Thinking about making the switch to FAR.

[capi3101]

AV-R8s make perfectly good canards.

Tried thems first. No joy.

Your delta needs a tail if the trailing edge of the main wing is too close to CoM for the control surfaces to work as elevons

Pretty sure that's what was happening in the case of last night's plane - the CoM was too far aft. It's what made me think of using canards in the first place - I figured the nose would've had more leverage there.

If you're rolling, the problem isn't lack of tail, it's insufficient dihedral and/or wingspan. Where are the red numbers in the FAR screens?

Got a red L_ß and M_δe at 10k/Mach 1, a red X_w at 15k/Mach 2, everything's green at 20k/Mach 3; everything's red at 30k/Mach 4 - by everything, I mean Z_w, M_w, Z_u, X_u and L_ß. Sim studies show decreasing oscillations at all levels except for 30k/Mach 4 - the L_ß is highly amped but still decreases in oscillation with time, while the lateral graph just shoots off. Main wing has ten degrees of dihedral as is.

It's really just a crummy design overall. I'm sure y'all will agree once I get a couple of screenies up.

[capi3101]

More cargo bay.

Makes my plane floppy without adding much mass forward. Would definitely take some pointers on how to go about strengthening up the airframe.

Inline docking port.

Above my current tech level, unfortunately.

More fuel.

Messes with where RCS Build Aid says my DCoM will be.

Passenger cabin with no Kerbals in

...you've got me there. I suppose the argument is that a large RCS tank has more mass than either a Mk2 Crew Cabin or Hitchhiker, and therefore makes a somewhat better counter-weight. Both of those, however, are more massive than a Mk2 Monoprop tank.

I don't mean to argue; I've had a pretty frustrating night and early morning with this, my latest craptastic design:

Trying to figure out how to fix it...I mean, the CoM needs to come forward for sure and the tailplane needs work. I'd say the wings need more span, except I think that the problem may be too much chord at the tip (that's one I really haven't figured out yet). She bucks the whole damn time despite FAR saying it's stable, and then when it gets over 20k she wants to fly backwards (which FAR says will happen).

The takeoff characteristics are lousy, too - FAR says Mach 0.4 with the flaps in takeoff setting before the AoA gets close to 5 degrees. I've taken off a couple of times with it by giving a quick light of the rockets, just enough to get the plane into the air.

Plane was originally a tail-less delta - and it was an ocean dart every damn time...

[Van Disaster]

What is the wing loading? I think you're on the money with lack of span, tbh. CoL needs to be considerably closer to CoM to get any decent pitch control out of that, I think, you don't appear to have much pitch authority.

As for floppy things, use KJR. You'll still need to strut more extreme structures but for simple things which really shouldn't bend you shouldn't have to plaster struts everywhere.

[capi3101]

Here...I was going to post this data earlier:

So, wing loading's just mass divided by the wing area, right? 38.301 tonnes divided by 102.223 square meters oughta be ~0.375 tonnes per square meter if I'm doing the math right.

Except that wing area figure would include the area of the tailplane...might behoove me to go back into the game and get the surface area of the wings alone. Unless that doesn't matter.

[eddiew]

Neither canards nor tail control surfaces sounds like a recipe for problems, tbh...

The obvious thought re the CoM is that you're using a pair of big nacelles to carry an extra couple of intakes - and you've aligned them to the back of the craft... I see no need for those to be at the rear, may be worth hauling them up to the leading edge of the wing? I'd also consider a bigger tail, with pitch control surfaces and a shorter, wider main wing; which would give you more space to put the engines without them moving backwards to make space for the big tanks.

Drones have a habit of looking a bit hunch-shouldered, a side effect of not having a heavy cockpit. Your overall shape lends itself much better to a crewed design - and might actually look really good with a cockpit

[Wanderfound]

Re: floppiness. Have a look at some of the Kerbodyne designs. I don't use KJR.

The small ones are often unstrutted; the medium size ones usually have a pair of struts at the front (lateral tanks -> docking port or cockpit) and another pair at the back (lateral tanks -> tail fuselage). That's generally enough to stop fuselage wobble unless it's a very long ship.

The only ones that get extreme strutting are the Mk3 designs; even that requirement should hopefully go away once the new wings and landing gear appear.

[eddiew]

Another thought; if the rockets are heavier than the jets, swap them to the foremost position. I prefer jets nearer the centre-line anyway, since if you do get an asymmetrical flameout it'll be less dangerous than if they're out at the wingtips

[Van Disaster]

That tail won't have much area by itself ( which might be part of any pitch control issue ) but that's plenty of actual wing. If you're having roll problems then those absurdly small ailerons won't help, no reason to have small control surfaces using pwings. For the B9 S2 plane I posted in here at some point the entire wingtips are the ailerons, they're not much smaller than one of the stock canards.

