[1.8.x] Oh Scrap!- A ScrapYard based Part Failure and Reliability Mod 2.0.1 (07/12/2019)

[dlrk]

Is 1.4 backwards compatible to 1.4.5?

[Beetlecat]

5 hours ago, jospanner said:

Just a quick question, does this enable catastrophic failures of engines? *boom*

I don't recall anything dramatic/traumatic -- usually just part or performance failures. The Boom could be an overheat, or fuel line leak, but it often just takes out the engine component. For anything more grand to simulate a chain reaction type failure -- Kerbal Launch Failure Revived might be what you're looking for. It seems to just use a set percentage chance of complete failure on launch, so nothing that can be safeguarded, etc. Just adds !FUN!!!

[DasValdez]

Well, determined today that the spamming still happens even if the craft is attached to launch clamps. Huh.

[BobTheRocketeer]

8 hours ago, DasValdez said:

Spamming on part failure rolls before launch. We're still encountering the part failure attempt spam when we deploy a plane to the runway. Scenario is we roll out a plane, Alt-F12 for the console, and see Oh Scrap spamming failure rolls twice per second. We can stop it by firing up the engines and getting the plane rolling. Can you help us understand why it would EVER/in what craft situation want to spam multiple rolls per second? Again, not sure if this is related to the "World Stabilizer" mod confusing the craft situation, but it isn't clear to me why Oh Scrap would ever roll twice per second.

I don't know if what I've seen is related to this, but I have had planes that will either launch with no failures, or get like five or six in the first thirty seconds or so before I even fire up the engines. It might be related to the parts reaching end-of-lifetime but all at once, but it's very odd to see about half a dozen failures go off in rapid fire.

Also I may have seen something like the KCT/Scrapyard vessel recovery system not updating the generation of the parts, but I def cannot confirm yet (and I seem to no longer have time to test anything either).

[Beetlecat]

32 minutes ago, BobTheRocketeer said:

I don't know if what I've seen is related to this, but I have had planes that will either launch with no failures, or get like five or six in the first thirty seconds or so before I even fire up the engines. It might be related to the parts reaching end-of-lifetime but all at once, but it's very odd to see about half a dozen failures go off in rapid fire.

Also I may have seen something like the KCT/Scrapyard vessel recovery system not updating the generation of the parts, but I def cannot confirm yet (and I seem to no longer have time to test anything either).

A beginning (takeoff) is a very delicate time...

[severedsolo]

10 hours ago, DasValdez said:

Spamming on part failure rolls before launch.

It's a bug - I had thought it was just rolling but not actually doing anything with it, apparently that's not the case. Regarding the "twice a second" thing - that's just how it works. Once every half second, it will check if it needs to roll, if it doesn't, it goes back to sleep again - will get this fixed up today (I know what it is, I just forgot about it in the last build, and thought it was harmless up to now).

10 hours ago, DasValdez said:

@severedsolo can you give us any details on the logic by which generations are increased? Also, I'm curious as to whether or not parts which are launched but not recovered (e.g. expendable stages) should be incrementing. 

The logic is is just either a) it pulls it from it's saved data if it's seen the part before, or b) how many times ScrapYard says that part has been built new. Expendable stages should increment yes - I will take another look at the KCT/Recover to Storage thing, see if I can reproduce it.

10 hours ago, DasValdez said:

Still struggling with part inventory/Scrapyard not displaying or grouping by the generations. Manageable but the larger our inventory gets, the more of a pain it is to find usable parts. Would be great it the UI (and this might be scrapyard feedback) allowed us to group by gen AND reliability rating... sometimes I don't care the gen as long as it's high reliability

Yes, I heard you the first 5 times. The answer is still the same - when I get to it. 

1 hour ago, BobTheRocketeer said:

I don't know if what I've seen is related to this, but I have had planes that will either launch with no failures, or get like five or six in the first thirty seconds or so before I even fire up the engines. It might be related to the parts reaching end-of-lifetime but all at once, but it's very odd to see about half a dozen failures go off in rapid fire. 

If the plane isn't moving yet then yeah it's the same bug.

