Seeking help with large SSTO spaceplane

[foamyesque]

I don't normally ask questions about plane or rocketbuilding around here -- prefer to answer them -- but I've been working on a fairly substantial plane and I'd appreciate some help tuning it to perform better:

 

 

It's 411.5 tonnes on the runway, and can piggyback carry a payload of approximately 180 tonnes into a 100x100 orbit, and then bring it back to a Kerbin landing. So, functionally, it does what I want it to do. But there's design kludges I'd like to eliminate or improve on. Most obviously, I wound up using ore tanks for ballast. There's a bunch of struts, but I haven't found any way to get rid of them since they all serve critical reinforcement needs (preventing wing, body, and engine mount warping under load). Also, I can't maintain more than about a 30 degree AoA on a loaded re-entry, which means I don't get the giant drag boost stalling out the wings would provide, which then causes an uncomfortable amount of thermal soak. And of course any advise that'd let me improve the payload fraction'd be helpful.

 

Craft file located here: https://www.dropbox.com/s/goatj8h1smol04h/Spruce Goose.craft?dl=0

 

[bewing]

Are you willing to accept a tiny bit less aerodynamic performance, in exchange for a significant payload fraction boost? Do you insist it has to land on the runway, or are you willing to ditch it in the water just offshore and drive it to the runway? Those MK1 crew cabins are twice as heat sensitive as the entire rest of the plane -- would you be willing to swap the two of them for one MK2 crew cabin? OK if I redesign the front end completely?

[foamyesque]

50 minutes ago, bewing said:

Are you willing to accept a tiny bit less aerodynamic performance, in exchange for a significant payload fraction boost? Do you insist it has to land on the runway, or are you willing to ditch it in the water just offshore and drive it to the runway? Those MK1 crew cabins are twice as heat sensitive as the entire rest of the plane -- would you be willing to swap the two of them for one MK2 crew cabin? OK if I redesign the front end completely?

 

Feel free. My thermal limit on ascent is the tailcones for skin temp and the Mk 1 crew cabins for core (secondarily). I currently am usually kicking on the rockets at ~1570-1580m/s surface speed and a 60m/s vertical rise.

S'far as the crew cabins go, sure. One of my earlier iterations had a Mk2 crew cabin in front of the cargo bay, structural fuselages where the Mk 1s are, and the cargo bay where the fuel tank now is. I altered it to get the dead mass further forward and shave the amount of ballast I was carrying to reduce CoM shifts. Wouldn't mind going back, but it'll make the thing heavier.

The killer for the aero is that I want to be able to recover and bring back down the payload fully loaded (or nearly so), which in my testing means being able to execute a glide landing for ~265t of craft, which is what has driven the large wing area.

Edited December 9, 2016 by foamyesque

[bewing]

I'm perfectly happy with wings -- they pay for themselves both on ascent and descent. So that's not where I'd be shaving anyway.

[FancyMouse]

Mount engines closer to the front? They're all mounted at the very rear of the plane which is probably the root cause why you needed Ore tanks to begin with.

About AoA - I usually don't bother worrying about that. Instead, I usually look at vertical speed during descent - KER can give the accurate number in flight. As long as I can maintain about -150~-200m/s in a controllable fashion, I know it's going to survive. The criterion might not work for arbitrary planes, but it works for all planes I've made.

Since you have quite a few elevons, there's one way to slow down high up - deploy two pairs of elevons, one pair normal, one pair inverted, so they provide drag but vertical forces cancel out, which makes it work even if elevons are mounted behind delta wings (which normally you wouldn't want to deploy flaps). Your decision if it's cheaty or not (I don't know if it's legit to deploy flaps on reentry), but it's quite effective in KSP (works better than airbrakes from my experience).

[foamyesque]

6 hours ago, FancyMouse said:

Mount engines closer to the front? They're all mounted at the very rear of the plane which is probably the root cause why you needed Ore tanks to begin with.

