Why does my rocket keep exploding?

[Clear Air Turbulence]

Hi all

I hope I am not posting in the wrong forum with this. I am hoping for some mentoring in this flying to minmus thing.

I am in sandbox mode.

Basically, I flew successfully to Minmus in an adapted version of the Kerbal X (just a bit more fuel for the lander) & landed safely.

But, (long story short) Jim Kerbal somehow bumped into the craft and tipped it over. Attempts to right it failed and the thing exploded. I have three surviving Kerbals on Minmus, and I need to send up an unmanned craft to rescue them.

So now I am trying to get a lander up there and back. My problem at this stage is building ze rocket.

Three components:

Launcher (asparagus staging, mainsail engines)

Transfer (Rockomax & Reliant)

Lander (with plenty of fuel to get back)

And because I am a n00b, I am still using mechjeb (blush)

Picture below.

My problem is this: the thing behaves oddly in flight.

Initially it got off to to a grand start and managed to get to the top of the atmosphere without exploding.

But at the last stage of the launcher, i seemed to lose steering authority, or whatever it is called. The nose just dropped and we headed back home, going back into the atmosphere at more than a km per second. Not good.

I added some fins to the central stage, making sure they were not tangled up with the asparagus trappings.

But now the thing explodes violently the moment the first stage (the boosters to right and left in the picture below) detaches.

I can't figure out why.

They cut off and fall away as they should, and then suddenly there is a huge explosion & then boosters 3 and 4 zoom off on their own, ram into the top stage, and blows everything to kingdom come.

Any advice would be appreciated?

Edited September 20, 2015 by Clear Air Turbulence

[The_Rocketeer]

Pic's broken.

[Signo]

The pic does not work (ninja'd), so I will give my blindfolded 2 cents: fins on the central stage are probably enough to solve the spin issue, if not you may want to try to lower the "asparagi".

The sudden unexpected disassembly may be due to a collision of the decouplers (or anything else). You may want to try different decouplers (the TT-70 gives you a little bit of extra room to operate) or you may try to rotate the new fins 180°, to clip them inside the stage. This is a kind of gamey solution but you will then avoid collisions while keeping your aquired stability.

I hope this could help.

[Clear Air Turbulence]

Sorry, fixed.

My one worry is about that long fuel line (see second pic)

[racnic0404]

Retro Separatrons on the boosters staged at the time of separation would be my suggestion to prevent collisions. Not sure about the other problem.

[Araym]

If i got the problem (half way to space your rocket spins wildly, and it starts to plunge back to kerbin), my 2 cents:

even if "stock aero" is plenty forgiving, some rocket configurations eat too fast their fuel.

If your lower stages become too light, and upper stages/payload is too heavy, it is inevitable.

How to avoid it: obviously, more "control autorithy" (more SAS/RCS) to take your rocket in trajectory could be useful, but looking to the stats you are displaying thru KER, I could see a common problem to a lot of rocket: TOO MUCH TWR.

Your first two stages goes from 2.00 upwards: this means that they eat A LOT of fuel much faster than it will be needed, imbalancing your stacks.

In VAB, reduce your engine by limiting their thrust (right click on them, then slide down their limiter) to stay at a good 1.2 TWR, for low atmospheric stages, and even to 0.9/1.0 for stages meant to operate in higher atmosphere (leave at full power the lander engine, at least in the 4/5 TWR range for the moon surface: even if you are going to land in a very low gravity moon like minmus, an extra kick could help you to slow down fast and good).

Returning to te Kerbin ascent problem: 1.2 TWR for atmo stages means for you (and also for mechjeb) more time to use the first stages fuel, less dV losses for drag (if you are using mechjeb ascent autopilot, set the max Q dinamical pressure between 20k and 30k psa - I stay lower for bigger rockets, higher for already light ones) so they stay longer in an heavy-bottom configuration in the ticker part of Kerbin atmosphere, meaning that they can ascent in a more, quietly way.

If instead you are flying manually, the lower TWR allow you to better manage power (take a look on KER stats during flights, and lower throttle during ascent to stay oat max 30m/s accelleration).