7 hours ago, jospanner said:

Just a quick question, does this enable catastrophic failures of engines? *boom* 

It's possible, but you get a bit of warning. One of the engine failure modes is a "fuel line leak" - if you continue to run your engines for a little while after it starts leaking (about 10s IIRC), it will blow up.

[BlueTiger12]

On 12/24/2018 at 6:59 AM, strudo76 said:

 

This could also be integrated with KCT, where pre-flight check would appear as an individual item what takes time to complete. That way it might lead to some interesting situations where you want to rush a launch by cancelling the pre-flight check item, and run the risk of a higher chance of certain parts failing.

I really like this idea. The thing about BARIS i don't like, is that this mechanic is disabled when using KCT.
Maybe this Pre-Flight test could cost money and take time, and based on the amount of both you spend, your reliability for this flight is increased (perhaps Part-based?) and maybe also there is a slight influence on further builds. Though i think it shouldn't increase a generation for all time.
With the combination of money and time, also for non KCT Users, there would be a tradeoff, but i personally think the additional time would have more impact then money.

[severedsolo]

On 1/8/2019 at 8:23 PM, DasValdez said:

Spamming on part failure rolls before launch.

Fix for this is now live on the build server - I want to investigate the KCT issue before I do a proper release

Edit: Scratch that. Apparently the build server isn't doing new builds any more, I guess magico turned it off or github is being stupid. Don't use the one on the server, it's nearly a month old. 

Edited January 9, 2019 by severedsolo

[dlrk]

@severedsolo

Not sure if you saw my previous comment, but is the current version backwards compatible to 1.4.5?

[severedsolo]

4 hours ago, dlrk said:

@severedsolo

Not sure if you saw my previous comment, but is the current version backwards compatible to 1.4.5?

Probably. AFAIK nothing broke between 1.4 and 1.5 so chances are it will work.

[DasValdez]

Is a Gen 8-9 Terrier (909) that been tested supposed to have a lower safety rating?

We have a Gen 11 terrier Tested: True with 1 previous use that only gets a Engine Safety Rating of 9.

Intentional or config issue? Thanks!

[severedsolo]

15 hours ago, DasValdez said:

Intentional or config issue? Thanks!

Uh, kind of intentional, but it's not supposed to drop off that quickly. Space engines have a very short ExpectedLifetime but a much better base failure rate  because I didn't think it was fun to have to test a space engine every time - but for some odd reason I hardcoded that... will slip it into a config on the next update.

[severedsolo]

Oh Scrap 1.5 Released

• Recompiled against KSP 1.6 (1.5.x should continue to work fine)
• Fixed issue where failure rolls would be spammed in prelaunch
• Exposed baseFailureChance and expectedLifetime for Space Engines in the MM config for EngineFailureModule
• Added bash script to automate release build process (only relevant if you are compiling from source, but it's in the Github Repo so mentioned it here)

 

[Agustin]

is it compatible with RP.1? it seems better than test flight

[Beetlecat]

Are static tests back on the menu (landing pad) yet, or do I need to get clever with test vehicles that launch then brake immediately to retain the equipment?

[severedsolo]

8 hours ago, Beetlecat said:

Are static tests back on the menu (landing pad) yet, or do I need to get clever with test vehicles that launch then brake immediately to retain the equipment?

Sorry mate I forgot about that... raised under #11 so I don't forget again

[Beetlecat]

12 hours ago, severedsolo said:

Sorry mate I forgot about that... raised under #11 so I don't forget again

Much appreciated-- sorry I did't think to do that already

[DasValdez]

23 hours ago, Beetlecat said:

Are static tests back on the menu (landing pad) yet, or do I need to get clever with test vehicles that launch then brake immediately to retain the equipment?

Are you talking about deploying a part to the pad and recovering it to get "tested" status for parts? That's working right now without getting an MET/2ms... do it all the time in our UHC playthrough on my stream.

If you mean increasing generation without an MET, that doesn't work. Feels fair the way it is right now... hopefully you don't want to increase generation just by deploy and recover without moving the craft at all?

