SSTO + FAR + Deadly Reentry Help

[Zeenobit]

I give up. Mostly.

I've been reading tons of tutorials on how to design, build, and fly SSTO space planes with FAR and Deadly Reentry mods. But after numerous failed attempts at getting my aircraft to fly, I think I need some help from the more experienced players.

For reference, this is my aircraft:

I'm using the Stock Revamp, Procedural Wings, B9 Aerospace, and KW Rocketry mods in this design. The rocket engines at the back are Poodle engines. The Jet Engines are stock TurboJet Engines. I do not have access to RAPIER engines in this playthrough.

From what I understand, the basic procedure for getting a SSTO into orbit with FAR is to:

1. Start with an ascent into a thinner atmosphere. I do this at about 30 degrees, going up until about 8,000 to 10,000 meters.

2. Lower AoA as much as possible, and start building as much speed as possible. Perfect. I can do this fine, and can get my velocity up to about 1,200 m/s to 1,700 m/s.

3. Once the jet engines start dying, switch to rocket engines, raise the nose, and start burning towards upper atmosphere and gain dV.

Step 3 is where it all goes wrong for me.

At step 3, one of the two following scenarios keep happening:

A. I raise the nose and flip over, losing control, and crashing.

B. I can't raise the nose, and I go into a dive, quickly losing altitude and finally crashing.

I can't get past step 3, it seems. I've tried multiple designs, and the one posted above has been the most successful so far.

Any tips? Any ideas? Links? Guides? Tutorials?

Any help would be appreciated.

Edited November 15, 2014 by Zeenobit

[Mechie_PhD]

I'm a noob too, but I'm guessing you cannot drastically change your AoA when switching to rockets inside the atmo. Doing so might cause 3A.

[Hodo]

Ok I see one thing that maybe going wrong for you on step 3. Because it is about the same procedure I follow on my SSTOs.

I think perhaps when you are pitching up you are pitching up to much on the burn stage. It is causing the airflow to become to strong at the front of the craft, moving the CoL forward and thus pitching you end over end.

If B is happening the issue is your CoL is moving back as you go faster, mach tuck. This issue is pretty common in SSTO space planes expecially ones with canards.

What is happening you are losing lift on your canards they are stalling out at those speeds and at the AoA you are attempting to get them to pitch to. This is causing you to lose all lift at the front of the craft, and thus pitching down no matter what happens. I try to keep from using canards for this reason.

You can fix most of this by reducing the amount of pitch authority this control surfaces have. Also by checking in the FAR diag screen to see how your lines look.

EDIT,

As seen here.

This is one of my latest designs I am working on. It flies fine but has some mach tuck issues at supersonic speeds. You can see it on the graph. But this is mostly ironed out by some minor wing and fuel tweaking.

Edited November 13, 2014 by Hodo

[funk]

It seems your CoM isn't aligned with CoT. I use RCS Build Aid for being sure that the plane isn't rotating in Space...

[Jerry060599]

I can post screen shots but that would not help as ships in ksp are far more than its appearance. I can give you some very usefull tips, however.

1. 10000m above ground is NOT nearly high enough, I would say atleast 23000 m and if you can mannage, 26000m is even better.

2. Don't go for big and heavy, go for a slim, small design.

3. Use saber engine from b9 parts pack. They really help.

4. Air breaks from b9 parts pack is a MUST. It is insanely hard to slow down in a sleek plane traveling at 2~3km/s.

5. Place all air break on the top of the plane and put other smaller parts(like RTGs, batteries and monopropellent. ) in a cargo bay from the b9 parts pack so that they wont burn up during reentery.

6. When returning from orbit, change your periapes to 14000m. It seems to be the perfect height for slowing down but not fast enough so that the plane burns up.

7. MOAR INTAKES. At least 2 intakes for every engine.

8. Struts don't hurt anything but maybe your fps. Strut the wings up to prevent wobbling.

Thats it, all the thing i can think up the top of my head of the lessons i have learned by trial and error.

Enjoy

[Wanderfound]

There are a few points to address here.

