Archive: 19 posts

---

Instead of starting a new thread, I'm just gonna' change what this one is about. Originally I was asking for ideas on how to make my mech able to detect the whether or not it's on the ground and I settled on having an emitter spit out mag keys below the mech so that the ground blocks them. So now I'm wondering if anybody has any ideas on how to make a decent jump activated switch for a vehicle. First off, I'm using a control pod like the kind demonstrated in my vehicle tools level. I'm going to be setting up a grab switch which will use the grab funcion to fire a weapon, tap-grab to toggle between weapons, and double tap to to exit the mech. Because of how the control pod wheel works, there isn't room for more than one grab switch so I can't just have the player press a different direction on the stick and grab (besides, I hate that kind of control scheme), so the only option left is to have the player trigger the switch by jumping. So here's the rub: the mech's movement jostles the player around so that if I just left an open space above him and used a sensor switch, it would be going off all the time. Same goes for an overhead button. In fact, I need something to hold the player down onto the control wheel so he can keep traction with it and maintain control. I made a bit of sponge and hooked it to a medium strength piston and spring to push the player down. The idea is that the player could jump hard enough to push the sponge up, but the jostling motions of the mech wouldn't push him up hard enough to trigger it. My initial tests were very nearly satisfactory, but when I implemented the jumping mechanic and hooked the switch to it, I found that it's not nearly as reliable as I'd like. It often takes two or three tries to get the jump switch to register and that would be intolerable in an actiony level. So my question is does anybody have any ideas for a better jump switch, or maybe a better control pod entirely? I've considered putting the player inside the wheel instead of on top (hamster wheel), but I would run out of space very quickly and it would mean having the control pod only accessible via layer change (can't enter a circle from above or below) and there's no room to do that. Also, being inside a wheel would reduce the player's jumping ability, making it even harder to get a jump switch to register.

2010-03-30 00:13:00 Author: Sehven Posts: 2188

---

Hmm sounds like a tough order for a thermo friendly idea. What does your mech's feet look like? Does it have a vertex (I think thats what it is) or joint?

2010-03-30 03:00:00 Author: Emogotsaone Posts: 1030

---

I'm not sure what you mean by vertex. A vertex is a point where two lines meet to form an angle. The feet pivot on a low strength wobble bolt and are held in position by elastic so they can pivot with the contour of the ground, but go back to their neutral position. There are also "toes" on the front and back with compressible pistons (makes the feet much less likely to get caught on bumps, but it's also one more thing that can break). I think the solution I've got is pretty good. It happens to work out so that the ground sensor is triggered just before hitting the ground, which lets the ground based stabilizers kick in, which slows down the mech just enough to keep the feet from breaking on impact (also, since the stabilizers kick in, the mech catches its balance much more quickly on landing than it did before, which is part of why the feet don't break now). At least, I haven't been able to break them in any of my tests since putting in the new sensor, but it was kind of rare before that, so....

2010-03-30 14:10:00 Author: Sehven Posts: 2188

---

Yeah i meant pivot, not vertex. But i was thinking of putting a key and switch one either side of the pivot point so when it flexes its triggers the switch!

2010-03-30 23:47:00 Author: Emogotsaone Posts: 1030

---

may i ask when is the level with this in it coming out? i want to know

2010-03-31 08:58:00 Author: YEAH_NAH Posts: 775

---

It'll come out as soon as I can get it finished. Work is going slow. Everything I add requires lots of testing and tweaking to get it working just right. I was about to start working on the jumping mechanics when I discovered some pretty severe flaws in the walking ability (it couldn't handle going uphill in reverse) and spent about a week tweaking and tweaking and then finally ripping the legs out completely so I could build new feet (I couldn't be happier with the new feet, btw: they solved ALL the walking problems). I added jumping a few days ago and since then I've been tweaking the flight mechanics and controls to try to give the player the most consistent and best possible control in triggering jumps and controlling directional movement in the air.