I am planning on a UHC static test stand that will require actually moving the part to it to perform the test. EG, launch a part with a docking port on it, use a truck to move it to the stand, and then actually test it on a stand. That would increase generation and feels like a fair time investment to increase generation without actually launching vehicles. Will roleplay it as we can only get to gen 5 or so that way, and anything higher requires actual use.

[severedsolo]

15 minutes ago, DasValdez said:

Are you talking about deploying a part to the pad and recovering it to get "tested" status for parts? That's working right now without getting an MET/2ms... do it all the time in our UHC playthrough on my stream.

Huh. Maybe I fixed it and don't remember doing it.... will take a look at it tomorrow

[BlackHat]

1 hour ago, DasValdez said:

Are you talking about deploying a part to the pad and recovering it to get "tested" status for parts? That's working right now without getting an MET/2ms... do it all the time in our UHC playthrough on my stream.

If you mean increasing generation without an MET, that doesn't work. Feels fair the way it is right now... hopefully you don't want to increase generation just by deploy and recover without moving the craft at all?

Ok What does MET and MET/2ms mean?

[Beetlecat]

 

2 hours ago, severedsolo said:

Huh. Maybe I fixed it and don't remember doing it.... will take a look at it tomorrow

Shoot-- this may be my bad for not even testing it out yet... :I

1 hour ago, BlackHat said:

Ok What does MET and MET/2ms mean?

Mission Elapsed Time(?) and 2ms means the craft has to achieve at least that speed, too. Essentially the craft needs to be "launched." I was asking, because I would previously run trial firings of rockets and boosters on the launch pad, letting them burn all the way, and run out of fuel before recovering them. A recent update of Oh Scrap! enforced the 2m/s rule, which meant a rocket strapped to the launch clamp would never go that fast.

2 hours ago, DasValdez said:

Are you talking about deploying a part to the pad and recovering it to get "tested" status for parts? That's working right now without getting an MET/2ms... do it all the time in our UHC playthrough on my stream.

If you mean increasing generation without an MET, that doesn't work. Feels fair the way it is right now... hopefully you don't want to increase generation just by deploy and recover without moving the craft at all?

I am planning on a UHC static test stand that will require actually moving the part to it to perform the test. EG, launch a part with a docking port on it, use a truck to move it to the stand, and then actually test it on a stand. That would increase generation and feels like a fair time investment to increase generation without actually launching vehicles. Will roleplay it as we can only get to gen 5 or so that way, and anything higher requires actual use.

This is a weird and brilliant way to be OCD about part / engine testing.   --Though only doable once docking ports are unlocked. And no, I wasn't meaning to cheat the system by just deploying/recovering. My usual engine tests run for a couple minutes and varying thrust levels to allow for the expected failure or explosion.

I'll have to use one of my rigs from the previous game and see if it still works with the current Oh Scrap! and part tested status.

Edited January 16, 2019 by Beetlecat

[TGApples]

Personally, I don't feel the need for static testing. I'm fine with that abstraction being rolled up in KCT.

For me the main good way this mod could go would be more failure modes: mostly geared toward softer failures rather than mission ending/kerbal killing failures. I may even download the source and have a tweak. For example:

• Parachutes that don't fully deploy rather than failing outright.
• ISP/thrust reduction/large heat generation in engines and SRBs. I notice from the source there's a thrust reduction failure mode for liquid engines though I've never had it in game. These would make it harder to reach orbit and may trigger an abort, but right now off axis engine failures (especially SRBs) are hard to abort nicely from. If it was just a 20% reduction in thrust it'd still really hurt, and might force a hard abort, but would be a lot more managable.
• Alternator failures on engines.
• Clogged pipes in fuel tanks. Right now fuel tank failures are really brutal on later stages or in cruise. During launch it's not too hard to just flip the priorities a bit to drain that tank quickly.... but after launch it's much more deadly. "Clogged pipes" could maybe limit fuel flow (but that sounds hard to mod), or just disable flow.
• I don't know if it's easy to modify reaction wheels by pitch/yaw/roll, but failing one axis may be interesting.
• Power consumption increases for SAS/antenna/probe cores due to "bad wiring" or something.
• Drills/refineries/fuel cells failures.