* The only reason to pull the nose up once you light the rockets is in order to climb to an altitude where there is less drag. It's not compulsory; so long as you maintain any climb rate at all, an increase in speed will eventually lift your apoapsis into space. Particularly with low-thrust high-efficiency rocketry (e.g. Poodles or nukes, rather than RAPIERs or Aerospikes), a shallow ascent is often best; let your wings do the lifting while the thrust is concentrated on increasing speed.

* Increasing your AoA by too much is going to cause your plane to want to go out of control. If you're getting stall warnings, you're increasing AoA too much and too fast. You need to lead the plane rather than force it; increase AoA by a bit, then wait for your prograde vector to catch up, then increase a bit more, etc.

* In the upper atmosphere, the thin air limits the grip available to your aerodynamic surfaces, reducing their ability to maintain stability. For a sufficiently well-built plane this isn't a problem, but for when it is: Vernors. Stick a few around the plane in relevant places (e.g. above, below and to either side of the nose and/or tail) and keep them turned off until you need them. Vectored thrust can compensate for a lot of aerodynamic flaws.

* Sorting out these issues is what the FAR analysis screens are for. It would help us help you if you could post some screenshots of the analyses. What would be relevant is an AoA sweep at the speed at which you tend to lose control (that will tell us how much AoA your plane can take while retaining stability) and a stability derivatives analysis at the same speed with the atmospheric density set to match the altitude at which it happens.

If you're not sure how to do these, have a look at post #3 in the Kerbodyne thread linked below. That'll also tell you how to translate altitude into atmospheric density.

The AoA sweep should look something like this:

And the stability analysis should look like this:

Edited November 13, 2014 by Wanderfound

[Zeenobit]

@Wanderfound, Here's the FAR analysis you asked for. The AoA Sweep looks very similar to yours. But I noticed a red value in the stability analysis. To be honest, I don't quite understand the stability analysis information.

@Hodo, FAR analysis at the settings you mentioned also seems very similar to yours. I mainly used the canards because without them I had trouble getting into that initial 30-degree ascent profile. I'll try to see if removing them would help!

[Zeenobit]

I'm a noob too, but I'm guessing you cannot drastically change your AoA when switching to rockets inside the atmo. Doing so might cause 3A.

That's true. But if I don't drastically change my AoA when switching to rockets, I end up either flying straight or nosediving. Eventually, I'd run out of fuel.

So I have to keep the nose up to get out of atmosphere. Which is proving to be the main challenge.

It seems your CoM isn't aligned with CoT. I use RCS Build Aid for being sure that the plane isn't rotating in Space...

My CoM is aligned with CoT. My CoL isn't. I use RCS Build Aid too. My CoL is just slight above CoM and CoT.

[Wanderfound]

Nicely built; those analysis screens are almost flawless. The AoA sweep suggests that the plane should tolerate fairly extreme (25°, maybe more) AoA while maintaining stability.

The touch of red in the numbers is not too big a deal; if you hover your mouse over it, you'll get a popup explaining what it's about (I don't have my PC on right now or I'd look it up for you; f you post it here, I'll try to give a FAR-to-English translation). The tooltip will reference assorted variables that relate to this image:

The lack of obvious flaws in your aerodynamic design suggest that the issue may be more piloting related. Just how sharply are you trying to lift the nose, and exactly what happens when you lose control? Does the plane pitch end over end, or yaw sideways, or roll upside down, or something else?

Also: what have you got the control surface tweakables (maximum deflection, pitch/yaw/roll, etc) set to?

Edited November 13, 2014 by Wanderfound

[NathanKell]

You also need way more vertical tail. Well, maybe not need, but it would be very helpful. Especially at high speed, [large] vertical tails become very important for stability. Consider how large (compared to wing area) the X-15's vertical tails are; it operates in the same velocity regime as KSP spaceplanes.

[DaMichel]

The following is basically what Nathan said. See Nbeta is one of the "important" values. From the FAR wiki

Mw is the primary static stability parameter: it tells you whether the center of lift is in the correct position. If this is negative (red), the plane is going to rapidly depart in a backflip (or nosedive) at the slightest uncorrected disturbance. In mathematical terms, it's the first derivative of the pitching moment over normal velocity - i.e. angle of attack, effectively.

