Ascent paths, gravity turns, SRBs

[CosmicCharlie]

If I increase my AoA past the default 5% (to 8%) my SRBs crash into my ship when I jettison them.  How are you people making gravity turns that start under 10km and shallower ascent paths without shedding your SRBs first?

[GoSlash27]

CosmicCharlie,

 Your SRBs colliding with your ship is a result of how you have them attached to your stack, not the flight profile.

 We would need to see some pics in order to help you troubleshoot your problem.

Best,

-Slashy

[Stone Blue]

How many srbs, and how is your craft oriented when you jettison them?

If only two srbs, make sure you are oriented so they are on the north/south facing side of the craft, not east/west (if you were flying horizontal, this would be ground/sky)....

If you need to rotate 90° so the srbs are off to the sides, and not under/on top of your craft, do the roll within the first 200m of launch...

Then, it depends on where you placed the decouplers... Too high on the srbs, and the bottoms will kick inward and hit the bottom of your rocket... Too low, and the nose of the srbs will kick inwards... I find placing decouplers just above the center of the srbs is best... Unless you use a pair of decouplers... then placement is not as critical...

Also, you can try decoupling the srbs just a split second before they run out of fuel... That will help keep them parallel to your craft till you outrun them, then they can go whatever direction they want without contact with your craft...

[Snark]

+1 to seeing a screenshot.  Also, can you describe your ascent profile?  From your description, it sounds like you may be waiting a long time to start it.

Gravity turn should start almost immediately right off the pad-- I start mine as soon as I'm going fast enough to have a reasonably stable prograde marker, like 20 m/s, barely even a ship-length or two off the pad.  And then gentle gentle gentle, AoA kept small all the way.

By the time you're at 10 km, you should already be at around 45 degrees.

Some suggestions for avoiding having your radial boosters collide with the ship when you jettison them:

• Have the boosters mounted as low as physically possible on the radial decoupler, i.e. you want the place where the decoupler attaches to the SRB to be as high as possible on the SRB.  This helps because it puts the decoupler above the SRB's CoM, which means the decoupler force will not just push the SRB away from the ship, but also rotate it outwards, so that aero forces will also help with the separation.
• Mount the SRBs low down on your ship, so that your ship doesn't have far to travel before it's in front of the jettisoned SRBs and in no danger from them.
• Make sure that your next stage (i.e. LFO) activates with the radial decouplers that separate the boosters, so that you're thrusting away from them at full power the moment they come loose.  This gets you clearance sooner.
• If you're using Kickbacks, use the big Hydraulic Detachment Manifold to separate them; the extra decoupler force is worth it.

 

[linuxgurugamer]

What @Stone Blue says, except for one thing.  IMHO, when you have two SRBs, fly so they are horizontal to the ground.  That way, they go off to the side, rather than up or down where they can possibly hit you

[Snark]

40 minutes ago, linuxgurugamer said:

What @Stone Blue says, except for one thing.  IMHO, when you have two SRBs, fly so they are horizontal to the ground.  That way, they go off to the side, rather than up or down where they can possibly hit you

+1 to this.  Not just for collision purposes:  flying with radial boosters above/below you rather than on left/right can also lead to aerodynamic instability due to asymmetric aero forces.  Makes it harder to control, easier to flip out.

[Stone Blue]

51 minutes ago, linuxgurugamer said:

What @Stone Blue says, except for one thing.  IMHO, when you have two SRBs, fly so they are horizontal to the ground.  That way, they go off to the side, rather than up or down where they can possibly hit you

Actually, thats EXACTLY what I was trying to say.... but I always have trouble clearly communicating my thoughts...Also, me being longwinded does not help...lol

[CosmicCharlie]

Thx for the responses.

So the boosters are much too high on the craft and too high on the decouplers I think, after reading them. Might've been able to figure that out myself if I wasn't so frustrated.

Quote

Also, can you describe your ascent profile?  From your description, it sounds like you may be waiting a long time to start it.

Well, I have been going vertical until it's time to turn to circularize, but I thought I'd try some more efficient flight paths.  Not as early as 100m, around 5km.  And, if you limit your AoA to 5% (do you?), you can't follow any kind of appreciably curved flight path. 

Quote

you can try decoupling the srbs just a split second before they run out of fuel...

I have tried this a very few times with explosive results, guess I'll try some more.

 

Thanks again.

