What makes this plane spin out?

[Red Iron Crown]

I'm getting better at spaceplanes and I think I've got the basics down, but I still have some flat spin problems with my current design between 25km and 30km as it slows its climb and starts to build speed. The CoL is just behind the CoM, and the plane has a twin tail arrangement with control surfaces for yaw authority, gimbals on the jets are enabled, plus there is a fair amount of torque available from the pod and probe core. It's stable at lower altitudes and speeds, but once the jets get into the meaty part of the power curve it starts to yaw in one direction or another and eventually spins. If I throttle back a bit it can limp up to 30 km and then fly fine. What am I doing wrong?

Aerodynamics are stock. Obligatory picture of craft in question:

Edited August 26, 2014 by Red Iron Crown

[goduranus]

Too much angle of attack? You can't tilt up too far above your airspeed vector

[Hodo]

The problem you are describing is caused by one of your engines throttling back due to air starvation. Check your engines thrust by right clicking on them and I am 100% sure that one of them is starting to lose power first.

[Red Iron Crown]

I don't think that's it, by that part of the ascent I'm only a bit above prograde. Just enough to keep a small climb rate.

[Wanderfound]

One: that's a short bodied plane. Yaw and pitch control come from the nose and the tail; the closer they are to CoM, the less leverage they have.

Two: off-vertical tails look cool and can work, but they complicate the aerodynamics. Try some properly vertical stabilisers. If you've got ground clearance enough, consider ventral fins.

Three: no cargo bay that I can see. Even on specialist speedsters, a small cargo bay is very useful. You can use it to shield all sorts of unaesthetic and unaerodynamic things. Stick one of the half-length bays behind the cockpit and fill it with batteries, SAS units, solar panels, lights, TAC-LS supplies, etc. Also may be handy for emergency crew transport or salvage.

Four: what are your control surface tweakables set at?

Five: if you want to do crazy aerobatic things, I highly recommend sticking Vernors all over the place and toggling them on and off when needed. Also good for stabilising occasionally temperamental ships.

Edited August 25, 2014 by Wanderfound

[Red Iron Crown]

Wanderfound:

1. I'm starting small and building up to bigger things, seems like the smaller the plane the easier it is to make work.

2. I like the canted aesthetic, but I'll try it with vertical ones.

3. Cargo bay is unnecessary for this ship and the accessories are hidden well enough (except the docking port). Though I have used the "service compartment" trick before.

4. Outer wing surfaces are elevons (roll and pitch), inner wing surfaces are elevators (pitch), tail surfaces are rudders (yaw), canards hidden by the intakes are fixed. It has good control authority in all three axes at lower speeds and altitudes.

5. Not really interested in aerobatics, just a stable ascent to orbit and descent.

Edited August 25, 2014 by Red Iron Crown

[Master Tao]

Since it's stock aerodynamics, I'm thinking the jets are throttling down due to lack of IntakeAir. That can be a little different for each jet – you'd need to watch the actual thrust on both. Another possibility is that the CoM may be shifting closer to the CoL as the jets burn fuel.

[Red Iron Crown]

Since it's stock aerodynamics, I'm thinking the jets are throttling down due to lack of IntakeAir. That can be a little different for each jet – you'd need to watch the actual thrust on both. Another possibility is that the CoM may be shifting closer to the CoL as the jets burn fuel.

The engines still have a surplus of intake air, and I'm using TAC fuel balancer to keep liquid fuel balanced among all five tanks. I should have mentioned that up front.

[Master Tao]

It's hard to tell just from the picture, but TAC FB may end up shifting the CoM back, especially since everything but the Radial Engine Bodies contain OX. Where is the CoM when you empty the tanks?

[Wanderfound]

Wanderfound:

1. I'm starting small and building up to bigger things, seems like the smaller the plane the easier it is to make work.

2. I like the canted aesthetic, but I'll try it with vertical ones.

3. Cargo bay is unnecessary for this ship and the accessories are hidden well enough (except the docking port). Though I have used the "service compartment" trick before.

4. Outer wing surfaces are elevons (roll and pitch), inner wing surfaces are elevators (pitch), tail surfaces are rudders (yaw), canards hidden by the intakes are fixed. It has good control authority in all three axes at lower speeds and altitudes.

