Tips & tricks for making a Cube

[Hotel26]

Quite difficult:

Filthy hard, in fact.

I've always suspected a bug with KSP and SAS and figuring out where the target is, particularly when many docks are in the neighborhood.  Doesn't make things easier.

I'd thought I'd gotten the first square to close but it had not.  I have no idea what Same Vessel Interaction is about.

In principle this could or should work.  Has anyone done it?  (Don't worry: I won't ask why!)  Any tips and tricks?

                                                                                     

My spokes have RCS at both ends, fueled/controlled by the little MP tug I'm using.  I'm using 1) spokes with hubs pre-fitted and pre-aligned (in the VAB) at both ends and I'm using 2) other spokes with no hubs.

(I'm aware this endeavor sounds pretty frivolous.)

                                                                                     

I can now add some background.  During the five years that I have been playing KSP, KSC Tracking Station has been receiving constant noise and/or signal jamming that turns out to have been a streaming message from a distant galaxy.

I have finally decoded this message and concluded that it probably contains the plans for an Alterkube.  Or maybe it is a Hyperkube or Enderkube.  It appears to be a simple but very large kube.  Sorry: cube.

Furthermore, when positioned rather precisely in geosynchronous orbit around a planet in the Kerbolar system, and properly powered, it has the curious property that: any vessel maneuvered into the cube, becoming completely enclosed by it, will be instantaneously transported into proximate rendez-vous with some other cube in the system, possibly determined (unclear on this) by which of the six faces used by the vessel to enter the cube.  The language about 'geosynchronous orbit' was not very transparent and another interpretation suggests that any Kerbolar orbit would do, making a close planet-trailing orbit the most convenient, but requiring some constant position-keeping.

For security, it's quite possible also that this mechanism will be locked for use only with a Veteran present.

Edited May 23, 2020 by Hotel26

[AHHans]

My suggestion is to first make two squares that will be two opposite sides of the finished squares; add the four connecting links to one, the other, or both (in theory it shouldn't matter, but I would probably add two on diagonally opposing sides to each of the squares); and finally dock the two parts together, making all four of the connections at the same time.

This worked well with my Nauvoo Asteroid Station, which launched as two identical parts that were then joined in orbit. But this station is a lot stiffer (it also has some extra struts that got jettisoned after assembly), and - which is probably more important - is has a control point in the center of each square. So I could tell SAS to point each part at the "target" and then I only had to make sure that the relative rotation is lined up during the docking. But otherwise the maneuver should be essentially the same.

[OHara]

5 hours ago, Hotel26 said:

Has anyone done it?  (Don't worry: I won't ask why!)

I've made smaller similar things, because it seemed like fun. 

Sir_frost has some self-assembling cages on kerbalx: https://kerbalx.com/sir_frost/craft

The fine control mode on RCS, with [Caps Lock] on giving you the blue indicators, will approximately balance the thrust from RCS ports so that translation commands mostly translate, with just a little rotation;  the approximation is better when the ports are well spaced out, as yours appear to be.

The SAS targetting bug you might be thinking of 20513 is fixed in version 1.9.0

With those long beams, I think it would be more enjoyable to first rotate them square, then translate into position.  The Navball Docking Alignment Indicator would help with the orientation if you like that mod.

Edited May 20, 2020 by OHara

[Vanamonde]

One of my favorite builds, from way back when. 

Was a bear to assemble, to such an extent that I haven't tried to re-do it, though I've often wanted to. What you can't see from this pic (which is why I chose it) is that the far corner there doesn't actually line up. This is just as close as I could get it to working. 

[Hotel26]

On 5/21/2020 at 3:08 AM, AHHans said:

first make two squares

Got it.  Fabricate the squares complete in the VAB.  Now the only problem with this will be that I want this cube to be quite large.  (I've updated the Original Post with my crackpot intentions behind this project.)  Appreciate the tip, though.

 

On 5/21/2020 at 6:00 AM, OHara said:

The SAS targetting bug you might be thinking of 20513 is fixed in version 1.9.0

This bug has been driving me crazy for at least 3 years!  For alerting me that this is NOT my imagination, I am forever in your debt.  Thank you!

I think with this, I can insert a control point into the middle of each spoke.  (I am still on 1.7.3 and will be for a while.)  And then I know I can dock.

The procedure I am using is to take a hub spoke +-----+ and then attach two spokes ----- to it.  Then attach another hub spoke +-----+ to join the square.  Then attach 4 spokes ----- to the square.  A hub spoke to one side of the top.  Then two spokes to that top edge.  Leaving the insertion of a final hub spoke which will have to slot into place and dock simultaneously with 4 ends.  It won't be a cinch, but with SAS holding things steady while I proceed very slowly at final approach, it should work.

I use/love/endorse the Navball Docking Alignment Indicator!!

 

On 5/21/2020 at 6:29 AM, Vanamonde said:

my favorite builds

I like this, too!

 

Thanks everyone for the helpful tips.  I'll post a shot of a working "Enderkube", if I get one.

Edited January 29 by Hotel26

[RoninFrog]

@cubinator has got to have some ideas.

[cubinator]

3 minutes ago, RoninFrog said:

@cubinator has got to have some ideas.

Boy, do I.

I'll get back to you on that.

[Hotel26]

Perfect square, joined at all corners:

I used OHara's AHHans' suggestion and placed control points at the midpoint of each spoke.  You can see the little MP tug is connected to one midpoint and aimed at the midpoint on the opposite side.  [KSP v1.7.3]

Putting 4 upright spokes on now is the same endpoint attachment process used for the middle spokes (2 & 3) and will be easy.

