Mun Walking - an attempt to circumnavigate the Mun on foot.

[Fengist]

Mun Walking

an attempt to circumnavigate the Mun on foot.

So, if you read the title of this thread and your first thought was, "this guy has lost his mind," you'd be correct. I would like to point out though, when you have little to lose, you don't miss it nearly as much as those who have lots.

I do however intend to at least attempt a circumnavigation of the Mun on foot. And I use that word attempt very loosely. As one astute observer of my preview to this thread pointed out, what I am mostly likely creating is Krakenbait. There's a very good chance this mission will be a dismal failure. If it's not, it'll take me, at the very least, weeks if not months to complete. This will be somewhat of a part-time mission and I will work on it as I have time (and patience).

I also use that word "foot" rather loosely. If you are assuming I'm going to walk some Kerbal EVA around the Mun, well, I haven't lost THAT much of my mind. My first thread in the mission reports forum is my ground-based circumnavigation of Eve. Usually though, you find me in the Challenges forum. Those of you reading this who have seen some of my creations in that forum may recall that I build walkers. For those of you who haven't seen my creations, before you immediately think, 'Oh great, another guy who thinks he can build walkers and ends up with a clunky tin can moving at 1 m/s," allow me to demonstrate.

This is Octo:

Ok, so it's more like 10 m/s rather than 1 m/s. I know. It's not blazing fast. But, Octo wasn't designed to be fast. It was designed to be rugged, durable and sturdy enough to traverse the Mun. Those of you who have been to the Mun know exactly what that terrain looks like. And, if you're like me, you've probably flipped more than one rover on it's roof. And you also know exactly what speed on the Mun does to rovers... it turns them into low flying space craft. Octo is designed to climb up, over and down the sides of craters, and stay on it's feet doing so.

Now that you know what Octo is and what it's intended to do, let me describe this mission. It's probably going to be unlike most other mission reports you read here. While there is a crew of 5 on board with supplies to make repairs if needed and they will be traversing the surface of the Mun as this thread progresses, the primary goal of this mission won't be a circumnavigation. It's my hope that this mission will pique your curiosity in walker design and encourage you to build your own. I've been designing and building walkers for months now and I can tell you honestly that I've only begun to scratch the surface of what's possible. By passing this knowledge along in the form of a mission, I'm hoping some of you will take up the challenge of creating you own and teach me things I haven't thought of.

Even though I have gotten... decent... in walker design, I'm far from a master. I can also tell you that Octo is far from perfect. I'm already aware of some flaws in the design that have and will cause problems. But, I don't think they're anything I can't work around. As long as I don't end up with yet another corrupt save from wandering parts, Octo should survive this. And, if it doesn't, since the Mun is only a hop, skip and jump away from Kerbin, repair parts and a replacement is close at hand.

So the basic format will be first, a mission report of our progress (and most likely failures) followed by a specific topic in walker design. It's my hope that you guys will also fire off as many questions as you have about walkers and I'll try to answer them as best as I can.

On with the mission -

Pre-Mission

In case you hadn't noticed, my walkers are not small. Getting Octo off the launch pad was a challenge in it's own right. First of all, it's heavy on the nose. Secondly, it has no self-righting mechanism which means, it had to be launched upside down so that it could land right side up. Now, imagine strapping a 38 ton dead bug to the top of a rocket with all 8 legs waving in the air... and launching it. I believe the part count was 540+ with a lot of that being struts. Thirdly, Octo is not short. In order to make a proper lander for it, I had to use 12 FL-T800 Fuel Tanks to lower the thrust below Octo's center of mass. By the time I was done building the lander, I had over 100 tons of dead bug and fuel. Landing 100 tons with 4 nukes was asking a bit much from the nukes so I ended up tweaking out most of the fuel.

The launch vehicle and orbiter was the same one I used to get Walkabout to Eve in my previous mission report. Though the screen shot below doesn't show them, I strapped on 18 srb's just to get the thing off the ground. This does show how to get a 38 ton dead bug off the launch pad though.

Once at the Mun I had tons of fuel left. But, I also had tons of mass. Since the lander only had 4 nukes, I used the 6 engine orbiter to help slow things down. I've pointed out a 10 ton counterweight I used to help move Octo's center of mass toward the center of the lander. The plan is, just before touchdown I would stage and separate the counterweight and simultaneously decouple the docking port on top of Octo from the lander. Octo would land gently on the Mun, I'd leave the lander thrust at whatever setting it was at touchdown and with any luck, the counterweight would be thrown clear of Octo and the lander, with it's center of mass restored, would fly off and land... somewhere preferably not on top of Octo.

