[WIP][1.8.x] SSTULabs - Low Part Count Solutions (Orbiters, Landers, Lifters) - Dev Thread [11-18-18]

[rasta013]

7 hours ago, Shadowmage said:

How beneficial are the current modular parts as far as frame-rate is concerned in 1.1?

My 2 cents is this is a bit of a trick question.  I see others mentioning that they have 140-150 part stations with no lag at all.  I on the other hand run an extremely large mod load (currently around 160-170) and many of them are graphics mods like Scatterer and EVE amongst several others.  Because of that, my stations begin to slow down noticeably as I start to approach the 100 part limit.  It's nowhere near as bad as it used to be and my slow downs typically entail a frame rate drop from my average around 60ish to down around 25-30ish.  As Nightside mentioned above this is largely a function of whether there's a planet in my view or not.  If there is, FPS tends to run around 25 for me on a system with 24Gb RAM on an I7 5820 3.2GHz with dual NVidia 770s.  If I remove the graphics mods from my setup that frame jumps drastically up to around 45-50 with a 100 part count station.  I suffer the frame rate drop willingly though because I want my game to look as nice as I can while playing even if I have to occasionally suffer that drop.

So, this is not really an easy question to answer regarding where the performance line should be drawn.  Those of us on large loads that include things to make it look pretty are going to suffer that drop.  So to me, the ideas of things like welded docking ports is very, very attractive.  Including other parts like RCS, adapters, antennas, solar panels etc. would only make that experience better.  The important thing to me though is that you find a way to do it that is comfortable for you and how you envision your mod.  I knowingly suffer the issues I do and can clear them up any time I want.

[JoseEduardo]

16 hours ago, Shadowmage said:

Indeed; I will likely be using them in my station investigations over the weekend.  Very well done looking parts.  I may even use some of the models for the initial prototype part design (through patches/etc; would not be distributing the parts/models).

I was going to ask if you would make the docking port part of the mesh, part of the .cfg file or stand alone mostly because of cx's APAS and CBM docking ports, but didn't in the end because I thought it was too soon 

I even added them through a patch to Orion and the APAS to Apollo (although it looks quite weird)

I didn't add for the Soyuz because I mostly use Bobcat's Soyuz 

btw, if you want ISS trusses you could use Habtech:

as for russian parts (if you wish to test them), Tantares is the only stock-alike I know, but there's Bobcat's Mir (+FGB/Zarya and docking covers to make a Zvezda) and Raidernick's Salyut and Almaz stations (but obviously, these aren't ISS material)

regarding styles, the only other two that I know are Tiangong and Skylab
Tiangong-1 is a resupply ship turned into a station


but Tiangong-3 is the final goal, however, the core looks a LOT like Mir


there's also the concepts, of course...
Freedom


Skylab II


Manned Orbiting Laboratory


OPSEK:


Exploration Gateway Platform


and Mir-2


Freedom and Mir-2 merged into the ISS, the module docked at the rear of Zvezda (which was intended for Mir-2, not ISS) is a Pirs/Poisk module, there's a second module with an arrow pointing saying "Bloc d'amarrage", while Freedom is quite self-explanatory 

the Exploration Gateway concept was proposed for a low cost using already-made modules (a.k.a. recycling) to save design costs

but yeah, there's mostly USOS and DOS/OPS/TKS based stations (even Tiangong-3 seems to be using a DOS core...)

[Kerbal01]

The latest release is hanging on Procedural Decoupler. I installed sstu manually, and other mods via ckan, so to trouble shoot I will nuke my current install, grab a new one and manually install all the mods.

 

Edit: a manual install worked.

Edited July 16, 2016 by DarthVader

[cxg2827]

@Shadowmage, if any of my blend or texture files would held you out with your experimentation, just shoot me a PM.

 

[blowfish]

I really, really hate to ask questions like this, but what ever happened to the RS-69?  It seemed like you had most of the geometry layed out, and that it could have reused most of the RS-68's UVs/textures.  Was it canceled or did you just shelf it and move on to other things?

Edited July 17, 2016 by blowfish

[Shadowmage]

14 hours ago, blowfish said:

I really, really hate to ask questions like this, but what ever happened to the RS-69?  It seemed like you had most of the geometry layed out, and that it could have reused most of the RS-68's UVs/textures.  Was it canceled or did you just shelf it and move on to other things?

I never found a geometry configuration that I was happy with, especially as I learned more about engine design in general I found that what I originally had would not have worked as well as I thought.

I might still revisit the concept in the future though; it would be nice to have an RS-68 class engine that isn't as tall, and a vacuum variant might still be useful for those times when you need to move a class-E asteroid around.

