Help with finding solutions for getting an AT-AT walk cycle like in the movies.

[BrickBear]

53 minutes ago, Nelson said:

Here are a couple more ideas that may help your cause. I too have the same issues with balance/leaning and sought to get some compensation in the feet and legs. I came up with a way of mating the external legs to their internal connecting rods with a small ball/bar piece wedged in between a stud-pin and round 1x1 plate to create a bit of negative camber (Image 1). I also utilized the ankle strut feature seen in the film models to get more negative camber from the leg to the foot. I made the struts just a bit too long for the distance so they push the feet out a bit crooked (Image 2). Also note that the feet/ankles on my model have very limited movement, probably less than 10 degrees. I found this essential to keep the model from pitching forward or backward and falling on its face or butt if the feet slid. Finally, you can see in image 3 that the model is quite rigid at all points during its cycle. If I hold the it sideways, the legs barely deflect at all. My early versions all started out pretty loose and I had to really lock things down. Of course, we are taking somewhat different approaches, which I find really intriguing, so these features may not be as necessary in your design. There was a post earlier in this thread that suggested a cam system that is very close to the concept I'm using (although I had already tried and failed with the parts he suggested and used a different method of creating cams.)

Mmm, I used a plate with rail to offset some of the lean. I’ve considered the inner rod as a point of support too, adjusting where the pins go too. If I may how exactly does yours work? 

Edited October 30, 2024 by BrickBear

[Nelson]

Here is my effort to explain the mechanics of my design concept.

 

[Davidz90]

Recently, I stumbled upon this video that may be relevant:

This design seems to be exceptionally stable.

[Nelson]

I've seen those walking designs based on Jansen's linkage. If you look closely, you'll see some gears in there that are only a quarter of 360 degrees. Unfortunately, there is no Lego analogue of this.

[Davidz90]

12 minutes ago, Nelson said:

If you look closely, you'll see some gears in there that are only a quarter of 360 degrees. Unfortunately, there is no Lego analogue of this.

Yes, noticed that as well. Thanks to this, only one leg is lifted at a time. While there is no direct Lego analogue, there may be some ways to create such an intermittent motion. Seems like an interesting problem, I'll look into that.

[Nelson]

7 minutes ago, Davidz90 said:

Yes, noticed that as well. Thanks to this, only one leg is lifted at a time. While there is no direct Lego analogue, there may be some ways to create such an intermittent motion. Seems like an interesting problem, I'll look into that.

I wonder if you could use part 78442 and reverse the way that 1/4 gear meshes. Might be fun to tinker with, even if it doesn't exactly correlate to the AT-AT proportions or walking style.

[BrickBear]

2 hours ago, Nelson said:

Here is my effort to explain the mechanics of my design concept.

 

Makes sense. Thankyou, very interesting. I’m very excited to get working on my version again eventually with all this inspiring talk. I probably need to order a few pieces first…

[Davidz90]

One option for slower-faster motion is this arrangement with two offse axles.

(I hope I'm not hijacking the thread, just throwing some ideas)

Edited October 31, 2024 by Davidz90

[Nelson]

4 hours ago, Davidz90 said:

One option for slower-faster motion is this arrangement with two offse axles.

(I hope I'm not hijacking the thread, just throwing some ideas)

Thanks for the share. I looked at some of these concepts on the 507 Mechanical Movements site (https://507movements.com/)

My main issue with gear-based designs is that they are:

A) not very stable
B) have a lot of lash so the accuracy suffers
C. can't take the torque applied to them in such a large and heavy model and tend to jump teeth.

But I'm always open to ideas. That's why I'm here. If someone can make a gear-based design work, I'm all in for adopting it.

[Davidz90]

 

53 minutes ago, Nelson said:

Thanks for the share. I looked at some of these concepts on the 507 Mechanical Movements site

Thanks for a cool link!

So far, I cobbled together a basic Theo Jansen style walker with no extra gearing. It is stable and walks, but by clumsily dragging feet on the ground. I got a new appreciacion of how much torque is involved in this.

 

[Nelson]

3 hours ago, Davidz90 said:

 

Thanks for a cool link!

So far, I cobbled together a basic Theo Jansen style walker with no extra gearing. It is stable and walks, but by clumsily dragging feet on the ground. I got a new appreciacion of how much torque is involved in this.

 

Ha! That's awesome. It reminds me of my first "successful" AT-AT build. It was pretty shuffly.

[BrickBear]

21 minutes ago, Nelson said:

Ha! That's awesome. It reminds me of my first "successful" AT-AT build. It was pretty shuffly.

I remember mine was taking my technic arctic truck and making that walk with the mechanism from the official lego motorised AT-AT. It kinda worked but it wasn’t very attractive and I don’t think my m-motor was torquey enough to move it. Hopefully an xl will be enough for the current iteration.

[Nelson]

This is my first attempt at embedding Flickr photos, so we'll see how it goes. I've spent several days experimenting with variations on my design, but I keep coming back to the one in the previous video. I honestly don't know why that version "worked" and my new attempts using the same concept do not. I've been using a more stable cam, albeit smaller. The one on the left in the photo is the previous version. It frequently came apart and used a tread from the 1980s. Perhaps it was the give in the rubber that made it all work. I don't know. My latest incarnation seems close. Unfortunately, I had to order a couple more 5x7 liftarm parts, so I really can't test it properly. I'm concerned about weight. The legs and chassis are already .82 kilograms, and the even though this is the most solid chassis I've ever built, it still flexes under the weight and torque and depletes the lift of the legs significantly. It will probably weigh twice that once all the cladding is installed.

old_and_new_cams by Nelson murray, on Flickr

latest_failure by Nelson murray, on Flickr

[Nelson]

I ordered a couple large Powered Up motors and a controller/battery pack yesterday. Has anyone tried using their programmability to help conquer some of the challenges of the one-leg-at-a-time AT-AT?