Just before touchdown. Well, things didn't go exactly as planned, but, as I'm sure you've guessed, Octo survived. I'll save the details for the next report.

---

A brief history of KSP walker design and how my designs came about.

Walker design started with one mod, Damned Robotics (now Infernal Robotics). This one mod gave KSP enthusiasts the ability to make things move. At pretty much the same time, walkers began to appear.

Most of them used counter rotating IR rotors in order to achieve movement.

Other designers took the route of pure hinges while others

. Even

was bitten by the walker bug. One look at this challenge and you'll see a whole myriad of contraptions built to resemble walkers.

When I started to get into walker design, I studied them all. The problems I kept seeing were, either the walkers were lethargic and hard to control or they required you to repeatedly jam buttons on the IR control panel. It seemed to me that there had to be a better way.

It wasn't till I saw

by Parallax that I realized, he had cracked one of the secrets to walker design, leg speed. If, you could find a way to rapidly lift and drop the walker's leg and by using lots of reaction wheels, you could keep it upright and stable and, move it forward. Parallax is still designing walkers and many of them are amazing. However, after studying his designs, I realized there was a key piece to his designs that limited their widespread usage. All of his designs require a script. Parallax uses Autohotkey to essentially program his walker's leg actions. While it obviously works and in a stunningly stable way, and while I do write code as a hobby and had no fear of learning Autohotkey, I made the assumption that most people didn't want to get into a scripting language to make a walker... walk.

And that's when I hit on the idea of cam operated walkers. The cam is an ancient mechanical device. If I could replicate one in KSP, there was a good chance I could convert rotary motion into linear motion. The cam in KSP is essentially an extension of the counter rotating rotors but with more lift and finite control. Plus, it wouldn't be attached directly to the walker leg. I could link that cam to the legs via struts and get linear motion. In order to achieve fluid movement, I would use IR's Free Docking Washers for the joints. It was immediately clear to me that if I could pull it off, I could create a 'fire and forget' walker. You turn it on, it walks. No buttons to jam to keep it walking, no scripts and no slow clunky movements. It would emulate actual walking with the cam and strut acting as a muscle. It took me several weeks of tinkering but I finally achieved that goal with a 12 legged version that looks a lot like Octo (link to the video below). The next logical step was to go biped. Roughly 2 weeks later, I created my first biped, Mun Runner (link to video below).

Since then I've tried to keep apprised of any new design theories in the field of walker design and, not to be egotistical, so far, the only two viable designs I've seen are the scripted walkers by Parallax and my cam walker designs. Perhaps, with the encouragement of this mission, some of you will yet surpass both of these design theories and force those of us in the field to re-think our own designs.

Links

Biped Cam-Operated Walker Tutorial w/Advanced Techniques [stock + Infernal Robotics]

Parallax's Youtube videos (not all are walkers)

---

Challenge

My challenge to YOU to circumnavigate an orbital body.

---

Parts Link

90 Degree 1.25

Right click and save the file. Unzip it to your GameData directory and restart KSP. It's just one part and nothing fancy but you should find it in your structure parts.

---

Mission Reports

Day 0 : Octo II: The saga begins again.

Day 1 - 100km... sorta.

Octo Tutorial Links

A brief history of KSP walker design - This post

Walker Nomenclature

Everything you never needed to know about cams

Edited March 26, 2015 by Fengist

[The Error]

Mun circumnavigqtion by walker?

..Engineer Bill won't be bored this run, I bet. (Unless it's unkerballed... but that's a tad boring, no?)

[Shania_L]

Wow,

Having read your Eve circumnavigation I am really looking forward to this, add to that instructions for my very own walker ... count me coloured impressed.

(P.S, your video is set to private so it cant be viewed)

[Commander Zoom]

When you announced you'd be walking around the Mun next, I honestly thought you were joking. I should know better by now.

[Fengist]

Mun circumnavigqtion by walker?

..Engineer Bill won't be bored this run, I bet. (Unless it's unkerballed... but that's a tad boring, no?)

It has 5 little green ones. But an entirely new crew.

Wow,

Having read your Eve circumnavigation I am really looking forward to this, add to that instructions for my very own walker ... count me coloured impressed.

(P.S, your video is set to private so it cant be viewed)

ACK! Thanks. Published now!