5. Not really interested in aerobatics, just a stable ascent to orbit and descent.

I must confess that I hadn't noticed the stock aero part. I fly stock so rarely that anything I say on this could be disastrously wrong.

Still, just in case there are some parallels between stock and reality: short bodied small planes are harder to get right than long bodied small planes. It's like a white water kayak vs a sea kayak; the short one is more manoeuvrable, the long one is more stable.

I don't see any SAS, and this is the sort of plane that in reality wouldn't fly without advanced fly-by-wire computer assisted avionics. With the Vernors, I understand if you'd rather avoid them as a bit too brute-force, but they do work just as well for rescuing spins as they do for intentional aerobatics.

[Red Iron Crown]

It's hard to tell just from the picture, but TAC FB may end up shifting the CoM back, especially since everything but the Radial Engine Bodies contain OX. Where is the CoM when you empty the tanks?

The red dot in the picture is the dry center of mass, it moves forward as the tanks empty. The center tank has no oxidizer in it during the ascent, I tweaked it out to compensate for the fuel the jets will use but wanted the tankage for it if I want to refuel in orbit for interplanetary work.

I must confess that I hadn't noticed the stock aero part. I fly stock so rarely that anything I say on this could be disastrously wrong.

Still, just in case there are some parallels between stock and reality: short bodied small planes are harder to get right than long bodied small planes. It's like a white water kayak vs a sea kayak; the short one is more manoeuvrable, the long one is more stable.

I don't see any SAS, and this is the sort of plane that in reality wouldn't fly without advanced fly-by-wire computer assisted avionics. With the Vernors, I understand if you'd rather avoid them as a bit too brute-force, but they do work just as well for rescuing spins as they do for intentional aerobatics.

Hmm. I was hoping to be able to stabilize it without resorting to SAS or RCS. You're right that there's no torque wheels beyond what the cockpit and core provide.

[goduranus]

what if you add a slight dihedral to the main wings?

also, be helpful if you could show a video or some screenshots of the spinout as it happens, or post the craft so that we can check out what happens.

[UmbralRaptor]

It looks like all of the (very draggy) intake area is ahead of the center of mass. Moving it back (most likely with ugly hacks involving radial intakes or cubic octagonal struts) may help. Especially if the yawing is also an issue at highish speeds at low altitudes.

[Claw]

Some things I can suggest without more pictures or being able to fly it myself.

1) Asymmetric thrust. Which I know you said you have surplus air, but sometimes the air distribution can be wacky. If the airplane always yaws and goes out of control in the same direction, then there are a couple things you can check just to make sure it's good. Do a run and watch the SAS input in the bottom left of the screen. When it starts to put yaw input in, right click on the motors to compare the thrust output. One of them may roll back a little even with excess air. Once the aircraft starts to yaw, you have even less intake air because of flying sideways, which makes the problem worse.

2) It might be a drag problem. Although it seems fairly uncommon to have drag problems in stock, they sometimes happen. If your craft goes out of control in more than one direction, you can investigate this. I think those big precooler looking scoops can cause some weird drag problems at high speed, but I'm not certain because I don't use them much. You can try to add more tail area to the back, or come up with some other way to offset the drag at the aft end of the craft.

3) You likely do not need to put your tails straight up. It might help, but that's because of two things.

- Make sure they are set to YAW only. (Which it sounds like you did.) If they are reacting during pitch and roll, then it will likely cause problems.

- You might simply need more vertical tail area. Straightening them would help this, but you can also put on other tail area to help it out.

I have plenty of designs with angled tails. While they can make designs a bit trickier, you don't have give them up completely.

4) For the same reason as the angled tails, your outer wing control surfaces could be causing problems. If the yaw starts while you're putting in pitch or roll inputs, then this could be it. The reason is because they are swept back. Putting in roll inputs will also cause a small bit of yaw input.

Just a few thoughts. Not sure if any of that will help.

Cheers,

~Claw

[Sirine]

A craft file upload, really helps.

[Jett_Quasar]

Generally control problems with delta wing craft can be resolved with the addition if canards.

JR

[Kasuha]

The distance between CoM and vertical stabilizers is very small, they have very little leverage. I also think that the fact that they are not vertical might cause some problems especially if you have any but yaw action switched off on them - they could have decent pitch and strong roll effect too.

I would probably battle it with reaction wheels, there's not much room to place another vertical stabilizer.