That might not actually be a terrible idea for this wing if you're looking to extend the span a little.

Edited April 22, 2015 by Van Disaster

[capi3101]

The obvious thought re the CoM is that you're using a pair of big nacelles to carry an extra couple of intakes - and you've aligned them to the back of the craft... I see no need for those to be at the rear, may be worth hauling them up to the leading edge of the wing?

Drones have a habit of looking a bit hunch-shouldered, a side effect of not having a heavy cockpit. Your overall shape lends itself much better to a crewed design - and might actually look really good with a cockpit

Okay - that makes sense; how will moving those tanks up towards the nose affect my drag?

I'd also consider a bigger tail, with pitch control surfaces and a shorter, wider main wing; which would give you more space to put the engines without them moving backwards to make space for the big tanks.

Well, I've read a number of articles that say supersonic planes have all-moving tailplanes to counter mach tuck. If that's not a consideration with FAR, I can certainly start using fixed tails with elevator surfaces; the tail I have at the moment is all-moving. Am definitely reconsidering re-shaping the wing. Let me run my thinking by y'all real quick:

Target wing loading is 0.3; I have 38 tonnes of plane, so I want 38 / 0.3 = 126.67 square meters of wing area. That's 126.67/2 = 63.33 square meters per wing. The fuselage length is 24 meters, so I want my root chord at 24/2 = 12 meters. If I say the tip chord is 12/10 = 1.2 meters, the wingspan I want should be = 63.33 / (0.5 * [12+1.2]) = 9.59 meters...

I think I see where I screwed up the math on my initial design - I didn't divide by 0.5, I divided by 2. 63.33/(2 * [12+1.2]) = 2.399 meters. So that would explain that. That's also been a recurring error...

That is how a trapezoidal wing area is calculated, right? One-half times (the sum of the root chord and the tip chord) times the semi-span? And then triangular wings are just 1/2 times the span times the root chord, right? So if I went pure delta (no tip chord), I'd want 63.33/(0.5*12) = 10.555 meters.

Another thought; if the rockets are heavier than the jets, swap them to the foremost position. I prefer jets nearer the centre-line anyway, since if you do get an asymmetrical flameout it'll be less dangerous than if they're out at the wingtips

Naw - I've got three tonnes of rockets and 4.8 tonnes of jets. But point taken.

That tail won't have much area by itself ( which might be part of any pitch control issue ) but that's plenty of actual wing. If you're having roll problems then those absurdly small ailerons won't help, no reason to have small control surfaces using pwings. For the B9 S2 plane I posted in here at some point the entire wingtips are the ailerons, they're not much smaller than one of the stock canards.

That might not actually be a terrible idea for this wing if you're looking to extend the span a little.

I've been putting four procedural control surfaces on my wings, trying to divvy them up as evenly as possible given the span of the wing (short span = small control surfaces as a result - this one only has three because the inside set would've mashed up against the side fuel tanks, which I'm going to move).

I've also been basing surface area of the tail off of surface area of the wing (40% of the wing area = tail area). Since I've established this fine rainy morning that I've been screwing that up for weeks and making the wingspan too small, I'm definitely of the opinion that the tailplane is too small...

Definitely some things to fix. Thank you all for your advice.

Okay...programming note for folks watching this thread: apparently KSP 1.0 is scheduled to come out this coming Monday (April 28) - I've somehow missed this fact up until this morning. I'm liable to do like I did when 0.90 came out and continue on until I can confirm new versions of all the mods I've used for 1.0 (and/or the fact that certain mods have become obsolete). My intent right now is to try out the new aero model before re-installing FAR just to see what its quirks are like, but that could change of course. Meantime I'll continue screwing up the math.

Actually, if the wing span is too short, that might explain why the tail-less delta didn't work in this case...

[Wanderfound]

Mach tuck is a thing that happens with FAR, and all-moving tailplanes can help to overcome it, but they are by no means necessary. If you have enough control authority to pull the nose up when you want to, you're fine.

[eddiew]

Naw - I've got three tonnes of rockets and 4.8 tonnes of jets. But point taken.

I was imagining swapping the two rockets for two of the jets - the other two jets would stay in place ^^; If you're moving 3T of rocket forward and 2.4T of jet back, that'll help shift the CoM (effectively you're adding 0.6 to the front and removing 0.6 from the back so it may do more than you think)

...that would however lend itself to having the wings and rocket engines both attached to the nacelles, in order to avoid unpleasant torques when in space... I've probably just suggested a total redesign without actually meaning to xD

[Hodo]

Makes my plane floppy without adding much mass forward. Would definitely take some pointers on how to go about strengthening up the airframe.