[blowfish]

3 hours ago, Shadowmage said:

I never found a geometry configuration that I was happy with, especially as I learned more about engine design in general I found that what I originally had would not have worked as well as I thought.

What do you mean?  Injecting the turbine exhaust into the nozzle is nothing new, and there's even precedent for having parallel exhaust ducts both injecting (Vulcain 2).

[Domfluff]

Since you were asking about inflatables, with mind to reducing parts but especially as you mentioned greater integration with other mods - I thought it might be useful to share something from my current career save.

(Relevant mods used here: SSTU, Ven's Stock Revamp, USI Life Support.)

At the moment, there are a paucity of parts that really suit long-term habitation. The following is a design-reference-mission sketch of a manned Duna or Eve flyby - conceptually similar to some of the Apollo Applications plans.

For reference: Apollo Manned Venus flyby - this wet workshops the third stage into a hab module, and replaces the lunar module with storage and life support.

Stowed. This is using the larger inflatable hab from Ven's Stock Revamp. Third stage not shown. Not supposed to be a recreation in form, but more or less in function.

This hab module consists of the following parts.:

Inflatable hab (10 seats)
Hitchhiker module (4 seats)
2x gigantor solar panels
4x nom-o-matic greenhouses
2x 1.25m fertilizer tanks
1x Cupola module
1x SSTU small docking port


USI life support provides hab time based on available seats. Some parts provide bonuses to this, and others provide multipliers. The hitchhiker pod provides a few months extra hab time, and I should write a MM patch for Ven's part as well, but the Cupola provides one of the multiplication effects, which is really important for creating long-duration missions.

Guidelines from the USI-LS docs:

Spoiler

//Suggested settings are based on part mass for consistency.  
// 
//For dedicated hab parts (no other generators, etc.):
//  Kerbal Months should equal mass * 5
//    ReplacementParts = 100 * crew capacity + 100 * Kerbal Months. 
//
//For parts that act as hab multipliers (dedicated or bundled with other functions/converters),
//a multiplier equal to the tonnage works well.
//
//For recyclers, their percentage should be mass / crew capcity (i.e. the UKS Pioneer Module at 3.75t = 75%)
//at crew capacity 5.  Increasing crew cap should result in an increase in mass.
//i.e. a 12-crew recycler that weighs 7.5 tons should have a recycler percentage equal to 7.5 / 12 = 62.5%
//Recyclers require (per crew capacity) 0.2 EC and 0.000002 ReplacementParts with a cap of 75%.
//If water is used as an input (0.0002 per crew capacity) the cap can be extended to 90%

Four greenhouses is one more than necessary here - each can process a little over one kerbal's Mulch - but this kind of (semi) closed-loop life support is vital for reducing mission mass as well.

The actual stowage here is awkward, because Ven's inflatable can't attach from the other end. You can fit that part radially, but (a) that feels a little like cheating to me, and (b) this is really just a proof of concept.



Deployed. This has almost two years of hab time, and the greenhouse processing provides just about enough supplies for a free return round trip (about a Kerbal year), but more could be added here. Those solar panels should be sufficient for power to Eve or Duna. Obviously the solar panels would be needed as the fuel cells won't last the duration.


This specific design has some minor issues - getting this out of the petal interstage is going to involve rather more RCS ballet than usual (attachment is on the wrong end), and the above is pretty much limited to a free-return (ish) style flyby, since the third stage would be used to send this away from Kerbin. The ship has more hab time than supplies (2 years to 1 year) right now, but that could be changed easily. There are also no recyclers in the above, and they could be added trivially.

Still, I think this is a fairly practical and minimal way of building a hab module, and is a practical example of some of the limitations that I've been running up against, and ones that I'd like to think that a dedicated part could help solve. It fits into something approaching the form factor at least.

Edited July 18, 2016 by Domfluff

[Shadowmage]

Did quite a bit of station building over the weekend; testing out different mods', trying out different station part concepts, and generally taking stock of what is currently available as far as station construction.  While there are some very nice parts' packs available for it (Tantares, CxAerospace were tested), they all result in fairly high part count for anything beyond basic stations.

Creating FGB/DOS/TKS styled station modules was quite possible, but they generally came in around 15-20 parts per module, which quickly starts to add up as more modules are added to a station.  The docking hubs are notable sources of high part-count as well, at 6-13 parts each.  These could both be made into single-part affairs, and hubs even potentially combined with the DOS modules, resulting in extremely low part count stations for when using these (think MIR in ~6 parts).

