Modular Fuel System Continued v3.3 (OBSOLETE)

[sidfu]

yes finaly deadly reentry will be updated then. its one of 4 mods i dont play without

[Tiron]

Okay, so I poked at things a bit and quantified the actual numbers involved.

In air breathing mode:

With the original MFS 1.3 ratios, the Sabre-M burns ~0.6 LH2 per second.

With new MFS v1a ratios, the Sabre-M burns ~2.35 LH2 per second. 391.7% of the original value. Which is funny, because the ratio only increased by 290% (37.7 versus 13)

The original, LF/LOX version of my Valkyrie SSTO had 920 units of LF to get up to speed and altitude, and retained around 2000 m/s delta-v on reaching orbit. The original switch to LH2/LOX reduced that value quite significantly: The lower weight allowed a more aggressive, more efficient ascent profile, but the reduction in total Delta-V more than made up for it. Adding an extra fuel tank got the delta-V almost back up to the original value, and was still quite lightweight compared to its original design weight. It also needed only about 600 units of LH2 to climb to speed and altitude. Really more like 500, but I tend to be a bit conservative.

With the new version it blows through that 600 unit allocation at right about 10000 meters. I estimate the the LH2 requirement to maximize speed and altitude to be at minimum 1750 units, and probably more on the order of 2000. The total volume available for fuel is 22,410 after adding the extra tank.

Edited October 15, 2013 by Tiron

[sidfu]

got a question how does your mechjeb stuff compare to the forked update someone has did?

[Tiron]

got a question how does your mechjeb stuff compare to the forked update someone has did?

He said it was based on the 2.0.9-80 dev version, which includes all the changes from Sarbian's temporary fork, which has been shut down due to r4m0n's return.

[NathanKell]

Carzum, thanks for catching that. The KW values are kinda bugged, apologies; they never got reset to stock from the modified values I'd been playing with. They'll all be fixed in the next patch. Also, it pointed out to me I had a bug in my thrust rescaling code (to lessen the pain of realistic sea level thrust I automatically upscale the thrust of lower stage engines so you'll retain your old TWR on liftoff, so the LV-T45, for example, has a max thrust of almost 250, so that you'll still get your 200 on liftoff.)

Tiron, I'll install B9 and play around. Something sure doesn't sound right.

sidfu, yup, it's the latest commits to the dev branch of MJ2 on r4m0n's github, which sarbian has uploaded the sarbian-branch fixes to.

[NathanKell]

Oh, right, regarding MonoPropellant. I'm considering it UDMH with a touch of stabilizer (the density fits that), so you could totally set up a MonoPropellant/N2O4 engine with a slightly lower Isp than MMH/N2O4. I mean, the Russians never used MMH, they just sacrified a bit of performance and went to UDMH.

(And if we added Hyrazine, we could even simulate Aerozine-50 as a tripropellant engine with MP/Hydrazine/N2O4...)

EDIT: Since Oxidizer is now (per ialdabaoth) used as Hydrogen Peroxide, here is the real-world equivalent of the MP/Ox combination: UDMH/H2O2. http://www.astronautix.com/props/h2o2udmh.htm

Edited October 15, 2013 by NathanKell

[Carzum]

Carzum, thanks for catching that. The KW values are kinda bugged, apologies; they never got reset to stock from the modified values I'd been playing with. They'll all be fixed in the next patch. Also, it pointed out to me I had a bug in my thrust rescaling code (to lessen the pain of realistic sea level thrust I automatically upscale the thrust of lower stage engines so you'll retain your old TWR on liftoff, so the LV-T45, for example, has a max thrust of almost 250, so that you'll still get your 200 on liftoff.)

Hey no problem! I'm just glad this mod is receiving some well deserved attention. Looking forward to the fixed version.

[rhoark]

N2O4 is a viable, if rare oxidizer for RP-1. It's a hypergolic oxidizer for any of the various hydrazine blends, including MMH which is a common monopropellant. The shuttle used MMH for RCS and MMH/R2O4 in the OMS. I think they are good candidates to equate with KSP's generic monopropellant and oxidizer.

[NathanKell]

Except the densities don't come close to working, and N2O4 is already in the mod as itself, as is MMH, at their real densities. Whereas the densities are great for UDMH and H2O2.

Look, I appreciate you hold this position strongly, and you're welcome to do this on your own end. And I am taking your advice and using MP as UDMH. But I'm not going to throw out all the work we've (and ialdabaoth before has) done for this set of fuels and restart with your new scheme. I mean, you don't object to LH2 and LOX being in, right?

[rhoark]

I mean, you don't object to LH2 and LOX being in, right?