When you announced you'd be walking around the Mun next, I honestly thought you were joking. I should know better by now.

Not to worry. Even my wife hasn't figured me out entirely. But you'll catch on.

Edited February 3, 2015 by Fengist

[Wabbit]

Ys! I guessed it right! Hahahahahahahahahahahaha!

[Space Peanut]

looking at that lift-off... thank god the aerodynamic update has not been done yet!!!

good luck with this new project

[Zuqq]

The Walkabout is one of my absolute favorite missions and I really look forward to seeing this one unfold!

[lajoswinkler]

This is insanity and I applaud it.

[Fengist]

Well, I'm glad you guys are excited about this one but I'm having to put it on hold and possibly cancel it entirely, through no fault of Octo. I seem to have found what caused the demise of Walkabout on Eve... saving. Already I have rotors on top of Octo getting badly out of part alignment and after spending a good bit of time experimenting, I know why.

We all know parts on our vessels tend to move around a bit. Mostly, this is caused by warp physics. We've all seen the flexing of parts. Octo can't warp. There's just too much going on for it to hold together as it warps. But, because these are rotating and moving parts, many of those parts have some extreme stresses on them. It appears that whenever you save the game, it doesn't save the vessel with the parts located where they were designed and should be, but rather where they are at the instant of the save. And, when you reload that save, the parts don't typically 'snap' back into their designed place or realign to the part creator's connection nodes. Wherever they are at the save are where they'll be when you reload. This... sucks.

It means, after multiple saves, if your parts are not perfectly aligned each time you save, they'll keep getting further and further out of alignment. The bad news for this mission is, trying to complete it in one setting simply isn't possible. The circumference of the Mun is on the order of 1.2 million meters. With Octo limited to 10 m/s max speed, I'm looking at 50+ hours to complete it. My intent was to break it down into 12 missions of 100 km each. And because Octo is no race car, it's taking me multiple saves just to make 50 km.

Here's an example of what I'm referring to. About half of the rotors on Octo have gotten badly out of alignment through saving while some of the legs were off the ground. This caused some of the rotors to receive all the weight of Octo and thus, push them upward. Once they're pushed up, there's no way to pull them back down.

I can't really claim that this is a design flaw as Octo does work and work well. It would appear that any long term mission that's subject to physical forces and moves is doomed to have mis-located parts due to game save mechanics. I already know that this is going to create corrupt saves with parts inside of parts or result in some bastardized Frankenstein of what I originally designed and I really have no desire to spend hours editing save files each time I want to log in.

Unfortunately, KSP is intentionally designed to have things go wrong. Whether this is part of that intentional mechanic, I can't say. But, it pretty much scrubs this mission.

When you push the envelope, you often end up with a busted envelope.

Edited February 9, 2015 by Fengist

[jake9039]

make an envelope save. I recomended this to some other people, but just make a hinge on top that has a hangar(that is the mod name) on top, that detaches, puts the ship in, save, walk, rinse, walk, repeat.

[Fengist]

Octo II: The saga begins again.

Ok, I've got lots of catching up to do so here's your infamous WOT. Dig out the bifocals.

Ok, it's been a couple months since I had to put this mission on hold due to misaligned parts. After much study, testing, contemplation and frustration, I think I now know why it happened on Eve with my circumnavigation and why it happened on the Mun with Octo: quicksaving.

Yes, as strange as that initially sounds, I'm pretty certain that's the cause. You see, I assumed that when you did a quicksave that everything was just dandy. And for most spacecraft that don't endure the effects of gravity, that would be a good bet. Ground based vehicles however, I'm fairly certain it's a bad idea to quicksave when they're moving. My Eve rover 'Walkabout' suffered some pretty badly misaligned parts. I had a habit of hitting the F5 button while still roaming around. I believe the first time I quicksaved while moving and then reloaded with F9, the parts went out of alignment. Every time I quicksaved while moving and reloaded again, the misalignment just got worse. I doomed the vessel to become a Frankenstein of misaligned parts by thinking I was protecting my investment.

Octo is MUCH more complex that Walkabout was. There are lots more stresses on it's parts than just gravity. Because I was quicksaving while walking with Octo, and because of the added part stresses and parts being misaligned while walking, it failed a lot sooner than Walkabout did. So, if I'm correct in this guess, it's not a design flaw in Octo. Rather, it's a habit flaw.

Armed with this new information I have developed a plan to test this theory and hopefully, restart, and complete, this mission.