[JDCollie]

-snip-

In addition to everything else everyone has said, you may try experimenting with your flight profile. I generally try and level out a bit at 20km, rather than 25, as a lot of my designs used to have problems maintaining sufficient lift at 27+km without some speed gain first.

[Dangerous_Beans]

I would probably battle it with reaction wheels, there's not much room to place another vertical stabilizer.

the nova-punch part pack has a long, low strake which i have found useful for adding yaw stability when attached to the ventral surface of the plane. it acts as a vertical stabiliser, but is short enough not to mess up your ground clearance.

you might be able to do something similar with the stock control surfaces, with all the control options disabled?

[Red Iron Crown]

Thanks everyone for the helpful replies!

what if you add a slight dihedral to the main wings?

also, be helpful if you could show a video or some screenshots of the spinout as it happens, or post the craft so that we can check out what happens.

It's my understanding that anhedral/dihedral only affects roll, am I wrong in this? Video is right out, my PC can barely handle KSP as it is. I'll take some screenshots tonight of the behavior if the advice given here doesn't resolve it.

It looks like all of the (very draggy) intake area is ahead of the center of mass. Moving it back (most likely with ugly hacks involving radial intakes or cubic octagonal struts) may help. Especially if the yawing is also an issue at highish speeds at low altitudes.

The intakes are fairly low mass, I would expect their contribution to drag to be overpowered by the control authority of the stabilizers. It handles well at low speeds, but it may be a case of drag being proportional to v^2 and lift only to v.

Some things I can suggest without more pictures or being able to fly it myself.

1) Asymmetric thrust. Which I know you said you have surplus air, but sometimes the air distribution can be wacky. If the airplane always yaws and goes out of control in the same direction, then there are a couple things you can check just to make sure it's good. Do a run and watch the SAS input in the bottom left of the screen. When it starts to put yaw input in, right click on the motors to compare the thrust output. One of them may roll back a little even with excess air. Once the aircraft starts to yaw, you have even less intake air because of flying sideways, which makes the problem worse.

2) It might be a drag problem. Although it seems fairly uncommon to have drag problems in stock, they sometimes happen. If your craft goes out of control in more than one direction, you can investigate this. I think those big precooler looking scoops can cause some weird drag problems at high speed, but I'm not certain because I don't use them much. You can try to add more tail area to the back, or come up with some other way to offset the drag at the aft end of the craft.

3) You likely do not need to put your tails straight up. It might help, but that's because of two things.

- Make sure they are set to YAW only. (Which it sounds like you did.) If they are reacting during pitch and roll, then it will likely cause problems.

- You might simply need more vertical tail area. Straightening them would help this, but you can also put on other tail area to help it out.

I have plenty of designs with angled tails. While they can make designs a bit trickier, you don't have give them up completely.

4) For the same reason as the angled tails, your outer wing control surfaces could be causing problems. If the yaw starts while you're putting in pitch or roll inputs, then this could be it. The reason is because they are swept back. Putting in roll inputs will also cause a small bit of yaw input.

I'm not making any control inputs when control is lost, though SAS is on and likely doing small corrections. I don't think it's asymmetric thrust as I was closely monitoring the jets for just that issue, ready to modulate the throttle if necessary. The jet engines and intakes were made as an assembly that was duplicated with symmetry, so to my understanding the distribution of intake air should be as symmetrical as possible.

A craft file upload, really helps.

Craft file. Has Spaceplane Plus parts and landing gear from B9.

Generally control problems with delta wing craft can be resolved with the addition if canards.

There are a small pair of canards by the intakes attached to the main body. Though I think they are likely not related to yaw problems.

The distance between CoM and vertical stabilizers is very small, they have very little leverage. I also think that the fact that they are not vertical might cause some problems especially if you have any but yaw action switched off on them - they could have decent pitch and strong roll effect too.

I would probably battle it with reaction wheels, there's not much room to place another vertical stabilizer.

That's a good point about the leverage, hadn't thought of that. I'll try moving them back as far as possible and making them vertical. The vertical control surfaces are set to yaw only.

I'd like to solve this without adding reaction wheels if possible, it seems to me like it should be possible.

In addition to everything else everyone has said, you may try experimenting with your flight profile. I generally try and level out a bit at 20km, rather than 25, as a lot of my designs used to have problems maintaining sufficient lift at 27+km without some speed gain first.

