Time to learn space planes... What's wrong with mine?

[ForScience6686]

18 hours ago, MitchS said:

Mk. 2 is largely the same, just with a cargo bay up front containing a simple 2t probe.

- Added payload bay
- Rebalanced some fuel--the precooler is further forward than the mk1 fuel tank to shift CoM further back 
- Slid the whole wing assemblies forward to offset the mass of the payload bay
- Adjusted canards
- Moved landing gear closer to new CoM... rotation speed was nearly 100m/s in the old position



The CoM is much harder to manage here. With the payload, it's too far forward, and SAS struggles to keep my desired attitude. Without the payload and low on fuel, the CoM slides way too close to CoL and I worry that it's not going to be aerodynamically stable. Is the solution here to just manually balance fuel before reentering? What about scooting the runway CoM back a little? This was a hard "happy medium" for me to find.

Also, why can't I alt+WASD trim anymore? Did that keybind get relocated for 1.2?

One tip I'll throw in.  I like to place an antenna or some other component on the top of the fuselage right at col.  This way you can manually check your com to col relation by zooming in on the craft.  Since the camera is set to the the com of the vehicle you can ensure that you are stable before re entry and adjust if needed.  This has become standard on all my planes to quickly check this in flight.

[AeroGav]

    

38 minutes ago, MitchS said:

Fascinating. So can one of you (or both!) boil down your last few posts into a couple of takeaways? What are your bottom lines, where do you agree with one another, and where do you disagree? Am I to understand that canards are a less stable design than one without them?

My takeaways :

• no difference in canard and tailplane handling qualities if both are adjusted to the same CoL/CoM position with an accurate CoL/CoP indicator.   Get CorrectCoL !
• shifts in CoM due to fuel burnoff and cargo unloading are much more likely to mess things up.   Get RCS build aid !
• you can tell when your craft is becoming unstable before it flips out - if your nose is above the horizon, SAS should be needing to apply UP inputs to keep it there.  If you see it maintaining a continuous nose down trim, watch out ! Transfer fuel forwards ASAP, and don't pull any manoeuvres.
• your craft has huge wing area and should not need to fly large pitch angles anyway on re-entry to avoid burnup. Still, always try to design stable and idiot proof airplanes.

[foamyesque]

19 minutes ago, ForScience6686 said:

One tip I'll throw in.  I like to place an antenna or some other component on the top of the fuselage right at col.  This way you can manually check your com to col relation by zooming in on the craft.  Since the camera is set to the the com of the vehicle you can ensure that you are stable before re entry and adjust if needed.  This has become standard on all my planes to quickly check this in flight.

 

This is also handy while you're building it if you don't have various mods; you can stick it at your dry CoM position, then fill the tanks to see what the movement is, and mark that as well. Makes balancing loading throughout flight much easier without recourse to modding.

[Starhawk]

2 hours ago, GoSlash27 said:

there are folks out there who (IMO) build better SSTOs than mine (lookin' at you, @Starhawk

Well, I don't know about that.

What I do know is that the vast majority of my knowledge of the underlying theory of aerodynamic design has come directly from you, Slashy.

You do me a great honour, Slashy, since I consider you one of the masters of spaceplane design around here.

FWIW, I never put rudders on any spaceplane I build.  I haven't for quite a long time now.  It was you that clued me in to that.

And, yeah, I tend to avoid forward canards.  Not to say that I never build with them.  I have and sometimes still do.  Overall, though, I leave them alone as it generally seems to work out a bit better that way.

Happy landings!

[MitchS]

Thank you all for the next wave of advice!

...Is it normal that it just took me 23 minutes to reach orbit?

[bewing]

1 hour ago, MitchS said:

...Is it normal that it just took me 23 minutes to reach orbit?

Yup.

Sometimes getting there a little faster saves a little fuel. But sometimes you just have to let it follow prograde the way it wants to.

[GoSlash27]

3 hours ago, MitchS said:

Thank you all for the next wave of advice!

...Is it normal that it just took me 23 minutes to reach orbit?

MitchS,

 It's within the expected range, but on the long side. I have built some intentionally underpowered demonstrators that took longer, like old "Bertha" here:

Still... taking that long to get to orbit is an indication that you could probably save some fuel with a different approach. Can you tell which phase of the flight was lagging?

Best,
-Slashy

Edited December 17, 2016 by GoSlash27

[foamyesque]

Is that fifteen minutes including a coast to apoapsis? My spaceplane's powered flight usually only lasts for about five minutes, some less.

