Jump to content

Rocket rotating during ascent - what can I do?


Recommended Posts

Ok guys and girls I looked on google for this noob question, there are tons of the question made with the same words I describe but all of them seems to describe a tip over and not really a rotation on its own axis. Soo I ended up asking here.

Here is the situatio: I launch, then I stage my boosters, then at around 350m/s and 7000m I start gravity turn toward EAST very gentle. As the prograde mark starts to follow the nose at just around 10 degrees, the nose have a tendency to point toward to NORTH, then I compensate to SOUTH and keep turning to EAST. There is a force that keep throwing the nose north and I compensate bringing nose south all this while it still incline to east as it should. Then the whole rocket start rotating in its own axis while I do this compensation. At first I was compensating with manual counter rotation, but made the piloting much more difficult, then I experimented just keeping compensating to south, when I did this after a few tension seconds the rotation stabilizes and I manage to point it directly east and achieve orbit.

Why is this happening?

Some people will say, if it is working why bother? Because I want to learn and make better rockets that is why.

Soo what you think?

Is there a mod that show me any deviation of center of mass/thrust/aerodinamics from the center of the rocket?

 

Thanks

UPDATE:

No autopilot, center of mass way ahead of center of thrust. However I do have just 4 chutes on just one side (because I want my main stage land on its side).

UPDATE 2:

I made a test eliminating all non radial parts, the rocket keep behaving the same way. 

Edited by felcas
Link to comment
Share on other sites

What's happening here sounds like a combination of a bunch of things. First off, it sounds like your rocket is not aerodynamically stable, which explains why it keeps deviating north. It wants to point in any direction other than prograde, and any deviation northwards, however small, will lead to an escalating precession in that direction. Exacerbating this problem is your gravity turn. You're starting it way too late-you should be starting it more around 100 m/s and 1 km up. Starting it as late as you do causes increased aerodynamic load on the rocket, which worsens the instability, and hence the precession. The rotation itself is probably caused by transfer of angular momentum to the least-moment spin axis (i.e. the longitudinal axis of the rocket). The spin stopping is probably due to SAS properly compensating for the spin once the rocket has left the denser parts of the atmosphere.

Or it could be something else entirely. Screenshots of your rocket, in the VAB and in flight, would be helpful in figuring out what's wrong.

Link to comment
Share on other sites

2 hours ago, felcas said:

However I do have just 4 chutes on just one side (because I want my main stage land on its side).

This is probably the root of the problem, leading to a small shift of the CoM and unaligned thrust. Aeroforces may also be at play. 

You may add something (e.g. Batteries) to balance the mass(and drag) of the parachutes.  

Link to comment
Share on other sites

Agree with the steely eyed goat - you've chosen one of the worst speeds to turn. Drag as a function of speed climbs and peaks from mach 0.8 to 1.2. By turning in the dense atmo at that speed, the ship get the most forces acting on it leading to instabilities SAS cannot compensate for. 

  • turn earlier or turn later
  • use a stronger SAS/add extra reaction wheels

The chutes - as long as you place them on top ('west') or bottom ('east') of the ship (looking at it horizontally) it should help with stability as you turn. Maybe put them lower down the part so the drag is effectively 'fin-like'. +/- add  a par of extra fins along that axis, just above your existing fins. No pics of your craft, so not sure if that advice is helpful.

Link to comment
Share on other sites

It's worth considering all of the offered advice, but like Kryxal said I have found that 3-way symmetry often causes this effect. You might try attaching your fins in 4-way symmetry instead.

With 3 fins only one of them can ever be aligned with EITHER the yaw axis OR the pitch axis, the other two will be trying to control a mixture of both and there will always be an uneven distribution of pitch and yaw control. And sometimes the pitch movement of one fin will counteract the yaw movement of another one. It's less than ideal.

With 4-way symmetry you will always have two opposite fins controlling mostly pitch, and the two other fins controlling mostly yaw.

Edited by HvP
Link to comment
Share on other sites

First of all thanks everyone for the prompt reply and help, and sorry to take some time to answer all you. 