I level off as my altitude gets to around 20km but the 20-25km change still happens quickly. Lift shouldn't be an issue as the plane has >1 TWR and could launch like a rocket.

Edited August 25, 2014 by Red Iron Crown

[Overfloater]

Of-course it will spin-out, but shutting down the jets won't be enough, you also need enough force from SAS or RCS to stop the rotation.

On my small plane which has 2 jets, in SPH i reduced thrust limiter of everything to "0", except for the jet - so i can point it exactly to the mass center. Then turn on symmetry and place the jet. This makes it spin-out proof.

Also, sorry if i'm stating the obvious... at 20,000m and above, make sure your vertical speed dial is no higher than 10.

This way you have enough time to respond to a flame-out. And just to be sure, right click the 2 jets and keep an eye on their thrust - as soon as one of them loses thrust immediately throttle down a bit. Keep doing that until you no longer gain speed by jets. (vertical dial still no more than 10!) At that moment turn on rocket and as soon as a spin out starts, shut-down jets. Then point to 30~45 degrees until apoapsis is in space.

Speaking of rocket, i suggest a different (more powerful)one, because in a setup like this i don't think you can circularize on jets alone. (but i haven't tried you plane so i might be wrong)

[mythic_fci]

Hmm, maybe I can help. Could you send me the craft file? If I have time, I'll try and fix it myself. In the meantime, try shifting some intakes more aft of the COM, and perhaps move the engines closer to the centerline of the craft. The SR-71 pilots faced severe problems with stability during shock wave capture intake failure (aka an "unstart") due to the fact that as well as the extreme aerodynamic drag this caused, the engine was in a mid-wing position, which is very hard to adjust for. In my latest 4-engine RAPIER design, I've incorporated a "auto/manual/manual/auto switch sequence". The engines farthest from the COM will auto-switch to rocket mode when low on intake air, because they will cause EXTREME yawing if they fail due to intake issues. The engines closest to the COM (aka attached directly to the fuselage) are manually switched using AGs, because if one fails, SAS can easily compensate.

[Red Iron Crown]

Of-course it will spin-out, but shutting down the jets won't be enough, you also need enough force from SAS or RCS to stop the rotation.

On my small plane which has 2 jets, in SPH i reduced thrust limiter of everything to "0", except for the jet - so i can point it exactly to the mass center. Then turn on symmetry and place the jet. This makes it spin-out proof.

I'm not sure what you mean by this exactly, can you clarify?

Also, sorry if i'm stating the obvious... at 20,000m and above, make sure your vertical speed dial is no higher than 10.

This way you have enough time to respond to a flame-out. And just to be sure, right click the 2 jets and keep an eye on their thrust - as soon as one of them loses thrust immediately throttle down a bit. Keep doing that until you no longer gain speed by jets. (vertical dial still no more than 10!) At that moment turn on rocket and as soon as a spin out starts, shut-down jets. Then point to 30~45 degrees until apoapsis is in space.

I'm reasonably sure it's not a flameout/differential thrust problem as I was monitoring it closely, but given your reply and the others suggesting it I'll try again and see if I can detect it.

Speaking of rocket, i suggest a different (more powerful)one, because in a setup like this i don't think you can circularize on jets alone. (but i haven't tried you plane so i might be wrong)

I'm trying to use the ascent profile where you push the apoapsis to the desired height on jets alone, the use rocket power to correct for drag as you coast to it and then circularize. With other planes that I've done this with the circularization burn is only 80-100m/s with tons of time to complete it, so the low thrust of the LV-N shouldn't be a problem (I hope).

Edit: FCISuperGuy, the .craft is here, has SP+ and B9 parts.

Edited August 25, 2014 by Red Iron Crown

[Kasuha]

One more take on the topic.

The question is your altitude-speed profile. With so few intakes per engine, you may need to be really fast while you're still relatively low. Otherwise you'll be getting flameouts even with correct intake/engine interleave, just because one intake set won't provide enough air to feed one engine above its flameout threshold.

Here's my slightly aged attempt with one intake per engine - normally I wouldn't try to fly that fast that low as drag loses are too high - in fact the plane is at about its maximum speed for that altitude. But ascending too quickly always led to flameout and spin.

[Red Iron Crown]

There are three intakes per engine, not sure if that's enough or not. 0.0195 units of intake area per jet engine, though I'm not sure how intake air is calculated.