Absolutely not, and I certainly don't want to be contentious about it. It's just a differing philosophy of what RealFuels should be I suppose. On the one hand, there's maximal realism: assign everything a real-world counterpart. My take on it is that different fuels are exciting inasmuch as they impact the vehicle design. You can't just swap your tanks from LiquidFuel to LiquidH2 - it takes an entirely different vehicle to account for the volumetric energy density. Likewise it would be very Kerbal to have fluorine/pentaborane configs with high Isp but significant chance of going up in a green fireball.

On the other hand, the differences in density and Isp between RP-1/H2O2 and RP-1/N2O4 is about 1%. There would be design considerations in a real rocket, but nothing we can model in KSP. Where it doesn't change gameplay, why not be parsimonious with the number of named resources?

The differences between the hydrazine blends are a middle ground - on the order of 10%. That's more reason for a division, but I'd argue still not one that changes designs in a big way. In any case though, the thorough research on spot-accurate Isp's will not have been wasted, even if attached to more generic names.

[NathanKell]

That's certainly a very fair way of looking at it. I tend towards the more-realism end myself (as if it wasn't obvious ) but I take your point.

How about this: I help you set up a conversion package that will take my setup and convert it to the one you want? And if other people like it they can use it too. It really shouldn't be hard, just a replacement ResourcesFuels.cfg, RealTankTypes.cfg, and a some search-replace in Engines.cfg.

[NathanKell]

v2 is now live: You can now change an engine's tech level from its present tech level up to the max tech level (currently 7). Note when I say engine I include SRBs.

The current TL (and available TLs) and type show in the info pane when selecting a part. You can modify the TL by going into the engine config panel as if you were changing its fuel supply.

It has the KWR fixes, and I tweaked B9 SABRE Isp. Still playing with it ingame and I'll try to reproduce your problems, Tiron...

All Isps are included in MFT/TechLevels.CFG as atmosphere curves and may be edited to taste. The types are as follows:

O: Orbital maneuvering engine

U: Upper stage engine

L: Lower stage engine

A: Aerospike

S: Solid

The versions with + in their type (like U+/L+/S+) sacrifice sea-level performance and gain in vacuum performance. So L+ engines are almost as good as U engines in vacuum, and almost as bad at sea level.

Edited October 16, 2013 by NathanKell

[Venturi962]

Thanks for all the hard work on this mod. I've downloaded and find it very interesting to play around with all of the Fuel / Oxidizer combinations. I much prefer the constant fuel rate model to the constant thrust model stock KSP uses.

If there are plans to add more Fuels & Oxidizers, I'd like to point out the Theoretical Isp reference in the book Modern Engineering for Design of Liquid-propellant Rocket Engines. Google Books allows you to view the table (located on page 20) here: Google Books Link

One thing I noticed is that this data makes me think that the Isp for the LVT30 & LVT45 are a bit low. Using RP-1 & LOX should give a Atmo Isp of ~270 (90% Efficiency) and a Vac Isp of ~305 (85% Efficiency) as opposed to 240/275 & 207/258.

Happy to discuss your methodology.

[rhoark]

For now I'll happily play as it is. It's if I ever get going with the resource transformation parts I'd like to make when I'll try to coerce everything into kerbalized chemical names like Propellium and Blutonium.

[regex]

What are all of the engine values based on? I find them terrible, is there any way to use the real fuels with the stock engine values without, say, recompiling the mod? I knee-jerked; on second look and realizing the fuel adjustments, they're probably quite good.

Edited October 16, 2013 by regex

[NathanKell]

The engine Isps are based on real engines using those fuel mixtures. KSP has really weird Isps: way too high for normal propellants (kerosene/liquid oxygen, or a Nitrogen tetroxide hypergolic mixture) and too low for liquid hydrogen/LOX.

As I said in the post _two above yours_:

All Isps are included in MFT/TechLevels.CFG as atmosphere curves and may be edited to taste.

Note that hypergolic Isp is 0.95x shown, and Hydrolox Isp is 1.3x shown. Kerolox is the baseline.

EDIT:

As to what the tech levels correspond to:

0: WW2 to 1955 (pre-Sputnik) [note that rocketry didn't advance that fast then]

1: Sputnik and the early stuff (57-60)

2: Mercury/Vostok/first probes era (61-63)

3: Mid 60s (Gemini and early Apollo) (63-67)

4: Mid-late 60s (somewhere between Apollo and the F-1A/J-2S upgrades) (67-72)

(end space race, so less development)

5: Apollo Applications (72-77)

6: Shuttle era (78-90)

7: Current/future

Edited October 16, 2013 by NathanKell
added TL description, isp scaling

[regex]

As I said in the post _two above yours_

I fear you're going to be answering that a lot. I was interested in whether you had derived statistics from the engines based on throat diameter and bell measurements because it's pretty obvious, to me at least, that KSP artists weren't even trying to be realistic in engine measurements (and thus balancing around that seems pretty damn silly). vOv I'll give it a try.