Yβ and Nβ are static yaw stability parameters. These will almost always be the correct sign at the same time, unless the dorsal fin has highly excessive area, and if correct basically make sure that the plane wants to fly straight rather than skate sideways.

Yβ is the first derivative of side force over sideslip angle, and Nβ is the derivative of yawing moment over sideslip. Nβ is exactly analogous to Mw, and must be always positive. Wings and the horizontal tailplane have a slightly positive effect on Nβ. However, the fuselage causes Nβ to decrease. Vertical tailplane, of course, strongly increase Nβ, even though it's possible to design aircraft without them (case in point: birds).

So your plane is statically yaw unstable. The problem is that the vertical tail is very close to the center of mass. So the tail has a small lever arm to turn the plane in the right direction, rendering the tail fairly ineffective. Either increase its size even more or try to put it further back.

It probably won't fix the pitching issue but it should at the very least make the plane easier to control.

One more thing, Zeenobit. Please do the static stability analysis again but this time enter 1 in the pitch field. It basically tells you how much pitch control you have. Look where the cm curve intersects the x-axis. This is the AoA you'll be able to reach at most. It should be well above 13 deg. 13 deg because this is the angle you need for level flight. It is shown in the stability derivative window, top right.

Edited November 13, 2014 by DaMichel

[Wanderfound]

So, as has been established so far: either more vertical stabiliser, a Vernor (or multiple linear RCS ports) either side of the nose, or shift the CoM forwards in order to get the tailfin more leverage. Removing the dihedral from the tailfins will also give them a bit more force, and you probably have enough roll stability already that you can afford to lose a bit.

Keep the FAR analysis window open and re-run the numbers after you move or change any of the aerodynamic surfaces; you're trying to get the red number to flip from negative to positive. Keep any changes that move the number in the right direction, ditch any changes that make it worse.

-

You may want to shuffle your CoM around for other reasons as well. Because your cargo bay isn't on CoM, adding and removing cargo is going to disrupt the balance of the plane. Whenever you add or drop cargo, you're going to substantially alter the CoM/CoL relationship.

You may want to try this: swap the positions of the cargo bay and the rear-centre fuel tank. This will increase your CoM/dCoM offset a bit, but you can balance that out by shifting the lateral tanks backwards. As well as putting the cargo bay onto CoM, this should also allow you to increase the distance between CoM and the vertical stabilisers, providing them with more leverage.

It may also be helpful to add a passenger cabin or somesuch behind the docking port, as a way of pulling both CoMs forwards without altering their relationship. The plane can certainly handle some extra weight; you've got a lot of wing on there.

Edited November 13, 2014 by Wanderfound

[Hodo]

As Wanderfound said about the CoM in relation to your cargo bay.

I often add empty tanks to my craft and use TAC fuel balancer to shift the CoM around in flight to maintain a very even flight profile at all speeds. This takes a bit of practice and is way overly complex and there is probably a plugin that would make that job easier.

Once you get the hang of it, you will find that you can get away with more and more wild designs. Things you never thought that should work actually work, and others will fail in fantastic ways.

[Zeenobit]

So I got a chance to implement all the suggestions here (lots of thanks, by the way!). In the end, I ended up almost re-designing the plane a bit.

Here are the results:

In terms of flight, this new design is much more stable and manageable at higher mach numbers. But I still couldn't punch through the atmosphere.

Here are a couple of new issues I'm having:

A. This design is proving to be a lot harder to take off. It does take off, but way late into the runway, with me pulling the stick all the way up. The original design needed just a slight nudge to take off. I think this is directly related to CoM being further towards the front and CoL further towards the back. I had to switch to taller landing legs to not smash my engine during takeoff.

B. At subsonic speeds it's a little harder to control this plane, I find. It needs a lot of effort to pull the nose up.

When I break through the thick atmosphere and go supersonic, everything suddenly becomes much better. The plane is much more stable and responsive.

But as I soon as switch to rocket engines, I lose control.

With regards to piloting, I mainly use trims to control the plane. It makes the flight much smoother and less jerky. And I turn off the reaction wheels, and SAS while in the atmosphere. With SAS, I tend to keep wobbling. The only assistance I do use is the "Level" assistance in the FAR menu just to keep the plane from constantly rolling side to side. So...

