Harnessing the Kraken: Eeloo in 11 hours!

[Boomerang]

A Kraken-powered warp drive? And all stock, no less? That is pure win.

I would love to see a huge expedition vessel using this technology, once it's been perfected. With a big crew, lots of landers, scientific instruments, etc.

You mean the Event Horizon?

[Azimech]

This all feels to me a bit like the Infinite Improbability Drive :-)

[Wooks]

One question: How do you stop it?

[Geschosskopf]

Inspired by the pics and vids a few pages back, I spent a little (too much) time last night experimenting with the Jenkens Drive in an effort to understand the operating principles and to try to nail down the essential parts and their best arrangement to maximize thrust and control, In the end, I gave up and downloaded Rune's ship just before I went to bed but haven't opened it up yet to see how it works, let alone fly it. Here are my findings, which hopefully will provide some verification of stuff already known/suspected and perhaps shed 1 or 2 extremely minor rays of light on furthering research. Sorry for no pics but I'm now at my work computer.

Anyway, I was using the basic method of extending the heavy landing legs through structural panels to push against a flat surface. I tried these in a number of different configurations, both of these basic elements and with additional elements like docking ports and all mixed in. Did not turn on part clipping, just used whatever clipping happens normally. All tests were performed by starting on the Launchpad and going from there. Never got to space. Here are my observations.

Landing Leg Arrangement:

All legs were attached radially to box girders. Variables tested on this were 1) mounting on centerline and opening outward vs. mounted on outriggers and opening inboard, 2) mounting legs on the inside vs. outside surfaces of the box girders, 3) if the legs clipped through each other or not, 4) whether the flat surface the legs push against was 1 part or several superimposed parts in the same place, 5) height to mount legs above the flat surface, 6) position along length of leg where leg clips through structural panel, and 7) number of legs used.

1. Opening Inboard vs. Opening Outboard

Doesn't seem to make any difference to the amount of force created. However, the outriggers necessary to allow opening inboard add mass to the whole drive system, which is a separate issue. Note, though, that opening outboard constrains how far above the flat surface you can mount the legs (variable 6) whereas opening inboard allows more freedom there for a given diameter of flat surface. All in all not a significant variable in terms of power provided, but it does seem to have an effect on controllability (see Observations and Conclusions below).

2. Struts Mounted Inside or Outside of Box Girder

When inside, the legs clip each other in the stowed position and clip the box girder in the deployed position. When mounted on the outside, neither of these clips happens. This appears to make no difference at all in the performance of the system. Thus, this variable is insignificant for performance and is more an aesthetic thing. However, if you mount the struts inside the girder, you can only move/remove them by clipping the camera into the girder with them, so editing is a bit more of a hassle.

3. Legs Clipping Each Other

As noted in #2 above, this makes no difference in the stowed position. In the deployed position, clipping legs into each other only happens if they're arranged to open inboard. This doesn't seem to make any difference, either.

4. Flat Surface is 1 Part or Superimposed Stack

I tried this to maximize glitchiness. I superimposed the 2.5m batteries in various numbers (you can fit up to 4 in the same place before the attach nodes quit working). In all cases, this made the drive fail. The result was always "stretching" the superimposed batteries out of each other without moving the ship in any way. Conclusion: only use 1 part for the flat surface.

5. Height of Leg Attachment above Flat Surface

Obviously, the legs have to be mounted far enough above the surface to have room to swing down and then extend at least a little. The idea here was to see whether you got more force by allowing less extension before the feet hit the flat surface. The answer there is "no". So this variable doesn't matter as long as the feet hit the flat surface somewhere in the process of extending.

6. Where along Leg it Clips Structural Panel

This clipping has to be somewhere in the telescoping part of the leg, which actually includes the very bottom end of the fat upper section. As long as the skinny part of the leg passes through the panel, however, it doesn't make any difference. However, if only the fat part of the leg passes through the panel, the drive won't work.

