Decoupler force and struts

[Jarin]

This may be more of a "I need a mod" question than a "how do I" question, but on the off chance that there's a good answer out there...

Decoupler force apparently drops to 0 when you strut between an outer stage and an inner one (strutting boosters to the main body, etc). Is there a way around this? It seems really strange that decoupler force is such a big thing in their stats/cost, but it becomes completely superfluous when you build something of any reasonable size where the decoupler alone isn't enough structural stability.

EDIT: Further testing shows that this is not the old strut/decoupler bug that I'm running into but a new one with decoupler momentum which appears to be entirely unrelated to struts.

Edited September 3, 2014 by Jarin

[Taki117]

This is what seprotrons are for. Place a couple on your boosters and have them fire at the same time as your decoupler, problem solved.

[Jarin]

I know how sepratrons work, they're just an added annoyance to work around what appears to be a bug, or at least a VERY odd design choice.

Edit: Let me put it this way: it makes the hydraulic attachment manifold completely useless, so I can only assume it's unintended. But it's been in for as long as I've been playing, and nobody so much as gives the oddity a second glance...

Separatrons are quite good at damaging the rest of the ship, too, unless you're very good at placing them. Stage-testing to make sure I aligned my separatrons just right is probably my least favorite part of rocket-building.

Edited September 2, 2014 by Jarin
clarification

[cantab]

IIRC the bug is because the force is applied when the struts are still present, and only on the next frame are the struts broken.

The only workaround I can think of, besides the common sepratrons, is to use quantum struts and disconnect them before you decouple.

[Jarin]

Okay, this has been around forever now (hell, I've seen plenty of asparagus designs that wouldn't work without it), but there's almost no information on it available, no bug report on the issue tracker, and almost no acknowledgment that it exists. It's reached the "not a bug, but a feature" state of treatment, despite the fact that it renders decoupler part progression completely meaningless in most cases.

With the latest devblog regarding the issue tracker upgrades, it occurs to me... should this just be submitted as a bug?

[Reactordrone]

Decoupler force apparently drops to 0 when you strut between an outer stage and an inner one (strutting boosters to the main body, etc). Is there a way around this? It seems really strange that decoupler force is such a big thing in their stats/cost, but it becomes completely superfluous when you build something of any reasonable size where the decoupler alone isn't enough structural stability.

Does it? I can't say that I've ever noticed it happening.

[Raveneye]

This may be more of a "I need a mod" question than a "how do I" question, but on the off chance that there's a good answer out there...

Decoupler force apparently drops to 0 when you strut between an outer stage and an inner one (strutting boosters to the main body, etc). Is there a way around this? It seems really strange that decoupler force is such a big thing in their stats/cost, but it becomes completely superfluous when you build something of any reasonable size where the decoupler alone isn't enough structural stability.

I just assumed, physics wise, that attaching struts to the body of inner stages was a REALLY bad idea (attaching two things with a hard point that are supposed to separate)... I always attach them to decouplers on the inner stages, and blow those with the ones holding the outer rockets. Sometimes I'll need a separatron to make sure they don't hit the inner ones, but they do blow out.

[Frozen_Heart]

Though I've had this issue since 0.17, its no longer happening to me. Did they change something in the latest version?

[Geschosskopf]

With the latest devblog regarding the issue tracker upgrades, it occurs to me... should this just be submitted as a bug?

I would do so. 0.24.2 definitely changed decoupler force and I've heard rumors here and there that it's on Squad's radar already, but it never hurts to emphasize the point. Besides, the rumors may be false .

[goldenpsp]

I would do so. 0.24.2 definitely changed decoupler force and I've heard rumors here and there that it's on Squad's radar already, but it never hurts to emphasize the point. Besides, the rumors may be false .

It already has been, if anyone wanted to take the time and go look

[DrMonte]

Can someone replicate to Reactordrone comment above? He provides quite obvious prove that everything works as it should.

I didn't noticed strange behavior either. (If you can say that about rocket launch in KSP)

[Geschosskopf]

Can someone replicate to Reactordrone comment above? He provides quite obvious prove that everything works as it should.

I didn't noticed strange behavior either. (If you can say that about rocket launch in KSP)

Things work differently on the ground and in flight for some reason. Don't ask me why.

In flight, with radial decouplers, especially the TT-70, there's not much if any net outwards force if you tie the booster to the center with struts, and you pretty much have to do that to keep it from wobbling and/or breaking off. The decoupler force is pretty much absorbed in breaking the struts, so the booster remains in the same position as before, just not attached, after which wind and gravity dictate what happens next. And that usually means the top end of the booster will pivot in and collide with the central stack, often destroying it.

This is definitely a change from previous behavior. Prior to 0.24.2, I didn't have to use Sepratrons at all on 1.25m boosters, SRB or LFO, as they were light enough for the decoupler to toss them aside provided it was mounted about at the booster's CoM. 2.5m boosters, however, needed a single pair of horizontal Septratons at about their CoM to get sufficiently clear.