[AeroGav]

My Firefox design i posted about on page 2 of the thread is very comparable to your ship Mitch - same number of engines, same cargo bay, mass also probably very close.

Takes about 7 minutes to get supersonic at 10km.  Main engine cutoff after about 15 minutes at an altitude of 42km and 2055 m/s.    After that, another 15 minutes of coasting then the circularisation burn, some of which you can time accelerate through.

Edited December 17, 2016 by AeroGav

[AeroGav]

@GoSlash27OK,  I've finally realised why my canard is a lawn dart and yours is wildly unstable.  

Spoiler

 

 

The Eureka! moment came to me in the bath,  I should probably have more of those.

There is no voodoo involved, and informational mods like correct CoL aren't relevant.

Your main wings have incidence but the canards do not.

Your canards are angled 5 degrees lower than the main wing in flight.  That's a large nose-down trim.  It causes the CoL indicator to move backwards,  but this does not give a true picture  - you actually have more lifting surface area ahead of CG than behind it.  

Spoiler

 

On the left I recreated Slashy's SSTO.   It looks stable, when fuelled, neutral or marginal when dry.      However when i angle the canards the same amount as the wing, the true picture emerges.  It is neutral when full and highly unstable dry. 

 

At zero AoA, pointing directly prograde, the craft appears balanced.  But as AoA increases, the canard's lift rises faster than the main wing's, because it is starting from a lower value.  Eventually the main (rear) wing stalls while the front is still generating lift , and the airplane flips.

 

To fix this - angle the canards up by the same amount as the main wing. This causes the CoL to move forward to its true position.  You will have to adjust your wings back some more to get the CoL touching the back of the yellow ball again.   This will be stable, but it still will not fly the same as your original craft.

Why? Because I bet the original SSTO has angled main wings and a tail plane that is not angled.  Tail plane angled 5 degrees below the main wing is a substantial UP trim.   It moves the CoL forwards forcing you to slide the wings further back to compensate, leading to an airplane with a very aft CoL.   And of course, the main (front) wings will always be at a lower AoA than the tailplane, and will stall first.

Spoiler

 

Original ship - tailplane at 0 incidence, wing at 5 degrees.  Wing stalls first, stable configuration.

 

Modified version of Slashy's canard with foreplane angled at 10 degrees and main wing at 5.  Thus, it has the same built in nose UP trim as the original tailed craft, which forces us to design a very aft CoL to compensate and results in the forward surface stalling first.

 

Spoiler

Notice how far back we need to move the wing to get the indicators to balance -

however, you'll see that the CoL moves aft, away from CoM as you pitch up.   This craft will be very departure resistant.

 

Conclusion 

Canards are OK

Adding incidence is OK

Canards + Incidence whilst forgetting to angle the canards too = flip happy

IF you are going to use incidence, lifting surfaces ahead of CoM should have incidence => incidence of surfaces behind CoM.

 

 

 

Edited December 17, 2016 by AeroGav

[GoSlash27]

AeroGav,
 I did rotate the canard to match the wing. I wouldn't have been able to get it to orbit otherwise. I keep tellin' you why it was unstable during reentry. The reason yours is not unstable is because your CoM is longitudinally centered and you have plenty of tailplane area to offset the drag of the canard at high AoA.

Best,
-Slashy

[GoSlash27]

I think I've got a way to help you visualize this. Imagine that you have a balloon shaped like the spaceplane. You tie a string to it where the CoM would be. Hang it out the window and start driving.

 

The way that balloon would behave is exactly the way your plane will behave during reentry.

With my bodged-together canard design, those canards want to be anywhere but in front. With nothing to counterbalance their torque, the plane will pitch away from prograde, putting the rudder out front. It doesn't want to be there either, so it begins to yaw the plane to get back where it wants to be and also induces a roll. The main wing is powerless to affect this because it's too close to the CoM to have any leverage. The two continue to fight with each other, tumbling the aircraft wildly, until lift becomes dominant over pure drag.

There are several ways that you can alter the shape of your spaceplane balloon so that this won't happen. You can put tailplanes on it to counteract the canards. You can move the CoG forward so that the main wing has enough leverage to counteract the canards. You can move the canards back so they have less leverage. You can even suspend the CoG between 2 equally sized wings. But the easiest way to cure it is to put the canards in the back instead of the front.

 What you don't want is a CoG way in the back surrounded by wing and a bunch of Mk2 fuselage/ canards hanging way out front.