About AoA - I usually don't bother worrying about that. Instead, I usually look at vertical speed during descent - KER can give the accurate number in flight. As long as I can maintain about -150~-200m/s in a controllable fashion, I know it's going to survive. The criterion might not work for arbitrary planes, but it works for all planes I've made.

Since you have quite a few elevons, there's one way to slow down high up - deploy two pairs of elevons, one pair normal, one pair inverted, so they provide drag but vertical forces cancel out, which makes it work even if elevons are mounted behind delta wings (which normally you wouldn't want to deploy flaps). Your decision if it's cheaty or not (I don't know if it's legit to deploy flaps on reentry), but it's quite effective in KSP (works better than airbrakes from my experience).

The engines are mounted at the back for a couple of reasons (one much more important than the other):

1. Structural sturdiness. Early versions of this had wing-mounted engine pylons ahead of the CoM to help balance out the rear-ward engines. However, the combination of the RAPIERs working their way up to 360kN apiece and aerodynamic load stresses gave me a long and frustrating series of structural failures and/or wing torqing, with struts only sometimes helping (and often making things worse) and being required in very significant amounts, wrecking the aerodynamics, and therefore needing more engines, which needed more reinforcing, etc. Death-spiralled pretty bad. Changing things so that they are in line with the load helped enormously, though the rear pods are still strutted to keep them from rotating as the thrust dials up. Haven't found a good solution to that yet.

2. It provides superior gimbal coverage. This is needed because, due to the piggyback layout, I have a significant vertical CoM shift as the fuel in the carry vehicle burns off. In atmospheric flight this tendency is controllable through the aerodynamic forces, but once the rockets are engaged a. the CoM shift accelerates dramatically as the oxidizer burns and b. the dynamic pressure drops like a rock as the ship climbs. This is also the reason for the lone Thud at the back; it acts as an orbital manouvering thruster and the high gimbal range allows it to do so regardless of the presence or absence of the payload.

 

If there was a aerodynamically clean, structurally sound way to get those engines further forward I'd be real interested in seeing it, but I haven't found one.

 

[FancyMouse]

5 hours ago, foamyesque said:

it acts as an orbital manouvering thruster and the high gimbal range allows it to do so regardless of the presence or absence of the payload.

How about one Rapier where you adjust its thrust limiter in flight to shift CoT? Yes it would mean no symmetry on that engine, and you'll need to manually adjust it in flight, but that can significantly reduce your gimbal requirements if you mount that special engine right.

Edited December 10, 2016 by FancyMouse

[foamyesque]

After thinking about @bewing's suggestions I made some tweaks and got significantly better performance.

 

1. I added a pair of small nosecones to the front of the plane (one radially attached, the other stack attached to the first so there's no exposed nodes). This came at a drag penalty, but the extra 200K skin temperature of the nose cones over the tailcones (and the fact that the nosecones effectively shield the tailcones) moved one thermal barrier well back;

2. I swapped out the Mk1 passenger cabins for a Mk2, and slid my wings backwards to keep the CoL steady. That bought me a little more lift area, but more significantly it fixed another thermal issue;

3. Because of the above two changes I can run the RAPIERs in airbreathing mode in a shallower ascent, which increased the amount of LF burned compared to oxidizer, so I swapped the front Mk2 LFO tank to just a pure LF one;

4. Changed the tri-stack of standard canards to a naked Big S Elevon 2. I don't much care for the look but it's unquestionably the most effective control surface in the game and I didn't like the canards being all clipped together either anyway.

 

Also made a few other fiddles (caught some misaligned docking ports and put them back in place; they docked on launch but wouldn't once the craft wasn't under load), but nothing prominent.

 

Net result: Landed with almost 8t more fuel than prior tests:

 

https://www.dropbox.com/s/4to0ve0397xtrlb/Spruce Goose (2).craft?dl=0

 

 

[bewing]

10 minutes ago, foamyesque said:

After thinking about @bewing's suggestions I made some tweaks and got significantly better performance.

