Need help to buid SSTO

[guitounet]

Hi Guys,

 

i'm trying to build a big SSTO to travel to far away planets, moons, asteroids and so on... My goal is to go far, do many things and come back safely (as everybody wants :D).

I have some background to launch small ssto to orbit around kerbin and it's moons but only with small ships (ie no payload or very few). I achieve this with many fly test and adjustments and "by chance" my ships were finally created.

Now getting things bigger chance is no more working and my tests flys are just crash test with no learning at all...... So i beg help to experienced engineers !

My goal is  :

  -  20-30t of payload
  - 1 crew
  - 1 passager cabin
  - a ramp to deliver some payload

Here is my actual design, almost working... With this I can go up to ~40km, but no more, i ran out of oxydizer for RAPIER.

 

Here is my fly plan that i try to stick

1. Full throttle
2. 10° pitch and getting to 10Km hight
3. try to keep altitude around 10km and wait until Mach 4/4.5 (usually i achieve this speed while i am around 20km)
4. change pitch to 30-45°
5. rapid ascent to 26km
6. change rapier to closeCycle
7. try to keep my pitch and heading after (30km it's very hard) and plane is almost uncontrollable
8. 40km.... oxydizer depletion... Get back to kerbin

What I have to to IMOO there are main objective :

  - set my plan to lighter weight
  - how to set it more manoeuvrable after 30km

Some more info about the ship :
  mass : 192 000/377 000
  liquidFuel : 38500 units
  oxydizer : 9000 units
  monopro : 1000 units

Screenshots :
https://www.flickr.com/gp/149071536@N08/8h6tHw

 

What's wrong with my ship ? Can someone help me ?

Thanks !

 

Edited March 24, 2017 by guitounet

[Jarin]

1 hour ago, guitounet said:

What's wrong with my ship ? Can someone help me ?

 

Largely your ascent being too aggressive. You want to burn as close to horizontal as possible while not losing lift. Gravity losses are a SSTO's worst enemy. You want to be pitching no more than about 10°, as well as running on air-breathing mode to 1500m/s. Your rapiers should only be there to give you a kick towards orbital speeds so your nukes can carry you the rest of the way. The less Ox you can get away with carrying, the more dV you have on the nukes once you get to orbit.

Speaking of nukes, I only see two in those screenshots. That means your space TWR is going to be abysmal with a craft that large. I assume that's also why "engage the nukes" was not step 6 on your listed ascent profile. Toss a couple more on there and drop some whiplash weight. With a good ascent profile, 14 rapiers should have NO trouble carrying you to the edge of space.

So, just to clarify, the details will change depending on craft and pilot, so I'll keep it to basic principles, but your general flight path should look like this:

1. Get the rapiers up to speed. Some people do this at sea-level, some at 10-15km. It doesn't much matter. Just run level at a convenient altitude until you're passing mach 2 or so. 

2. Shallow ascent. You want to hit 1500m/s before 25km altitude, which is where the rapiers start faltering. A 35° ascent is for rockets and hot-rods, not utility SSTOs. This should also help your stability; I suspect you're stalling and dragging through the upper atmosphere.

3. Engage the nukes before rapier changeover. They're plenty useful at 20km on up, and that extra thrust can let your air-breathers get a bit more run-time. 

4. Keeping at 10° AoA or less, accelerate on nukes and LFO rapiers. If you have KER, keep an eye on the AP on the HUD. As long as it's going up, you're fine. Don't worry if your Ox runs out around 40km; just keep riding those nukes; you should have enough speed that your AP will still be climbing. 

Your major design concerns are balancing how much Ox you carry, and how much engine weight with nukes. Without a .craft file to tweak I can't give the best advice here.

[bewing]

When you change pitch, you lose a lot of speed. So I think that is one mistake you are making -- you do not want to steer at all if you can avoid it. Any pitch higher than 10 degrees while you are still below 30km altitude is very draggy. Also, this is a very heavy plane, and it is hard to put enough wing area onto such a plane. So you probably do not have enough.