The new quicksave plan.

1. Stop all motion.
   
2. Retract all landing gear (toes) so there's no gravitational stresses on the legs and struts.
   
3. Deploy the LLL 'Skids' I mounted under Octo (originally, the idea was if Octo had problems with craters I could 'slide' down them. Fortunately, it doesn't).
   
4. Center all rotors and rails.
   
5. And finally, quicksave.
   
6. 
   

And so far, after testing several saves and reloads, parts are staying where they should be.

So with that in mind, I announce the Second attempt at:

Mun Walking with Octo II.

Since there were zero design changes to Octo (I did add 4 more nuke engines to the lander module because it took forever to slow it down with just 4), I'll spare you all the details up till the landing as they were pretty much the same as the first attempt. With those 4 extra nukes, this landing when a lot smoother and took a lot less time than the first landing. And, this time, the counterweight did exactly what I hoped it would, ejected out the back. Because Octo's center of mass is different from it's physical center, I used a 10 ton counterweight to bring the lander+Octo's center of mass in line with the center of thrust. Once landed, the counterweight was discarded so that I could power up the lander and decouple it from Octo. When this technique works correctly, your payload lands safely and the lander vehicle flies off... hopefully... to a safe place to crash. This lander, the last I saw it, was heading for orbit.

With all the excess discarded and with high hopes that we were finally off on this mission, Pomeroy drew the short straw and got to plant the flag. As you can tell from the description of the landing spot in the screenshot, Pommy isn't the most articulate Kerbal. But the flag got planted and we began heading east across, according the the map, Farside Crater. You would think that would imply there's a Nearside Crater, but we haven't found it yet.

Anyway, we've made it 30km from our starting point and we're still walking with no issues. We did have to stop once to reattach our 'rover' because we forgot about it and it was flopping around a bit. I know, but what good would a walker be if it didn't have it's own rover? Pfft. I'll describe it and how it attaches in a future post.

100 tons of dead bug landing on the Mun.

Touchdown. I took a chance here and used the engines for landing gear. Since I had pretty good control, I hit the surface at 1m/s and nothing broke. That junk out behind Octo is the 10 ton counterweight.

With the engines throttled up to 30pct and Octo decoupled from the lander, the lander becomes an unguided missile. Fortunately, it flew straight and headed for orbit.

Pomeroy displaying his literary genius while planting our Day 0 flag marking the landing spot.

An orbital view of our landing spot on Farside Crater. From here, we head East.

Hill Climbing - Octo seems to navigate hills with ease. The only place were we have to slow down and take it easy is on the steep edges of craters, both going down and coming up. Octo's speed is controlled by adjusting how far the rails are extended.

A quick shot showing the LLL Skids mounted under Octo. Originally, they had a dual purpose, to allow Octo to slide down hills if needed and to lift it to the proper height for when the rover needs to be deployed. Now, they're a quicksave safety mechanism.

---

Octo Nomenclature.

I promised to provide some guidance on building walkers and here's the next installment. While everything here may at first appear blatantly obvious, I want to make certain that when I get into more advance tutorials that you know what I'm talking about. So that there's no doubt as to which part I'm referring to, I'll point out the major ones and give a brief description of them.

Feet - This may sound a bit odd, but Octo's feet are the most critical part of the entire vessel. On any walker design, the feet determine whether you propel your walker forward or if you come to a screeching halt and fall flat on your face, or just as bad, go nowhere. I've spent, God knows, how many hours trying to perfect walker feet. And I'm still not satisfied. About the only option for creating feet are to use some sort of landing strut. I've seen other parts used, and I even used RBI tracks on my Mun Runner to go roller skating... which ended in disaster, but nothing yet matches the balance of landing struts. Feet need brakes. I know this sounds odd, but most parts in KSP don't have brakes. If you put a fuel tank on a slope, even full of fuel, it'll slide down that slope. If you add landing struts or wheels with brakes, it won't. Any part that doesn't have brakes, slides. Want proof? Here's the master... or disaster... of KSP design (no offense Scott), Scott Manley and his 'Munwalker'

The big thing you'll notice is that he's using girder's for his feet. If you pay close attention, you'll also see those girders slide across the surface of the Mun rather than propel the walker forward. Why? Brakes. Girders don't have them. Landing struts do. And that's why Octo has 4 LT-2 Landing Struts on each foot as it's toes. There's also a LT-1 Landing Strut attached to each of the outer feet pointed outboard. Those add a bit of stability when turning.