HTHs,
-Slashy

 

Edited December 17, 2016 by GoSlash27

[MitchS]

Okay, everyone! Busy couple of days, but I'm back to the game. I've had some success with my first design, and can deploy two small 2t probes to an equatorial LKO with a couple hundred delta-v left over to get home. I also built a version that replaces the cargo bay with a crew cabin and some additional fuel. I can do equatorial LKO rescues with it no problem, except that the powered flight still takes so long.

I've started a new design to broaden my skillset--an unmanned SSTO built around a crew cabin, to be able to do rescues in various orbits of Kerbin. I swapped the whiplash out for a RAPIER, but have had no luck whatsoever. I figured that by streamlining the design and making it smaller than my first one, I could get away with one engine, but I'm thinking this one is just too underpowered. I capped out below 10km and around 310m/s. Power dives ineffective. Is a one-RAPIER configuration possible? What do I need to be doing differently?

[OrbitalBuzzsaw]

1 minute ago, MitchS said:

Okay, everyone! Busy couple of days, but I'm back to the game. I've had some success with my first design, and can deploy two small 2t probes to an equatorial LKO with a couple hundred delta-v left over to get home. I also built a version that replaces the cargo bay with a crew cabin and some additional fuel. I can do equatorial LKO rescues with it no problem, except that the powered flight still takes so long.

I've started a new design to broaden my skillset--an unmanned SSTO built around a crew cabin, to be able to do rescues in various orbits of Kerbin. I swapped the whiplash out for a RAPIER, but have had no luck whatsoever. I figured that by streamlining the design and making it smaller than my first one, I could get away with one engine, but I'm thinking this one is just too underpowered. I capped out below 10km and around 310m/s. Power dives ineffective. Is a one-RAPIER configuration possible? What do I need to be doing differently?

Yaw control surface, also lose the RCS. You don't have a docking port so no need. Go for a pair of RAPIERS on the wings with a nuke on the back.

[AeroGav]

3 minutes ago, MitchS said:

Okay, everyone! Busy couple of days, but I'm back to the game. I've had some success with my first design, and can deploy two small 2t probes to an equatorial LKO with a couple hundred delta-v left over to get home. I also built a version that replaces the cargo bay with a crew cabin and some additional fuel. I can do equatorial LKO rescues with it no problem, except that the powered flight still takes so long.

I've started a new design to broaden my skillset--an unmanned SSTO built around a crew cabin, to be able to do rescues in various orbits of Kerbin. I swapped the whiplash out for a RAPIER, but have had no luck whatsoever. I figured that by streamlining the design and making it smaller than my first one, I could get away with one engine, but I'm thinking this one is just too underpowered. I capped out below 10km and around 310m/s. Power dives ineffective. Is a one-RAPIER configuration possible? What do I need to be doing differently?
 

The Rapier has a less generous thrust curve below mach 1 so it can be harder to get through sound barrier than on the Whiplash.  Of course it's a whole lot better at high speed.  Just think of it like you bought a Stage IV turbo / Race cam kit for your car,  top end power will be amazing but you'd expect to loose a bit of drivability around town.

As for that passenger design, its covered in what appear to be warts,  maybe it isn't feeling well?   Alternatively, you might want to reconsider putting so many rcs on the thing, they do add substantial drag.    In fact, if you attach them to the cargo bay, you will probably find they refuse to fire when the bay doors are shut because the game considers them to be inside, but the good news is they will make no drag. Any ports that do in fact still fire when the doors are closed can be sunk inside the bay slightly with the offset tool until they no longer work.

Ideally you'd have the cargo bay right on the cg,  but you are packing a lot of reaction wheel there so it can probably offset the torque they generate when translating.   

How are you controlling pitch there btw?  Engine gimbal and reaction wheel will do much i assume, but the elevons will struggle being so close to CG.  Those big S ones are very powerful  just hope they don't generate too much trim drag or downforce working from their disadvantaged position.

Also how you doing for electricity?    Rapier got no alternator, you were kind of spoiled for choice on the other design i bet, two LV-N and a Whiplash is enough to run an AM Rock radio service.

[AeroGav]

Also,  I just remembered a nasty cargo attachment issue that can create vast amounts of drag.

It is quite easy for stuff to get attached to the front/rear of the fuselage pieces ahead of and behind the cargo bay rather than the cargo bay's attach node itself.   The game will then consider the contents to be outside the ship and give you a drag penalty accordingly.  If you do somehow make it to orbit expect these items to overheat and explode on reentry.