Your plane may have too much drag in the parts that you used -- we will need screenshots or a .craft file to be sure. To have control, you need big enough control surfaces in the right place -- and you need to balance the CoM just right.

Other than that, you may need to reduce the weight. Maybe get rid of all that Monoprop you have. If you have monoprop, you probably also have RCS thrusters, which have weight and drag -- so maybe you can get rid of them, too.

 

[AeroGav]

A long range SSTO needs to have a fairly high fuel fraction and carry a minimum of oxidizer.  That means keeping engine mass down to a minimum, which means boxing clever with your flight profile, as well as designing for low drag and high lift  (in stock aero, wings generate very little drag, but large wing area  reduces fuselage drag by raising your altitude and lowering your AoA).

As a rule of thumb,  I'd say you want one jet engine per 25-30 tons.    The limit is getting supersonic.  Once you are supersonic, ramjet effect makes engines much stronger, and you only need one RAPIER per 50 tons to hit 1500 m/s in level flight at 22km.    So, half the jet engines should be Whiplashes (lighter, much bit better performance below mach 1) , half should be RAPIERs.  BTW, you can even use nukes to help go through sound barrier if you're struggling with that.  They produce 75% of max thrust from 6km upwards.

For rocket mode, you want 60kn of thrust per 15-20tons of launch weight.     One RAPIER produces 180kn in closed cycle mode, so if you took my previous advice, you already got way more closed cycle thrust than you need.  However RAPIERs in  closed cycle have terrible ISP.     Generally I switch over only some of the RAPIERs to close cycle when air breathing acceleration peters out, and leave the rest running on air breathing until they quit completely at 30km.   I only bring enough oxidizer to boost my speed 300 m/s.    By that point orbital freefall effect is supporting 85% of your weight, so combined with residual lift from the wings you'll soon be cruising at over 35km,  at which altitude drag is negligible.   One nuke per 30 tons is easily enough to carry you the rest of the way to orbit.

Some of mine - 150tons, 3 rapier, 2 whiplash, 5 nukes https://kerbalx.com/AeroGav/Diamondback-gav2  (this was someone else's design i fixed up, has a drill if i remember correctly, but isn't finished)

 

Interplanetary cargo ramp , all my own work this time !

https://kerbalx.com/AeroGav/C7-Galaxian

Interplanetary CR100 bay ship, 5 R 6 Nuke -

As regards your design - 

1.  I prefer not to use large fuselage tanks for my liquid fuel.    Big S wings and strakes can do the same job while contributing useful lift.  Fuselage pieces are always draggy.

2. Quad couplers should have only 33% more drag than tri couplers as they only mount 33% more engines, but from what i remember they have something like 50 or 77% more drag.  2.5m Tri couplers are the sweet spot

3. The X wing config isn't optimal.  Flat wings generate 100% lift, dihedral wings sacrifice a bit of lift for extra roll stability.  Anhedral sacrifices lift ot make negative stability.   Just stick to a combo of flat and diheddral wings.

4. Cones on the back of nuke and rapier engines to reduce the drag from their  rear attach nodes. like in this video.

https://www.youtube.com/watch?v=XXhIb5OlSRY

Finally, what you are trying to do is very hard.    Ramps add a lot of drag and weight.   Rovers tend to flip over at lower speeds than an aircraft capable of carrying one.  Why cargo when you have interplanetary spaceplane ?  Unless you're trying to build  a base i suppose.

you can help yourself a bit by compromising on 100 % re-usability.   mount some Whiplashes radially on decouplers as well as drop tanks, to help you get off kerbin fully loaded.

[A_name]

I second the prior advice all of which is good.

However, I do want to emphasize on this point:

1 hour ago, bewing said:

When you change pitch, you lose a lot of speed.