Hips and Knees - While these are both pretty basic in their design, a single free docking washer with some 90 degree parts, you'll find they're not that easy to create. Reason being, very few mods make a 90 degree part. And, for making a walker, it's a lot simpler if you have one 90 degree part rather than making one out of several parts. That's why you'll see something rather unique on Octo, a part I designed and created myself. The knee uses 2 of these and the hip uses one. For those interested, as soon as I get a few moments between posts, I'll set up a zip file where you can download this part and link it in the OP. While you can make a walker without them, they're damned handy.

If you'd like to learn how to make your own using a really simple 3D modelling tool, I recommend this tutorial. It taught me what I needed to know to make that part using Wings 3D, which is a LOT easier to use, but a lot less powerful, than Blender.

The Cam - This is essentially the heartbeat of any walker using my design techniques. Without it, you're destined to constantly press buttons to open and close hinges or slide pistons. Or, if you prefer the designs by Parallax, scripting in AutoHotkey. This simple cam design allows you to turn your walker on and watch it walk. As a matter of fact, the whole time I've been writing this post, Octo has been strolling across the Mun on it's own. The theory is simple. To take a rotary motion and turn it into a linear motion, to go from round and round to up and down. The rotary motion is created by an Infernal Robotics (IR) Rotatron. Attached to that is an IR gantry rail. On the gantry slide I have two FL-A5 Adapter's with an IR Free Docking Washer sandwiched between them. The outer FL-A5 Adapter has a strut connected to the knee of Octo. When the Rotatron is turned on, the entire process begins to spin. But, because the outer FL-A5 Adapter is attached to the Docking Washer, it doesn't spin. When the gantry rails are slid, the rotary motion turns into a linear motion which in turn, moves the knee up and down. The more the rails are slid, the more the knee moves up and down and the more forward motion Octo achieves.

So now, you have a very basic idea of the major parts used to build a walker. In future posts, I'm going to go much more into detail on each of these specific features.

Edited March 25, 2015 by Fengist

[The Error]

So it begins...

Good luck with this! But without SuperEngineer Bob, I fear to see how parts will break.

[Fengist]

So it begins...

Good luck with this! But without SuperEngineer Bob, I fear to see how parts will break.

I lol'd

[Fengist]

The link to the 90 degree mod I made for Octo's joints is now listed in the OP.

[Elowiny]

Yay ! Another circumnavigation attempt made by you ! I saw your Eve trip, and I loved it !

Good luck !

[Fengist]

Yay ! Another circumnavigation attempt made by you ! I saw your Eve trip, and I loved it !

Good luck !

Glad to have you back aboard! Hope you enjoy this one as well. If you don't, I've linked in the OP to my challenge wereby YOU circumnavigate something and tell us about it.

One entrant, Slam_Jones, is off and running on his attempt to circumnavigate Kerbin.

[Specialist290]

Well, this certainly looks ambitious. I'll be keeping an eye on your progress.

[Fengist]

Well, this certainly looks ambitious. I'll be keeping an eye on your progress.

Men are basically smart or dumb and lazy or ambitious. The dumb and ambitious ones are dangerous and I get rid of them. The dumb and lazy ones I give mundane duties. The smart ambitious ones I put on my staff. The smart and lazy ones I make my commanders. - Erwin Rommel

I think I'm in that first category.

[Fengist]

Day 1 - 100km

Ok, so we didn't quite make it 100km in one day but... we're 100km from where we started. This is sorta in the middle of day 1. Our first day on the Mun was spent dodging craters. If you haven't been to the Mun in a while, the craters are still there. Ludfal, our braniac scientist claims it's because there's no geologic activity on the Mun. Pommy asked when was the last time Ludfal saw a volcano on Kerbin. So, that argument ended rather quickly. While still stomping around in Farside Crater we quickly remembered a design flaw that I sorta forgot to fix before I launched. I designed Octo with those silly square lights. Every Kerbal knows round is better. So, we couldn't see where we were going when it started to get dark. And, we ran into another minor problem. With the new method of quicksaving while lifting all the legs off the ground, the skids under Octo turned out to be a really bad idea. They... well... skid. And they don't stop skidding till you find a flat spot. Have you ever tried to find a really flat spot on the Mun? Mostly, we've had to turn sideways to the hills. While this works for the moment, there are going to be some spots where we're going to have to stop and turning sideways to a hill is going to be more dangerous than quicksaving while moving. So, the boys fired off a message to KSC and in no time, a care package arrived. The fellas back on Kerbin loaded up KAS container with four round headlights, a grappling hook and a ground pylon.