You should then arrive in orbit without problem (obviously, if you do not point to low your rocket, tipping upside down because you make a too inclined gravity turn )

[Goomblah]

Here's my guess - Boosters 3 and 4 are not strutted adequately. I'm not sure, but that fuel line may be giving the 3 and 4 boosters some structural integrity, which they lose when the first two boosters are staged and the fuel line is severed. I've had similar issues with mysterious explosions upon staging, and MORE STRUTS generally solves the problem.

Also, that rocket is really overbuilt for what it's trying to do. The 3-Kerbal command module is insanely heavy (though the thermal and impact resistance are nice), and I don't think you should need monopropellant for anything, since you aren't docking and MJ is probably going to land it for you (right?). Even without RCS, MechJeb can land on target quite nicely, easily to within a few meters of the target. That 750 RCS is a lot of mass that you really don't need, and the same applies to the command pod. My suggestions:

-Use a Hitchhiker module instead of the command module. MJ is providing control anyways, and the Hitchhiker weighs half as much as that command module, and seats one more (which won't matter). You could alternatively use the Mk2 passenger module, which is even lighter, but that is a lifting body and can do weird things to your CoL. You will need to put dedicated reaction wheels on your craft, though, since the Hitchhiker and passenger modules don't have any.

-Halve the size of your transfer stage, at least. You really only need a Rockomax 16 (quarter orange tank), plus a Poodle engine for your transfer stage. Those weight savings on your lander will really help.

-Your ascent stage can be a lot smaller, simpler, and cheaper (though it looks like you're in Sandbox, so that won't matter). Replace your asparagus with an orange tank and a quarter orange tank with a Mainsail, and then stick 6 Kickback (the largest) SRBs on it. Put 6 tailfins in between the boosters, ditch the second set you have now, and you should be set. Ride the boosters with the mainsail off, stage the boosters and activate the mainsail, and then you should make it to orbit without using the transfer stage. Even if you do dip into the transfer stage's dV, with all the weight savings your lander stage should have an insane amount of dV, and should be able to do the round trip no problem.

[wumpus]

Transfer (Rockomax & Reliant)... uh,oh.

This happens when you drop the orange tanks and mainsails and are firing *all* remaining rockets? Quick guess is that you are trying to re-create the pendulum fallacy. Your center of thrust gets ahead of you center of mass and your rocket is no longer stable. This may be masked by the fins in the atmosphere and when they they are no longer stabilizing the rocket...

Some quick fixes: replace the reliants with LV-909 "terriers". If you manage to circularize, turn the terriers off before making your transfer burn (assuming the center of mass stays behind the center of thrust for the entire circularizing burn).

If that doesn't work (and the issue is indeed the pendulum fallacy: make sure this is the case by removing the booster stage in the VAR, lowering the fuel level to about where you lose control and check if the center of trust is above the center of lift), you can either just replace the terriers with a rokomax poodle (this will fix a bunch of aero problems as well) and work from there or you can try to burn until you are about to become unstable (many kerbals died to bring us this information...) and then turn off the terriers and fire up mechjeb.

Description of the pendulum fallacy here:

PS. Struts are a good thing. Even spaceX knows they need "Moar [unfailed] struts!". But my guess is that strut failure would occur well before leaving the atmosphere if that was the problem.

[Scarecrow88]

Have you thought of simplifying your design. I needed a rocket for exactly the same mission, rescue 3 Kerbals on Minmus, and came up with a SSTO, 2nd stage transfer and 3rd stage lander/return craft, and it works with no problems.

Engine on the 2nd stage is a Skipper, and on the lander/return stage a 909 Terrier.

[Superfluous J]

That ship is really large. You don't need mainsails (and surely don't need 5!) to reach Minmus with anything less than a Grand Tour mothership, and even then it may be too much.

Make your rescue ship a lot smaller, and the lifter's size will decrease in kind. I see you have KER. Build a smaller lifter (I see you're using Terriers, they're great. Try to make your lander with only 1 Terrier) and then build a lifter/transfer vessel RESTRICTING YOURSELF TO 1.25M PARTS. Get your m/s about to what you have on that guy, and marvel at how much smaller the new ship is. Just remember to not sacrifice control or power generation and you'll be fine.

Regarding flipping, Mainsails are very heavy and the front of your rocket is a bit draggy. Regarding the side boosters exploding, the only thing can think is that perhaps the mainsails are so powerful that they're causing some sort of stress breakage in the struts or decouplers, once you toss the first 2 and lower the total mass of the rocket significantly.

