[Minimum KSP version - 1.11] Kerbal Attachment System (KAS) v1.12

[Dominiquini]

On 7/6/2021 at 2:05 AM, IgorZ said:

Yeah, looks like a bug. Created a ticket. It shouldn't break anything critical, though.

I see the ticket, but this message appears with no KAS parts installed in the vessel. Neither I have any inactive vessel in range, with or without KAS parts! This message appears with my first launch in Carrer and every time I perform  the first decoupling event (planned or unplanned) of some vessel I'm controlling. Subsequent decoupling events are OK!

Thanks.

[ElonsMusk]

I am reusing fairly large boosters and while KAS provides welcome QoL/Part Reduction, it takes a long time to fuel large vessels.

Is it possible for a user to manually change min/max transfer speeds from 4/20 to something greater?

[IgorZ]

On 7/22/2021 at 5:05 PM, ElonsMusk said:

I am reusing fairly large boosters and while KAS provides welcome QoL/Part Reduction, it takes a long time to fuel large vessels.

Is it possible for a user to manually change min/max transfer speeds from 4/20 to something greater?

Yes, it can be adjusted via the config in RTS part:

		// KASLinkResourceConnector
		maxTransferSpeed = 20.0
		autoSpeedTransferDuration = 4

 

[ElonsMusk]

On 7/26/2021 at 6:43 PM, IgorZ said:

Yes (. . .)

So helpful! 's what I needed to scale it for, have to avoid all time warp. Docking things so heavy and with springs on the vessel is just kraken bait I greatly appreciate KAS.

Spoiler

[IgorZ]

44 minutes ago, ElonsMusk said:

So helpful! 's what I needed to scale it for, have to avoid all time warp. Docking things so heavy and with springs on the vessel is just kraken bait I greatly appreciate KAS.

  Hide contents

Wow! Are you playing a fuel dealer role in the game?  

[ElonsMusk]

4 hours ago, IgorZ said:

Wow! Are you playing a fuel dealer role in the game?  

Reusing rockets requires refuelling, so yes one hat I put on is fuel hauler. Trying to build out a whole stock infrastructure for my bridgecrane

[ss8913]

Question for everyone.. I remember that this mod used to include an electromagnet part, and indeed there are still references to that part when one googles for "KSP electromagnet" (referencing the HE-3(?) part, from KAS)... anyways, what happened to it?  Was it somehow broken?  I used it quite a bit back in the day, have been doing other things with KSP lately but wanted to get back to doing things that needed such a part, but I find it to be missing.. is there a KAS+ or something that adds the old part back?

[IgorZ]

On 8/4/2021 at 7:45 PM, ss8913 said:

Question for everyone.. I remember that this mod used to include an electromagnet part, and indeed there are still references to that part when one googles for "KSP electromagnet" (referencing the HE-3(?) part, from KAS)... anyways, what happened to it?  Was it somehow broken?  I used it quite a bit back in the day, have been doing other things with KSP lately but wanted to get back to doing things that needed such a part, but I find it to be missing.. is there a KAS+ or something that adds the old part back?

https://github.com/ihsoft/KAS/wiki/Legacy-parts-destiny

TL;DR; There are no electromagnet parts as of now.

[Hohmannson]

On 8/5/2021 at 5:45 AM, ss8913 said:

something

Another mod has magnet PartModule and two sample parts that have magnets - have fun https://github.com/Angel-125/KerbalActuators

[The Aziz]

I have good news everyone. Using KAS cables, winches etc connected to 1.12.2 ground anchor and anything else, you can stop them from sliding. I am currently testing it on Moho and movement of my lander is none with winch attached.

[shimmy00]

I was trying to build an orbitally-assembled craft in KSP 1.12.2 now using KAS to try and keep it from bending and wobbling, and when I build the craft sometimes it works, but others like just now for some reason with this one the joints seem to not provide any support and it just wobbles the same - this happens even with the "Fixed Telescopic Joint" when "DOCKED". While turning or thrusting the vessel, either the joint expands and contracts telescopically, or else the joint ends seem to act like they aren't attached to the structural elements and move around. I've noticed this in earlier version combos but it seems to be here again and I'm not sure what's going on. It seems it may be related to the order in which stages are assembled but that shouldn't make a difference. I use KAS v1.9.