7. Number of Legs Used

I tried combinations of 4 and 8 legs aimed at the same flat surface. 4 legs would work, 8 would not, somewhat contrary to expectations. I need to test 2 and 3 legs still. It could be that less is more. In addition, this variable includes whether or not 1 "unit" of legs is better than 2 (or more) "units" pushing on the same or different surfaces in the same or opposite directions. I have not tested all possible combinations. However, I saw enough here to make some general conclusions about how the whole Jenkens Drive works, which I'll now discuss.

General Observations and Conclusions

NOTE: Only applicable in surface/atmosphere situations. May not be applicable when in space.

1. Operating Principle

My theory, which I'm pretty sure of, is that the Jenkens Drive works by turning off Newton's 3rd Law in interations between various parts of the same ship. IOW, it's using physical parts to create the same effect that happens when you pump fuel from end to end of a rotating ship. Normally, the 3rd Law is in effect for physical parts, so that making a landing leg push against another part of the same ship creates equal and opposite forces and nothing happens. But by clipping the leg through the panel, you create a situation where for SOME calculation iterations, the force on only 1 end of the leg gets taken into account, resulting in a net force on the ship as a whole and thus movement. The interesting thing is that force retained is USUALLY the one pointing towards the nose of the ship, regardless of whether the leg creating it is pointed up or down.

2. Oscillations

Because joints can flex, and because the net force created is only intermittent, the Jenkens Drive creates oscillations along the length of the ship. This turns the ship into a mass + spring system. It is thus possible to amplify these oscillations by incorporating docking ports somewhere along the length of the ship. The only place they can't be is between where the legs attach and the flat surface they push against, which kills the drive. Hewever, their placement (and the number of joints made with them) is a matter of trial and error, depending on the mass of the ship, the location of its CoM, etc. IOW, it's a matter of tuning and harmonizing. Because docking ports have seemingly fixed amounts of springiness and periods of oscillation, they only help in cases where those harmonize with the rest of the ship. Also, only regular-size Clamp-O-Trons work. Both the SRs and JRs serve as dampers. Likewise, adding mass to the system can damp out oscillations to the point that the drive won't work.

3. Direction of Net Force vs. Gravity

This is usually unpredictable and may be a function of your computer. In a standard drive "unit" of 4 legs, each time you run the drive, anywhere from 0 to 4 of the legs might provide thrust. Whether this is truly random or dependent on hardware is unknown. However, I seem consistently to get only 0 or 1 leg providing thrust most times, occasionally 2. Never yet 3 or 4. With 0, the ship doesn't move. With 1, the ship either spins in place if CoM is at its center point or goes in circles (actually vertical loops) if CoM is off-center. Sometimes, a 2nd leg will kick in briefly causing the spinning or looping ship to move laterally a short distance but mostly it just stays in the same place. With 2 legs providing constant thrust, the ship pitches over to point E with its nose about 30^ above the horizon, but then moves SE away from KSC at 5m/s. A VERY limited amount of control over direction is possible while in this state but not enough to change the basic situation of moving SE while pointing E and pitched slightly up. Attempts to alter this usually result in the ship going into an unrecoverable spin or nosedive (see below). I have been unable to replicate Rune's feat of going straight up from KSC, which I suspect requires all 4 legs providing balanced thrust.

It also appears that the ship alternates between being under the influence of normal gravity or the Kraken, supporting the idea that the thrust is only intermittent. If the CoM is not centered, gravity wants to pull the heavier end of the ship down. Thus, for operating on or near the surface, designing the ship with the Jenkens Drive on top of the payload would make the ship a pendulum and thus more stable. It's definitely a bad idea to get the ship pointing down while the drive in operating, because you're unlikely to be able to pull out.

4. Launching Payload

The force provided by any 1 leg is effectively rather small, at least while "locked with gravity" (as Rune calls it). Outside gravity, the force is apparently huge given the 20G acceleration Rune claims, so perhaps inside gravity it doesn't last long enough to act very big. Anyway, in my own experiments, where usually only 1 leg provides any thrust, , the Jenkens Drive is barely above to lift its own weight, let alone a useful payload.

Anyway, that's what I've got so far with my own stuff on the ground. Next chance I have, I'll try them in space and also try Rune's ship from the Spacecraft Exchange and see if that makes any difference. If not, then I'm thinking your computer hardware and/or settings might have a lot to do with how well the Jenkens Drive works.