1. Should I enable SAS during low atmospheric flight? Is there something wrong with the design that makes SAS wobble the plane constantly?

2. Should I enable SAS and/or reaction wheels during high atmospheric flight?

I also noticed turning off the "Level" assistance in FAR menu when switching to rockets keeps things from spinning out of control. I didn't have time to test this more, but I have to check it.

[Wanderfound]

Are you adjusting your wing mass/strength tweakables? If they're still set to default, you're carrying much more weight than necessary; wind them down to about 0.4.

Shifting your landing gear a smidgeon further forwards (or lowering the rear gear) will ease rotation. Tailstrike issues can usually be solved by careful piloting, but if you don't want to have to worry about it:

The SAS wobble is a consequence of the SAS code being (a) designed for rockets and stock aero, and ( not very good. There are several solutions, but you've already hit the simplest: turn SAS off and use the FAR stability aids instead.

What do your analysis screens look like at temp 0 / density 0.5 / speed 0.8?

Again, when you lose control after turning the rockets on, exactly what happens? And do you have access to Vernors?

Edited November 14, 2014 by Wanderfound

[DaMichel]

What Wanderfound said. And in addition to that:

I think your pitch authority is okay. Beware holding full pitch up will make your plane flip (because cm>0 all the way). So use trim with smaller deflections and tap w/s for fine control.

Not sure what happens in rocket mode. Do you increase your AoA too much? There is probably a limit where your plane will become unstable and flip.

The switch to rocket mode can easily bring your plane out of balance because of the change in thrust. If it is out of alignment with the CoM just a bit this incurs a change in torque. So you'd have to re-trim the craft. SAS can help with control here. Also try to throttle down slowly, stabilize the craft, switch to rockets, throttle up slowly and stabilize.

[Wanderfound]

Also, BTW: the SAS wobble is due to the SAS overcorrecting because it's expectng rocket levels of control authority rather than the much stronger forces a plane generates. As the air thins, your aero surfaces reduce in power, which in turn reduces SAS wobble. By the time you get to rocketry altitude, it's probably safe to turn the SAS back on.

[Hodo]

There was a mod that I still use by Regex called the PID, it let you limit the force of the SAS module. It is quite handy when you fly with SAS on.

You can also reduce this oscilation by reducing the max control angles of the control surfaces. Bring them down to about half of the default settings.

[Jovus]

I'm willing to bet good money that your failure mode when switching to rockets/in space is having the plane nose uncontrollably up, and that the reason for that is those huge, heavy vertical stabilizers putting your CoM above your CoT.

[Zeenobit]

Are you adjusting your wing mass/strength tweakables? If they're still set to default, you're carrying much more weight than necessary; wind them down to about 0.4.

Shifting your landing gear a smidgeon further forwards (or lowering the rear gear) will ease rotation. Tailstrike issues can usually be solved by careful piloting, but if you don't want to have to worry about it:

... <snip> ...

The SAS wobble is a consequence of the SAS code being (a) designed for rockets and stock aero, and ( not very good. There are several solutions, but you've already hit the simplest: turn SAS off and use the FAR stability aids instead.

What do your analysis screens look like at temp 0 / density 0.5 / speed 0.8?

Again, when you lose control after turning the rockets on, exactly what happens? And do you have access to Vernors?

I'm not lowering my wing mass. I will try that tonight.

The rear landing gear to avoid tail-striking seems like a very useful thing. I'll try that as well.

More details on what happens during the switch below.

I do have Vernors unlocked. But I'm not sure how they would helpful to my situation. How would I use them in this scenario?

The switch to rocket mode can easily bring your plane out of balance because of the change in thrust. If it is out of alignment with the CoM just a bit this incurs a change in torque. So you'd have to re-trim the craft. SAS can help with control here. Also try to throttle down slowly, stabilize the craft, switch to rockets, throttle up slowly and stabilize.

Interesting. My CoT is aligned with CoM. But the jet engines are slightly (just a tiny bit) forward with respect to the rocket engines to balance the CoM. I wonder if that's a contributing factor.

Also, BTW: the SAS wobble is due to the SAS overcorrecting because it's expectng rocket levels of control authority rather than the much stronger forces a plane generates. As the air thins, your aero surfaces reduce in power, which in turn reduces SAS wobble. By the time you get to rocketry altitude, it's probably safe to turn the SAS back on.