[chris-kerbal]

2 hours ago, shimmy00 said:

I was trying to build an orbitally-assembled craft in KSP 1.12.2 now using KAS to try and keep it from bending and wobbling, and when I build the craft sometimes it works, but others like just now for some reason with this one the joints seem to not provide any support and it just wobbles the same - this happens even with the "Fixed Telescopic Joint" when "DOCKED". While turning or thrusting the vessel, either the joint expands and contracts telescopically, or else the joint ends seem to act like they aren't attached to the structural elements and move around. I've noticed this in earlier version combos but it seems to be here again and I'm not sure what's going on. It seems it may be related to the order in which stages are assembled but that shouldn't make a difference. I use KAS v1.9.

Not sure, but I don't think the intention of the KAS parts was to reinforce joints or keep crafts from wobbling. The telescopic joints were more for towing things like a trailer?!

[IgorZ]

19 hours ago, shimmy00 said:

I was trying to build an orbitally-assembled craft in KSP 1.12.2 now using KAS to try and keep it from bending and wobbling, and when I build the craft sometimes it works, but others like just now for some reason with this one the joints seem to not provide any support and it just wobbles the same - this happens even with the "Fixed Telescopic Joint" when "DOCKED". While turning or thrusting the vessel, either the joint expands and contracts telescopically, or else the joint ends seem to act like they aren't attached to the structural elements and move around. I've noticed this in earlier version combos but it seems to be here again and I'm not sure what's going on. It seems it may be related to the order in which stages are assembled but that shouldn't make a difference. I use KAS v1.9.

Could you please make a short video and provide logs captured during this session?

[ss8913]

On 8/26/2021 at 1:43 AM, chris-kerbal said:

Not sure, but I don't think the intention of the KAS parts was to reinforce joints or keep crafts from wobbling. The telescopic joints were more for towing things like a trailer?!

if you combine KAS with quantum struts, however.. as soon as you set the KAS parts to 'docked', then the quantum struts take effect, and they're pretty strong.  If you just want general structural reinforcement, KJR is the way to go.

[shimmy00]

On second inspection, I am not able to reproduce the seeming behavior and am also not sure if what I observed was actually a bug or just that the connections have a bit of "give" to them even though they are much more rigid than without - I compared with the vessels connected and not connected (i.e. not even latched in place) and there's definitely a major difference so they must be working (it's also possible the rest of the structure may be weak in some way, too). The log also doesn't seem to show anything anomalous, either . So perhaps there's no bug here, just a misunderstanding of how it is supposed to work.

ADD: On a closer look, yes the KAS joints still flex. But it's quite less than with no joint at all. So it could just be that there's more "stress" when connecting a massive spaceship vs. two little land buggies

ADD 2: If I spot something really anomalous, I'll post more, but I'm not sure now.

Edited August 28, 2021 by shimmy00

[darthgently]

2 hours ago, shimmy00 said:

On second inspection, I am not able to reproduce the seeming behavior and am also not sure if what I observed was actually a bug or just that the connections have a bit of "give" to them even though they are much more rigid than without - I compared with the vessels connected and not connected (i.e. not even latched in place) and there's definitely a major difference so they must be working (it's also possible the rest of the structure may be weak in some way, too). The log also doesn't seem to show anything anomalous, either . So perhaps there's no bug here, just a misunderstanding of how it is supposed to work.

ADD: On a closer look, yes the KAS joints still flex. But it's quite less than with no joint at all. So it could just be that there's more "stress" when connecting a massive spaceship vs. two little land buggies

ADD 2: If I spot something really anomalous, I'll post more, but I'm not sure now.

