withstand the load in the center, while not becomming dense and heavy pack of parts that will spring and bend anyway :)

I haven't made any more progress, but I have been thinking about your comment and maybe I need to make a single base frame, instead of going for the modular approach. It will take a lot of bracing to make it sturdy enough and what are the actual benefits?! It's not that I will be dismantling the wheels every other day or something.

And I like to position the wheels a bit more to the center, where much of the load of the upper body will be transfered to the base. I can obviously achieve this very easily by rotating the wheels 180 degrees.

On 2017-5-5 at 8:14 PM, DrJB said:

One thought comes to mind. Depending on how much load you want the robot to carry, do you think one single motor per wheel will be sufficient? The 'weakest' part is typically the axle from the motor (and thus the wheel) and there is so much load it can carry, but not more.

And maybe I should use two motors per wheel in this new single base.

On 2017-5-4 at 9:41 AM, The_Icestorm said:

One question: how much clearance is there between the ground and the wheel-frame?

I'd say about 4cm.

[Jim]

Base-v4.1a by Jim van Gulik on Sketchfab

I have been testing Sketchfab and it's very cool. I had to import my LDD file to Stud.io, which resulted in some illegal connections. But uploading the model to Sketchfab can be done directly from Stud.io.

I must say I am impressed! 

[syclone]

Unfortunately it requires more than my laptop can give, the page freezes completely while loading and then it runs at barely 2fps. 

[Eddie_Young]

Your photos are beautiful  I usually only do nice photos of the finished product because I'm lazy...

 I've had a 7kg robot in the past and since you said you're going for around 20 kg, you really have to be careful about estimating the gear ratios, the wheels are enormous! And I definitely agree with two motors per wheel.

Sketchfab looks cool but what does one use it for? 

[Jim]

On 6/20/2017 at 6:44 PM, Eddie_Young said:

I've had a 7kg robot in the past and since you said you're going for around 20 kg, you really have to be careful about estimating the gear ratios, the wheels are enormous! And I definitely agree with two motors per wheel.

Sorry for bumping my old topic, but since we discussed it in another topic, I was going through it again. And I found it worth mentioning that I recently acquired two more EV3 sets. So for the base I will have a setup of two Large Motors per wheel and two EV3 units. 
I have rechargeable batteries for all my EV3 units, but I do need to get myself an extra power plugs. Or three extra plugs, so I can charge simultaneously.

I am aiming at making a rotating body, using the BWE gearracks in black (Rough Terrain Crane). I have two EV3 units for the upper body.
After the reviews of the 2H 2018 Technic sets, I will try to reinvigorate this project. I simply can't let all this EV3 potential go to waste.

By the way; some design decisions;

1. I have abandonded the modular concept for the base, because I will probably never make any other modules and the base needs to be strong enough. There's no need for it to be modular.
2. I will not think about the exterior/bodywork of the robot, until I have finished the chassis.
3. I will try to use mostly the same elements as in the original EV3 set. Which means black beams, red connectors, white panels etc.

[Jim]

So I took the bull by the horns and started working on a new design. Slightly smaller, so less empty space around the wheels. But with two motors, so still pretty bulky. And I did end up with a somewhat modular design. But it's simply easier to design one wheel element and copy it three times. I just need to make sure the center will be sturdy enough. Adding a layer on top will probably make it sturdy enough. 

[Erik Leppen]

If you're going for sturdy and rigid, I wonder why you're working studless. I would expect it to be much easier and more effective if you took the 75192 approach to frame building...

Personally, I find studded frames take less space for the same rigidity, and mainly for that reason I find them look more elegant.

Studless frames meant to be very rigid, quickly look like a giant block of parts. And to be honest, that's the first impression I get when I see your photos. There's lots of double layers, but mostly, it's relatively small/short parts (5x7 beams are great for three-dimensional work, but they are only 7 long). Every joint adds more play.

Also, on your last picture, the left and right halves aren't connected, except for four pins between pairs of 5x7 frames.

As a final advice...triangles. Your whole design seems to trust on the rigidity of parts themselves (I know studded frames usually do this too), mainly 5x7 frames, but you don't seem to have built in much other rigidity in the structure of those parts. There's no long continuous beam covering the whole length that will absorb the tensile force at the bottom and the compressive force at the top. Secondly, adding those little red connectors along the sides will hardly make things stronger (I think they only make it harder to build); but do add weight and complexity. I think the best way to achieve stiffness is having two horizontal layers of studded "planes", one bottom, one top, connected by angled beams forming triangles. The larger the triangle (and the more "equilateral") the better. If there's little play in the joints, the forces will have to be abosrbed by beams as stretch/compress, which beams are very good at.

See this bridge model. It consists of mostly triangles, because triangles are rigid. Filling up a triangle with more stuff wouldn't add anything - what matters is that the 3 sides are robust and connected firmly with as little play as possible. Even if the sides of a triangle could hinge freely, the fact that it's a triangle will prevent it from doing so.

I think I would approach your base structure as two parallel bridges, crossed by two other parallel bidges perpendicular to them. Like a tic-tac-toe grid (seen from above) of bridges.

Yes, your structure may become (much) taller than 5 studs. But you may need that to achieve the strength. Strength-wise, it's best if the force-bearing elements are far apart (because a larger arm results in a lower force), so they form a large hollow structure. Any free space between the frame and the bodywork is space that could have been used to add a really strong extra triangular link.

(Also, why would the base need to be flat at all. I think a more pyramid-shaped design, where the center sits higher, would give you much more space to build a strong base.

PS I'm not an engineer, but I did play Bridge Building Game a lot in the past.

Edit: here's a quick example I fiddled in MLCAD:

More diagonals will still need to be added (yellow lines as suggestions; probably with diagonals too). This whole thing (so far) is only 140 beams (and, if my count is correct, 280 pins). Of course, you'll need a bunch of connectors at the crossings. It's 13 tall because that allows to use the 5-12-13 triangle (12 being vertical). As a bonus, you could combine this with 12-12-17 triangles which are an almost-exact fit (you won't notice it's not exact).

Another edit: here's a quick test with 12-12-17 triangles:

Edited July 14, 2018 by Erik Leppen
Added suggestion

[Jim]

Hi Erik,

Wow, thanks for the elaborate post/answer!!

Studful is a no go for me, sorry. I have a love for studless Technic, so I basically want to stick to that. And preferably using mostly elements from the EV3 sets, but that's not a rule. So thanks, but no thanks 

You are absolutely right about triangles and the bridge design though. It's not a big deal that the design will become more than 5 studs high. I was planning on adding another layer to the base anyway. This 5 stud design was just one layer to begin with. Since it will be fairly big, I like to work with easy to use/design chunks of the model. Of course, this should allow for rigid building.

I do like your setup btw. I will see if I can use your design and sort of slide in the wheel hubs. You definitely got me thinking 

Btw the current wheel hub setup is indeed not finished and it should be placed inside another assembly, because there is no rigidity at the moment.

[Jim]

My previous setup is resulting in a rather large footprint, which might not be so convenient. That's what I have started a new approach, which might be better, structural and size wise. Of course this is just a concept, which needs to be worked out further.

The wheels are closer together, which might be beneficial for supporting a high(er) load.