Everything You Know About Kraken Drives Might Just Be Wrong...

[dnbattley]

(Cross post from the Reddit board)

So I came across some YouTube videos about Kraken Drives which seemed to re-iterate a few misconceptions about them and their capabilities.

Myth 1: Limited carrying capacity

A kraken drive relies on the non-Newtonian wheel "reaction" force to propel a craft forward. A single wheel will have a certain carrying threshold; more force required = more wheels can be placed in parallel.

Myth 2: Hard to control

Kraken drive force is no different to an un-gimballed engine: put the force along the centre of mass and it will work fine. There are some complexities, primarily around where the centre of force on a wheel/lander leg is generated, but these are simply things to be aware of in the vessel design. There is a maximum amount of force you can put through a wheel before it deforms, creating less predicable effects, and also wheels can only be compressed by a given amount before they warp and twist, also reducing their predictability and effectiveness.

Myth 3: Cannot be used in time warp

Prior to self-interaction, this was true. These days Kraken drives can be time warped without any issue. Interestingly they only work at full power in single physics time, however. Physics warp appears to maintain their output but over a longer time frame (i.e. reducing their actual force), and this can even be used to make more precise manoeuvres

Myth 4: Cannot be used in interstellar space

This myth is based on a true fact: applying kraken force to a wheel in Kerbol SOI (particularly further out than the Kerbin orbit) generates a tremendous amount of wheel stress which can cause the wheel to buckle a few seconds after being turned on. This can be alleviated by reducing the spring and damper strength to 0, but the real trick is to add a KAL which "cycles" any kraken drive wheels, toggling their self-interaction on and off over a period of about 3-4 seconds (from my testing this only needs to be off for <0.5s, so the power loss is limited, and if this is staggered over all wheels then the effect is smoothed) as this stops the wheel damage from occurring while allowing thrust to be maintained indefinitely.

To provide evidence that these myths are indeed just that I knocked together a rocket which took 10500 units of Ore from KSC to Dres.

Having spent many dozens of hours investigating their use, I have already shared some of my creations here and on KerbalX, but could also think about making a longer tutorial, if there is interest.

Edited February 1, 2020 by dnbattley
typos

[dnbattley]

Some further more quantitative investigation from testing the above lander in other orbits:

Myth: (and to my surprise, given my anecdotal experience) excluding low-level effects near a planetary surface, kraken force is NOT subject to the local gravity well/body of influence: a k-drive generates the same acceleration in Low Eve Orbit as it does in high Pol orbit. What does vary, however, is the wheel stress experienced around each body. Wheel stress varies by local body, and is a function of craft mass and inverse function of distance (i.e. it is higher further out), but it is NOT related to the level of leg/wheel compression, nor can it be "shared" across multiple k-drives (i.e. each k-drive experiences that same amount of stress resulting from the full mass of the craft). However, per myth 4, and as a further update, this stress is only applied approximately 8.4 seconds after the drive is engaged.

Information: K-drive power is relative to the wheel used. The LY-05 is my "go to" leg as it is easiest to build with, is robust (except to temperature), physically small, cheap and lightweight generates a force per wheel of up to around 500 kiloNewtons* (i.e. a craft of c.123T can be accelerated to about 3g using a parallel build of 8 LY-05 K-drives with a "high" compression, without distortion). Higher values may be possible but at the risk of distortion and consequently uncontrollable acceleration. Lower values are possible, though fine control can be very challenging.

* My calculations are going very wrong here: either I've forgotten my high school physics, or the Kerbal universe throws normal force calculation out the window. If anyone can correct me here I'd appreciate it. I'm calculating 50,000 (units) from the data, but then in practice I'm finding the acceleration equivalent to 4 Vectors (which should be 4,000 kN, or 500kN per engine) - AND THEN (and this is truly bizarre) when I stick a Mammoth engine onto the rocket (with infinite propellant to control my variables and reduced ore to keep the launch weight to 123T), the K-drive alone generates more than double the amount of thrust than it did before the Mammoth was attached, now making 7.6g acceleration!

Edited February 8, 2020 by dnbattley
Actual weirdness in calculation...