For those who asked, here is a few pictures of my rocket. I will try to make the video tomorrow

One of them show in red the only components that are NON-RADIAL. With the exception of a remote control (the little orange box that extend airbrakes 10 seconds after staging) are the 4 parachute modules preciselly installed to make the rocket land on its belly (because if it lands standing it always roll, drop and explode).

I tried to take them off and make mroe tests but this seems to only aleviate the problem and not eliminating it.

fiiDBwF.png

LBlmalj.jpg

Al5ip9q.jpg

Forgot this one with the center of things

A9CXFLp.jpg

On 2017-6-17 at 10:11 PM, Kryxal said:

Pictures would help.  Incidentally, if you have things attached in 3-way symmetry, that can be a bit touchy.

I posted a few pictures. I am not using 3-way symetrical only quad sym. thanks

Link to comment
Share on other sites

On 2017-6-17 at 10:27 PM, IncongruousGoat said:

What's happening here sounds like a combination of a bunch of things. First off, it sounds like your rocket is not aerodynamically stable, which explains why it keeps deviating north. It wants to point in any direction other than prograde, and any deviation northwards, however small, will lead to an escalating precession in that direction. Exacerbating this problem is your gravity turn. You're starting it way too late-you should be starting it more around 100 m/s and 1 km up. Starting it as late as you do causes increased aerodynamic load on the rocket, which worsens the instability, and hence the precession. The rotation itself is probably caused by transfer of angular momentum to the least-moment spin axis (i.e. the longitudinal axis of the rocket). The spin stopping is probably due to SAS properly compensating for the spin once the rocket has left the denser parts of the atmosphere.

Or it could be something else entirely. Screenshots of your rocket, in the VAB and in flight, would be helpful in figuring out what's wrong.

Thanks for all the info.

1-How do I know if my rocket is aerodynamic stable? Are there guidelines? Can you send me one or 2 links?
2-Early on my adventures in KSP I tried to learn as much as I could from the basic and I read a lot people saying the best was to start between 8000 to 11000km but somepeople startin as early as 5000 to make the turn smoother and more professional, soo I am trying this way, that why I start at 7000 and I am actually getting good results despite the deviation. 
3-The 3rd part of your commentary I couldn't understand  much, sorry too technical for me :) 
4-Perhaps you are right about the SAS but I feel like the rocket is clearly rotating in such a way that it always end with a specific point of the rocket at the same heading (90), I mean lets say we draw a compass over the nose and look from top of the nose, the specific number (which I dont know) always end locked with heading 90.

I posted a few shots, tomorrow I plan to install FRAPS and make a video for you guys.

 

On 2017-6-17 at 11:14 PM, Spricigo said:

This is probably the root of the problem, leading to a small shift of the CoM and unaligned thrust. Aeroforces may also be at play. 

You may add something (e.g. Batteries) to balance the mass(and drag) of the parachutes.  

As I explained I made tests removing the non radial parts and the problem aleviated but didn't go away.

 

23 hours ago, Weywot8 said:

Agree with the steely eyed goat - you've chosen one of the worst speeds to turn. Drag as a function of speed climbs and peaks from mach 0.8 to 1.2. By turning in the dense atmo at that speed, the ship get the most forces acting on it leading to instabilities SAS cannot compensate for. 

  • turn earlier or turn later
  • use a stronger SAS/add extra reaction wheels

The chutes - as long as you place them on top ('west') or bottom ('east') of the ship (looking at it horizontally) it should help with stability as you turn. Maybe put them lower down the part so the drag is effectively 'fin-like'. +/- add  a par of extra fins along that axis, just above your existing fins. No pics of your craft, so not sure if that advice is helpful.

Thanks for the reply with lots of info.

As I told to other member I do that because I read a lot of people doing that as well, but I can try diferent, will do tomorrow. However I had other rockets with diferent designs even tri rockets for the same task (12pax tourism to orbit), using same conditions and I don't get this kind of problem (I get other problems but not this one :) )

I will try your suggestiong tomorrow!

Chutes are installed in the direction of the turn yes, can't install them in the bottom or I will not achieve what I want (land main stage on belly) 

I placed a few pics, pls have a look, also I will try to make a movie tomorrow.