Regarding landing sideways on Minmus, I feel for you We've all been there (I still do it). A well-made rocket should be albe to right itself on Minmus though, without using the engines. Just rotate it with WASDQE so it's upright.

I'm assuming you stranded 3 guys on Minmus? You may have bitten off more than you can currently chew. I suggest making a one-man rescue vehicle to get your space legs, and then a 2-man rescue vehicle to pick up the other 2 guys. KSP is a game of (a near-infinite amount of) baby steps.

Edited September 17, 2015 by 5thHorseman

[xtoro]

What does it say after the explosions when you press F3 to bring up the flight log? My guess is that you're separating the tanks on top and bottom first, and right after separation, MechJeb is changing your pitch causing one of the tanks to collide with either the center tank or its engine, or both. MechJeb sometimes makes drastic pitch changes after separating large stages because of the big change in TWR. It can also overcompensate for the sudden change in CoM causing the rocket to pitch downwards.

[Thegamer211]

I think you should place aerodynamic caps...

[Clear Air Turbulence]

Wow, thanks for all the help.

I have used the TT-70s but it still blows up.

As far as I can see the empty booster tanks fall away and then hit the main ship on the way down. Then everything goes pear-shaped, if pear-shaped means tons of liquid fuel igniting at the same time.

Scarecrow, I like your ship. I don't like aspartamine staging; so finicky.

I will look at your design.

But how do you ensure your lander has enough delta-v to come back? According to my calculations, I need about 1950 m/s to land on Minmus, pick up my lost Kerbals, and return them home to the Mother Queen.

- - - Updated - - -

What does it say after the explosions when you press F3 to bring up the flight log? My guess is that you're separating the tanks on top and bottom first, and right after separation, MechJeb is changing your pitch causing one of the tanks to collide with either the center tank or its engine, or both. MechJeb sometimes makes drastic pitch changes after separating large stages because of the big change in TWR. It can also overcompensate for the sudden change in CoM causing the rocket to pitch downwards.

"TT-70 Radial Decoupler collided into Rockomax Fuel Tank

Rockomax Fuel Tank Collided into Rockomax Fuel Tank

AV-R8 Winglet collided into Rockomax Fuel Tank

Structural failure on linkage between Rockomax Fuel Tank and Rockomax Fuel Tank"

and so it goes

Edited September 17, 2015 by Clear Air Turbulence

[Padishar]

The first stage puts him in orbit , the second does the 900 odd m/s burn to minmus (and probably the capture burn into minmus orbit and possibly the deorbit burn too) and the landers stage just does the rest. It should take a lot less than 1950 to land, take off and return to kerbin.

[Clear Air Turbulence]

Regarding landing sideways on Minmus, I feel for you We've all been there (I still do it). A well-made rocket should be albe to right itself on Minmus though, without using the engines. Just rotate it with WASDQE so it's upright.

I'm assuming you stranded 3 guys on Minmus? You may have bitten off more than you can currently chew. I suggest making a one-man rescue vehicle to get your space legs, and then a 2-man rescue vehicle to pick up the other 2 guys. KSP is a game of (a near-infinite amount of) baby steps.

I actually landed dead on, and without mechjeb. Just put the thrusters on really low and concentrated.

But then

(1) I found out the ladder did not reach the ground

(2) So Bill could not climb back up after planting flag

(3) So he ended up underneath the ship, jumping up and down

(4) And ship was a bit top heavy...

It was all very painful after the glowing feeling of accomplishment.

[ghpstage]

I actually landed dead on, and without mechjeb. Just put the thrusters on really low and concentrated.

But then

(1) I found out the ladder did not reach the ground

(2) So Bill could not climb back up after planting flag

(3) So he ended up underneath the ship, jumping up and down

It was all very painful after the glowing feeling of accomplishment.

You don't really need a ladder at all on Minmus, the jetpack has more than enough TWR to get you to and from a landers hatch.

[Padishar]

The gravity on minmus is low enough to get into orbit with the EVA jet pack (and even back to kerbin though you will burn up on reentry since 1.0) so you can just hit r to enable the jet pack and fly up to the hatch. It is still quite easy to knock your ship over if you hit it too hard so get some practice with the jets first.

[I_Killed_Jeb]

Wow, thanks for all the help.

I have used the TT-70s but it still blows up.