Preferably, after you begin your speed run, you shouldn't be maneuvering AT ALL.

This is my go-to profile with Rapiers:

1. Have some incidence on your main wing. This will allow you to maintain a positive AoA and therefore gain vertical speed without having to pitch up the nose, thus maintaining the prograde vector.

2. Start speed run between 7km and 10km altitude. The lower your TWR, the earlier you'll want to start.

3. To start the speed run, point your nose between 5 to 10 degrees above the horizon. Again, the higher your TWR, the higher you can afford to point your nose during the speed run.

4. Don't touch the controls. At the speeds you'll be hitting, the curvature of Kerbin will give you all the pitch you need. That is to say, you'll always be pointing in the same direction but your nose will "drift upwards" as the ground curves away below you.

5. As soon as you stop gaining speed, switch to closed cycle. You should be going AT LEAST 1,200 m/s, preferably more. If you're going slower than that by the time you begin losing speed, it means you started the speed run too late or you pointed to high initially.  Don't touch the controls.

6. When your vertical speed exceeds ~300 m/s, or your time to Ap exceeds ~ 1 minute, or your Ap height exceeds ~35 km, switch your SAS to prograde lock. Don't touch the controls other than that. If you did the prior steps right, at this point your nose should be pointing at ~30 degrees.

7. Cut engines when your Ap is at ~75km (or whatever your target orbit is).

[AeroGav]

14 minutes ago, A_name said:

When you change pitch, you lose a lot of speed.

That is because by default, most control surfaces have a max deflection of 30 degrees, and because most people fly with keyboard (and ksp joystick support is poor) when people nose up they are getting all 30 degrees deflection,  which acts like airbrakes.  Also,  these "all or nothing" pitch adjustments send the whole aircraft into a large AoA change and you'll see 5 or 6 G being pulled when you;re making the adjustment.

I do make pitch changes in my flight profile,  because during the speedrun i'm keeping AoA very low to stop us climbing to an altitude where the engines loose power, but once we go closed cycle mode you want to run the aoa at the one that gives best lift / drag ratio.  You can reduce the drag that comes from pitch changes 

• reduce the authority of the control surface in tweakables or use one with a lower max deflection angle.  Eg. use an elevon with a 30deg max deflection at 30% authority limit, or  an advanced canard at 100% (max deflection 10 degrees)
• Change flight path by using pitch trim  instead of WASD keys.    Increase nose up trim with ALT and S, decrease it with ALT W.  Note this doesn't work with SAS on

example here -

or, you can use fine rotate tool on elevons/canards to make your aircraft fly naturally at 4 degrees AoA or whatever is optimum for lift drag ratio. During the speedrun i'm using nose down trim , when the speedrun ends i use alt + x to reset trim to neutral, which causes the aircraft to return to it's built in trim angle 

 

Or you can have some trim flaps bound to an action group to make the change , with the "authority limiter" of the trim flaps adjusted to give the right angle (i use RCS action group for trim flap here)

 

[Jarin]

1 hour ago, AeroGav said:

 

I really don't like this image. That climb to orbit line makes people think they need to haul back on the stick and climb like a rocket.

[Spricigo]

1 hour ago, Jarin said:

I really don't like this image. That climb to orbit line makes people think they need to haul back on the stick and climb like a rocket.

Actually they need to climb like a rocket in the rocket phase, just that at the time a rocket will be going mostly horizontal.

Also, I see your point but this is because of people misreading the graph. People will think in distance, while the graph clearly is about time.

[guitounet]

2 hours ago, Jarin said:

I really don't like this image. That climb to orbit line makes people think they need to haul back on the stick and climb like a rocket.

Exactly that's what I was thinking befaore reading all post from this thread.