2010-03-31 09:24:00 Author: Sehven Posts: 2188

---

WHOA< i cant wait for this LEVEL!!! sounds AWESOME!!!

2010-03-31 09:25:00 Author: YEAH_NAH Posts: 775

---

Ok I'm pretty sure I get the gist of what you're saying. I've come up with a control pod that I use for everything and anything. I can't quite remember but I'm fairly sure it was you that didn't really like the point and grab method of control pods, if I am correct just here me out. the best why I can descride my pod is through an image so I drew this up. http://i875.photobucket.com/albums/ab311/SR20DETDOG/Other%20Junk/Controlpod.jpg NOTE: Each block is 1X2 small grid spaces. This is just the basic layout modifications can be made to allow sackboy/girl in from the top side etc but the gabbable material must stay in the origal locations. Obviously the position of the motion sensor and it's radius and angle are just an example. It takes a little bit of fiddling and testing to work out the right settings but it really is very hard at all. Depending out how you seet it you can make it more or less sensitive which would help your jostling problems. Now yes this does look quite unreliable I know but trust me on this it does work. An example of how I used this is the player has four grab switches to use then when they jump (activating the sensor switch) those for switches perform different funtions. Hope this helps Sehven.

2010-03-31 12:36:00 Author: SR20DETDOG Posts: 2431

---

Instead of starting a new thread, I'm just gonna' change what this one is about. Cos that's not gonna confuse people. So now I'm wondering if anybody has any ideas on how to make a decent jump activated switch for a vehicle. The fundamental problem you have, regardless of whether you use a sensor switch or a button to detect sackboy jumping is, presumably, when the mech goes into free-fall. At this point, from the switch's POV, it's as if sackboy has jumped. I can see two possible solutions. 1. If the mass of the mech is less than the mass of sackboy, then sackboy *should* fall faster than the mech, meaning that the switch won't be activated in free-fall. Even if the mass of the mech is marginally larger than that of sackboy, then, theoretically, there's a point where sackboy's relative acceleration in free-fall is less than that of when he jumps, therefore simply putting a 'platform' above sackboy connected via a weak spring to the ceiling, when tweaked correctly, should be able to prevent sackboy from activating the sensor in free-fall, but still be weak enough to allow sackboy's jump to activate the sensor. 2. If you have already created a device which can detect when the mech is in free-fall, which I believe is the case from looking at your control pod level, can you use this to deactivate the jump sensor, so it can only be operated when the mech is stationary? After all, you can't actually jump in free-fall anyway, since you have nothing to push against. Option #2 would seem to be more elegant, but depends on the reliability of your motion sensor.