As far as I can see the empty booster tanks fall away and then hit the main ship on the way down. Then everything goes pear-shaped, if pear-shaped means tons of liquid fuel igniting at the same time.

Scarecrow, I like your ship. I don't like aspartamine staging; so finicky.

I will look at your design.

But how do you ensure your lander has enough delta-v to come back? According to my calculations, I need about 1950 m/s to land on Minmus, pick up my lost Kerbals, and return them home to the Mother Queen.

- - - Updated - - -

"TT-70 Radial Decoupler collided into Rockomax Fuel Tank

Rockomax Fuel Tank Collided into Rockomax Fuel Tank

AV-R8 Winglet collided into Rockomax Fuel Tank

Structural failure on linkage between Rockomax Fuel Tank and Rockomax Fuel Tank"

and so it goes

The solution is relatively easy in this case. Use separatron boosters when staging the tanks, they'll fall quite far away, clear from the main stack

[Scarecrow88]

Padishar is correct about my rocket. The first stage is enough to get to an 80k orbit at Kerbin. The second stage handles the transfer to Minmus, capture to orbit once there and has enough fuel for the de-orbit burn. The lander doesn't need much to land on Minmus and return to orbit, the largest burn required is to get back to Kerbin. If you set your Pe at Kerbin to 35k or so, you don't even need to burn to get captured, just let aerobraking do all the work. The lander has fuel to spare to complete the tasks assigned to it.

Edited September 17, 2015 by Scarecrow88

[Goomblah]

But how do you ensure your lander has enough delta-v to come back? According to my calculations, I need about 1950 m/s to land on Minmus, pick up my lost Kerbals, and return them home to the Mother Queen.

The thing with returning from a moon is that you don't need all the dV that you used to get out there - you just need enough to get back to the elliptical encounter orbit that you used to transfer to the moon.

Looking at a dV map, your Minmus mission should look like this:

Arriving at Minmus:

3500m/s to achieve LKO > 300 to match planes, assuming you start equatorial > 930 to raise Ap to Minmus orbital altitiude > 160 to brake into a low Minmus orbit

=4890 m/s

Landing and Takeoff:

180m/s minimum, plus 50m/s safety margin plus 50-80m/s margin for plane change and course alterations (MJ isn't too efficient at landings) = 280-310m/s > 180m/s to reach orbit again (MJ is much better at ascent)

=490 m/s

Return:

160-180 to escape Minmus > 10ish for course corrections > aerobraking into a slower orbit, or even simply landing straight from the elliptical orbit

=190 m/s

Total:

5570 m/s

Your current rocket has 7833 m/s, which is 2,263m/s more than you need, even with the decent amount of wiggle room this estimate provides. The landing-and-return stage doesn't have to do everything the transfer stage did in reverse; there's no need to Hohmann transfer back to Kerbin when you can aerobrake or simply land straight from your return. I generally make a few aerobraking passes to keep anything from getting too toasty; passing through the atmo at 40 at 3km/s should be safe, as long as you are facing retrograde (to make the engines and fuel tanks act as a heat shield). You may be able to dip a little lower even, but re-entry makes me nervous when Kerbals are involved, so I generally take it slow.

I was able to reach Minmus, land, and return to Kerbin with 1000m/s to spare using this:

It rides the Kickbacks to 10km, at which point you manually activate the Poodle while the boosters are still going; stage as necessary until you reach orbit. It takes a little while, given the low TWR of the Poodle stage, but MJ handled it just fine. The Poodle handles circularization, the plane change, and half of the Hohmann burn; the Terrier above it does everything else. I was able to safely return by passing through the atmosphere at 35km twice, without touching the 1km/s in the last stage's tank at all. Overall, it's much smaller, and for the purposes of Career, much cheaper.

Edited September 17, 2015 by Goomblah

[Clear Air Turbulence]

Thank you! Very useful. It's interesting how many ways there are to build a rocket

[MalfunctionM1Ke]

Well... you could launch 3 of These and bring your Crew back to Kerbin

Edit: You can also launch your Kerbals into an Orbit around Minmus with their Jetpack alone.

Edited September 18, 2015 by MalfunctionM1Ke

[Clear Air Turbulence]

Hi Goomblah, thanks for those ideas.

I have built a two-stager. Just a big central booster & the two kickbacks, plus my Even Heavier Lander, which I have adapted from the stock Superheavy.

