Realism -- putting it all together (?)

[carolyn]

Hello! Long time since I've posted, but I want to bring up a topic that, from what I've managed to find, hasn't seem to have been addressed...

My kids and I have been KSP players since version 0.13, and to make things interesting/challenging again (as well as educational), I've recently have been trying to assemble the right set of mods for (somewhat)realistic modelling of our solar system and real-world rocket technology. Sure, we could jump over to Orbiter, but there's nothing like KSP for creativity and having fun while learning.. Science!

Although I'm in IT (a C# developer, btw) and no slouch in such things, I found this to be quite a challenge to set up, to put it mildly. So I can imagine that for most players who haven't spent many hours tweaking KSP and/or reading the forums, getting going with this would be just too difficult now.

From what I see, one main issue is that there isn't a place online where you can go and get step-by-step instructions on what to do to get a good realism setup. Information is scattered, and the info that's there is full of lingo and acronyms that people (like me) who haven't been forum regulars have to muddle through to figure out.

The other main issue is that the realism mods, particularly those with scaled rocket parts, are now woefully inconsistent in scale and performance with respect on one-another. (No offence to the mod developers -- you've done marvelous work.) Most seem to be either still on "kerbal scale" or at 100% scale, which is just too big to make anything practical within the confines of the game.

So I've started to take it upon myself to go through each part (stock,KW,Nova,etc) and re-scale them based on real-world engines specs, etc., to about 60%-by-length, and 20%-by-mass/thrust. It's a lot of work.

It would be great if a group got together (especially if important mod developers got on-board) to come up with some standards for scaling/etc to agree on and start to work towards... and also to put-together a friendly easy-to-find instruction page for installing a standard set of realism mods known to work together correctly, for people who just want to play realism without spending hours on setup with inconsistent results.

Has anything like this been done? If not, would anyone like to join me in starting to organize this, or at least helping with the part scaling work? Thanks!

[123nick]

http://forum.kerbalspaceprogram.com/threads/59207-0-23-Realism-Overhaul-ROv3-Modlist-for-RSS-1-10-14 this may help you. using most of the mods in the ksp forum with the word "realism" in it should give you a realistic KSP.

[NathanKell]

"which is just too big to make anything practical within the confines of the game. "

Um...what? People have been making rockets that will put 1000-1500t into LEO, rockets that mass in the 20 thousand ton range...it's been fine.

snjo has already changed the size of the SPH; I'm sure someone will get around to rescaling the VAB soon, but for now you can just move the rocket around inside it.

Here is an installer in progress for realism mods: http://forum.kerbalspaceprogram.com/threads/67668-Mod-bundler-for-Real-Solar-System-%282014-02-03%29

Edited February 9, 2014 by NathanKell

[regex]

It would be great if a group got together (especially if important mod developers got on-board) to come up with some standards for scaling/etc to agree on and start to work towards... and also to put-together a friendly easy-to-find instruction page for installing a standard set of realism mods known to work together correctly, for people who just want to play realism without spending hours on setup with inconsistent results.

So basically what you're saying is that all those guys in the Realism Overhaul thread aren't already doing this? You should check that thread out, introduce yourself, ask for help. They're a friendly bunch who are pretty passionate about their project.

[carolyn]

http://forum.kerbalspaceprogram.com/threads/59207-0-23-Realism-Overhaul-ROv3-Modlist-for-RSS-1-10-14 this may help you. using most of the mods in the ksp forum with the word "realism" in it should give you a realistic KSP.

Well, that's what I had gone by mostly, as well as searching the word "realism". But imho, it ought to be easier for people who just want to install-and-go and get consistent parts, and I want to find who is working on that and help if there is something I can contribute.

"which is just too big to make anything practical within the confines of the game. "

Um...what? People have been making rockets that will put 1000-1500t into LEO, rockets that mass in the 20 thousand ton range...it's been fine.

snjo has already changed the size of the SPH; I'm sure someone will get around to rescaling the VAB soon, but for now you can just move the rocket around inside it.

Thanks for the tip about the VAB, but what about the kerbals themselves, have they been scaled so they aren't 2.5 feet tall? And it's great that people have managed huge rockets, but I don't think I'm alone in having trouble with heavy things just collapsing on themselves. It just seems, to me at least, that there's a lot of advantage to scaling to around 50-60% length and 12-20% mass, keeping things in the range Squad intended.

Is there consensus on this? Is 100% the standard for realism, or the goal? I've heard about 64% as well.

Here is an installer in progress for realism mods: http://forum.kerbalspaceprogram.com/threads/67668-Mod-bundler-for-Real-Solar-System-%282014-02-03%29

Thanks for that -- besides the scaling thing, a bundler is the other half of the project I'd hoped was out there. I'll check it out, and contribute if I can.