You might try using the Autostrut and Rigid options in the part menus for the parts on the craft.  I think you may need "Advanced Tweakables" enabled in the Settings for these options to show up.   I find setting autostrut to "heaviest part" helps quite a bit (there are some cases you don't want to do this involving asteroid capture and "klaws").  The mod Editor Extensions has buttons to autostrut and rigidify the entire craft.  There are also some design and piloting guidelines that are best followed that are informed by real-world physics and engineering.  Like a long thin rocket is going to be more vulnerable to bending and stressing joints than a stubbier one

Mostly though, the physics engine in KSP gets more fragile as craft part count gets higher and the longer the program has been running between game exits and resumes.  Even if every mod were coded to avoid all memory leaks and overly-interesting techniques to wedge their functionality into KSP (that wasn't initially designed with mods in mind I assume), then KSP, the Unity engine, and the Mono environment it all runs in is not completely stable, especially when under high craft part count conditions and a zillion other conditions involving floating point rounding errors and interesting design decisions.  As fun as KSP can be, it is not a great physics simulator when compared to what the likes of the Jet Propulsion Laboratory runs simulations on, I'd guess.  It didn't't have the same goals, budget, or expertise behind it, so the outcome was different.  But it is still fairly amazing for what it is - even if it does act out in a cobbled-together frankenstein-monster kind of kraken-spawning manner on occasion

So quicksave (F5) a lot and don't consider it "cheating" to use F9 if you want to enjoy the game

There are reasons KSP 2 is starting fresh in many ways, I'd imagine

[jefferyharrell]

2 hours ago, darthgently said:

I find setting autostrut to "heaviest part" helps quite a bit (there are some cases you don't want to do this involving asteroid capture and "klaws").

"Grandparent part" is a good choice too, especially for parts attached to decouplers.

But really I think the best advice is use Kerbal Joint Reinforcement Next.

Edited August 28, 2021 by jefferyharrell

[shimmy00]

11 hours ago, darthgently said:

You might try using the Autostrut and Rigid options in the part menus for the parts on the craft.  I think you may need "Advanced Tweakables" enabled in the Settings for these options to show up.   I find setting autostrut to "heaviest part" helps quite a bit (there are some cases you don't want to do this involving asteroid capture and "klaws").  The mod Editor Extensions has buttons to autostrut and rigidify the entire craft.  There are also some design and piloting guidelines that are best followed that are informed by real-world physics and engineering.  Like a long thin rocket is going to be more vulnerable to bending and stressing joints than a stubbier one

Mostly though, the physics engine in KSP gets more fragile as craft part count gets higher and the longer the program has been running between game exits and resumes.  Even if every mod were coded to avoid all memory leaks and overly-interesting techniques to wedge their functionality into KSP (that wasn't initially designed with mods in mind I assume), then KSP, the Unity engine, and the Mono environment it all runs in is not completely stable, especially when under high craft part count conditions and a zillion other conditions involving floating point rounding errors and interesting design decisions.  As fun as KSP can be, it is not a great physics simulator when compared to what the likes of the Jet Propulsion Laboratory runs simulations on, I'd guess.  It didn't't have the same goals, budget, or expertise behind it, so the outcome was different.  But it is still fairly amazing for what it is - even if it does act out in a cobbled-together frankenstein-monster kind of kraken-spawning manner on occasion

So quicksave (F5) a lot and don't consider it "cheating" to use F9 if you want to enjoy the game

There are reasons KSP 2 is starting fresh in many ways, I'd imagine

Thanks. Yes, I am aware the KSP physics engine is not perfect(*). However, I decided to test that possibility by trying this out with a craft similar to the one I was making but with many fewer parts (maybe like only 1/8 the number of parts or so - the original had several hundred, most of which were struts I thought were necessary to keep it from wobbling itself apart!). Aside from the game running a lot smoother and less laggy, it did not affect the behavior I talked about (If anything, the simpler craft seems to wobble worse). I also created a couple of videos so you can see it in action. The way to get the wobble is to, after joining the craft pieces, turn the craft using SAS reaction wheels. The turning isn't so bad - it's when it slows down that the wobble starts in earnest.