I got into orbit, so I guess that was my first successful SSTO. Although something went wrong with the separator; the one kickback did not disconnect but stayed stuck on the launcher all the way up. Somehow we kept flying straight and we had exactly enough to circularize, so it served its purpose.

I now have a lander with just about enough d-V to get to minmus and back again, so it will be a test of my flying skills from here on. Much obliged!

[Friend Bear]

Agreed with Horseman. Though I did not read his entire post, if I repeat what he says, it all the more affirms that conclusion.

The rocket is too tall. You can do a couple things, but I would suggest taking out one of the inside orange tanks. if you want to make up for that lost delta v, toss in an extra non-orange tank on the first and second stages to increase that overall initial thrust duration.

Next, you are probably turning down too suddenly. I do not know mechjeb (and I suggest playing without it. I imagine it is nice when you know all about the game and want to get lazy, but it fails to teach you how to play it better. Trial and error will help you with these analyses) but it sounds too me like you do not have enough vertical velocity OR you have too much vertical velocity and you are burning out those asparagus too early, forcing a climb at an altitude under 60k with the single, internal, 3rd stage mainsail. This is not always a problem, but when your rocket is hugely tall and your Thrust to Weight ratio is very low, it causes issues. Mainsail is buff, yes, but two orange tanks in dense atmosphere is rough, especially with a large final stage sitting atop.

As you launch, immediately, and VERY DELICATELY, aim the rocket a degree or two East. From there on, let gravity turn you and use about half throttle. The idea behind the best, most stable climb to orbit is not speed (although you need it to qualify the climb since too slow a climb destroys any chance of horizontal acceleration a minute into your climb to orbit) but instead burn time. (as for speed; what I mean is that a low speed reduces the usefulness of your overall deltav, while a high speed can do the same. If you are going fast at low altitude, even near horizontally in preparation for orbit, you lost a lot of potential delta v because you fight atmosphere. On the other hand, if you go too slow and lose your asparagus without burning any horizontal, sure you got into space, but now you have to burn 2000+ delta v horizontally to get an orbit in a short window with a single engine because once you are at the peak of your climb, it becomes harder and harder to stay out of the atmosphere with a low horizontal velocity). But careful; this means your gravity turn will be more aggressive. You do not necessarily need to aim into your prograde the entire time. Back off a bit and aim more vertically if it is too aggressive. 45 deg is ideal at 35-45k altitude, for example. This slow, natural turn prevents the bend you are seeing.

Lastly, if you have to have a long rocket, use struts. Make the entire thing as solid as you can throughout.

Also, I suggest losing some final stage fuel. You only need a few hundred delta v. Use Engineer Redux. It does an excellent job at telling you per stage Delta v, weight, etc., without giving autopilot.

-----------------------

Just read your post about asparagus breaking away and tearing apart your ship. Couple things:

First, that lower throttle is important here, too. It gets you higher with less fuel so when you do break away, it is with less atmosphere resisting the ship and the falling away stages.

Second, this is where you need to point as directly in prograde as you can. If you do not, then the pieces will not drift delicately and directly retro to the climbing rocket. You could use sepratrons. Place them on either side in the center of mass (ish) of each asparagus and have them initiate at decouple. They cause the stage to shoot outward and away (or whatever direction you set) at decouple and get them out of the picture.

-------------------------------

Hi Goomblah, thanks for those ideas.

I now have a lander with just about enough d-V to get to minmus and back again, so it will be a test of my flying skills from here on. Much obliged!

I have always wanted to see if they have enough monopro to orbit from being stationary on minmus. Maybe you could try that? See if you can get a kerb to orbit minmus. That way, you don't have to land the shuttle and you can just rendezvous in orbit!

I don't think he is using infinite monopro cheat?:

Edited September 18, 2015 by Friend Bear

[michaelsteele3]

Hi all

Basically, I flew successfully to Minmus in an adapted version of the Kerbal X (just a bit more fuel for the lander) & landed safely.

But, (long story short) Jim Kerbal somehow bumped into the craft and tipped it over. Attempts to right it failed and the thing exploded. I have three surviving Kerbals on Minmus, and I need to send up an unmanned craft to rescue them.

Well obviously Jim was carrying out his plan to take over and rule Minmus but didn't think it through very well. Any person named Jim is always a shady person.