Dave's Lego AT-AT

Building the AT-AT

The Feet

This was the biggest problem I had with the 2nd AT-AT. The feet were horridly inaccurate-- they looked like scrawny bird feet. But the problem was that doing an accurate foot meant a lot less stability. After all, the ankle works on a track, with the ankle rotating on the track for foward/backward rotation of the foot, and with the track bending for left/right rotation of the foot. To top it off, the track (and really the entire leg) is pretty darn thin. Hence, for the 1st and 2nd AT-AT, I had chickened out and just built my own foot design. But this time I was going to do it right, I had decided.

Before I even looked at scale, I started playing around with tracks, and seeing how the track could be made, what pieces worked best, etc... That's about when I made first contact with Shaun Sullivan. He suggested his own method: using 2 1x5x4 half arches in grey to make the arch. My first thought was that this would be too wide, but when I measured the actual thing, it turned out that being 12 studs wide was almost perfect, so using 2 half arches (10 wide total) wasn't even quite wide enough! I also rounded the top of the arches out with 1x2 and 1x3 grey slopes to make it a bit better looking, even though this made the track a bit too tall.

The other thing that frightened me was the diameter of the feet themselves. They needed to be about 12 studs wide and that seemed rather large. I ended up using 6x6 "wings" to make an octagonal foot, instead of actually making a stepped or round foot.

The Legs

For the most part, the legs were pretty straightforward. The only problem areas were the joints. I wanted it to be poseable ultimately, with bendable knees, and hips (and upper hips), but problems arose.

My first attempt was with the Star Wars style hinge pieces for the knees, since they could bend up to 90 degrees, and seemed fairly sturdy, especially if I used two of them. This was phenomenally easy, so I tried it. Fortunately, I hadn't dismantled my 2nd AT-AT's body, so I still had something to test the legs out on, that could pretty easily simulate the weight of the final body. The result? Not even close. Those knees simply wouldn't do.

So I played around with a couple different designs. The one that I finally came up with (and only implimented with one test leg) was something similar to the technique used to lock the arms of the large Spyrius robot (see Kevin Loch's instruction database for 6949). It wasn't actually freely able to rotate, but it locked the legs in place pretty well, and was capable of nearly completely concealing the locking mechanism (being that it was only 2 bricks thick) I repeated this design for the hips, and felt pretty confident that it would do the trick.

The failure of the locking system came when it was time to do the upper hips. The locking mechanism just didnt seem to work. I actually fixed a technic beam in place and attached the upper thigh segments to test if it was even possible. The result? Nope. The weight of the body would either severely twist or completely break a technic axle (actually 4 technic axles) when done "Correctly".

But why should this be true, you ask? Well, the problem is actually related to balancing the AT-AT. The upper-hip joint serves two purposes: to lift the leg higher (not really NECESSARY) and to help in balancing the body. Without them, the weight of the AT-AT rests on 4 points at the "corners" of the body (actually not quite). The problem with this is that lifting one leg up in the air to take a step results in the same problem you'd have in balancing a rectangular table on three legs-- it's very unsturdy. To help with this problem, the upper hip joints are added closer to the center of gravity, to act more like a pivot point. Also, it allows the legs to "extend" (kind of) to compensate for these problems, making walking a splendid display of complex motion. This allows the body to remain more or less stable, too, as opposed to tilting and swaying as much.

So I pretty much gave up completely on the locking mechanism. However, I'm not quite beaten yet... The other idea is to essentially have the AT-AT balance on the three corners (and have the upper hip joints essentially usesless), and lock them in place, while tilting the body away from the raised leg. Possible? Maybe. But I haven't tried it yet... so stay tuned...

Meanwhile, I kept the Star Wars style joints on the knees, and turned the other joints into more or less straightforward pin-style joints. The knees still lock, though, with a 2x2 brick placed on the top and bottom of each knee joint.

The Head

The head is probably the part I'm the most happy with-- mainly because I was able to somewhat imitate all the angles in the "real" thing. Hence, even the "eyes" bend inward and upward so you can see them when viewed from the front.

Also, the interior is pretty accurate... near the top of the door there is horizontal "padding/panelling" (see the movie), in addition to the placement of the terminals, and the miscelaneous junk and panelling here and there (as much as possible) around the interior of the head.

The cannons are actually pretty good as well (In my own humble opinion)... the chin cannons actually have the "notch" like the ones in the movie (despite the fact that they are a little narrow). I did this (of course) with a 1/2 technic peg placed in the "hole" end of your standard classic space antenna/pole. The eye cannons are also slightly undersized, but for the most part, the reflect the actual cannons used in the movie..