Chad, our rover driver, was all excited when the care package landed 100m away. That was all the excuse he need to dig the rover out and go for a spin. He even got mechanically creative and mounted the KAS box onto the back of the rover and drove it back. When he pointed out that carrying a grappling hook on their back was not something anyone else really wanted to do, everyone nodded in agreement and decided he wasn't playing, he was working. I don't think you could have wiped the grin off his face with a sledgehammer.

Anyway, Chad brought the goodies back to Octo and since he was already out there, he turned the jet pack on and installed 2 lights on the forward rotor pod and two more on the sides of the cab. We can see now. We've already got two KAS boxes on the roof so Chad mounted this one under the nose of Octo. He figured that would be the perfect place. The plan now is, if we're in a situation where we have to stop on a hill, we can attach the grappling hook to the winch and then attach it to the ground. Octo may still spin around a bit, but it won't slide down the hill. And, if that doesn't work, we'll try the ground pylon and some piping. And if all that fails, we'll just jam the rover into the side of Octo like a wedge. Chad doesn't like that plan though.

Since we were stopped, Chad decided to go ahead and mark the spot with a flag even though we were only 60km from where we landed. We'd hoped to make 100 before planting a flag but, we took the first day pretty cautiously. We only had the rotors running at speed 6 so our ground speed stayed around 7-8 m/s. Not horribly fast but horribly safe. After that flag planting though, we decided to push the envelope a bit and wound Octo up to speed 8. Now that may not sound like a huge change but our ground speed jumped to 12-13 m/s. And that's about as fast as we want to go. Even that speed is a bit iffy. Even on some of the little hills, Octo leaves the ground for a bit.

We finally crawled up and out of Farside Crater and began our trek across the highlands. There's a few anomalies on our path and we'll probably stop to check them out, but they're several days away near the North West Crater.

So far, everything it working pretty well. Our part alignment while moving is nothing unexpected and the new method of saving seems to be working. Nothing is where it shouldn't be.

Chad playing with his toys... erm rover.

The care package from KSC. It's the KAS box under that lander.

Chad back in the rover with the care package on the back.

Chad on the roof installing Octo's new lights.

And the mounting spot for the new KAS box. Here's a good shot of the rover. Note the Clamp Jr. on the top of it. There's one on the rear as well.

This is a look at the rear of Octo with the rover door lowered. You'll see the two clamp Jr's inside... along with mounting for the legs. Building the rover as an assembly and trying to attach it to the roof of Octo proved... impossible. We wanted to dock the top of the rover to the roof because, if you dock to the end, while Octo is walking, the other end bounces up and down. So, we built it with the docking port on the end. Once we landed, we could then undock it from the end port, and as luck would have it, it was the perfect size to dock to the roof. Now when it bounces, it doesn't go very far.

Chad giving you some perspective of Octo's size.

Oh damn. Just to the up and left of Octo... a Shiney Spot. We're doomed.

Crawling out of Farside Crater

Octo airborne in mid-stride.

Oh damn, another shiney spot right under the cab. This is Octo in 'sleep mode' with all of the feet retracted and the skids deployed. The rotors and rails are also 'zeroed out.'

Kerbinrise on the Mun

Lodtrey, our pilot. He's still smiling because we haven't gone that far yet. Give it a few days.

---

In my next post. Everything you never needed to know about cams and rotors...

Edited March 26, 2015 by Fengist

[Kieve]

Oh no, not the shiny spots! Brace yourselves, the Kraken is nigh!

Just one question - is tucking the KAS box under Octo's chin wise? Seems like it might risk getting bumped off a crater lip or something, and then you'd be out a pylon / grappling hook.

Paying close attention to that upcoming Cams & Rotors post too. The more I learn, the less explodes on me!

[Fengist]

Oh no, not the shiny spots! Brace yourselves, the Kraken is nigh! http://i50.photobucket.com/albums/f321/Kieve/PWI/tigers/horror.gif

Just one question - is tucking the KAS box under Octo's chin wise? Seems like it might risk getting bumped off a crater lip or something, and then you'd be out a pylon / grappling hook.

Paying close attention to that upcoming Cams & Rotors post too. The more I learn, the less explodes on me!

I see what you did there. I'm taking a walker around the moon and you're asking if I'm wise. I love rhetorical questions.

