Thoughts on decouplers

[ppboyle]

Have some thoughts on decouplers that I thought I'd put down.

Current issues:

• Weight is inconsistent between different decoupler options. Even things like different fairing bases have different weights for a particular diameter.
• It's easy to cheat two different things with decouplers. You can very easily have sub-sized decouplers - say, .5m decoupler on a 2m stage - and the rocket still holds together.
• You can also using crossfeed on decouplers to do asparagus staging, booster crossfeed, or drop tanks, even without fuel lines..
• Integration costs should factor the cost of having more stages, but because of inconsistency with the decouplers, one of the easiest options for figuring out a staging integration cost - decoupler mass/total craft mass - would be wildly inconsistent

Nice to haves:

• Decouplers care about the weight of things attached to them
• Pyrotechnic decouplers should have increased weight for more 'oomph'
• An option to integrate small SRBS with decouplers instead of single-impulse explosions
• Decoupler reliability through TF - delayed or failing decoupling
• Enabling crossfeed came with weight/cost/reliability penalties

Ideal System: To me, the ideal decoupler system would do the following:

1. Look at the weight of part tree attached above or below it in the stack - taking the minimum of that, determine the weight & cost from multiplying that mass * a g-load the player sets with a slider.
2. This determines a load factor - exceeding that load factor causes the decoupler to suffer a reliability penalty when it comes time to decouple - or total failure if its exceeded by a wide enough margin.
3. Size of the decoupler might be constrained to a certain minimum based off of this load factor, but otherwise the cost of increasing decoupler size is negligible . The cost & weight come from the structural strength required + extras like stronger pyrotechnic separation or crossfeed plumbing.
4. In flight, decouplers borrow the heat systems 'danger sliders' to show if they're exceeding their load margins.
5. In flight the masses are constantly recalculated so a good steely eyed rocket man may be able to consider a lower load factor if he knows the part weights would be lower.
6. Offer progressive upgrades, like tanks do, that allow players to trade off weight for cost or tech investment, and unlock crossfeed capabilities eventually

The above would create incentive for players to reduce the # of stages they use, or suffer reliability issues, and also reduce their max g-loading, applying some of the concerns real rockets have with structural strength.

Stretch goal

1. A pump rate slider, with attendant costs, would really be fantastic. Probably impossible without rewriting KSP's resource system (or at least, implementing a parallel system)

[PhineasFreak]

I can definitely agree on the TF front: separation systems are one of the most frequent points of failures in space flight (from stages to fairings to simple protective covers).

Considering the structural strength: i believe this can be adjusted via the breakingForce and breakingTorque of the part (lower values make the nodes have less structural strength) but i am not sure if KJR completely overrides fields like these in-flight.

[ppboyle]

I've worked with Unreal physics before, and Unity uses PhysX under the hood as well - I've also seen how Rockets behave without KJR. Both of those make think that we'd MUCH MUCH rather use a home cooked system for calculating structural failure than really on the joint system built into PhysX

[NathanKell]

So, todos:

1. Someone needs to go through and remass decouplers.
2. Someone needs to go through and re-cost decouplers.

For now, I'm happy with proc interstage bases' masses/costs. Others should probably be brought in line with that. Note that the thinner PLF base / interstage bases have lower masses but higher costs.

I agree about upgrades. However that will have to wait for 1.3, since they need to cost funds.

I obviously agree about TF, but TF needs rewriting before I can contemplate adding it to more parts...

For vehicle complexity / launch costs, what I am thinking is this: add two new KCT variables, total number of vessel stages and total number of staged parts. We can then use those to increase launch costs based on complexity (this will be another reason to not cluster engines btw). It might be something like, 0.5 + (num stages x 0.04) + (num stageable parts * 0.0625)

[Bornholio]

Will make a table of decouplers as a start for RP-0 work (Should this actually be a RO point?)

[Bornholio]

https://docs.google.com/spreadsheets/d/1UQdcSof0cj4b6ddnii54C1rblAYErff4kkO40Y9irug/edit?usp=sharing

[linuxgurugamer]

I'm wondering if decouplers are really needed. In real rockets they don't have separate decouplers. The decouplers are built into the rocket tanks themselves. I'm thinking of writing a mod to add integrated decouplers to all tanks semicolon default would be not included, would use a the right click menu to enable the decoupler, which would also add weight to the tank. Does this make sense? Would you be interested in the development of this mod?

[linuxgurugamer]

The advantage to having the decouplers integrated would be a reduction in the park count, which everybody usually approves up. Also, it just looks weird when you decouple a stage to have this decoupler just floating around there, and this would eliminate that weirdness.

[ppboyle]

Tbh how decouplers look never bothered me that much - having leftover pins there after it decouples also doesn't bug me, though the fact that they're always black/grey while rockets are mostly white is a bit weird.

The only thing that truly bugs me about how decouplers look is that there's no reasonable way, other than with struts(which are also always black/grey), to couple something at two points - usually boosters are coupled at at least two points near the top and bottom for structural stability, afaik

[Theysen]

I'd get rid of the flat shaped decouplers alltogether and go with the interstage plates as major part for decoupling.

[ppboyle]

Theysen, how would you do parallel boosters then?

[PhineasFreak]

@ppboyle I assume that @Theysen was talking about the in-line decouplers and not the radial ones.