These days, however, I need Sepratrons for all sizes of side boosters. Further, they have to be mounted horizontally at the very top end of the booster to keep that from pivoting inwards. And if the booster isn't mounted vertically for some reason (say on the tapered SXT 4-5m lower stage), they need a 2nd horizontal pair on the bottom end as well.

[Kasuha]

I have not make any specific tests with radial decouplers and struts yet but I trust reactordrone's results. It's a good thing this has been fixed as it was kind of annoying.

The current problem (which might be related to the fix) is that the decoupler loses all speed as soon as it decouples. If you decouple just the decoupler, you can see that it does not follow the rocket, it just stays in place and starts falling. That's why the booster turns inwards - the decoupler gives it rather strong impulse against its motion. This is the case all the time until krakensbane kicks in when the ship exceeds 750 m/s - after that everything is fine again.

A wild guess is that the issue might be fixed if we mark the decoupler as massless? Haven't tried it yet. But since the decoupler becomes root part of the "ship" I'm afraid it won't help.

Another viable workaround might be using docking ports instead of decouplers. The downside is that their mass stays with the rocket, not with the booster. And you need to mount the booster axially to them, not radially - so best approach would be to put a cubic octagonal strut or girder there and attach the booster radially to that. And of course you have to stage using action groups. Pretty annoying IMO.

Edited September 3, 2014 by Kasuha

[Jarin]

The current problem (which might be related to the fix) is that the decoupler loses all speed as soon as it decouples. If you decouple just the decoupler, you can see that it does not follow the rocket, it just stays in place and starts falling. That's why the booster turns inwards - the decoupler gives it rather strong impulse against its motion. This is the case all the time until krakensbane kicks in when the ship exceeds 750 m/s - after that everything is fine again.

This seems to describe the problem perfectly. I had thought it was just weird aerodynamics causing my boosters to veer inwards instead of being pushed away, but on further testing, it does behave completely differently in motion from how it does on the launch pad. Note the 500m/s test veering much harder inwards (several explosions just after that screenshot that I didn't catch), and the 800+ m/s working perfectly again. Good to know!

Javascript is disabled. View full album

Edited September 3, 2014 by Jarin

[Claw]

Yes, this bug was found almost right after the release of 0.24.2. It's like Kasuha describes. I believe the problem is that the speed given to the decoupler at the time of decoupling is somehow wrong. The effect gets worse as you approach 750 m/s, then goes away after that.

The bugs thread in my signature has a link to our previous discussion on the subject.

Cheers,

-Claw

[Renegrade]

I scribbled up a little MM fix for radial decouplers - prior to 0.24 strutting across them completely negated any ejection force they had, and now that doesn't seem to happen (at least not in all circumstances), and instead a weird torque is applied at speed.

The fix will make 'em fall away as in 0.23.5; straight down. Sepratrons are recommended (although not strictly required).

[Geschosskopf]

Yes, this bug was found almost right after the release of 0.24.2. It's like Kasuha describes. I believe the problem is that the speed given to the decoupler at the time of decoupling is somehow wrong. The effect gets worse as you approach 750 m/s, then goes away after that.

You know, prior to 0.24 there was a speed-passing bug with the 3.75m stack decoupler. If you fired it below 750m/s, it got zero speed so collided with the engine above at whatever speed < 750m/s the rocket was doing at the time, usually resulting in a big explosion. It worked fine about 750m/s, however. Anyway, this got fixed in 0.24, which is when the radial decouplers went haywire. So if the radials have a speed-passing problem, maybe the fix for the TR-38-D went too far.

[cantab]

That .23.5 bug was due to said decoupler being physicsless. It was worse on Linux, where with certain strutting arrangements past the decoupler firing it would simply crash the game. I spent a bundle of time investigating only to find out the issue was already known and the fix was to make the decoupler not bleeding physicsless dang you Squad what were you thinking argh!

[Jouni]

I feel that decouplers that don't push the booster away should be the default option, as they make staging so much easier. Almost all of my rocket designs assume that the decoupler just drops the booster without pushing it outwards. As long as I use the TT-70 decouplers instead of the smaller ones, there is enough clearance between the rocket and the booster that I basically never hit the boosters.

[Claw]

You know, prior to 0.24 there was a speed-passing bug with the 3.75m stack decoupler. If you fired it below 750m/s, it got zero speed so collided with the engine above at whatever speed < 750m/s the rocket was doing at the time, usually resulting in a big explosion. It worked fine about 750m/s, however. Anyway, this got fixed in 0.24, which is when the radial decouplers went haywire. So if the radials have a speed-passing problem, maybe the fix for the TR-38-D went too far.

Yes, this is basically along the correct lines. There was a fix pushed in to 0.24.2 to fix broken stack decouplers. It resulted in breaking radial decouplers. Stack decouplers are actually still broken too but since they are on the stack, it doesn't cause parts to tip and collide.

The physicsless decoupler problem of the TR-38-D is actually a completely different problem not related to the physicsless status. It was because the part caused some internal vessel structures to go wacky when decoupling. For some reason turning the physics on bypassed whatever was broken.

Cheers,

~Claw

[DrMonte]

You know..., I've built up to 1000t lifters but I've never used a single Sepratron.