USOS styled cylindrical parts were a bit lighter on part count at only 3-per module (module+2 docking ports); but that is of course for a barebones module with no accessories.  Could easily combine the docking ports and a few accessories into a single modular part for these, and could potentially include a 'hub'/'node' style for these modules as well.  Not the best target for part-count reduction; but even just integrating the docking ports and some optional adapters would be useful.

The rest of the initial Station Core parts' concepts all seem to still be sound ideas so far.  Welding docking ports - definitely useful.  Multi-docking hubs, check (also integrated into the modular parts, see above; though these would be more 'truss' styled).  Inflatables, yep have uses for those as well.  So no real changes in concepts for those parts so far, but I haven't really tried setting up any stations to use them yet.

 

As far as performance is concerned, I had no problems with a ~120 part station, but didn't get to test it out past there; going to add a few more modules to it and see where the performance limitations are at for my hardware setup... but even at 120, that is at least 50% more than I could handle in 0.90 / 1.0.x.

 

Got the 'Modular Station Core' module mostly written up over the weekend; mostly just needs a bit of code to handle updating of attach nodes, crew capacity, and IVAs.  The attach node bit is just a matter of getting it written up; mostly just need some proper test model definitions to get the node location/orientation data from.  Will likely start working up some testing configs that use Tantares/Cx models to test out the modular part concept.

Dynamic crew capacity doesn't appear to work while in-game; whatever crew capacity is specified in the prefab part appears to propagate across to the in-editor and in-flight parts.  I -thought- that this was supposedly supported as mentioned in some of the recent patch notes, but cannot figure out how to get it working so far.  Will likely take a look at how RoverDude is doing it for his inflatables as I believe he was the one responsible for that bit of code in stock, and likely knows how it is all supposed to work out.

Have not yet tried/looked into the dynamic IVA swapping; that is waiting until I can figure out the crew-capacity issues (does not good to swap IVAs if I can't change crew sizes).

18 hours ago, blowfish said:

What do you mean?  Injecting the turbine exhaust into the nozzle is nothing new, and there's even precedent for having parallel exhaust ducts both injecting (Vulcain 2).

Mostly I mean that 'I never found a geometry that I was happy with'.  Even if it would have worked (and probably would have, just not as well as I had originally thought), I just didn't like the look of it.

Edit:  I do still have the original geometry around, pretty much just at the state it was last shown.  Perhaps I'll look into finishing it up after I get the current WIP engine geometry done (LR-81's, LMAE, LMDE).

Edited July 18, 2016 by Shadowmage

[tater]

Up the thread I mentioned a part idea---a full diameter hub part that includes a crew common area (since full diameter means 2.5m). Basically like 3 crossed hitchhikers with 2.5m docking ports (or welding docking ports). That's assuming that you think 2.5m docking ports should be a thing, else 2.5m welded, and standard size regular ports as an option, perhaps you could pick which are which via right-click. 

 

For interplanetary ships, I wonder if there could be a novel sort of docking port... Some of the Mars DRA stuff shows laterally mounted stages (like a Delta IV Heavy, but in space). You can build those in KSP, but trying to assemble such a thing in orbit is pretty non-trivial, you usually want 2 docking ports per tank, and then dock both at once to keep it lined up. Would a linear "mating" segment be possible, perhaps it's gendered, and the "female" side is a truss part of some sort, and the tank gets the male side. They'd snap together linearly in only one way, making assembly on orbit possible.

Edited July 18, 2016 by tater

[ComatoseJedi]

When it comes to space stations, the longest running one we can even compare it to is the I.S.S. Which is way past it's prime and was constructed by the lowest bidder. Granted, I have nothing against I.S.S. like station parts, but there's nothing else we can compare to a real world equivalent, besides the conceptual inflatable habs, which the BEAM module was installed on the I.S.S. not to long ago. There is a lot of conceptual freehand with stations. The only limit is your imagination and how you'd do it with current or near current technology. But, when you think inflatable habitation, you also include inflatable planetary habitation in a sense. 

Spoiler

 And just for some eye candy and future thoughts. Next logical steps:

Spoiler

Sorry! I get carried away with topics of this nature.  

[tater]

You can see the lateral Trans Mars Injection stage modules in the crew vehicle.

BTW, I think a non-spinning transit habitat is a showstopper for a Mars mission. The current known issues regarding visual problems, and the well-established bone loss problems are crippling for typical Mars mission durations, IMO, and we have no idea if the sojourn on Mars itself even helps all that much (we have no 1/3 g data at all). All that is aside from the fact that a Mars Hab really needs about 3m of soil covering it (I'm thinking ahead to building cool, realistic crafts with these parts  )

Edited July 18, 2016 by tater

[JoseEduardo]

@ComatoseJedi What about Mir? it was until not long ago the longest-running station, and Salyut-7 at the end was a testbed for modular stations using a docked TKS module that remained there until de-orbit of the station

[Shadowmage]

Very early concept/prototype geometry for some of the DOS-based parts.  No real detail work done, just figuring out the basic shape and geometry.