 

I guess several of these are a bit more complex than just switching something to "off", but it would add a lot more flavour to the mod, and increase the "every mission is different" and "plan for failure" feelings the mod currently does a great job at adding.

Edited January 16, 2019 by TGApples

[severedsolo]

42 minutes ago, TGApples said:

Parachutes that don't fully deploy rather than failing outright.

Doable (I think, fairly certain I can just rejigger the existing code) and I like the idea, so will probably happen.

43 minutes ago, TGApples said:

ISP/thrust reduction/large heat generation in engines and SRBs.

Definitely doable and on the list (if/when I ever get time lol)

43 minutes ago, TGApples said:

Alternator failures on engines.

Eh... I keep thinking about this, but the conclusion I come to nearly every time is "not worth it" - alternator failure is really only a problem at launch, and chances are whatever you are launching has enough charge to get it to orbit in the 2 minutes or so it takes anyway, at which point your alternative power source would kick in. (If you disagree, please say so though - I would like a reason to change my mind)

45 minutes ago, TGApples said:

Clogged pipes in fuel tanks. Right now fuel tank failures are really brutal on later stages or in cruise. During launch it's not too hard to just flip the priorities a bit to drain that tank quickly.... but after launch it's much more deadly. "Clogged pipes" could maybe limit fuel flow (but that sounds hard to mod), or just disable flow. 

Disabling flow is possible, I think DangIt does something similar. Like this one too.

45 minutes ago, TGApples said:

I don't know if it's easy to modify reaction wheels by pitch/yaw/roll, but failing one axis may be interesting.

Uh... not sure this is possible but I'll look into it

 

46 minutes ago, TGApples said:

Power consumption increases for SAS/antenna/probe cores due to "bad wiring" or something.

Really like this one. Will be looking into this too.

Cheers for the suggestions (and for those of you who are going to point out that you already suggested SRB thrust failures etc, thank you too if I didn't acknowledge it)

[DasValdez]

20 hours ago, Beetlecat said:

This is a weird and brilliant way to be OCD about part / engine testing.   --Though only doable once docking ports are unlocked. And no, I wasn't meaning to cheat the system by just deploying/recovering. My usual engine tests run for a couple minutes and varying thrust levels to allow for the expected failure or explosion. 

Test stand test roleplay is actually really fun, and as it is it's got a pretty cool balancing effect: I think it's going to come more into play as rockets get bigger and more expensive. Early in career, it's easy to just advance gens by launching missions, which tend to be small/rather inexpensive, not a huge penalty on (uncrewed) failure. As you move to larger, more expensive rockets, it becomes more prohibitive to just throw rockets away for generations, so I'm all about a static fire test stand.

Already tested deploying a "stand"out of the SPH using launch clamps, but way off the runway so it doesn't get cleaned up. Stand is rooted to the ground, you can still deploy to runway, drive over, and integrate the part for testing. Hardwired to KSC electrical grid too haha 

Edited January 16, 2019 by DasValdez

[Beetlecat]

2 hours ago, DasValdez said:

Test stand test roleplay is actually really fun, and as it is it's got a pretty cool balancing effect: I think it's going to come more into play as rockets get bigger and more expensive. Early in career, it's easy to just advance gens by launching missions, which tend to be small/rather inexpensive, not a huge penalty on (uncrewed) failure. As you move to larger, more expensive rockets, it becomes more prohibitive to just throw rockets away for generations, so I'm all about a static fire test stand.

Already tested deploying a "stand"out of the SPH using launch clamps, but way off the runway so it doesn't get cleaned up. Stand is rooted to the ground, you can still deploy to runway, drive over, and integrate the part for testing. Hardwired to KSC electrical grid too haha 

I'm really digging this idea! Do you mean you were able to place launch clamps off the runway-- or an actual craft you moved out of the way to avoid clean-up?

This goes really well with another thing I started to do, which was to house Kerbals in stationary basses/craft at the KSC, which forced me to deal with USI resources, and part failures of batteries, etc. and incur penalties using the Crew R&R mod, so I had to do crew rotation, too. Anything to make it wildly complicated even before making it to orbit...

Edited January 16, 2019 by Beetlecat