2010-03-31 15:33:00 Author: Aya042 Posts: 2870

---

SR2, I appreciate the suggestion, but you're right: it was me who said I hated multi grab controls. I hate hate hate hate HATE!!! them. But thank you for taking the time to draw it up for me like that. Year_nah, thanks for saying that. I hope it turns out awesome. Every mech I've built so far has been so-so and they've each had some pretty severe flaws. This one is meant to address all of those flaws and to hopefully not introduce any new ones. Aya, the mech's mass is considerably greater than sackboy's. Also, since it's not a graceful ballerina, there is quite a bit of jostling whenever it's moving, so sackboy needs to be held down onto the wheel anyway: I'm using the platform idea you said, but with a weak piston instead of a spring. Pistons work better for shock absorbers because you can set their min and max length, and you can hook a directional switch to them to make them actively push back into their neutral (extended) position. And they're less bouncy. The problem is that sackboy doesn't seem to always jump with the same force, so getting the switch to trigger every time is proving troublesome. For some reason, it seems to work best when he's running to the left. I've tweaked and reshaped the control pod over and over, but still can't figure out why that's the case. As for the motion sensor, it works decently to detect higher speeds, but it's nowhere near perfect: especially at the apex of the jump when the mech has very little vertical movement. I'm using one for a speed governor during flight, though: if the mech gets moving more than a certain speed forward, the forward rocket is cut off. Last night I decided to try to build a hamster wheel control pod with a jump switch. It seems to work perfectly (though I'll have to throw together a pretty wild vehicle to test it on... wild as in bumpy and bouncing all over the place. The problem is that it's much bigger than what I'm currently working with in my mech, and you can only access it via layer change which means I'll have to completely redesign the mech's upper body if I decide to use it. Still, I'm really liking the way it worked in initial tests, so it might be worth it. http://www.lbpcentral.com/forums/attachment.php?attachmentid=14695 http://www.lbpcentral.com/forums/attachment.php?attachmentid=14696 The green part is thack glass. The blue is thack sponge glass (meaning it's grabbable and weighs the same as sponge, but it looks like glass--provided by Aya). The red part is theck pink floaty glitch to look like glass, and the yellow part is thick wood glitched to look like glass. It's not necessary to make everything glass, but I like to do it that way, so that a window can be put in front of it and it'll all be invisible. I don't actually have it all color coded like that, but I thought it would make it easier to see what's going on. The green part is glued directly to the thin layer backplate, so it's static: it holds sackboy in the middle of the wheel so he doesn't move around inside the pod. The red part is loosely bolted to the backplate (yes the green is glued THROUGH the red, but the red can still move freely) and the yellow is glued to the red. When sackboy runs, it makes the yellow/red part spin and the teeth on the outside of it push the pointer up or down, which triggers the directional switches. When he stops, the pointer goes back to neutral, and the vehicle stops moving. The blue part has the grab switch used for all the grab functions, and it pushes sackboy down so he can maintain traction during slight bumps and when in freefall. The piston is weak enough, though, that when Sackboy jumps, he can push it up and the mag key on it will trigger the switch next to it. The problems are, as I said, you can only access it be being in the layer in front of it and jumping in, it's huge compared to my other pods (I'll have to completely rebuild the mech's body to accommodate it, should I choose to use it), and like my other pods, it will only be able to handle a certain degree of tilt before the player loses the ability to spin the wheel or push the plunger up. I have a similar pod that RobotCrash and I built: it's smaller and doesn't have a jump switch, but it can tilt quite a bit further and still work fine.

2010-03-31 20:51:00 Author: Sehven Posts: 2188

---

Aya, the mech's mass is considerably greater than sackboy's. I realise that. I was implying that you could stick a load of peach floaty in it to make it lighter, but that has other disadvantages. As for the motion sensor, it works decently to detect higher speeds, but it's nowhere near perfect: especially at the apex of the jump when the mech has very little vertical movement. Well, my theory runs along these lines... Imagine if your mech has two identical control pods, one containing sackboy, and the other containing a block of some material which weighs the same as sackboy. Now, in theory, whenever the block of material is not touching the 'floor' of the control pod, then you can assume that the same would be true of sackboy too, therefore if you disable the jump sensor whenever this block is not touching the floor, with a simple mag key/switch setup, that should be a reasonable way to stop the jump sensor from activating in free-fall, or in other situations where sackboy is pushed up artifically. This is roughly how I assumed your motion sensor worked. Obviously I'm not suggesting that you build two control pods, but use that theory in order to build a device which can detect when sackboy is being pushed upwards against his will. Last night I decided to try to build a hamster wheel control pod with a jump switch. Interesting, although I wouldn't have thought that this would provide any more accuracy with jump detection than your current design, but not having tested it out, I'm just speculating.

2010-03-31 21:09:00 Author: Aya042 Posts: 2870

---

The relatively flat area you stand on makes for better jumping. Obviously, it's not flat cuz it's the inside of a wheel, but because the circle is much bigger, the narrow section that sackboy stands on is flatter. A similar effect could be acheived by making the wheel sackboy runs on top of bigger, but that would take up even more space. It would allow me to keep the top-loading cockpit, though, so I'll have to think about it s'more. It's too bad there's no way to build an actual conveyor belt so I could have the best of both worlds: flat running/jumping surface, and small form factor.