FGB-TKS-VA (2.5m -> 1.875m -> ??)

 

FGB-Tapered (2.5m -> 1.875m -> 1.25m)

 

FGB-Cylindrical

[JoseEduardo]

the third one looks like Almaz



(and no, I don't speak the language in the pic)

[tater]

What is the purpose of the highly tapered parts on TKS-VA?

Honestly, I google the Almaz stuff, and the tapered bits look like a launch escape tower.

Ah, the tapered bits on the ends are the return vehicles.

 

So on your TKS, that would be an entirely different part as a reentry vehicle, leaving it really just a cylinder.

Edited July 18, 2016 by tater

[JoseEduardo]

from what I know, it goes like this, from top to bottom:
Antennas
LAS/Retro-rockets
RCS
VA pod

the parachute goes between the VA capsule and the RCS pod

EDIT: ah, almost forgot, just like the Soyuz the TKS/VA has autopilot, and it's 4 flights were unmanned, no passengers cosmonauts ever flew aboard it because it was under testing and then it was cancelled

Edited July 18, 2016 by JoseEduardo

[SirKeplan]

2 hours ago, JoseEduardo said:

from what I know, it goes like this, from top to bottom:
Antennas
LAS/Retro-rockets
RCS
VA pod

the parachute goes between the VA capsule and the RCS pod

EDIT: ah, almost forgot, just like the Soyuz the TKS/VA has autopilot, and it's 4 flights were unmanned, no passengers cosmonauts ever flew aboard it because it was under testing and then it was cancelled

LAS has already been jettisoned in the images above(the de-orbit block does look a bit similar), this shows it with the LAS

As far as i can tell this is the order, top to bottom:
Launch Escape System
De-orbit rockets
Parachute
RCS/Soft landing rockets
VA capsule

 

Edited July 19, 2016 by SirKeplan

[JoseEduardo]

the parachute goes between the VA pod and the RCS pod, kinda like Apollo's chute, where it is under a case at the top, but yeah, I totally forgot that LAS, but I think in an emergency both would be activated, wouldn't they?

[StickyScissors]

11 hours ago, ComatoseJedi said:

 

 

I'm curious how it is proposed to land the BA-330's on the surface without any noticeable propulsion, and assemble them together without wheels, like that? I want to replicate it, but have no idea how to go about it while still looking realistic.

[tater]

^^^then they need to be buried in a few meters of soil   .

[Domfluff]

I believe the intent is to assemble modules in orbit, then land them vertically. I'm trying to find a concrete reference, but I'm not sure if there are thrusters under the legs there, or these would detach. Roverdude's upcoming Konstruction mod will give you some other options (aliens-style power loader claws, forklifts, etc.). In KSP or reality that would imply a certain amount of infrastructure already in place, At the moment, assembling in orbit and landing vertically probably makes the most sense, but clearly this is more suitable for the Mun/Moon than anything with atmosphere.

[blowfish]

13 hours ago, Shadowmage said:

Mostly I mean that 'I never found a geometry that I was happy with'.  Even if it would have worked (and probably would have, just not as well as I had originally thought), I just didn't like the look of it.

It looked fine to me, but there are a couple of other ways it could look (this and this).

[ComatoseJedi]

7 hours ago, StickyScissors said:

I'm curious how it is proposed to land the BA-330's on the surface without any noticeable propulsion, and assemble them together without wheels, like that? I want to replicate it, but have no idea how to go about it while still looking realistic.

In a real life concept, I would imagine they would be stowed and assembled on site. The one on the I.S.S. was stowed in the trunk of the Dragon cargo vehicle and took a little while to retrieve and place. The photo had no reference to how it got there and is only a model. I wouldn't 

7 hours ago, tater said:

^^^then they need to be buried in a few meters of soil   .

Yes, they do. But, how would you accomplish this? I don't think there will be construction crews on Mars (or any solid planetary body) with heavy machinery to cover the habs. Unless you land a Bobcat at the hab site. Would be a funny sight to land a Bobcat months ahead of the human crews to sit there and make it's own fuel until the crew got there. Radiation mitigation is something that hasn't really been addressed ad nauseum.

[tater]

Yeah, it's not really a KSP thing since we cannot interact with the terrain in the ways needed.