Actually, the highest lift points on Octo are the front and back, that's where the feet meet the ground. Plus, Octo has another lifting capability. All of the legs are IR pistons. While the one between the knee and the hip isn't used as it tends to bend things, it's there to match the size of the one between the foot and knee. By extending those, Octo gets some really good ground clearance (and a lot higher COG). And, when crossing the pointy lips of craters, that's a good thing. The nose and tail pretty much never get close to the ground until I lift the landing gear. It... should... be ok there. The other point to those pistons is a longer stride and thus, more speed. I was just goofing off with them while running Octo's rotors at 8 and by extending the legs, I managed 18 m/s. Not too shabby for an 8 legged walker. But, there's also a lot of sideways twisting motion in the legs and I didn't design Octo with a lot of clearance between the feet. They almost touch when passing each other. At those speeds with the legs extended, one wrong turn or badly planted foot and they'll get tangled up pretty quick.

If you watch the video in the OP again, most every shot is with the legs extended. The shots where I'm reversing and neutral steering, and the very last shot, are with the legs retracted. Since that video, I've become more comfortable with increased rotor speed rather than increased leg length. Just looks and feels safer.

[Elowiny]

Do you use TAC ?

Also, what mod do you use to get the IVA seen in the last picture.

There is shiney spots on the Mun too ? YOU ARE DOOOOOOOOOMED ! The underground monsters will destroy your ship easily in low gravity ! Be careful !

[Fengist]

Everything you never needed to know about Cams

On our planet, the cam has been in constant use for over 5,000 years. For Kerbals, this is a striking new invention. Capable of converting rotary motion to linear motion, those of us living in both worlds are only beginning to understand it's implications for Kerbals. In this tutorial, I'm going to delve into the cam and how it works and how to build one in KSP.

Here's an animation from Wikipedia that I like to use to demonstrate just how a cam should work.

The cam is basically a wheel with a bump on it. As it goes around, that 'bump' is used to push and pull on what's known as the cam follower. In the image above, the yellow part is the 'follower.' In order to understand cam's a bit more in detail, you'll need to understand a few of the parts of the cam. Here's an image I created with the various sections named.

The lobe of the cam is the part that does all of the work. It determines how far the follower is pushed and for how long it stays pushed. The lift is the measure of how far in distance, and the duration is the measure of how long in degrees. The heel determines the position of the follower when it's at rest and no work is being done. There are lots of different shapes of cams to perform different kinds of work. A spiral cam will slowly build up lift and suddenly drop to no lift. An 'S' cam is shaped like the letter S and has two lobes. It will slowly lift twice per rotation and rapidly drop off.

The cams I use are essentially double lobed or, offset like a crankshaft. They both push and pull. In the image below the crankshaft is red.

For the purpose of walker design, I'm going to show you how to build this basic cam.

There's lots of different way to build a cam in KSP. If you look my OP , the first two links in the list, you'll see my very early attempts at making a cam. They were basically adapters from the B9 mod with docking washers and another adapter on top of the washer in a fixed position. Whenever I turned the rotors on, the cams began spinning and producing work. While this cam did it's job, it had one major flaw, it wasn't adjustable. That meant that whenever I turned the rotors on, it was at full speed. It also meant that stopping was a chore. Rather than gradually slowing down those walkers came to an abrupt stop and I had to use a hopping method to completely stop moving. If I needed to adjust the lift of the cam, I had to go back to the SPH and move the washer location. This became a chore so I began looking for a way to make an adjustable cam. Below is what I came up with.

Here's an assembled image of a cam I use on most of my walkers.

And here's an exploded view with the parts labled.

The basic theory is, when the gantry rail is at the center position, there is no lift and no duration. Even with the rotor spinning, it produces no work. As the rail is moved outward from center, the lift increases, thus producing work. While neither of the FL-A5 adapters are truly necessary, they allow me to visually determine how much lift my cam is producing. With traditional cam's the lobe pushes against the follower. While that's doable in KSP, I've found it more beneficial to physically attach the cam to whatever I want to push against. The docking washer allows the outer adapter to spin freely so that as the cam rotates, the outer adapter can always remain oriented toward the follower and push or pull against it. Without the washer, whatever you attached to the cam would flail around in a big circle. For my walkers, I use a basic strut connector. If you want to have a bit more visual appeal, B9 has invisible struts.

So now, you should have a pretty good idea of how the cam should work. In building a walker you're going to need to manage a few things like stopping and starting. Here's the basic process I use.