The Neck

The neck is again something that I'd ideally like to work out a little better. For the moment, the neck bends at two pivot points at each end. And like the legs, I tried technic locking mechanisms, but a single technic beam just won't cut it, due to the leverage of holding the very heavy head out at roughly 90 degrees.

Ideally, I wanted it to be as accurate as possible, and this means having each ring segment bend individually. I had an idea for this at one point which I may yet consider trying. Basically, the neck would bend in only 1 direction, bending at each ring segment. However, the neck would be free to rotate at both ends allowing the direction of bending to vary. This would also allow the head to be locked simply by gravity when positioning it to bend upwards or sideways.

Of course, seeing Shaun's solution to the problem has intreuged me-- he plaecd a series of technic beams in a latticework held together with friction pegs to help hold the massive weight of the head. I'm not sure if his exact design could support my AT-AT's head (mine is a bit bigger than his) but I am still curious to try it-- especially in conjunction with the individual ring segment bending idea.

The actual ring segments on the neck are passed down to me by Paul Baulch, who suggested the idea, I think exactly as I built it... although his method for making the neck bend was vastly different-- I'm not sure I understand it even now :) But in any case, the ring segments are his idea (Thanks Paul!)

The Body

The body was largely difficult to construct. Why, you ask? Well, the main reason was the sidewalls. They slope in multiple directions. This makes for a lot of annoyance. The 2nd (and 1st) AT-AT featured boring straight walls that were simply at 90 degree angles to the base of the AT-AT, which was pretty boring. I wanted to simulate all the angles that I could, and still have it look good.

I tried several methods, but finally I arrived at one that I felt worked best for tilting the walls inwards. Each wall was hinged on technic pegs, allowing it to slope further in. Originally, I had 5 or 6 peg connections (depending on the wall segment), but because of the incredible amount of time that took to both assemble and dissasemble, I removed most of the connections, and reduced it to two, one at each side of the base of the wall. While this meant that the walls were free to bend outwards at will, it made it easier on me to make modifications.

But the largest problem with having slanted walls was designing around them. Initially, when I put in troop benches with seat backings, they were JUST tall enough to be obstructing the inward bending walls. In other words, at their base, they were still 1 stud away from the wall, but at the top of the bench, they hit the wall, making the bench bend forward and in effect, break it. This meant that eventually, I went to boring style benches (I would have liked to have enough black seats (or grey or dark grey), but alas, I had nowhere near enough-- I only own about 2 to 4 of them...

This problem also came into play when designing the engine room and the speeder bike garage. I liked the placement of the garage in the Star Wars Technical Journal, since it appeared more like the speeder bikes could come out flying from a hatch in the roof, rather than in the Incredible Cross Sections book, in which I assume they are brought out of the side hatches. Upon thinking about it later, my design is actually still rather impractical due to human accessibility, but I'm not about to go modifying it THAT much again anytime soon...

At any rate, because of the wall sloping problem, I had to essentially restrict my speeder bike garage to fit within about 10 studs (that's about how wide the narrowest part of the opening is). The other problem with the garage is that the speeder bikes released by Lego are WAY to long. Hence, I re-designed the speeder bikes to be a lot more realistic size-wise. Instead of being 4 studs wide (about 4 and a half feet wide?!), I cut it down to 3, which is still too wide, but it's better... I also cut the length of the speeder bike from ? studs to ? studs to make it still just a bit too long, but considerably better scale-wise.

The "final" speeder bike garage has a carriage with 4 bikes inside that slides backwards and forwards, tilts upwards, and allows the bikes to "shoot out flying". Before this I had a crane arm that locked onto the speeder bikes and would actually raise the bikes out of the garage before releasing them. I'm still not content with the design (although it looks neat). I'd ideally like to have the bikes attach to the walls with space in the center of the bay from which the bikes are launched. But that's still to come.

Another problem I encountered was the escape hatches, of which there seemed to be 6 (3 on each side). Each of these seemed to be placed right in the way of where a troop bench should go, or other interior device, such as the speeder bike garage. In the end, I actually left out the middle escape hatch entirely (which I would like to fix assuming I can get the necessary pieces!) Otherwise, thanks to the fact that the troop benches have no backing, and the speeder bike garage assembly is in the center of the rear section, the hatches are currently accessible for minifigs fairly easily.

Next Page

Intro | History | AT-AT Specs | Building | Features | Pictures | Thanks | Lego Page