4 hours ago, AeroGav said:

 

As a rule of thumb,  I'd say you want one jet engine per 25-30 tons.    The limit is getting supersonic.  Once you are supersonic, ramjet effect makes engines much stronger, and you only need one RAPIER per 50 tons to hit 1500 m/s in level flight at 22km.    So, half the jet engines should be Whiplashes (lighter, much bit better performance below mach 1) , half should be RAPIERs.  BTW, you can even use nukes to help go through sound barrier if you're struggling with that.  They produce 75% of max thrust from 6km upwards.

 

For rocket mode, you want 60kn of thrust per 15-20tons of launch weight.     One RAPIER produces 180kn in closed cycle mode, so if you took my previous advice, you already got way more closed cycle thrust than you need.  However RAPIERs in  closed cycle have terrible ISP.     Generally I switch over only some of the RAPIERs to close cycle when air breathing acceleration peters out, and leave the rest running on air breathing until they quit completely at 30km.   I only bring enough oxidizer to boost my speed 300 m/s.    By that point orbital freefall effect is supporting 85% of your weight, so combined with residual lift from the wings you'll soon be cruising at over 35km,  at which altitude drag is negligible.   One nuke per 30 tons is easily enough to carry you the rest of the way to orbit.

[.....]

As regards your design - 

1.  I prefer not to use large fuselage tanks for my liquid fuel.    Big S wings and strakes can do the same job while contributing useful lift.  Fuselage pieces are always draggy.

2. Quad couplers should have only 33% more drag than tri couplers as they only mount 33% more engines, but from what i remember they have something like 50 or 77% more drag.  2.5m Tri couplers are the sweet spot

3. The X wing config isn't optimal.  Flat wings generate 100% lift, dihedral wings sacrifice a bit of lift for extra roll stability.  Anhedral sacrifices lift ot make negative stability.   Just stick to a combo of flat and diheddral wings.

4. Cones on the back of nuke and rapier engines to reduce the drag from their  rear attach nodes. like in this video.

https://www.youtube.com/watch?v=XXhIb5OlSRY

Finally, what you are trying to do is very hard.    Ramps add a lot of drag and weight.   Rovers tend to flip over at lower speeds than an aircraft capable of carrying one.  Why cargo when you have interplanetary spaceplane ?  Unless you're trying to build  a base i suppose.

you can help yourself a bit by compromising on 100 % re-usability.   mount some Whiplashes radially on decouplers as well as drop tanks, to help you get off kerbin fully loaded.

Thanks, i'll will redesign engines and wings configuration And post a .craft file if i need more help (and sure when plane is over).
To be more precise of usage of this craft, i'll tell it's just for fun carrying big rover and to have a new goal, maybe will be used to build some station later !

5 hours ago, Jarin said:

Largely your ascent being too aggressive. You want to burn as close to horizontal as possible while not losing lift. Gravity losses are a SSTO's worst enemy. You want to be pitching no more than about 10°, as well as running on air-breathing mode to 1500m/s. Your rapiers should only be there to give you a kick towards orbital speeds so your nukes can carry you the rest of the way. The less Ox you can get away with carrying, the more dV you have on the nukes once you get to orbit.

Speaking of nukes, I only see two in those screenshots. That means your space TWR is going to be abysmal with a craft that large. I assume that's also why "engage the nukes" was not step 6 on your listed ascent profile. Toss a couple more on there and drop some whiplash weight. With a good ascent profile, 14 rapiers should have NO trouble carrying you to the edge of space.

So, just to clarify, the details will change depending on craft and pilot, so I'll keep it to basic principles, but your general flight path should look like this:

1. Get the rapiers up to speed. Some people do this at sea-level, some at 10-15km. It doesn't much matter. Just run level at a convenient altitude until you're passing mach 2 or so. 

2. Shallow ascent. You want to hit 1500m/s before 25km altitude, which is where the rapiers start faltering. A 35° ascent is for rockets and hot-rods, not utility SSTOs. This should also help your stability; I suspect you're stalling and dragging through the upper atmosphere.

3. Engage the nukes before rapier changeover. They're plenty useful at 20km on up, and that extra thrust can let your air-breathers get a bit more run-time. 