All of my rotors are in action groups and I toggle them on and off. All of the rails are in a single action group to start and stop them sliding. Depending on when I want the lift to occur, I may rotate certain rails 180 degrees from others. If you look at Octo's video, you'll see that all of the cams are rotating in the same direction when moving forward, but half are lifting while the other half are pulling. This allows me to stagger the leg movements. You'll also want to play with action groups that cause your cams to spin in different directions. Reversing the direction of all of my cams allows Octo to walk backward. Reversing the cams on one side or the other allows Octo to neutral steer.

By playing around with rotor direction and rail orientation, you can have your walker doing any number of motions from running, spinning in circles, and a gentle stroll to hopping like a frog.

The first thing I do is start all of the rotors spinning. If the rail is centered or zeroed out, there's no lift and I have no motion. As I hit the action key to slide the rails, I start to develop lift. Until you learn how your walker reacts, do this slowly. As I increase the rail slide distance, the legs get more lift and more travel distance and Octo picks up speed. Stopping is just the reverse. I gradually click the '0' button for the rails on the IR control panel. As they start moving toward the center, there's less lift and thus less leg travel distance. Once I'm at a slow enough speed, I center the rails. The rotor speeds never change. With the rails centered, all motion stops and then... I can stop the rotors.

Cam direction, cam speed and rail slide direction are going to be very dependent on how you design your legs. Every single walker I've ever designed has it's own personality and prefers these settings a certain way. In my tutorial, the stock walker rotor speed was 15. If I set Octo's speed to 15, the legs would fly apart. It prefers speeds between 6 and 8. The only way to really determine the 'perfect' speed and settings for you cams is experimentation.

Octo uses a single cam per leg. Other walkers I've designed use 2 cams per leg. If you look at the video of Mun Walker, you'll see that the cams are of different sizes, the rear one producing much more lift than the forward one. In that design, I wanted to lift the foot a lot higher than I did the knee so, it used two different cam sizes. As you learn where to place your cams and how to attach them, you may discover you can get by with 1 cam for 2 legs or you may need more.

And one final note on cam's. If you physics warp, expect bad things to happen. With most craft, the physical forces on them are relatively small. With cams running, you're constantly yanking and pushing on parts. Pieces will wobble around. Engaging physics warp seriously magnifies these forces.

Now, here's a thought experiment for you. Octo uses a single strut per cam to lift the legs. It's duration is essentially zero degrees. Suppose, I wanted to create more duration on my cams. How would I accomplish this?

Feel free to ask any questions you may have if I failed to describe this properly.

- - - Updated - - -

Do you use TAC ?

Also, what mod do you use to get the IVA seen in the last picture.

I do not use TAC. My machine is quite old, running Win 32 bit and only 4g of ram, so I have to be pretty selective in what mods I use. The IVA is from the LLL mod pack. It's the same pack I used for the basic design of Walkabout. I kinda like them because they're 2x1 rectangular rather than KSP round. They allow me a lower center of gravity on most of my craft.

There is shiney spots on the Mun too ? YOU ARE DOOOOOOOOOMED ! The underground monsters will destroy your ship easily in low gravity ! Be careful !

I thought exactly that when I saw the first one.

Edited March 26, 2015 by Fengist

[Fengist]

And, it has happened again. Part misalignment.

I'm now convinced that it's due to saving and I have no way to fix it. It's a choice the dev's made. Basically, if I save a game with Octo's feet on the ground, gravity pulls down on the center of the walker. This causes parts to become misaligned. If I save, then the restore is misaligned. If I save with the feet off the ground, then gravity pulls down on the feet, causing... parts to become misaligned. Once a part is misaligned, it just gets progressively worse with each save and restore.

In order to have perfect saves, either this game needs to be fixed so that part misalignment is no longer an issue or I need to find a way so that gravity has 0 effect on every single part of the craft, and I'm really not please with the idea of having to hack gravity every time I want to save my game.

This also means, that any other rover that doesn't have 'shocks' absorbing gravity when you save, will eventually suffer the same fate. Now you know why all 'wheels' have and/or need shocks.

This time, I made it roughly 200km before I started to notice rotors being misaligned. So, the design works, the game doesn't.

Sorry to lead you all on this goose chase but I'm throwing my hands up in utter disgust.

If you have any question on walker design, feel free to PM me... and I'll advise you against it until the game is fixed.

Edited March 31, 2015 by Fengist