[SpaceSphereOfDeath]

It's the weekend, and I've got some spare time, so I might as well join in!

[Rune]

SCIENCE!

Well, that must have been a lot of fun! And it is an interesting and detailed account, so good job! I hope all goes well flying the Far Star, she is a bit temperamental (hint: always quicksave before engaging the drive above the atmosphere).

I would add that the stability of the drive seems to be related to how it is connected to the rest of the ship. No additional support, and it prone to misalignments that create torques and lateral fores. Too much bracing, and the assembly is too rigid and the legs all break. It0s a delicate balance. Also, I believe each leg creates a direction of thrust, but the total thrust is independent of the number of legs. Short of calculating the resulting vector, only then scaling it to some arbitrary quantity. Only the real effect must somehow be based on errors calculating positions and not forces.

Rune. Now go tame those krakens!

[SpaceSphereOfDeath]

I've discovered that rotations can effect my drives performance, rolling left seems to disable it entirely, if I yaw left it speeds up, and pitching down makes it explode, and SAS causes it to float, I tested it and it only fell 1m after about 20 seconds, nearly perfect!

EDIT: I have done it, may I present to you, the kartoffen II:

Not only is it locked in space, but also rotation. It's a single craft and is unfortuantly non-toggable, it can be turned off, but not back on again. I will test to see if it can be moved with rockets

Edited February 7, 2014 by SpaceSphereOfDeath

[Penguinhero]

all i see is an Eve launcher...

the possibilities are unfolding...

[SpaceSphereOfDeath]

While experimenting with the 'warp test 6' craft I discovered a way to increase the operational life time, more landing legs!, I added another 8 and it still worked fine, a slight rotation problem. But still functional

[Geschosskopf]

I would add that the stability of the drive seems to be related to how it is connected to the rest of the ship. No additional support, and it prone to misalignments that create torques and lateral fores. Too much bracing, and the assembly is too rigid and the legs all break. It0s a delicate balance. Also, I believe each leg creates a direction of thrust, but the total thrust is independent of the number of legs. Short of calculating the resulting vector, only then scaling it to some arbitrary quantity. Only the real effect must somehow be based on errors calculating positions and not forces.

Thanks for the tips on strutting. I'll have to experiment with them some. I wasn't using any at all for the above. Maybe struts would give a higher chance of all legs working.

In my experiments, during the leg deployment animation, the feet would extent clear through the flat surface but once the animation was over, they'd pop back in so the feet would be on the flat surface. If this ship was going to fly at all, this would be when it started. But I noticed something else when I had the legs on outriggers to open inwards. The ship only moved when the joints attaching the leg to the structure were stretched. This was much easier to see when the leg was on an outrigger to open inwards, because the outrigger would be at an angle to the central core, and the leg would be at an angle to the outrigger. Usually, this would only happen for 1 leg, and that leg was almost always on the outside of the curve if the ship was going in circles. The other leg joints would be in their usual positions.

The stretched joints are indicative of force being applied by the leg to its attachment point. In fact, if you delete the structural panels clipping the legs, you always get that same sort of deflection when legs attached to 1 part of the ship push against another part of the ship, or against the ground when you've got the ship so it can't move (like with launch clamps). This, and my ship usually going in a circle, is why I think the movement produced by the drive is from a force and not a positional error.

So it seems to me that clipping the extended leg through the panel screws up the force calculations. For each leg, normally there are 2 forces: 1 where the leg attaches and 1 were the foot touches. Most of the time for me, it seems the clipping destroys both forces. But sometimes it only destroys the force pointing down, resulting in the ship going up. Note that most times, even though graphically the leg is drawn in a compressed position due to the flat surface in the way of full extension, the attachment joints aren't stretched, indicating no force exerted on them.

Your mention of legs breaking got me thinking. Legs break (either stop functioning or snap off) when they encounter excessive force. In my tests, I never had a case of legs breaking off the ship. However, several times the legs did stop working. This happened when starting the drive caused the ship to tear itself in half. The top section would go up a short distance and then fall back down under normal gravity. And the legs would not function afterwards. At the time, I wrote those tests off as total fails because I was looking for the slow float upwards. But now that I think about it, those could have been the closest I came to success. All the legs must have functioned, creating a total force sufficient to destroy the ship and kill themselves. So I guess I'll have to go back and tweak those designs (like by adding struts) to see if I can control that

[Comrade Jenkens]

Geschosskopf: Its really useful you doing this detailed investigation of the physics involved. I've not had a single breakthrough today though.