2010-03-31 21:15:00 Author: Sehven Posts: 2188

---

It's too bad there's no way to build an actual conveyor belt so I could have the best of both worlds: flat running/jumping surface, and small form factor. The only option I can think of there would be a similar technique to how you do the vertical detection, i.e. have weak springs/pistons pushing from either side, and have them activate when sackboy is pushing hard enough to the left or right.

2010-03-31 21:26:00 Author: Aya042 Posts: 2870

---

So another flaw with my current pod designs is how much space the flipper takes up. As you can see in the above images, a full third of the width of the pod is just the flipper, so now I'm wondering if there's an alternative. Whatever it is, it can't have another layer: so far the pod is two thin layers (the front thin layer will be a window or wall that comes down in front of sackboy to keep him from flying out of the pod) and a thick (theck+thack) layer. Also, it can't introduce any more latency than what there already is, so even though Rtm's wheeley good switches rock, they wouldn't be useful in this situation... besides, they'd need another layer. Any thoughts?

2010-03-31 22:47:00 Author: Sehven Posts: 2188

---

I'm probably not welcome here, but couldn't you use a tap/grab mechanism so it will activate when jumped but not during freefall? It's probably no use to you so I'll leave now

2010-03-31 22:56:00 Author: Kern Posts: 5078

---

Of course you're welcome, and I appreciate the suggestion. The problem with the grab switch is I'm already using grab, tap, and double tap for other functions. Otherwise, a tap-grab would be great for jumping. I think I've got it. I used Rtm's wheeley good switch design (the one he used on the drawbridge on the end of his level) but made the two relay wheels into theck and thin gas (one each) and was able to squash the whole thing into the same layer that the control pod wheel already occupies... mostly. I did have to add a front thin layer, but it's gas, so it won't get in the way of anything else. The last hurdle was to make sure it was possible for sackboy to get into the pod since the thin layer gas wheel in front of it completely blocked him. I cut out most of the wheel, leaving two ridiculously thin spokes (it's gas so it can't break and its mass is irrelevant) so sackboy can easily get in now. I'd post pictures but everything I've changed since the previous ones is invisible so there'd be nothing to see.

2010-04-01 00:39:00 Author: Sehven Posts: 2188

---

I didn't mean a grab switch I meant a spring and mag like yours currently but rig it to a tap grab instead so if prolonged activation it shuts down?

2010-04-01 11:24:00 Author: Kern Posts: 5078

---

While reading your aerodynamic properties of wheels (https://lbpcentral.lbp-hub.com/index.php?t=23824-Aerodynamic-properties-of-wheels)thread I gathered that your primary goal was a better stabilization method than jets. I started thinking about the fact that your "up sensor" has to detect that the mech is off balance before it reacts. I was convinced there had to be a more active balancing method. In researching (googling) aerodynamics I came across this wikipedia page on flight envelope protection (http://en.wikipedia.org/wiki/Flight_envelope_protection) and thought maybe you could somehow just make it impossible for the player to do anything that would make it fall over. any way... my point is: Maybe you could make the jump button only operational when the mech is in prime jumping position (i.e. no sack folk tossing about in the control pod) or only when you want them to be able to jump. I've only been testing on the buggy in your copyable level so far since I don't have a mech. I've decided to make a simple mech so I'm more familiar with everything you have to deal with. And I still think a jet pack is going to end up in there some where