Thanks

 

Link to comment
Share on other sites

19 hours ago, HvP said:

It's worth considering all of the offered advice, but like Kryxal said I have found that 3-way symmetry often causes this effect. You might try attaching your fins in 4-way symmetry instead.

With 3 fins only one of them can ever be aligned with EITHER the yaw axis OR the pitch axis, the other two will be trying to control a mixture of both and there will always be an uneven distribution of pitch and yaw control. And sometimes the pitch movement of one fin will counteract the yaw movement of another one. It's less than ideal.

With 4-way symmetry you will always have two opposite fins controlling mostly pitch, and the two other fins controlling mostly yaw.

Thanks for the reply.

I always use Quad Sym. I also had a older rocket for the same purpose and I pilot in the same circunstances and it fly actually better then this one and it is a TRI-SYM and TRI-Rocket design. Go figure :wink: 

Posted a few picts, have a look

Thanks

7 hours ago, Physics Student said:

Yes, RCS Build Aid has also a feature for that. I highly recommend it!

Thanks for the tip I will check it for sure!

Link to comment
Share on other sites

3 minutes ago, felcas said:

Thanks for all the info.

1-How do I know if my rocket is aerodynamic stable? Are there guidelines? Can you send me one or 2 links?
2-Early on my adventures in KSP I tried to learn as much as I could from the basic and I read a lot people saying the best was to start between 8000 to 11000km but somepeople startin as early as 5000 to make the turn smoother and more professional, soo I am trying this way, that why I start at 7000 and I am actually getting good results despite the deviation. 
3-The 3rd part of your commentary I couldn't understand  much, sorry too technical for me :) 
4-Perhaps you are right about the SAS but I feel like the rocket is clearly rotating in such a way that it always end with a specific point of the rocket at the same heading (90), I mean lets say we draw a compass over the nose and look from top of the nose, the specific number (which I dont know) always end locked with heading 90.

I posted a few shots, tomorrow I plan to install FRAPS and make a video for you guys.

To address your points in order:

1. In short, your rocket is aerodynamically stable if the center of lift is below the center of mass. Looking at the screenshot you posted, your fins are too high up. Nominally, you want fins as far down as possible.

2. It would seem you've been fed bad (or at least outdated) information. Before the 1.0 update overhauled the aerodynamics model, the optimal time to start your gravity turn was anywhere between 5 to 10 km up. However, since 1.0, the optimal time to start a gravity turn is typically closer to 1 km.

3. Sorry about that. I got a bit carried away with the physics, and anyways it's all conjecture based on some effects I've noticed while flying rockets that completely lack roll control in Realism Overhaul, combined with what I know of classical mechanics. It may or may not reflect what's happening in your case.

4. You're right. It does sound odd. As I said, all my conclusions were conjecture-it could be something else entirely that's causing the rotation you're experiencing.

Hope the advice helps, such as it is.

Link to comment
Share on other sites

1 hour ago, felcas said:

Posted a few picts, have a look

Alright, the pictures are really helpful now. If you still get problems when you aren't using radial parts and it's not a symmetry issue then I still have a few ideas.

I made a replica of your craft and tested it with no radial parts attached except for the fins and solid rocket boosters on radial decouplers. I found that the fins you have are a little too high to give you good control authority once you start to pitch over. By dragging the solid boosters and their decouplers down to the level of the first engine and then placing the control fins at the base of the bottom tank I had much better control.

Obviously, this doesn't look realistic. If you have access to the Thumper solid motor I would suggest using four of those instead, which gives you room to place the control fins in the gaps between them. Or, alternatively it is actually possible to fit the eight Hammer engines around your rocket without colliding with the fins. You just have to grab the decoupler with the solid boosters attached and then hold down the Shift key when sticking them back on, or turn off angle snap.

I would also suggest turning down the thrust some on those solid rocket boosters too, or using fewer solid boosters. You don't need such a high thrust to weight ratio as that. Ideally you want to start easing over as soon as you reach only about 100 or 200 m/s and make a consistently smooth gradual turn so that you are at 45 degrees somewhere around 10km. Then continue that gradual turn until you are aiming at the horizon line as you exit the atmosphere. Having such high thrust makes it difficult to achieve this gradual turn. A TWR of about 1.5 at launch is usually sufficient.