[SpaceSphereOfDeath]

I've made short video demonstrating the warp drives, if any lurkers want to know how they work/look

[andrew123]

Hmm... I managed to slightly stabilize my slipspace drive, but have started to make anti gravity drive unsc ships. I need to make it more stable though before testing with larger ships.

[rholtsla98]

One question: How do you stop it?

That sir is a completely different issue

[TheCanadianVendingMachine]

If this does work, doesn't that mean that SQUAD will have been lying about no warp drives!!!

/sarcasm

[Comrade Jenkens]

If this does work, doesn't that mean that SQUAD will have been lying about no warp drives!!!

Maybe the entire plan was for us to discover the kerbal laws of physics and have to construct one from scratch!

[thomasjkenney]

Has anyone tried this underwater yet? (I'm only half-joking...would be very cool to have a WORM-drive submarine!)

[Tw1]

Has anyone tried this underwater yet? (I'm only half-joking...would be very cool to have a WORM-drive submarine!)

Testing is underway.

It makes a fairly average water engine, traveling at about 5m/s.

I think I'll stick with the propeller mod.

So far, any attempt to dive results in disassembly.

However, this may just be a question of piloting technique. I'm going to keep at it.

Left unattended, it seems to go around in circles.

Update:

Some degree of underwater success can be achieved when diving vertically while stationary.

However, it takes a fair bit of the force to prevent it coming back up again, meaning the dive angle must at max, 5 degrees off vertical.

Deviations from the angle you dived at tend to make the whole thing turn and fly upwards. So you can't turn.

Edited February 8, 2014 by Tw1

[A Fat Pokemon]

Ahh yes! The day has arrived where we can put away our jet and rocket engines, and use a Kraken drive on our boats!

[andrew123]

I finally found the secret sauce behind making a powerful slipspace drive, and it seems to increase exponentially with each application. I may be able to go to speeds measured in lightyears if this drive decides to meet projections... Ion engines also actually multiply the effect if applied in certain ways. (Hint: the secret involves clipping... lots of clipping.)

Edited February 8, 2014 by andrew123

[Tw1]

I'm tempted to abandon current plans, and start a grand tour with this engine.

Edit: Nope. Not reliable enough to move manned spacecraft it seems.

Wrote about it here.

Edited February 8, 2014 by Tw1

[astropapi1]

I've been following this thread since it started, and Damn... You guys have made a giant discovery. I remember when the FTL Egg was first posted, but that wasn't that cool because it wast a simple and (mostly) uncontrollable bug. But this thing is really great. The ability to toggle it makes it even more useful. All I can say is I'm already building a manned ship for a Moho fly-by using this drive, and designing a new K-drive module for my station.

I know this is like playing with fire, because the Kraken might backstab me at any moment and nuke my save, but it's completely worth it. People discovering and designing warp drives isn't a common event, is it? Scott Manley must do a showcase on this.

[Geschosskopf]

WARNING WARNING WARNING

If you want to experiment with this stuff, DO NOT, under ANY circumstances, use the tweakable on the landing legs to put them in the deployed position in the VAB to check parts fit. This will break what you've broken in the game, so that the legs will no longer clip through the structural panels but instead will stop at them like you'd normally expect. You'll have to exit and restart the game several times to get things back to the desired state of brokenness.

In other news, I'm unable to replicate the results of others even using the exact same designs. For instance, Rune's Far Star. This approximates the advertised performance but doesn't duplicate it. When I engage the drive on the launch pad, it does climb smoothly and stays pointing straight up, which is way better than my own designs have accomplished. However, it doesn't stay directly over the Launchpad. Instead, it accelerates slowly to the east while climbing steadily at 5m/s. Because it's moving east, it's definitely moving horizontally. If it was geostationary it wouldn't move sideways and if it was letting Kerbin rotate under it, it would appear to be going west. And then there's how it works in space. Engage the drive and all legs instantly break after providing only a split-second of thrust, not enough to change the orbit more than a couple of meters. And following the repair instructions Rune posted doesn't work--the legs still instantly break again even with ESC pushed.