That's what I try to do as-is when I get that high, assuming I get that high.

With the recent design, this is what I'm doing and what's happening during the jet -> rocket switch:

1. I wait until my jet engines start losing thrust. I'm going at about 1,200 m/s at this point, in a 22-25 km altitude.

2. I start adjusting my trim slowly until I have minimal trim.

3. I press Alt + X to cancel any remaining trim.

4. At this point, I'm losing a bit of altitude due to loss of trim.

5. I turn off FAR flight assistance.

6. I turn on the cockpit reaction wheels.

7. I turn on SAS.

8. I switch to rockets. (My AoA at this point is near 3-4 degrees)

9. Immediately, I start experiencing the plane wanting to destabilize. Usually, this happens with a sudden change in pitch. Sometimes, it also changes roll or yaw.

10. SAS tries to correct this, minor wobbling happens. I'm either losing altitude, or going into a large AoA/stall.

11. Loss of control.

I'm willing to bet good money that your failure mode when switching to rockets/in space is having the plane nose uncontrollably up, and that the reason for that is those huge, heavy vertical stabilizers putting your CoM above your CoT.

I don't. It's actually very difficult to pull the nose up in that situation. If what you're saying is correct, reducing the weight of the vertical stabilizers should fix this. More reason to try that theory tonight. It makes perfect sense!

There was a mod that I still use by Regex called the PID, it let you limit the force of the SAS module. It is quite handy when you fly with SAS on.

You can also reduce this oscilation by reducing the max control angles of the control surfaces. Bring them down to about half of the default settings.

Are you referring to this mod?

http://forum.kerbalspaceprogram.com/threads/96654-PID-Tune-%28No-Longer-Available%29

It seems like it's no longer available. It sound very useful even for rockets. Any alternatives?

Edited November 14, 2014 by Zeenobit

[Hodo]

-snip-It seems like it's no longer available. It sound very useful even for rockets. Any alternatives?

Unfortunately no other alternatives. My only other suggestion is to reduce the control range of the control surfaces that will help reduce some of the oscilations from the SAS.

As for your craft suddenly losing control when you switch over to rockets, sounds like you are having an action group failure, where it is not activating everything at the sametime. This is usually caused by moving a paired part after you set the action group. Go back into your action group settings in the SPH and click on the action group that engine is set to, then click on each on of the engines individualy and see if they are set to action group.

Hope that helps with that a bit.

[Wanderfound]

O

I'm not lowering my wing mass. I will try that tonight.

The excess wing mass is probably a large part of your problem: at the moment you're carrying extreme high-G aerobatics wings. Those things aren't intended for space.

I do have Vernors unlocked. But I'm not sure how they would helpful to my situation. How would I use them in this scenario?

Vernors are super-powered linear RCS ports that consume LFO instead of monoprop. What they give you is the ability to put a small thruster anywhere you want it, and have the RCS system handle the task of controlling them for you. So, if you're having trouble lifting the nose: stick a Vernor on the underside of your cockpit and/or the top of your engines. If you're having yaw stability issues, stick one either side of the cockpit and/or engines. If you're getting uncontrollable pitch-up, stick 'em on the cockpit roof or under the tail. Etc.

They won't consume any fuel so long as you keep the RCS toggled off, but as soon as you toggle it on it will start adding its force to any manouevres you make (or will try to hold your direction constant if you aren't maneouvring and have SAS on). Stick enough Vernors on, and you can stabilise a brick. Keep them turned off as much as possible, but flick 'em on for a bit any time the ship starts to misbehave.

See the things just in front of the canards on this ship:

Those are Vernors, which I have turned on (by toggling RCS on) in that shot because I'm flying at an altitude where there isn't a lot of aerodynamic grip.

I frequently put them on my craft even if they don't need them for stability, because they also make good low-gravity VTOL jets and may allow you to recover from an otherwise incurable spin or stall.

They're also handy for stretching the envelope of possibility when indulging in extreme aerobatics. Just don't try to dock with 'em on; they're too powerful for the fine control needed during docking. If the ship is intended for docking, add conventional RCS as well and use an action group to toggle the Vernors off.