Lastly, the stack is a little bendy. There's no harm in using the autostrut function to reduce the wiggle in those fuel tanks.

Link to comment
Share on other sites

Are you thinking something like a set of 4 at 100% and a set at 30-50%?  I can second that idea, and it may be worth shifting the fins so they're on the cardinal points and making sure fuel is drawn from the bottom tank first.  As-is, all control surfaces move for all single-axis maneuvers.

Link to comment
Share on other sites

13 minutes ago, Kryxal said:

Are you thinking something like a set of 4 at 100% and a set at 30-50%?  I can second that idea, and it may be worth shifting the fins so they're on the cardinal points and making sure fuel is drawn from the bottom tank first.  As-is, all control surfaces move for all single-axis maneuvers.

Not having a thrust curve for the solid rocket boosters does make life interesting doesn't it?

I wasn't thinking of using two sets of four with lower thrust on the second set, but it could be worth a try. I think @felcas can get away with just turning the thrust down on all of them together though, since the mainsail (I assume it's a mainsail engine?) has enough thrust to sustain the flight once the boosters run out at about 7km.

And YES I always prefer keeping the four fins oriented north/south east/west. Much less likely to fight with each other. And if nothing else, disable roll control for those fins.

Edited by HvP
Link to comment
Share on other sites

2 hours ago, HvP said:

since the mainsail (I assume it's a mainsail engine?)

It's gotta be a Skipper... you can see the dark black bell of it in the first and last pics, darker than the boosters, where the mainsail has a grey bell, lighter than the boosters. A Mainsail would also stick out quite a bit below the boosters and would be much more visible at those angles.

Link to comment
Share on other sites

6 hours ago, felcas said:

One of them show in red the only components that are NON-RADIAL. With the exception of a remote control (the little orange box that extend airbrakes 10 seconds after staging) are the 4 parachute modules preciselly installed to make the rocket land on its belly (because if it lands standing it always roll, drop and explode).

Did you place those parachutes with mirror symmetry, or manually?

 

On 6/18/2017 at 1:58 AM, felcas said:

However I do have just 4 chutes on just one side (because I want my main stage land on its side).

Actually... I think I can see that you have also added 2x2 chutes on the lowest stage in the same way... only on the 'back' side. I draw that conclusion from the first picture, where one can tell that the chutes are not on the back/left side of the rocket, but also from seeing 2x2 chutes instead of 1x4 (which would be normal for a 4x radial symmetry) in your staging sequence, in stage 6. So that's 8 chutes in total that are only on one side, and that would certainly have an asymmetric influence on the aerodynamics. All bits add up.

Btw, I tried to re-create your rocket, and I get to just over 94t with 95 parts; your info screen in the VAB claims 96t with 97 parts. I play pure stock so obviously I am missing the remote control and MJ parts, which seems to account for two missing parts, but I've looked and looked and I can't really see what could cause 2t of difference in mass. For just the remote action and the MJ part (is it at the back of the Mk1 pod? It's hard to tell but something seems to be there) 2t seems excessive. If either of those parts are really that heavy, since they are placed asymmetrically and KSP calculates drag based on part mass, that could account for a lot of the effect you are seeing.

 

Pics of my stock re-creation:

Spoiler

MSbgq8C.png

94t and 95 parts. Remote control and MJ parts missing, but otherwise identical as far as the pictures permit. The 2t mass difference is a mystery to me...

Zlr7UGQ.png

 

 

Edited by swjr-swis
pics added
Link to comment
Share on other sites

7 hours ago, Kryxal said:

it may be worth shifting the fins so they're on the cardinal points

 

7 hours ago, HvP said:

I always prefer keeping the four fins oriented north/south east/west.

From the description of the ascent, I suspect it's this. The problem comes on at about the same time as the rocket gains significant speed, and then gets back under control as the rocket leaves the lower atmosphere. I get such better results with NESW control parts that this has become one of the design habits that I don't realize I know.