[NathanKell]

Well, swapping kerbals for humans is indeed on my todo list, and IIRC some of the more artisty folks are also working on that. So while nothing is there yet, it *is* coming.

Are you using KJR? I have yet to have a rocket collapse on itself from weight since I started using KJR, and now even launch clamps works well in RSS due to KJR.

64% was a consensus back when we thought we couldn't rescale planets, but since we can, and we have KJR, we might as well go 100%. That's certainly where all the realism work has been for the last three months or so.

Also, the issue is if you scale down the size of things, you mess up the ballistic coefficient (and atmospheric flight/ascent/reentry gets weird) since surface area is quadratic not cubic or linear. Nor does scaling mass to 41% (.64^2) work all that well either, because while you can change payload masses as desired, fuel masses are...fuel masses, and not really subject to rescaling unless you *also* start messing around with Isp.

(64% derived from looking at the size of the Mk1 and Mk1-2 pods, and seeing them as Mercury and Apollo, IIRC).

Anyway, if things are breaking for you at 100%, and you're using KJR, post on the KJR thread and we'll try to figure out what's going wrong.

[carolyn]

So basically what you're saying is that all those guys in the Realism Overhaul thread aren't already doing this? You should check that thread out, introduce yourself, ask for help. They're a friendly bunch who are pretty passionate about their project.

Sorry if I sounded negative, because they've done great work! But just going by what I found after installing it and looking through the parts, it appeared that the scaling issue had not been addressed much, and that (along with a bundler) was the issue I was talking about.

I know it's a WIP, but since I couldn't find good information about where things stand on this issue, I asked about it here, with the intention of helping with scaling if help was needed.

[carolyn]

Are you using KJR? I have yet to have a rocket collapse on itself from weight since I started using KJR, and now even launch clamps works well in RSS due to KJR.

64% was a consensus back when we thought we couldn't rescale planets, but since we can, and we have KJR, we might as well go 100%. That's certainly where all the realism work has been for the last three months or so.

Yes, KJR helped greatly. But because of those earlier collapse issues (and VAB/Kerbal issues), I had gone to 60% where I still get some disconcerting wobble with KJR (but no collapse). So I was concerned about going to 100% where the already-large mass/thrust would be multiplied by 5. I didn't want my kids and I to take time building nice spacecraft, only to find they would just go splat. Glad to hear KJR is up to doing 100% scale.

Also, the issue is if you scale down the size of things, you mess up the ballistic coefficient (and atmospheric flight/ascent/reentry gets weird) since surface area is quadratic not cubic or linear. Nor does scaling mass to 41% (.64^2) work all that well either, because while you can change payload masses as desired, fuel masses are...fuel masses, and not really subject to rescaling unless you *also* start messing around with Isp.

Yes, I realized the atmosphere would have to be thinned or something to make reduced-scale work. But wouldn't engines, fuel, etc., be fine if you scaled all part masses and all engine thrust to the cube of your length scale, leaving fuel and Isp alone? By my math, that would work. Of course that's moot if 100% is where everything is going, but for "academic" purposes, I'd like to know if I'm wrong.

[NathanKell]

carolyn: no worries; apologies if my first response was curt!

And help with scaling is *always* needed!

For some examples of glorious excess with RSS/RO, see the examples on this subreddit: http://www.reddit.com/r/RealSolarSystem

With KJR you do have to be a bit careful how you build (although 2.0 is *really* nice; Ferram added a lot of improvements). In particular, KJR has a hard time dealing with featherweight parts attached to massive ones (like, say, a MechJeb control part attached to a 2500t stretchy tank!), so try to avoid that when possible.

Regarding scaling--if you scaled part masses down by the cube, then yes, fuel volume and mass would be correct for the deltaV, and tank density wouldn't change, but then you'd get lower-than-RL ballistic coefficients (vs. the stock game's higher-than-RL). That's fixable by Ferram, say, adding a surface area multiplier, but anyway. You'd get the same TWR (although if Kerbin were not earth-size your ascent would take a different amount of *time*, which would be an issue, let alone the sharper gravity gradient...but let's assume RSS with 64% parts and ignore that)...but the issue with fuel is, if your part is 64% the size and you *don't* scale masses by 1/.64^3 (as you originally suggested), then your fuel density will be wacky.

Basically, if you rescale mass by the cube, then ballistic coefficent gets messed up; if you recale mass by the square, then BC stays broadly correct but density goes way up (you're fitting more liters of fuel into a tank than there are liters in the tank).