Interesting. My CoT is aligned with CoM. But the jet engines are slightly (just a tiny bit) forward with respect to the rocket engines to balance the CoM. I wonder if that's a contributing factor.

<snip>

With the recent design, this is what I'm doing and what's happening during the jet -> rocket switch:

1. I wait until my jet engines start losing thrust. I'm going at about 1,200 m/s at this point, in a 22-25 km altitude.

2. I start adjusting my trim slowly until I have minimal trim.

3. I press Alt + X to cancel any remaining trim.

4. At this point, I'm losing a bit of altitude due to loss of trim.

5. I turn off FAR flight assistance.

6. I turn on the cockpit reaction wheels.

7. I turn on SAS.

8. I switch to rockets. (My AoA at this point is near 3-4 degrees)

9. Immediately, I start experiencing the plane wanting to destabilize. Usually, this happens with a sudden change in pitch. Sometimes, it also changes roll or yaw.

10. SAS tries to correct this, minor wobbling happens. I'm either losing altitude, or going into a large AoA/stall.

11. Loss of control.

Okay, first up, you're switching from jets to rocketry much too early. You want to run the jets on their own for as much of the flight as you possibly can. So, don't switch to rocketry as soon as the jets start losing power; save the rocketry for when the jets are completely incapable of producing thrust at all (i.e. flameout). At the time the rockets come on, you want to be doing at least 1,500m/s at an altitude of at least 30,000m. Have a read through the piloting guide in post #2 of the Kerbodyne thread or listen to my piloting tutorial on the latest episode of the Kerbal Podcast.

The loss of control immediately post rocketry is most likely the plane reacting to the sudden change in thrust and the alteration of gimballing ability. Use the right-click tweakables to turn the thrust limiter to zero on the jets, then use RCS build aid to check your engine-induced torque from the rockets. Then turn the jets back to 100% and the rockets down to zero and do the same for the jets. Just because all of the engines together are balanced doesn't mean that the jets or rockets are balanced on their own.

Are you referring to this mod?

http://forum.kerbalspaceprogram.com/threads/96654-PID-Tune-%28No-Longer-Available%29

It seems like it's no longer available. It sound very useful even for rockets. Any alternatives?

Yes. I haven't tried this one yet, but it also includes a PID controller: http://forum.kerbalspaceprogram.com/threads/100073-0-25-Pilot-Assistant-0-2-Nov-15

Edited November 14, 2014 by Wanderfound

[endl]

from what i can tell those intakes are for subsonic you might want to put on some radials that perform better at hypersonic, or invert that if you have better parts for the inline intakes. your basic grasp of the flight path is correct wonderfound told me to climb to 15k though and stay above 25k when my speed gets high enough, you might want to just build speed until things start to get hot (since your playing with DR) then climb to cool them off. depending on your TWR you might have to "bounce" (do your speed run/climb/speed run) a few times to pick up speed. some times i do 2 or even 3 bounces before i get the required speed i need.

[Zeenobit]

I finally did it! Thanks for all the help!

I think the major issues that were causing me problems were:

1. DaMichel's point regarding thrust alignment. I checked this, and the thrust vector for my jets were far ahead of my rockets. More than I thought. I carefully aligned these.

2. I reduced the wing mass to 0.4 as Wanderfound suggested. Shaved off about 4 tonnes.

3. The idea with the small landing gears at the back to assist with take off was absolutely genius.

4. I waited till the jets were down to single digits in thrust, and throttled down before switching to my rockets.

5. I also switched the rocket engines from Poodles to Rockomax "Cutter" Linear Aerospike Engines from RLA Stockalike mod. Higher atmospheric ISP.

Incorporating all these points, the transition from jets to rockets was seamless and very smooth this time.

I did add Vernors, but I didn't really need them. Just enabling the cockpit reaction wheel tipped my nose just high enough to gain altitude.

I did fail to re-enter safely though. I got close, but lost control at the end. That's another issue for another day. I think it just needs a bit of practice.

Regardless, here is my first SSTO under FAR:

Thanks again for all the advice!

[DaMichel]

Well done! Welcome to the wonderful world of aerodynamically realistic space planes