[klond]

 Thanks for posting.  I'm not sure I'm ready for a discussion this deep right now but I want to share what I have.  Although I have little evidence to back up my claims.

 On the subject of mass effects on legs @Bubbadevlin I think was the person to point out to me/us a year or two ago when he made a big leg-powered fan VTOL that leg forces are indeed in direct relation to craft mass.  The answer at the time was to separate the propeller/turbine from the main craft (as one would do on a turboprop), but put the legs on the spinning part so that the main craft could be edited without effecting leg performance since getting power sometimes meant running the legs right up against their boom limit.

 

 Warning: rambling

 I have had problems with K-drives the last few patches, mostly with my blimp.  When I ascend to 200ish meter above the KSC I lose the gentle push the k-drives give and the they go hyper and out-of-control.  I'm assuming this is the force that is wanted by others, but not for me and a gentle blimp.  The blimp still works at low altitutes, but this has me shy away from any further k-drive stuff for now.

Multiple k-drive in action pointing in different directions don't act as I would expect.  If I put 3 pointing forward and 3 pointing up, I should go forward and up, but weirdness takes over.  I can't explain it better than that.  More testing needed.

 

 I was not able to replicate the 8.4 second wheel stress you mentioned at the runway.  Must not count in the low-level-effects zone.

 

On 2/1/2020 at 7:34 AM, dnbattley said:

There is a maximum amount of force you can put through a wheel before it deforms, creating less predicable effects, and also wheels can only be compressed by a given amount before they warp and twist, also reducing their predictability and effectiveness.

 Tru dat.

On 2/8/2020 at 7:12 AM, dnbattley said:

kraken force is NOT subject to the local gravity well/body of influence

 But you can hang other stuff from the craft that will be pulled by gravity, by stock hinge or unpowered robotics parts.  That's how the blimp stays upright.  The whole thing hangs from the top of it's center of mass.  Don't tell anyone tho  

 

 That's all I can think of for now.  Good stuff man.

[dnbattley]

Thanks for adding this: the low level vs high level effect is one I haven't really explored scientifically, but have some anecdotal observations that might help here. You are right that for an object launched from the KSC the transition point is around 200m, and it dramatically increases velocity above that, but it seems to be more fundamental than that: when approaching from the other direction (i.e. at speed from space) the k-drive effect works differently, and is able to convert almost any vertical velocity to its "upward speed", generating extraordinary forces to do so. As a case in point, I generated g forces in excess of 10000g utilising this effect for a recent challenge.

My theory here is that the physics engine generates the speed (and the force required to do so is a derived value) at low altitudes, but at high altitudes it generates a force (from which the speed is derived).

As to the transition height: landing the vessel in some locations can dramatically change the transition height (and touching down vs landing - or switching craft - may not be the same thing in the game's logic here), and sometimes I have experienced transition heights of 1000s of metres, even in close proximity to the KSC. My theory there is that the transition height is X metres above a value for the "baseline location" for the craft, though whether that is a function of biome or locale is something I haven't properly explored.

For your purposes you might want to try flying over grasslands/water and then switching/landing craft to see if that helps consistently create the effect you want. Let's see if we can unwrap this effect further

4 hours ago, klond said:

leg forces are indeed in direct relation to craft mass

There is definitely a deeper effect here too: identical mass vehicles can have different speeds based on different configurations of the non-critical parts. It is definitely odd!

[dnbattley]

4 hours ago, klond said:

Multiple k-drive in action pointing in different directions don't act as I would expect

I've noted the same thing, but again haven't really bothered to explore further. One thing I have found is that a K-drive generates an "inertia" in its direction of force, which can be difficult to overcome, even by a high powered reaction wheel, but disengaging the drive, then turning, then re-engaging works...

[klond]

On 2/12/2020 at 11:07 PM, dnbattley said:

One thing I have found is that a K-drive generates an "inertia" in its direction of force, which can be difficult to overcome, even by a high powered reaction wheel, but disengaging the drive, then turning, then re-engaging works...

 Yes!  I forgot about this.  On early blimp versions I had upwards of 100 large reaction wheels to turn the blimp while being supported in the air by k-drives.  This extra part count hurt me at the time and I ended up using sideways k-drives and jet engines to turn instead.  Something def. going on there.