I'm starting to think I need to come at this with a totally new approach. I've come so close with my cam designs, even having limited success, but I'm hitting a wall on how much weight and torque that design can handle.

[BrickBear]

16 hours ago, Nelson said:

I ordered a couple large Powered Up motors and a controller/battery pack yesterday. Has anyone tried using their programmability to help conquer some of the challenges of the one-leg-at-a-time AT-AT?

I'm starting to think I need to come at this with a totally new approach. I've come so close with my cam designs, even having limited success, but I'm hitting a wall on how much weight and torque that design can handle.

I’ve seen a video where someone tried it with ev3 (or whatever the latest incarnation of mindstorms was) but I can’t find that video anymore and it walked in the same way as the official motorised AT-AT’s anyway. I think it could be done though. If you used smaller motors actually in the legs to bend them at the right timing which takes you to 4 connections (I think the limit is five or six per hub). Then you’d only need something to lift them up like a cam and move them forward and back like a piston which could probably be done with one motor.

Maybe…I’m a powerfunctions guy myself and only ever made one walker with ev3

[Nelson]

I think I’ve rung the final bell on this walking concept. It works at some level on lighter models, but just can’t handle the weight and torque at this scale. I’m pretty bummed because I’ve invested literally years of time in the design and it theoretically addresses all the engineering hurdles the challenge of a one-leg-at-a-time walker presents. It would likely work if I could use other materials than Lego.

Back to the drawing board.

 

[BrickBear]

11 hours ago, Nelson said:

I think I’ve rung the final bell on this walking concept. It works at some level on lighter models, but just can’t handle the weight and torque at this scale. I’m pretty bummed because I’ve invested literally years of time in the design and it theoretically addresses all the engineering hurdles the challenge of a one-leg-at-a-time walker presents. It would likely work if I could use other materials than Lego.

Back to the drawing board.

 

It might not have reached the state you wanted it to but it’s still very much alive and exciting to look at. Every attempt is also a page in our book of experiments.

[Nelson]

I found this experiment that I came up with a couple years back. 

 

[oracid]

Can you tell me what its weight and the dimensions between wheel axes (length and width) ?

[Mikdun]

On 11/12/2024 at 3:28 AM, Nelson said:

It works at some level on lighter models, but just can’t handle the weight and torque at this scale.

It seems to me like either there is high resistance somewhere in the mechanism, or the motors are too weak to move it.

[Nelson]

There is a difficult balance when trying to make a model like this. I could certainly make a lighter model that walks one leg at a time and looks vaguely like an AT-AT. Or I could make a very screen accurate model that has difficulty walking at all. If you haven't tried to walk that line, give it a go and see where you end up on that spectrum. It's a challenge you won't soon forget. @Oracid, I can only estimate what the final weight might be. The chassis and legs alone are 928 grams. I'd say the full model will end up somewhere close to 1.25-1.5kg. I've posted some photos where you can see the model with some studded beams in front to show length and width. I had a look at your walking models. Do any walk one leg at a time? @Mikdun, yes, the design is susceptible to binding up due to the weight resistance. It's not so much the power. The two large Power Functions motors have more than enough torque to snap the driveline axles. The parts themselves just don't cooperate. I can't get the perfect geometry for the various cams and levers and have to settle for something close. Ultimately, I don't think this current concept will work for the detail and weight of the model I hope to make. I've posted a photo of what I'd like the final AT-AT to look like.

at-at-length by Nelson murray, on Flickr

at-at-width by Nelson murray, on Flickr

at_at_assembly_desktop_2 by Nelson murray, on Flickr

 

Edited November 13, 2024 by Nelson

[LegoTT]

On 10/30/2024 at 10:12 PM, Nelson said:

I've seen those walking designs based on Jansen's linkage. If you look closely, you'll see some gears in there that are only a quarter of 360 degrees. Unfortunately, there is no Lego analogue of this.

https://www.bricklink.com/v2/catalog/catalogitem.page?P=41667&name=Technic, Arm 2 x 5 with 1/4 Gear 8 Tooth Double Bevel&category=[Technic, Gear]#T=P

[Nelson]

10 hours ago, LegoTT said:

https://www.bricklink.com/v2/catalog/catalogitem.page?P=41667&name=Technic, Arm 2 x 5 with 1/4 Gear 8 Tooth Double Bevel&category=[Technic, Gear]#T=P

There's no real way to use this part that I've found. Correct me if you've found a way. The gear's centerline is a pin hole, so there's no easy way to drive/power it, and the appendage opposite the gear teeth hits anything the gear is meant to engage during the second half of its rotation.

Edited November 14, 2024 by Nelson

[BrickBear]

This one would be pretty damn close if it weren’t for the awkward bar

https://www.bricklink.com/v2/catalog/catalogitem.page?P=44810&name=Bionicle Matoran Torso, Gear 9 Tooth with 3 Axle Holes and 2 Pin Holes&category=[BIONICLE]#T=C

edit: Now i look at it i’m not sure it’s a quarter?…

Edited November 13, 2024 by BrickBear

[Stereo]

2 hours ago, BrickBear said:

edit: Now i look at it i’m not sure it’s a quarter?…

It looks like a 28 tooth gear to me, so 8 would be 90 degrees apart and then it has a 9th one.