Original craft (~600 - 700 parts incl. struts):

Simplified craft (~70 - 80 parts incl. struts):

FWIW these are not stock craft; I am using the KSP Interstellar Extended parts mod as well. Not sure if there's some interaction going on between the two or not. The first video is of two parts of a "proper" craft, the second is a stripped down version with few struts and connection spam and a dummy weight to simulate the fuel tanks. In both cases the wobble is there. IT's also much less severe than if the KAS joints are not engaged at all (i.e. not connected - no video). So they do have an effect But it's enough to mess with some things. The tank/ballast mass is about 600 t, the other part of the craft is about 60, so a 1:10 ratio of mass on each. Hence why I wonder to what extent this is a bug at all and not just a misunderstanding on my part - the joints are under lots of stress, so they may be "giving" even if not completely failing. I could, I suppose, prepare experiments with smaller masses.

---

(*) I also agree with you that it's amazing - if one thinks the kraken glitches in this are bad; other games I've tried have a lot worse - a bunch of physics objects often can't even sit nicely upon each other.

[darthgently]

34 minutes ago, shimmy00 said:

Thanks. Yes, I am aware the KSP physics engine is not perfect(*). However, I decided to test that possibility by trying this out with a craft similar to the one I was making but with many fewer parts (maybe like only 1/8 the number of parts or so - the original had several hundred, most of which were struts I thought were necessary to keep it from wobbling itself apart!). Aside from the game running a lot smoother and less laggy, it did not affect the behavior I talked about (If anything, the simpler craft seems to wobble worse). I also created a couple of videos so you can see it in action. The way to get the wobble is to, after joining the craft pieces, turn the craft using SAS reaction wheels. The turning isn't so bad - it's when it slows down that the wobble starts in earnest.

FWIW these are not stock craft; I am using the KSP Interstellar Extended parts mod as well. Not sure if there's some interaction going on between the two or not. The first video is of two parts of a "proper" craft, the second is a stripped down version with few struts and connection spam and a dummy weight to simulate the fuel tanks. In both cases the wobble is there. IT's also much less severe than if the KAS joints are not engaged at all (i.e. not connected - no video). So they do have an effect But it's enough to mess with some things. The tank/ballast mass is about 600 t, the other part of the craft is about 60, so a 1:10 ratio of mass on each. Hence why I wonder to what extent this is a bug at all and not just a misunderstanding on my part - the joints are under lots of stress, so they may be "giving" even if not completely failing. I could, I suppose, prepare experiments with smaller masses.

---

(*) I also agree with you that it's amazing - if one thinks the kraken glitches in this are bad; other games I've tried have a lot worse - a bunch of physics objects often can't even sit nicely upon each other.

Are you under 100km in altitude?  Below 100k the underlying coordinate system is tied to the surface and that might be triggering a glitch.  I've found that under 100k things get sketchier on big craft.  And docking can get harder.  The only other things I can think of is placement of the reaction wheels and the amount of reaction wheel control authority.   But that looks like some of my craft when I don't use autostrut and rigid.  KJR works for a lot of people also but I removed it when trying to find the source of another issue and didn't reinstall when I found the autostrutting and rigid did fine; so one less mod to deal with, right?  I'm curious where your reaction wheels are placed and what happens if you increase/decrease wheel control authority but I don't think that would cause the "noodle" effect.  That looks like a classic joint issue.  One last thing is clipping of parts can (apparently) cause strange tensions in a craft if you don't allow them in the MOD+F12 cheat menu.   I don't see any clipping in the vids, but thought I should mention that.  It is possible to unintentionally clip things in the editor, so maybe something to check by using the clip cheat on another try

Edited August 29, 2021 by darthgently

[shimmy00]

1 hour ago, darthgently said:

Are you under 100km in altitude?  Below 100k the underlying coordinate system is tied to the surface and that might be triggering a glitch.  I've found that under 100k things get sketchier on big craft.  And docking can get harder.  The only other things I can think of is placement of the reaction wheels and the amount of reaction wheel control authority.   But that looks like some of my craft when I don't use autostrut and rigid.  KJR works for a lot of people also but I removed it when trying to find the source of another issue and didn't reinstall when I found the autostrutting and rigid did fine; so one less mod to deal with, right?  I'm curious where your reaction wheels are placed and what happens if you increase/decrease wheel control authority but I don't think that would cause the "noodle" effect.  That looks like a classic joint issue.  One last thing is clipping of parts can (apparently) cause strange tensions in a craft if you don't allow them in the MOD+F12 cheat menu.   I don't see any clipping in the vids, but thought I should mention that.  It is possible to unintentionally clip things in the editor, so maybe something to check by using the clip cheat on another try