I also shamelessly copied SpaceSphereOfDeath's design, as I have previously copied all others posted here. Again, not what's shown in the videos. With SSOD's, the ship is a bucking bronco, spinning and tumbling for all it's worth until all the legs break. While doing this, it does change trajectory but there's no way to make it go in the desired direction.

So I'm beginning to think maybe a lot of this stuff is hardware-dependent. Like if you have a slower machine, the game does more fudging on the calculations. I dunno, just a guess. Here's what I've got. It's a couple years old but hasn't yet felt its age:

• Intel Core i7-2500K at 3GHz overclocked to 3.7GHz
  
• 8.0 GB RAM
  
• Win7-64
  
• NVIDIA GeForce GTX 590
  

Anyway, more SCIENCE! to report.

1. The ONLY legs that work for me at the big ones. The small and medium legs always stop on the structural panels instead of clipping through them, even before I did the tweak on the big legs.

2. I made the usual drive unit and then constructed radial arms sticking out above and below the drive unit (legs and flat surface). On the ends of these arms, I put various bits of structure so I'd end up with facing pairs of Clamp-O-Trons held just barely apart from each other to either side of the drive unit. The idea here was that if the legs kicked 1 end of the ship away, the magnetic force between the docking port pairs would pull the other end along after it. This produced interesting results. When the docking ports were close enough together to dock on spawning the ship (even though you could still see space between them), the drive produced only torque, not linear thrust. IOW, the ship would just spin in place, whether on the pad or in space. However, if I manually undocked the ports, or built it so they couldn't dock on spawning, then the drive didn't work at all.

3. When inboard-opening legs break, fixing them restarts the drive, tossing the Kerbal off. You have to quickly switch to the ship and disengage the drive before it leaves the Kerbal too far behind to catch up. But the bottom line is, I can actually fix inboard-opening legs. I cannot fix outboard-opening legs. As soon as I do, they break again. But on the plus side, the ship goes nowhere. It gives 1 jerk that throws the Kerbal off but then stops right there so getting back aboard is no problem.

4. I also tried using Gigantor solar panels instead of lander legs. This failed. The panels extended through the flat surfaces like they weren't there, then functioned perfectly normally.

5. One more thing about Far Star. One time I engaged the drive on the Launchpad and let it climb a ways, then turned the drive off. I let the vertical speed build up to -50m/s, then reengaged the drive. The result was that the ship exploded. Apparently, the full up thrust kicked in, resulting in massive tension in the ship's structure. Pretty cool .

Anyway, for whatever, I don't seem to be able to make any of this stuff work very well. I'll give it a bit more time but I given how even proven designs don't work for me, I doubt I'll have any better luck.

[andrew123]

I used have a laptop with a powerful i7 along with a 6770m (wimpy, I know, but I also have 8 gb of ram mixed with a 1 tb hdd which were also fried when my hp laptop decided to have a psu overload while in sleep mode), and I managed to get a warship to over 180000 m/s... but I think it depends on the type of "warp drive" you make. I can easily imagine using my drive for extreme acceleration, but using it to slow down may be a terrible idea... but my drives exploit a "perfect clipping" bug in the game.

[astropapi1]

WARNING WARNING WARNING

If you want to experiment with this stuff, DO NOT, under ANY circumstances, use the tweakable on the landing legs to put them in the deployed position in the VAB to check parts fit. This will break what you've broken in the game, so that the legs will no longer clip through the structural panels but instead will stop at them like you'd normally expect. You'll have to exit and restart the game several times to get things back to the desired state of brokenness.

Anyway, more SCIENCE! to report.

1. The ONLY legs that work for me at the big ones. The small and medium legs always stop on the structural panels instead of clipping through them, even before I did the tweak on the big legs.

Oh, so that explains it...

I was going to make a long post asking for help about my mini design, which I tested several times, including extending the legs in the VAB, but it turns out I got it all wrong.