 

Cool. I've been messing with your plane for hours -- my designs mostly work, but have pitch control issues as I'm sure you well know, and I'm tired of dealing with it now.

First off, you are a total nutcake for trying to launch this lander thing full of fuel. Any rational person would launch it with just enough to land on Minmus, and the dV savings from the fuel weight would easily get you to Minmus. And it would fix most of your vertical CoM shifting issues.

Some things I found that you may try on your new plane:

In your nacelles, at the back, you have 4 rapiers. 3 are attached to NCS adapters, with one NCS adapter having an unattached node. *AND* you have a 4 way coupler with only one node used, and 3 nodes open. So you need to reattach those properly. You only need one or the other -- the NCS or the 4 way -- not both.

The 2.5m nosecones can be replaced with a Rockomax Adapter + a 1.25m nosecone -- they weigh less and are somewhat less draggy.

I took the 3 batteries out of your MK2 cargo bay and replaced them with a 4k battery in your main fuselage. And then I took the 2 advanced reaction wheels out and replaced them with another large reaction wheel next to the new big battery. Then I took the rtg and fuel cell out and put them in a 1.25 service bay, and the antenna I put on the outside and then deleted the cargo bay.

 

 

[foamyesque]

1 hour ago, bewing said:

In your nacelles, at the back, you have 4 rapiers. 3 are attached to NCS adapters, with one NCS adapter having an unattached node. *AND* you have a 4 way coupler with only one node used, and 3 nodes open. So you need to reattach those properly. You only need one or the other -- the NCS or the 4 way -- not both.

Hang on, what? The rapiers should all be attached to the quad coupler, and the NCSes attached to the rapiers to smooth the visual transition because the quad coupler is vertically offset from the tank in front of it. There should be no open nodes at all.

EDIT: As for why I'm launching the lander full of fuel, that's easy: It's meant to simulate a bulky, heavy payload that can't fit in a cargo bay or fairing, the only reason you'd ever need a spaceplane this big in the first place

Edited December 10, 2016 by foamyesque

[bewing]

Well, to do that you need to attach the NCSes to the back end of the rapiers, and then rotate and move them to being in front. But you can only use the front node on the rapier once, so you can't attach the rapier to both things with just one node.

 

Oh, one other thing I did that saved a bit of weight and drag that you might test out. I downgraded all your wheels one level. This put steerable medium gears in the back, which is a heck of a lot better than no steering anywhere. But I was able to take off hundreds of times with no issues on the lighter wheels -- and I filled up my versions of your planes with lots more fuel, so it was heavier than yours most of the time.

 

Edited December 10, 2016 by bewing

[foamyesque]

25 minutes ago, bewing said:

Well, to do that you need to attach the NCSes to the back end of the rapiers, and then rotate and move them to being in front. But you can only use the front node on the rapier once, so you can't attach the rapier to both things with just one node.

 

Oh, one other thing I did that saved a bit of weight and drag that you might test out. I downgraded all your wheels one level. This put steerable medium gears in the back, which is a heck of a lot better than no steering anywhere. But I was able to take off hundreds of times with no issues on the lighter wheels -- and I filled up my versions of your planes with lots more fuel, so it was heavier than yours most of the time.

 

For the RAPIERs: I know; that's what I did. If you detach the quad couplers all the rapiers should come with them, and you should find yourself unable to stack attach anything to either the rapiers or the quad part of the quadcoupler. (Except on the top engine pods, which didn't need the visual blending).

 

As far as the gear goes, my concern is not so much takeoff as it is landing. I might try a wheel downgrade, but I was operating on the theory that the additional spring strength and suspension travel would be needed to handle the ground shock on touchdown (or bounces, in the event of rough field). Did you try a loaded landing?

Edited December 10, 2016 by foamyesque

[bewing]

Yes on the landing, but I assumed that I was a sensible person and burned or transferred off most of the 18 miilion units of fuel in both craft before I landed in the piggyback configuration.