It is under 100 km but I think it also wobbles anywhere from my experience. Haven't noticed anything difficult with docking. The reaction wheels might be a trick - they're in the front of that long part coming off the tanks. Does placement actually affect things? (It would make sense in real life; didn't know about in KSP though.) I also do not clip parts - I generally almost always stick to nodes as much as possible. So it seems the joint issue is the most likely one.

[darthgently]

Would there be a way for the mod to temporarily disconnect KAS connected vessels during scene load then re-establish the connections after the loaded scene has settled?  I get a often get a lot of unrecoverable base twisting when loading ISRU base scenes with a few refueling tankers connected via resource transfer stations to the ISRU base.  I imagine that also keeping the rotation/translation relationship between the craft constant during load would be required while they are detached

[darthgently]

What is the source of the kinetic energy when detaching vessels  docked via KAS?  Could it be nullified in the KAS mod somehow in the next release?

 

[IgorZ]

6 hours ago, darthgently said:

Would there be a way for the mod to temporarily disconnect KAS connected vessels during scene load then re-establish the connections after the loaded scene has settled?

In the "UNDOCKED" mode its already working this way:  the vessels are loaded as detached, and then the connection is restored. It's different in the "DOCKED" mode. There, there is just one big vessel even though you see it as a set of vessels connected via RTS. In this mode detaching before loading and attaching afterwards will kill the ISRU functionality, since it distributes the accumulated resources on vessel load.

1 hour ago, darthgently said:

What is the source of the kinetic energy when detaching vessels  docked via KAS?  Could it be nullified in the KAS mod somehow in the next release?

It happens due to shifting of the CoM. In the docked mode CoM is roughly located in the middle part. When you undock, the side tanks become separate vessels and each gets its own CoM while the middle vessel gets its mass significantly reduced. Nothing can be done here.  I can only suggest making more strong foundations under the big tanks to make the rapid mass change less disturbing.

[darthgently]

7 minutes ago, IgorZ said:

In the "UNDOCKED" mode its already working this way:  the vessels are loaded as detached, and then the connection is restored. It's different in the "DOCKED" mode. There, there is just one big vessel even though you see it as a set of vessels connected via RTS. In this mode detaching before loading and attaching afterwards will kill the ISRU functionality, since it distributes the accumulated resources on vessel load.

It happens due to shifting of the CoM. In the docked mode CoM is roughly located in the middle part. When you undock, the side tanks become separate vessels and each gets its own CoM while the middle vessel gets its mass significantly reduced. Nothing can be done here.  I can only suggest making more strong foundations under the big tanks to make the rapid mass change less disturbing.

So it really comes down to how the physics engine handles undocking, which appears to be fairly unrealistic.  Because the physics engine is CoM based while reality deals with distributed mass in analog parallel, lol.  But I'm going from "DOCKED" to "UNDOCKED" mode, then pausing and then doing "DETACH".  The jolt happens on detach.  So confusing at times

Edited August 31, 2021 by darthgently

[shimmy00]

Update on the last post: Apparently, the "problem" was indeed due to excessive wheel authority and perhaps placement. I had 4 reaction wheels on that thing - turned out it only needed 1. After trimming those down and adjusting authority appropriately the wobble went to a very tolerable level. I was able to fly the mission and the vessel performed great; just took a bit more time to turn. So it seems the trick was excessive torque being applied by those monster wheels, and thus the links are not perfectly rigid. They're rigid up to a point, but then they start to yield, even before breakage. So it's arguable whether this is a true bug or not: it seems to make some sense, the KSP reaction wheels are crazy powerful and if you put a crapload of torque on something, you can expect it to yield. One should also not expect to be able to turn something like 600 or 700 tonnes of craft when fully fueled and loaded nimbly on its gyros any more than you can likewise easily turn a giant mining dumptruck on Earth.

Edited September 1, 2021 by shimmy00