4. Keeping at 10° AoA or less, accelerate on nukes and LFO rapiers. If you have KER, keep an eye on the AP on the HUD. As long as it's going up, you're fine. Don't worry if your Ox runs out around 40km; just keep riding those nukes; you should have enough speed that your AP will still be climbing. 

Your major design concerns are balancing how much Ox you carry, and how much engine weight with nukes. Without a .craft file to tweak I can't give the best advice here.

Ok i'll will try this kind of flight plan as soon my plane will be redesigned !

Thanks for your help !

[AeroGav]

9 hours ago, Jarin said:

I really don't like this image. That climb to orbit line makes people think they need to haul back on the stick and climb like a rocket.

Yeah I need to fix this.  It is time/vs altitude from one of my recent flights.    I was only climbing at about 5 degrees pitch angle, but because we were going so fast horizontally, the vertical speed was huge.

Anyone who has better skills with Excel and MS Paint, feel free to jump in.

[ForScience6686]

8 hours ago, AeroGav said:

Yeah I need to fix this.  It is time/vs altitude from one of my recent flights.    I was only climbing at about 5 degrees pitch angle, but because we were going so fast horizontally, the vertical speed was huge.

Anyone who has better skills with Excel and MS Paint, feel free to jump in.

The graph is fine, people's misunderstanding of it is on them.  It is clearly marked and explains the procedure well.

[Kryxal]

It's probably the sharp discontinuity in the graph ... that implies an angle change instead of accelerating to orbital velocity.

[Jarin]

4 hours ago, ForScience6686 said:

The graph is fine, people's misunderstanding of it is on them.  It is clearly marked and explains the procedure well.

A graph that is widely misunderstood by the inexperienced is still a graph that could be improved, as others including its creator have recognized. There's no superiority points to be won here with saying that you understood it on first glance.

[AeroGav]

1 hour ago, Jarin said:

A graph that is widely misunderstood by the inexperienced is still a graph that could be improved, as others including its creator have recognized. There's no superiority points to be won here with saying that you understood it on first glance.

Distance from KSC would have been a better X axis , but how do i get the game to display that for me?  Also, it would have to be distance "as the crow flies" not "as the worm burrows" if you know what i mean...

Edited March 25, 2017 by AeroGav

[Jarin]

12 minutes ago, AeroGav said:

Distance from KSC would have been a better X axis , but how do i get the game to display that for me?  Also, it would have to be distance "as the crow flies" not "as the worm burrows" if you know what i mean...

Would distance traveled in the f3 menu do? Just have to track it every couple seconds.

[bewing]

1 hour ago, AeroGav said:

Distance from KSC would have been a better X axis , but how do i get the game to display that for me?  Also, it would have to be distance "as the crow flies" not "as the worm burrows" if you know what i mean...

If you open the context menu of the cockpit, it shows a CommNet first hop distance -- which is the precise distance to the Tracking station, to high precision.

 

[guitounet]

Hii Guys,

 

is there a rule of thumb for wing area Vs plane mass to optimize building ?

[AeroGav]

I use CorrectCoL,  which helps improve the accuracy of the blue marker , but it also has a stability analysis function.   You give it the height and speed you want it to simulate and it plots a graph of how strong your nose up/down tendency is across the range.

Crucially, it also plots a blue vertical line which shows how much AoA you need to sustain level flight at that speed and altitude.

So,  I give it the altitude of 22km, and the speed of 1125.   I'm looking for an AoA of less than 5 degrees, because more than that is draggy.   This is a typical speed for the start of a speed run, and 22km is about the max height you'd want to do your speedrun at (after that, loose a lot of power).   I also simulate 1350m/s , which is mach 4.3, when the rapier starts to loose power rapidly with increasing speed.    I don't want my level flight aoa to be less than 2 degrees, that means i probably have too much wing.