Edited December 26, 2015 by CosmicCharlie

[Snark]

2 hours ago, Stone Blue said:

If you need to rotate 90° so the srbs are off to the sides, and not under/on top of your craft, do the roll within the first 200m of launch...

Then, it depends on where you placed the decouplers... Too high on the srbs, and the bottoms will kick inward and hit the bottom of your rocket... Too low, and the nose of the srbs will kick inwards... I find placing decouplers just above the center of the srbs is best...

I'd suggest that it's never a good idea to do a roll like that on launch. It uses precious time that will unnecessarily delay the start of the gravity turn, and it's easy to introduce wobbles or misalignments that will exacerbate any control problems your rocket might have.  Instead of doing a roll, just rotate the whole rocket in the VAB before launch, so it's facing the right direction (east, not north).  No roll needed. [Insert "why doesn't the game make this the default" grumble here.]

As for SRB attachment point, I would contend that it should be as high on the SRB as possible, to kick the nose outward. If you have a problem with "the bottoms will kick inward and hit the bottom of your rocket", that means you're mounting the SRBs too high on the rocket. You want them as low as possible, both for aero stability and for collision avoidance.

 

 

 

[quyxkh]

Jettisoning them just before they run out of fuel can work too.

[Stone Blue]

59 minutes ago, Snark said:

I'd suggest that it's never a good idea to do a roll like that on launch. It uses precious time that will unnecessarily delay the start of the gravity turn, and it's easy to introduce wobbles or misalignments that will exacerbate any control problems your rocket might have.  Instead of doing a roll, just rotate the whole rocket in the VAB before launch, so it's facing the right direction (east, not north).  No roll needed. [Insert "why doesn't the game make this the default" grumble here.]

Interesting... I usually dont get wobbles or "drifting" until I get some speed up, and usually not till around 1000m.... I also dont usually start my grav turn till I DO have some speed built up, and usually around 1000m... So I usually dont have issues having time to do a 90° roll in the 1st 200meters...500m at most...

And yeah, why DO they start you facing North instead of East?... I'm sure I've seen it discussed before, but I dont remember the explanation...

[linuxgurugamer]

Well, for myself, if I need do roll, I usually do it about 5 seconds after launch, when my speed is around 50m/s.   I start the turn immediately after the roll is completed, around 100m/s.  I try to be at 68 degrees around 25 km, and at 45 degees at 35 km.  Seems to work well for me.

[GoSlash27]

CosmicCharlie,

 Looking at your pic, the SRBs are mounted too high in relation to the core and too high on the decouplers. Either will cause collisions.

 Not related to your collision problem... aerodynamics matters now. You'll have better results with a series stack design instead of a pancake design. It will lose less DV to drag and also be more stable. In addition (okay, so it *is* sorta related to your collision problem).. series stacking would eliminate the potential for collisions because the jettisoned stages are behind you instead of next to you.

 Finally... it appears to me that you don't need SRBs at all in this design. Your core stage seems to be able to easily lift itself. SRBs should really only be used for 1) adding thrust to get a weak core off the pad and moving or 2) acting as a cheap disposable first stage. You'd have better luck adding fuel rather than boosters. This would also eliminate the collision problem, since there would be no SRBs to collide with.

Best,

-Slashy

Edited December 26, 2015 by GoSlash27

[Stone Blue]

1 hour ago, CosmicCharlie said:

Well, I have been going vertical until it's time to turn to circularize, but I thought I'd try some more efficient flight paths.  Not as early as 100m, around 5km.  And, if you limit your AoA to 5% (do you?), you can't follow any kind of appreciably curved flight path.

Holy Cow!!... you mean you've been going straight up to +70km, THEN turning??? ... So basically you've been doing it WITHOUT a gravity turn at all..???

By keeping the AoA at 5°, doesnt mean you set it at 5° and leave it there... By 5°, you keep it there while slowly pitching the nose just in front of the prograde marker... SO you do a CONSTANT turn, slowly, and keeping the center of the Level Indicator (gold "V"), on the end of one of the three little yellow lines (about 5°) sticking out from the Prograde circle.... So by 10-15km, you should be pitched over, overall, between 30°-45°, depending on how big, heavy, and stable your craft is.... By 45km-50km, you should be pitched over, overall, almost horizontal...

And, ok... You have more than just a pair of opposing boosters, so you dont need to roll THAT craft....lol
Yeah, the decouplers are too low on the boosters (I usually place just a bit above center on the booster), AND also, the boosters themselves are much too high on the tanks...
I try to keep the nozzles on boosters even with the nozzles on the rocket, if not even a hair below... (So if the whole thing was sitting with the nozzles on the ground, ALL the nozzles would be on the ground...)
 