Align the controls (aerodynamic surfaces, RCS) with the UI keys/buttons. Something about splitting control inputs leads to unwanted angular momentum, and this shows up as roll in normal rockets because the pitch/yaw axes have more positive stability. I don't fully understand why it happens: from what I think I know of computing, I would have expected the initial effect to be miniscule, but perhaps the Unity engine and the SAS PID tuning amplify it over time.

Link to comment
Share on other sites

7 minutes ago, CSE said:

perhaps the Unity engine and the SAS PID tuning amplify it over time.

That is almost a certainty. There are very common situations in which the current SAS code gets into a self-amplifying oscillation - often triggered simply by changing from stability hold to follow prograde in mid-flight.

 

 

I think the winglets add very little to the control authority of the rocket to begin with, roll being the worst off. The Delta Deluxe winglets have the lowest control authority of all options (only 20% of their wing area is used, 0.13 - compared to 0.18 for the elevon 4), so most of their effect ends up being just stabilizing drag keeping the command pod pointing at space - and even that is poor due to very low initial CoM.

A central stack engine gimbal has negligible control authority over the roll-axis. That leaves roll almost entirely up to the reaction wheels/SAS, which is weak in the atmosphere for this rocket. This is easily shown by disabling the winglets for all three axes: you will hardly notice the difference in flight behaviour. Now try disabling the reaction wheel torque and leave it all up to the winglets instead: flipping rocket and horrible roll control.

Link to comment
Share on other sites

I want to thanks EVERYONE for your very rich input and thoughts on this. I really didn't expected soo much effort to help me.

I learned a lot with all the comment, be it directly or just pointing me to look for more info.

And there are TONS of new things I can try to understand the root of the problem and who knows make my designs and flying even better. Soo I ask a bit of patience on all of you, because I need to make a list of all the points you made and try them (this is going to be my homework for today) before coming back here with my results.

Again thanks everyone! And if anyone have more ideas please let me know, I am not closing the subject, with this message.

:)

Link to comment
Share on other sites

33 minutes ago, swjr-swis said:

Snip

 

Small asymmetries, low CoM, SAS induced oscilations, little control authority ...I think you are right on the spot.

 

@felcas what seems to be the case its that your craft is stable....but not much. Its a issue composed of several factors:

Crew cabins are considerable less dense than (loaded) fuel tanks, thus the tendency its to have a CoM too far behind, aeroforces near the top of rocket have more leverage, so you need stronger aeroforces on the rear to compensate and once you deviate a bit from prograde the drag increase pushing you further away. Moving fuel tanks around may alleviate the problem...or create new ones.

Asymmetries cause torque (If I'm not mistaken KER reports it, in VAB/SPH, by default, if not check KER's display settings), if this is small enough your reaction wheels and control surfaces can deal with. Try to make the rocket as symmetrical as possible, take advantage of the offset tool and angle snap for better results.

Control surfaces may act oddly when trying to respond to more than one kind of control input (pitch, yaw, roll). Try to setup each group to only respond to one kind of input, often is advantageous to have limited authority or no control surfaces at all and (winglets only for passive stabilization). Also there is instance where movement along one axis cause torque around a different axis because of the geometry of your craft (is more an issue for planes, but someitmes is noticeable for rockets too).

On the 'unsolicited advice' category: there is some thing about ship design and piloting that may be helpful to you. Those don't seems to be direct causes of the issue at hand, just 'good practices' for you to consider.  On top of my mind those thread my be a good reading:

 

Link to comment
Share on other sites

12 hours ago, swjr-swis said:

It's gotta be a Skipper... you can see the dark black bell of it in the first and last pics, darker than the boosters, where the mainsail has a grey bell, lighter than the boosters. A Mainsail would also stick out quite a bit below the boosters and would be much more visible at those angles.

Yes I see you're right, of course. That introduces the additional possibility that the thrust-to-weight ratio might actually be a little low after dropping those solid boosters depending on how the gravity turn is done. It's possible the nose is trying to dip and if the rocket is not well aligned along the pitch/yaw axes then trying to bring the nose back up is inducing movement in the other axis as well.