The thing is  though, if you've got nuke engines and are using wings that contain fuel, it's hard to have "too much wing" in a monoplane design, that doesn't wobble apart on the runway.

If you're doing a 20 ton challenge, or 30 ton challenge /whatever,  too much wing could eat into payload to orbit.  If you're doing a "4 engine challenge" , to lift as much as you can with just 4 engines, then more wing helps you lift more weight with less engine.

So don't sweat it too much, but your design should be able to keep AoA below 5 degrees and still get enough lift,  if keeping the AoA to less than 5 degrees causes overheating on ascent, you don't have enough lift.

I recommend CorrectCoL is just an awesome mod for all sorts of reasons.

[guitounet]

20 hours ago, AeroGav said:

I use CorrectCoL,  which helps improve the accuracy of the blue marker , but it also has a stability analysis function.   You give it the height and speed you want it to simulate and it plots a graph of how strong your nose up/down tendency is across the range.

Crucially, it also plots a blue vertical line which shows how much AoA you need to sustain level flight at that speed and altitude.

 

I recommend CorrectCoL is just an awesome mod for all sorts of reasons.

Can someone give some more info for understanding graphs ?
I do understand what are the user editable value.

 

edit : of course, i red the correctCoL page

Edited March 26, 2017 by guitounet

[Spricigo]

On 25/03/2017 at 2:09 AM, AeroGav said:

Yeah I need to fix this.  It is time/vs altitude from one of my recent flights.    I was only climbing at about 5 degrees pitch angle, but because we were going so fast horizontally, the vertical speed was huge.

Anyone who has better skills with Excel and MS Paint, feel free to jump in.

 

No, you don’t need to fix anything if the info it shows is correct. No matter what you do with the graph people will still get it wrong if they keep reading "bananas" when you write "watermelon" (which is exactly the issue there).

That said :

-the legend, I suppose, is meant to tell "vertical axis is altitude (in meters), horizontal axis is time (in minutes)"  but actually says "blue line is time, orange line is altitude". 

-I think a Altitude vs Speed graph would be a better way to present a flight profile. Another possibility is Attitude vs Speed. Time is not as much relevant IMHO

-There is a few ways to read the distance from KSC, some mods may be usefull (waypoint manager, KER, etc). Or you can calculate it from time and speed.

 

[Wanderfound]

On 26 March 2017 at 8:54 AM, guitounet said:

Hii Guys,

 

is there a rule of thumb for wing area Vs plane mass to optimize building ?

It depends on what you're after.

There are two schools of thought, that we might call U-2 vs F-104.

The U-2 school calls for large wings that continue to provide good lift even in ultra-thin air. This allows you to float up to altitude with a very efficient, shallow ascent. The downside is an increase in airframe mass (often counteracted by a reduced need for high TWR, allowing a reduction in engines), fragility and high-speed drag.

The F-104 school substitutes speed for lift. Small wings, just enough to get you off the runway. The benefit is that you minimise airframe mass and high-speed drag; the price is an inability to fly at extreme altitude without excessive AoA (which can more than counteract the drag reduction benefits of the smaller wings, although building with wing incidence can reduce the issue substantially) and increased takeoff and landing speeds.

F-104s tend to be built with higher TWRs than U-2s, but this isn't always the case.

Both approaches work, although they demand different flight profiles.

U-2s aim to float through the 20-30km region at minimum AoA, slowly building speed in the thin air while supporting the plane on the wings.

F-104s instead want to crank it up to 1,500m/s down around 12km, then pull into a steepish zoom climb that tosses the plane above 30km on a ballistic trajectory. You'll lose 100m/s or so initiating the climb, but if you do it low enough and have decent TWR you can regain that speed before the jets fade out.

@AeroGav is an enthusiastic partisan of the U-2 school; I'm more of an F-104 guy. But, as I said: both approaches work, so go with whatever looks like more fun to you.

Edited March 28, 2017 by Wanderfound