Edited December 26, 2015 by Stone Blue

[CosmicCharlie]

Quote

By keeping the AoA at 5°, doesnt mean you set it at 5° and leave it there

Finally understood that you follow prograde with the controls, thought it pitched over by itself. The reason it doesn't do it by itself is SAS?

I was using MechJeb and leaving "Limit AoA" to the default 5%. So that's just completely wrong and it should be more like 45%?

[Stone Blue]

21 minutes ago, CosmicCharlie said:

Finally understood that you follow prograde with the controls, thought it pitched over by itself. The reason it doesn't do it by itself is SAS?

Not sure exactly what you mean... No, SAS wont do the maneuvering itself.... Also, leaving it by itself with SAS off wont do the meaneuvering... You have to do it manually...

UNLESS:
 

Quote

I was using MechJeb and leaving "Limit AoA" to the default 5%. So that's just completely wrong and it should be more like 45%?

Did you mention before you were using MechJeb?... You should have mentioned that from the beginning...lol

No, Heaven's no!, Do NOT set "Limit AoA" to 45°....

Lleaving "Limit AoA" to 5° is fine, and even now recommended, in MJ...

I'm probably not gonna end up explaining this clearly, but:

MJ isnt "all-knowing", and if its NOT checked, but you have a craft design that ABSOLUTELY needs to limit the AoA, MJ wont know the difference, and it may manhandle the gravity turn how it THINKS it should be done... And it may crash the craft... Basically, if its unchecked, MJ may decide its fine to pitch over 45° mid-gravity turn... I have no idea how MJ does the calculations for autopilot...

I never really bothered with Limit AoA until 1.0 came out with the new aero model... Keeping a shallow AoA wasnt a big issue, since the (stock) drag and aero model was hugely forgiving...

I could be wrong, but I thought 1.0.5 brought in fine-tuning/stricter enforcement of drag & aero physics... ??

ANYWAY, with MJ, you can actually also tweak the ascent profile... Theres a button that opens another little window, that graphs the gravity curve as you tweak it...

Edited December 26, 2015 by Stone Blue

[Waxing_Kibbous]

+1 Snark and Slashy- I generally keep the bottom of the SRBs either level with or below the LFO engines, and the decoupler is about 1/4 of the way down on the SRB, with a strut holding at about 3/4 of the way down. *If you don't know how to do this use the translate tool in the VAB to move the SRBs down on the decoupler because it's tough to attach them properly anywhere but the middle

Edited December 26, 2015 by Waxing_Kibbous

[Pecan]

3 hours ago, CosmicCharlie said:

Well, I have been going vertical until it's time to turn to circularize, but I thought I'd try some more efficient flight paths.  Not as early as 100m, around 5km.  And, if you limit your AoA to 5% (do you?), you can't follow any kind of appreciably curved flight path. 

I have tried this a very few times with explosive results, guess I'll try some more.

 

Thanks again.

ARE you using MJ?  If so IT should be doing the turning, not you.  Engage autopilot on its Ascent Guidance module before you fire the engines to launch.  Without getting into the technicalities it will attempt to fly the programmed gravity-turn, but limit the AoA to the set limit (5 degrees is fine) if your vehicle is aerodynamically stable enough to do so.  As a rule you should be aiming for a TWR of around 1.4 - 2 across the burn-time of each stage.  Any higher than that and you'll accelerate too fast for the turn to i) be aerodynamically stable or even, ii) not to be possible (hardly turning at all before you're many km high).

As a general comment - that's a huge and complicated design, would you mind telling us what you're trying to do and what successes you've had already?  It is best to start with small and simple designs in order to learn how it all works before trying something like this.

[fourfa]

Just to add a comment to the above...  If you set an AoA limit of 5 degrees but don't limit acceleration, then in general you won't be able to follow an efficient ascent path. As the craft burns fuel and gets lighter, it will accelerate harder and harder and get to space before Mechjeb has turned very much at all.  (Again in general, highly dependent on design and staging.)