Two ways to check (quick)

• make the cargo bay the root part with the re-root tool (4) and temporarily detach the fuselage section behind.  Does your cargo come away with the passenger cabin or stay with the cargo bay? Do the same to the front.   Save your ship first or Ctrl Z to undo after.

Thorough method

• alt f12 to bring up debug menu
• physics tab, aero submenu, tick the "enable aero data in action menus" box
• in flight, you can now see how much drag each component is generating, and crucially, "Shielded : True / False"
• On the runway, open the bay, right click each item in the bay and pin the menu in place. Close the bay.  They should all toggle to "Shielded : True" when the doors shut - if not you have an attachment problem.

 

Anyway, I could speculate all night , but the only true way to know is to share the craft file, unfortunately i'm not going to have much time to play till 22nd Dec !

[MitchS]

Aw, man... I like having RCS capability for rescue missions... It takes a lot of the tedium out of setting up precise intercepts on the first shot--I'd rather puff around with a little bit of translational RCS to get a 0.0km intercept than rotate, rotate, fire engines gently, rotate, rotate, rotate, fire engines a little too hard, ad nauseum. I never thought they induced drag... I guess I wouldn't pay much of a drag penalty for that convenience though. I'll take 'em off. 

The Big S control surfaces are, again, vestigial organs from a previous design. Whoops. I used to have the wings much further back. I'll remove them and opt for a conventional tailplane. 

I've flown a few ships with and without a yaw control surface, and I'm pretty convinced by those of you who were saying they aren't needed. The only reason I find myself wanting to use them is because "it's the right thing to do," but I don't really think that's very compelling anymore. Especially with the torque wheels I've been using on my spaceplanes. KSP doesn't need rudders, and rudders cause drag. No rudders.

I'm off to go have dinner with the fiancee and her family, so maybe later tonight I'll ditch the RCS, install the tailplane, and check the cargo bay connections, then do a test flight to see if my drag penalty is reduced.

In the meantime, if you guys could share some insight on how you've best implemented RAPIERs into your light SSTO designs, I would appreciate the reading material!

[AeroGav]

Like i say, you can have rcs so long as you  bury the ports in the cargo bay.  But i couldn't see a docking adapter on the ship so wondered why  you were taking it up there.  BTW two tanks is a lot of RCS !  I tweak my RCS ports to respond to translate command only, not rotation , since the wheels can do that. 

Tailplanes for yaw suppression are great, but as Slashy et al have said , the actively steering rudder tabs on the ends of them are optional.   The only time they do good is if you have prograde set on a craft with angled wings and are flying it on prograde assist, when they act like a rather effective yaw damper.   But i only started angling my wings recently, so for the first 1500 hours of KSP playtime i was dragging them around as payload.

re: single rapier ship

getting through the sound barrier is the main difficulty.  After that you'll have more thrust and less weight (no nukes) than your other plane.  In rocket mode you have 50% more power than the nuke bird with less weight, really high twr.  But ISP is nearly three times worse and the fuel gauge will unwind like a broken clock.     So make sure you take advantage of the high airbreathing top speed of the new rapiers , and lessen the delta v gap to be bridged.

Edited December 18, 2016 by AeroGav

[zolotiyeruki]

On 12/16/2016 at 4:39 PM, ForScience6686 said:

One tip I'll throw in.  I like to place an antenna or some other component on the top of the fuselage right at col.  This way you can manually check your com to col relation by zooming in on the craft.  Since the camera is set to the the com of the vehicle you can ensure that you are stable before re entry and adjust if needed.  This has become standard on all my planes to quickly check this in flight.

That's funny--I use the exact same trick!

[GoSlash27]

MitchS,

 The ship I posted upstream was a single RAPIER design, both the easy version and the bodged canard version. It's a larger and heavier ship than you've posted here, so clearly it is possible. In fact, a single RAPIER is capable of handling much more cargo than I loaded it with there.

If you can't clear Mach one, then your overriding concern is drag. You must look for ways to reduce the parasitic drag to get you past Mach 1.2. Once you achieve that, drag reduces and thrust increases making the rest of the journey a cakewalk.

You need zero AoA through the Mach barrier and minimal parasitic drag. Those control surfaces are also probably overkill. Make sure you're riding the air-breathing mode for all the velocity you can get.