Three more spokes into the top square should be more of the same.  The 12th and final spoke will be the final test.

Thanks all; and thanks @OHara!, especially about the KSP bug.

Edited May 23, 2020 by Hotel26

[JoeSchmuckatelli]

How in the world are you getting it all together? 

 

Waaaay back when, before I knew to use struts for stability I tried using two radial decouplers to solve the problem of my long SRBs wobbling all over the place (with disastrous results).  I presumed, then, that KSP only likes a single connection between pieces. It also did not work to have radial decouplers on distinct "core" section fuel tanks attached to a single long SRB - and this reinforced my perception. 

Clearly I was wrong.  (or did this change at some point?) 

[Rhomphaia]

53 minutes ago, JoeSchmuckatelli said:

How in the world are you getting it all together? 

 

Waaaay back when, before I knew to use struts for stability I tried using two radial decouplers to solve the problem of my long SRBs wobbling all over the place (with disastrous results).  I presumed, then, that KSP only likes a single connection between pieces. It also did not work to have radial decouplers on distinct "core" section fuel tanks attached to a single long SRB - and this reinforced my perception. 

Clearly I was wrong.  (or did this change at some point?) 

This is why it has to be built in orbit. docking ports docked together in flight connect differently than parts connected in the editor, it is more like the way that Struts work and can loophole around the no loops limitation, that is why docking ports have actions for both "undock" and "decouple"

[AHHans]

45 minutes ago, Rhomphaia said:

This is why it has to be built in orbit. docking ports docked together in flight connect differently than parts connected in the editor, it is more like the way that Struts work and can loophole around the no loops limitation

Indeed! The first dock that connects will work like a "normal" dock, i.e. somewhat like a regular connection made in the editor. The other docs that connect will act more like struts.

One interesting feature of "multi-docking" is that if you undock the primary dock without having undocked the other docks, then one of the other (previously strut-like) docks will become the new primary dock.

[OHara]

13 hours ago, Hotel26 said:

I used OHara's suggestion and placed control points at the midpoint of each spoke. 

You are confusing me with @AHHans there, who by coincidence also was the first to note on the bug tracker that your former nemesis the targeting bug had been resolved.

( It occurs to me now that you might like to have the docking ports engage only when the rotation is square.  I don't think you really need it unless you care exactly how the docking ports at midpoints are oriented.  KSP has a mechanism to wait until the ports are rotated ('clocked') correctly before engaging the docking clamps, that we can enable using the patch in this post :

[Hotel26]

12 hours ago, OHara said:

You are confusing me with @AHHans there

I did and apologies to both!  Although I really did intend to mark your response, due to its confirmation of the existence of my Nemesis Bug.

You and AHHans are going to have to tacitly share the Nobel Prize here, because I can't mark both of your responses.

Thank you to both of you and all who have replied.

 

Edited May 22, 2020 by Hotel26

[peteletroll]

16 hours ago, OHara said:

It occurs to me now that you might like to have the docking ports engage only when the rotation is square.

You can use DockRotate's "Rotate to Snap" feature to fix orientation after docking.

[Hotel26]

On 5/21/2020 at 2:02 PM, Hotel26 said:

I used AHHans' suggestion and placed control points at the midpoint of each spoke

Just an update on this.  When attempting to place the last "spoke", the distance between the mid-point control points across the cube dimension was great enough to diminish accuracy beyond that required.

I think my only hope now is to return to propelling from a mid-point, but controlling from an endpoint and using manual alignment with a final tweak on rotation to have the opposite end-point aligned well enough to achieve magnetic capture.

I've done this successfully once to mate two pair of docking ports.  The acid test will be the final four pair simultaneously..

[Hotel26]

I've returned to this project and have some success to report.  [click & arrows for slides]

  

 

 

I'm aiming to use 16 segment spokes, but decided to run a pilot with 6-segment spokes as a proof of concept.  I've made a new,very light-weight MP tug (Midge) and two of those will be included in the package.

In the fourth shot above, I can testify that every join in this cube is perfectly made.  Q.E.D.

You will also observe that the result looks like a Normandy Beach on D-Day, due to the ugly guide rails.  Nevertheless, I plan to make a Midge attachment that is capable of straddling the light-weight hub connectors and performing the same function, but is also then removable.

Guiding the final spoke into position was simply a matter of approaching on a diagonal from above, stopping, squaring everything up to ensure the guides would engage and then moving into position for magnetic forces to finish the job.

The full payload is mounted on one of my Methuselah lifters.  I believe that, after LKO refueling and with an Escort interplanetary transit injector lit up under it, it has the legs to go to any destination in the system.

Completing a Hyperkube is a marathon docking exercise, but very satisfying when a perfect result is obtained.  I'll report back here again when I have published a fully-fledged, tested machine.

 

[Hotel26]

OK, final report, I think!

First some shots.  Arriving in geosynchronous orbit:

 

Only the last piece remaining for placement:

 

Voila!

 

I forgot to select Same Vessel Interaction on all (2 x 6) ports involved in positioning the last piece...  But after some magnetic, autonomous movement, all but one join was made completely, but it was pretty close:

I have now set SVI in the VAB on all the beam ends, which might be enough.

The construction ship, after separation:

 

And the final result:

 

I had a lot of fun with this project!!  And published the result: Hyperkube

And opened a mission topic : Project Andromeda: the Hyperkube Docking Marathon

Yahoo!