You can work the throttle by hand, to give it time to turn.  This is one of the factors experienced players are watching while piloting to orbit by hand.  Or you can have Mechjeb do it, by setting an acceleration limit.  If you have the typical 1.4TWR on the launchpad, limiting to 20m/s^2 (~2.0TWR, on Kerbin) works pretty well.  But you need to change or uncheck this box once you get out of the thick atmosphere - AoA matters little in the upper atmosphere (even while still showing lots of red glow), and your circularization burn is much more efficient at full throttle.  And I should point out - SRBs don't have throttle control after launch (you can monkey with thrust limiting in the VAB but that might be outside the scope of this question...)

Or you can allow a higher AoA limit - I often allow 15 degrees on my launches, with no acceleration limit (or rather, just a global 90m/s^2 for RP purposes, that being a reasonable limit for human pilots).  Unbalanced rockets with strange drag will NOT like this though, and it's a fraction of a percent less efficient than the perfect ascent.  But you don't have to watch the throttle and adjust or uncheck acceleration limits, can often just fire and forget.

With very careful attention to design, staging, drag, TWR at beginning and end of every stage, you can definitely make a rocket that makes a perfect gravity turn at 5 degree limit, no accel limit, and gets to space very efficiently with no flight profile changes on the way up.  That challenge alone is probably a lot of the fun for many players.  Personally I just slap the payload on an oversize V-SSTO lifter (which I recover at KSC on chutes) and get on with the mission.  Spending time babysitting every load to orbit to get there for 3300m/s instead of 3400, when my usual lifters have plenty to spare anyway, is personally not worth my game time.

Edited December 27, 2015 by fourfa

[Snark]

Having seen the screenshot:  No wonder you're getting SRBs hitting the central stack.  They're mounted way, way too high.  They should be far down your rocket, so that you're leaving them behind you almost instantly after jettisoning them.

+1 to comments about the overall rocket design, won't add to them here, they've already been well addressed.  Basically, "avoid pancakes."

5 hours ago, CosmicCharlie said:

Well, I have been going vertical until it's time to turn to circularize, but I thought I'd try some more efficient flight paths.  Not as early as 100m, around 5km.  And, if you limit your AoA to 5% (do you?), you can't follow any kind of appreciably curved flight path. 

If you wait until you're already 5 km to start your gravity turn, that's far too late.  Not fatal, it's not you-will-not-go-to-space-today too late, but it's very inefficiently late.

Starting the turn really early, practically right off the pad, means you can stick very very close to prograde the whole way, i.e. AoA stays close to zero the entire time.  It doesn't keep you from following a curved flight path-- you naturally follow a curved flight path, at near-zero AoA the whole time, because gravity does the work for you.  That's why it's called a "gravity turn."

(Although if you're already 5 km and going hundreds of meters per second before starting your turn, then I can see why you would think you need a high AoA to follow a curved path-- because then you would.  But then you're not doing a gravity turn, you're using your engines to muscle your rocket into a curved path.)

[OhioBob]

23 hours ago, Snark said:

I'd suggest that it's never a good idea to do a roll like that on launch. It uses precious time that will unnecessarily delay the start of the gravity turn, and it's easy to introduce wobbles or misalignments that will exacerbate any control problems your rocket might have.  Instead of doing a roll, just rotate the whole rocket in the VAB before launch, so it's facing the right direction (east, not north).  No roll needed. [Insert "why doesn't the game make this the default" grumble here.]

Just to add to the above..  If you are going to rotate the rocket in the VAB, disconnect the launcher from the payload, rotate the launcher 90 degrees, and then reattach to the payload.  Don't rotate the pod or probe core because this will change the orientation of the pitch and yaw axes on the launch pad.  For example, suppose you rotate the pod 90^o clockwise in the VAB.  Now to perform an eastward gravity turn you have to pitch down using the W key rather than yaw right using the D key, as is customary.  There is nothing wrong with this per se, but it can be disorienting if you're not familiar with it.

Also note that the default line of sight in the VAB is different than on the launch pad.  When you are in the VAB and looking out the door toward the launch pad, you are looking east.  When you click the launch button and move your rocket out to the launch pad, the default line of sight is now looking north.  So if you want your SRBs on the north and south sides of the rocket, they need to be on the left and right sides of the rocket when in the VAB and looking out toward the launch pad.  This is a quarter-turn different then the way they will appear on the launch pad.

 

Edited December 27, 2015 by OhioBob

[CitizenVeen]

If you have trouble with your SRB's colliding into your ship, it can help alot to mount 2 Sepratons (http://wiki.kerbalspaceprogram.com/wiki/Sepratron_I) on the SRB, and turn them (using QWEASD) so that the nozzles face the main body of you ship. Mount them next to the separtor, and have them fire at the same stage you set your decouplers. It causes the SRB's to be jettisond away from your ship. Rather then just gliding off it slowly when you only use the decoupler.