The one thing that sticks out for me is all the reaction wheels. They make a nice work- around for a fundamentally unstable design, but they can hold you in an aerodynamically inefficient flight profile when SAS is engaged. A truly clean design can work aerodynamically without reaction wheels or with reaction wheels without control surfaces. I've seen it (and done it) done both ways, but IMO *minimal* control surfaces work better because they naturally revert to a low drag state.

Good luck,

-Slashy

 

 

 

 

Edited December 19, 2016 by GoSlash27

[foamyesque]

13 hours ago, MitchS said:

Aw, man... I like having RCS capability for rescue missions...

 

<snip>

In the meantime, if you guys could share some insight on how you've best implemented RAPIERs into your light SSTO designs, I would appreciate the reading material!

 

I like RCS capabilities as well. Being able to dock or perform precision orbital adjustments without needing to change the spacecraft attitude is always nice, and the extra control oomph for reentry can fix or cover up very nasty control in high-angle flight (and in transitioning out of it and into a glide). The trouble you're hitting is that the lighter a craft is, the more significant drag from features like a set of RCS thrusters (or external solar panels, landing gear, etc) becomes. A 400 tonne spaceplane hardly notices those things but they can completely break a 4 tonne one.

So for small ships you have to optimize really heavily for drag. The critical feature of RCS is the ability to translate on all axis; since KSP reaction wheels are so powerful the rotational capabilities are often (though not always) overkill. That means you need only five Place-Anywheres -- dorsal, ventral, left, and right, all on the longitudinal CoM, and one on the nose to provide aft-translation. Fore-translation can be achieved with the main engines or a sixth place-anywhere. Alternatively, two place-anywheres and two quads, assembled into five-way thrusters, will also provide full translation capabilities.

 

I will disagree about putting them inside cargo bays. The skin drag of cargo and service bays is not trivial and they are significant extra mass. If you already have one for mission critical items, then sure, take advantage, but otherwise I'd stay away.

Edited December 19, 2016 by foamyesque

[Palaceviking]

I vote ditch the rcs completely,  you don't need it for rescue missions. 

[GoSlash27]

By way of illustration, here's the biggest single RAPIER design I've done:

Needless to say, it was on the ragged edge and I had to work to get it past Mach 1. I have found that anything that can exceed 380 m/sec can make it to hypersonic speed.

Best,
-Slashy

Edited December 19, 2016 by GoSlash27

[Wanderfound]

You're getting there.

Most of the basics are already covered, but to recap:

1) Enough jet thrust to reach take off speed and crack transonic. Any more is a mass-wasting luxury.

2) Enough post-oxygen thrust to reach orbit without losing excessive energy to drag. If you're going interplanetary, nukes are worth the hassle; if not, dump the nukes and mount an oxidising rocket instead. Pure LF planes are possible, but LF/O designs are much easier.

3) For high-tech orbital, just RAPIERs. For low-tech orbital, Whiplashes plus either Swivels, Thuds or Terriers. For beyond orbital, RAPIERs plus a nuke or two.

4) If you're going orbital, enough LF tankage to run the jets until switchover, but no more. If you're going nuke/RAPIER, enough oxidiser tankage to lift the apoapsis above 70km, but no more. A nuke ship should circularise on the nuke, and carry as much LF tankage as it can lift.

5) Go easy on luxury goods, especially drag-inducing ones like RCS. If you don't need 'em, don't take 'em. This includes things like using bulk SAS to compensate for aerodynamic imperfections. If you use the thrust limiter, even high-power engines are fine for rendezvous.

6) Think about heat tolerance. Low-tech aircraft parts are not designed to handle reentry. If you're careful enough, even a low-tech plane can manage it, but it's a lot easier if you stick to high temp tolerant parts.

 

SSTA demonstrator:

 

Edited December 25, 2016 by Wanderfound

[Wanderfound]

On 19 December 2016 at 8:50 AM, MitchS said:

In the meantime, if you guys could share some insight on how you've best implemented RAPIERs into your light SSTO designs, I would appreciate the reading material!

Examples do it best:

Album at http://imgur.com/a/tsf3M

Craft file at https://www.dropbox.com/s/zgpi3ed6efvlg7w/Kerbodyne Unorap.craft?dl=0

Things to note:

1) Minimal drag; no unnecessary surface clutter.

2) Good heat tolerance on the leading parts.

3) Enough fuel to get up, plus a bit of extra LF in case you miss your reentry.

4) Minimal weight shift as the fuel drains.

5) You could easily add a small service bay to this if you want to stash some science gear.

Edited December 25, 2016 by Wanderfound