[NathanKell]

Heh, yeah, probably. Sorry for being snippy; it's just I spent a ridiculous amount of time trying to get the Isps right.

HAHAH no.

I came to that same conclusion too, that's why all I tried to do was class engines based on my best guess as to what the nozzle geometry was or was supposed to be.

That's why the Skipper is an upper stage engine, for instance, why the Mainsail is an L+, and why (given the bolded part) I sucked it up and counted the LV909 and Poodle as O.

I put the classes and tech levels in my previous post in an edit, if you haven't seen them yet.

So alas no, it's just "is this a boost engine, an upper stage engine, or an engine that doesn't care about vac performance and pays a penalty in TWR for being restartable" (the latter being the only way I could make sense of LV-1/LV-909/Poodle, let alone the fact that their nozzles are all wrong).

[regex]

Heh, yeah, probably. Sorry for being snippy; it's just I spent a ridiculous amount of time trying to get the Isps right.

Yeah, sorry for calling your stuff terrible. I had a second look and realized I'd seriously knee-jerked, my apologies.

[NathanKell]

No probs. :]

[NathanKell]

Just a note. ferram's Isp Scaling doesn't work with this yet. We're working on compatibility.

[Tiron]

Just a note. ferram's Isp Scaling doesn't work with this yet. We're working on compatibility.

Could that be my problem then, since I use FAR?

I mean, the Sabre-M is a seriously big engine (640/860 kN depending on mode), although this is a FRACTION of what the real one it's based on is actually supposed to do (maybe that's why it's -M instead of -L?). The 1960/2940 kN it's supposed to be capable of would be SERIOUSLY overpowered, if it didn't just flat out break things.

My 'problem' is mainly that the fuel consumption roughly QUADRUPLED going from MFS 1.3 to MFSc 1a. I'm sorry I wasn't clearer about this, but I pulled those numbers by changing the ratios in the .cfg file with everything else being v1a standard. The original MFS used 13/87 as the LH2/IA ratio on the Sabres, and changing it to the 37.7/62.3 from v1a (roughly tripling the LH2 side of the ratio) quadruples the fuel use, with all else being equal.

Honestly, my personal suspicion is it's one of those cases where so much else about the engine varies from reality that it's hard to use any of the 'realistic' numbers, especially given the fact the entire game's scaled down.

I am curious as to where you got that '23 to 1' number from, because when I was poking around their site they listed a fuel/air ratio of 0.08 (or roughly 1 LH2 per 12.5 air). They don't specify how they arrived at that, but when talking about the LACE further up the same paper they mention the stoichiometric ratio being 0.029, which according to some other stuff I found is actually a rounded-off version of the stoichiometric ratio for hydrogen/air by mass. From what I can tell from the paper, it uses more fuel than air because they use up some of the LH2 cooling the intake air and powering the turbocompressor for it, although less than a LACE did. And it sounds like they shove that excess hydrogen into the outer part of the engine and burn it in that outer-ring bit, which is supposed to be some kind of weird pseudo-ramjet or something. Or somesuch. I'm just a computer tech, I'm learning this crap as I go.

Edit: Ooo. One thing is certainly much better in v2. The folders are much cleaner! I only had to remove one 'source' folder, it took like maybe 10-20% as long to get it installed as a result. Kudos!

Edit2: Fuel use on Sabre-M now about 2.09/s in air-breathing mode with v2 setup. let me just quickly pull FAR out and try it to see what that does...

Edit3: Not a bloody thing. I really think it's just the ratio change. :|

Edited October 16, 2013 by Tiron

[sarbian]

As promised, here is a build of MechJeb that shows Sea Level TWR (and has its figures corrected for all other regimes too). TWR and Max TWR show vacuum TWR; SLT shows sea level TWR. Atmo time shows time at sea level, and it and vacuum time are both correct (i.e. identical).

Could you share your patch code ? If it's useful it should me merged with main MJ2.

I wish ialdabaoth replied to my PM too ...

Edited October 16, 2013 by sarbian

[NathanKell]

sarbian, I did post my patch code, some days ago, on your thread.

But here is the link again. https://www.dropbox.com/s/wl3nmfbylx6mgpg/MechJeb2_thrustscaling.zip

Tiron, as I said above, I installed B9 and am taking a look. In v2 I tweaked the SABRE Isps slightly, and that might help a bit, but stay tuned.

I do plan on keeping Source in the plugins folder--it's as good a place as any, a number of mods do it, it makes compiling and updating easy, and it doesn't do the game any harm to have CS files in GameData.

[asmi]

Good stuff here - MFT update was long overdue.

It would be great to add support for non-restartable/finite-times-restartable and ground start/airstart/no-zero-G-start for engines to bring it even closer to RL.

Oh, and if you could come up with patch to KER similar to what you've done for MJ, it would be awesome!