 

Edited December 27, 2015 by CitizenVeen

[Snark]

56 minutes ago, OhioBob said:

Just to add to the above..  If you are going to rotate the rocket in the VAB, disconnect the launcher from the payload, rotate the launcher 90 degrees, and then reattach to the payload.  Don't rotate the pod or probe core because this will change the orientation of the pitch and yaw axes on the launch pad.  For example, suppose you rotate the pod 90^o clockwise in the VAB.  Now to perform an eastward gravity turn you have to pitch down using the W key rather than yaw right using the D key, as is customary.  There is nothing wrong with this per se, but it can be disorienting if you're not familiar with it.

Also note that the default line of sight in the VAB is different than on the launch pad.  When you are in the VAB and looking out the door toward the launch pad, you are looking east.  When you click the launch button and move your rocket out to the launch pad, the default line of sight is now looking north.  So if you want your SRBs on the north and south sides of the rocket, they need to be on the left and right sides of the rocket when in the VAB and looking out toward the launch pad.  This is just the opposite of the way they will appear on the launch pad.

I would contend that doing the separate-and-rotate-just-the-rocket that you suggest would be more confusing, at least to me.

The case where it mattes how the rocket is facing is when the rocket has either bilateral or radial-two symmetry, e.g. has two radial attachments that are on the left and right (not top and bottom) sides of the rocket as it heads east.  Essentially, this means you're flying the rocket like a plane-with-really-crappy-glide-ratio.  For me, at least, this is far less confusing to have a command pod (and navball) that lines up with the left and right boosters, so I'm doing a pitch-down maneuver rather than a yaw-right maneuver.  Maybe other folks would do better, but for me, a procedure such as you suggest would get stuck sideways in my brain and cause cognitive dissonance.  It would be like trying to fly an airplane that has the cockpit mounted such that it's rolled 90 degrees relative to the rest of the plane:  physically just as flyable, but mentally hard to cope with.

It also means that I have a consistent experience with those cases where I actually am launching a spaceplane vertically off the pad-- that's one where you really do want to be doing pitch-down rather than yaw-right.  Planes handle pitch maneuvers much more gracefully than yaw.

Following what is (for me) the less confusing path of simply rotating the whole rocket kit-and-kaboodle so that it faces east rather than north, it does mean that I have to remember that for this ship, the gravity turn needs to be "pitch down" and not "yaw right".  But I find that the lesser of two evils, here.

As for your other point, concerning line of sight in VAB versus the launch pad:  You make a good point in terms of initial camera angle, but in the interest of avoiding potential confusion, it's worth noting that the physical orientation is the same for both.  The orientation of the ship is exactly the same in the VAB as it is on the launch pad; the only difference is which way the camera initially faces.

I think Squad would save players a lot of confusion if they would just change the behavior so that,

1. the default orientation of rockets is facing east, not north
2. the default initial camera view on the launch pad is facing east, not north

This would mean that VAB default camera, launchpad default camera, and ship default orientation would all match up, and bilaterally symmetric rockets (which are a very common design) would have the naturally favorable orientation as well.  Doing this would mean that all existing KSP players would need to get used to the idea that "I pitch down after launch" rather than "I yaw right after launch", but that would just be a one-time thing and I think wouldn't be that big an adjustment.

Edited December 27, 2015 by Snark

[Stoney3K]

With bilaterally symmetrical rockets (and, more specifically, shuttles), I always have the problem of my boosters catching on body lift and slamming into the flat parts of my craft when I decouple them. I've lost quite a few shuttle wings because my SRBs didn't want to jettison away from the orbiter, but instead, slammed themselves into the very fragile wings on the bottom.

[Pecan]

1 hour ago, Snark said:

...I think Squad would save players a lot of confusion if they would just change the behavior so that,

1. the default orientation of rockets is facing east, not north
2. the default initial camera view on the launch pad is facing east, not north

...

Pods used to face East (ie; hatch on the West) in the VAB and North, as now, on the pad, with the camera as now - East in VAB, North on pad.  Since that meant the rocket turned 90-degrees between building and launching it the arrangement was changed so the orientation is the same in the VAB as on the pad.  I don't expect much effort to be put into it again because it's nothing to do with Kerbal Flight Simulator, so all the 'plane fliers aren't impacted by it.