Link to comment
Share on other sites

Ok guys I tested all the theories you threw at me and I have to say the problem was really those parachutes. It happens that there was still 4 other parachutes mounted on non radial standards and they was hiden between the small SRBs. Those stand-up guys are really heavy around 200kg each, soo 4 of them is almost 1 ton, I had 8 on a total of 1.6t pointed due north. There was also a remote control that weighted 50kg and yes it does make a diference, I tested! 

So I remade my rocket with all the parachutes pointing east now. I also incorporated a lot of new suggestions all of you sent to me.

  • Placing fins and SRBs on the botton.
  • Exchanging the 8 small SRBs to 4 Thumper SRBs - this actually saved me $5.600 - this also downgraded the TWR from 2.28 to 1.66 another suggestion.
  • I learned about fuel flow and I defined the botton tank to consume first.
  • I rotated the chutes to face east, it still pull the nose down but at least it is in the right direction, it is now working for me and not fighting with me.
  • Initiating the gravity turn at 1Km also resulted in great economy, I am ending up there with 50% more fuel then before, it gives me a feeling of more professionalism :) I need to know what I am going to do now, use a smaller tank to make my rocket even more efficient and cheaper or just have fun using the surplus to speed break on re-entry.
  • Fins aligned with cardinal points (E/W/S/N)

Also, some suggestions did not solve the problem I brought to you (I tested all of them individually), but I thank you anyway because even not solving my problem this made me experiment and understand lots of configurations and I learned a LOT, plus, as I wrote above, after solving the problem the suggestions helped me build a better rocket.

  • Initiating the gravity turn after 30km with less denser atmosphere - didn't solve problem and consumed much more fuel, couldn't get in orbit at all.
  • Using speed between 100 to 200m/s, actually helped tip over the rocket.
  • Nullify wheels
  • Use more wheels - aleviated the problem but did not solve
  • Use bigger SRBs instead of smaller ones.
  • Placing fins at bottom - made the problem worst but made the rocket more maneuverable.

Thanks a lot for everyone!

Link to comment
Share on other sites

@felcas Yay problem solving!

Just a suggestion for the future, don't be afraid of asymmetry. You can always balance out those parachutes with a counterweight on the other side. The RCS Build Aid mod that Physics Student mentioned earlier is something I consider an essential addition to the game and is a huge help in balancing craft in this situation.

I'm glad we could be of some help. Happy flying and have fun!

Edited by HvP
Link to comment
Share on other sites

My new vessel on KerbalX, inspired by this thread:

https://kerbalx.com/Spricigo/FourteenAgain

14 tourists to orbit and back.

Fun fact: it has m some of the flaws (but not to the same extent) reported in OP's original rocket (since it reach orbit almost by itself, with more than enough deltaV to deorbit and took like 15min to design that's good enough)

 

 

Link to comment
Share on other sites

12 hours ago, Spricigo said:

My new vessel on KerbalX, inspired by this thread:

https://kerbalx.com/Spricigo/FourteenAgain

14 tourists to orbit and back.

Fun fact: it has m some of the flaws (but not to the same extent) reported in OP's original rocket (since it reach orbit almost by itself, with more than enough deltaV to deorbit and took like 15min to design that's good enough)

 

 

LOL that one looks like my 1st tourism rocket! Have a look! :) Have in mind this is early on career, soo no 2.5m tanks and rockets available at that time, only those smal 1.25. I loved the design. I wonder...maybe I can make this one better too with everything I learned and the new techonology.   

KWu6QMH.png

ayI3smF.jpg

Edited by felcas
Link to comment
Share on other sites

This thread is quite old. Please consider starting a new thread rather than reviving this one.

Join the conversation

You can post now and register later. If you have an account, sign in now to post with your account.
Note: Your post will require moderator approval before it will be visible.

Guest
Reply to this topic...

×   Pasted as rich text.   Paste as plain text instead

  Only 75 emoji are allowed.

×   Your link has been automatically embedded.   Display as a link instead

×   Your previous content has been restored.   Clear editor

×   You cannot paste images directly. Upload or insert images from URL.

×
×
  • Create New...