2010-04-01 20:27:00 Author: IStwisted Posts: 428

---

but rig it to a tap grab instead so if prolonged activation it shuts down? Ah, I see. That's actually not a problem. The cockpit does occasionally receive enough of a bump to push sackboy up a little, but he's never held up for an extended period. In fact, the way I've currently got it, the only time an accidental jump is triggered is after a really hard landing on angled ground. But even when that happens, sackboy is immediately pushed back down, so putting in a tap timer on the jump switch wouldn't help. (incidently, after I managed to get a decent ground detection system in place, the risk of accidental jumps dropped to near zero since the ground based stabilizers kick in before landing). Actually, it may be possible to do the opposite: make it so a tap doesn't trigger a jump, but a longer press does. I'll have to experiment with that. Thanks for the suggestion. ...flight envelope protection... I'm actually sort of already doing this. There really isn't a situation where jumping needs to be prevented: the player can jump on any grade surface, and in my test course it's even possible to jump within half a second of walking off a cliff to recover and put yourself back on top of it (any later than half a second and the rockets won't be able to overcome your downward momentum before they shut off). Using the motion detector I showed in my vehicle tools level (the part that detects movement, not tilt) I have a governor connected to the forward rocket, so if the mech goes over a certain speed, the player can't push it any faster. I haven't got around to building a governor for reverse yet, but I will. As for a better stabilizer than rockets... yeah, rockets are crap for stabilizing, especially when they're reacting AFTER the vehicle has become unstable. What I'm using instead is invisible dark matter emitters. Above the mech, there's an invisible, rectangular, theck-layer superfloaty "balloon" (superfloaty: using the overlap glitch to pack 15 blocks of peach floaty into one spot, so it's 15x as powerful as regular floaty) to help hold the top of the mech up. Unfortunately, a balloon that could actually support the mech completely would also make it much harder to control: it would increase drag to a ridiculous degree and it would make the mech feel too floaty (if you've ever driven my Mark II mech, you'll know what I mean about the mech being floaty). So I've got the ground based stabilizer which is the invisible dark matter emitters: when the mech leans forward, an emitter continuously spits out tiny theck dark matter just ahead of the balloon to keep the mech from leaning any further, and the same for backwards. Obviously having mid-air collisions with dark matter would be disastrous for stability and getting any kind of distance from a jump, so the emitters are shut off during jumps and replaced with a huge superfloaty balloon that gets emitted wrapped around the normal balloon. This reduces the mech's overall weight to the point the rockets can actually lift it and it makes the mech much more bottom heavy so it's near impossible to over-tilt it. The lifetime on the emitter is set so that the balloon disappears about two seconds after landing on a normal jump (normal meaning the landing area is the same elevation as the takeoff area) and about a second or so before landing when jumping from an an elevated position but landing on a lower spot. Just before landing, the ground sensor kicks in and reactivates the dark matter emitters so even if you're coming at the ground at an angle, the mech rights itself (this has the added bonus of reducing speed so that the mech's feet aren't broken by the impact). This is all well and good in theory, but there are still problems. The flight speed governor, for instance, will cut off the forward rocket if the mech is moving too fast so it can't go any faster, but it doesn't do anything to slow it down. Depending on the angle the mech launches at, it's possible to make a football field length jump or to struggle to make it forward even 10 feet. I've experimented with using retro-rockets to slow down a fast moving jump, but even at full blast, they're not especially effective. I'm not really sure what to do about this. It's also still possible to hit the ground hard enough to break the mech's feet (specifically, you can break the toes on the feet) if the mech is level when it hits (yes, ironically, a better landing is MORE likely to cause damage) because the tilt sensors don't kick in when the mech is level, so you don't get the deceleration that you do when coming down at an angle. This is rarely a problem since landing in a level position distributes the force over the entire foot... unless the ground you're landing on is angled. Last night I rebuilt the cockpit with the hamster wheel controls. It wasn't nearly as difficult as I'd thought it would be. Unfortunately, the new jump switch isn't as reliable in a real world setting as it was in the testing phase... but it's much better than the old one.

2010-04-01 21:33:00 Author: Sehven Posts: 2188

Posts: 1077139    Threads: 69970    Members: 9661    Archive-Date: 2019-01-19

Aus dem Archiv wurden alle persönlichen Daten wie Name, Anschrift, Email etc. - aber auch sämtliche Inhalte wie z.B. persönliche Nachrichten - entfernt.
Die Nutzung dieser Webseite erfolgt ohne Speicherung personenbezogener Daten. Es werden keinerlei Cookies, Logs, 3rd-Party-Plugins etc. verwendet.