Above my current tech level, unfortunately.

Messes with where RCS Build Aid says my DCoM will be.

...you've got me there. I suppose the argument is that a large RCS tank has more mass than either a Mk2 Crew Cabin or Hitchhiker, and therefore makes a somewhat better counter-weight. Both of those, however, are more massive than a Mk2 Monoprop tank.

I don't mean to argue; I've had a pretty frustrating night and early morning with this, my latest craptastic design:

http://i.imgur.com/1hotYlp.png

http://i.imgur.com/r3wlcGz.png

http://i.imgur.com/hddADCY.png

Trying to figure out how to fix it...I mean, the CoM needs to come forward for sure and the tailplane needs work. I'd say the wings need more span, except I think that the problem may be too much chord at the tip (that's one I really haven't figured out yet). She bucks the whole damn time despite FAR saying it's stable, and then when it gets over 20k she wants to fly backwards (which FAR says will happen).

The takeoff characteristics are lousy, too - FAR says Mach 0.4 with the flaps in takeoff setting before the AoA gets close to 5 degrees. I've taken off a couple of times with it by giving a quick light of the rockets, just enough to get the plane into the air.

Plane was originally a tail-less delta - and it was an ocean dart every damn time...

First I would like to say that flaps are part of your problem on take off, your flaps are to far back away from your CoM thus they are acting like elevators and pushing the nose down instead of creating lift. This isnt helping your take off speeds at all.

As for your CoM and CoL, I think they are to far apart, right now your CoL indicator is set roughly at the center of your wings which is correct for subsonic flight but as soon as you start going super sonic it will move back towards the rear of your wings, further away from your CoM. This will make the nose VERY heavy at supersonic speeds and it will only get worse at higher mach speeds.

I suggest moving your CoL closer to your CoM. Ideally you want it just behind the CoM.

About like this craft, granted this craft it was a bit to close but its CoM shifted forwards as fuel drained.

To fix your problem I would do two things. Add a Leading Edge Extensions (LEX) to the front of the delta wing... like the F-18 Hornet. I have done it on several craft that had issues with the CoL being to far back and canards weren't an viable option.

Like this craft, and this craft also utilizes leading edge slats and flaps for take off.

[Van Disaster]

The best way to deal with mach tuck is just to emulate Concorde & shift some fuel - if you start cranking up pitch control angle to counter it you're causing excess drag and you've reduced your overall control authority. If you go check up on CoM for real airliners you'll find a huge pile of info about how CoM position affects fuel consumption due to trimming. The other way is to take off with CoL deliberately forward & manage the instability until you go supersonic; on the way down you'll have burned off fuel so if your tanks are generally aft then CoM will move forward.

CoL ball is inside CoM, basically right in the same place. That's not a bad example of the ratio of control surface I like too, although the all-moving aileron wingtips are a bit more excessive than I usually do.

[Hodo]

The best way to deal with mach tuck is just to emulate Concorde & shift some fuel - if you start cranking up pitch control angle to counter it you're causing excess drag and you've reduced your overall control authority.

.

This is what I usually do. I will sometimes design the craft with an empty fuel tank just for shifting fuel around during flight.

[capi3101]

Alright - design changes implemented:

Some pretty major changes, obviously. I decided to ditch the inline rocket fuel tanks and put them out into a fuel fuselage, getting one set of Turbojets off of hardpoints and onto the fuel fuselage. I kept the other pair of jets out on hardpoints. I took out the two Mk2 Rocket Fuel Fuselages from the base fuselage, which shortened the plane lengthwise; I then recalculated the necessary size of the wing. Sticking it on the plane, it looked short, so I added a cruciform all-moving tailplane, and changed the functions of the wing control surfaces to aileron - spoileron - flap/spoiler - flap. I adjusted the wing position to cancel the engine torque. FAR gave me green lights except for L_ß below Mach 1, and that was a stable oscillation, so off I went.

Last night was actually a night of a bunch of little screw-ups. The first flight failed because I hit D instead of S and sent the plane careening to its death. I hadn't add sufficient strutting to the frame anyway, so that gave me a chance to fix that problem. Second flight I reverted because the starboard inboard engine flamed out at 24k - I had forgotten to push the magic "fix everything" Intake Build Aid button before I took off. Third flight......

Of all the damn things to fail the mission entirely - the probe's RCS tanks were clipped into the fuselage. That's what I get for not redesigning the probe - I've launched dozens of these things in stock and the carrier plane was designed as an open frame around the side tanks...

Actually I had one more screw-up before the night was over; getting back to the SPH to take screenies of the plane, I realized that the engine torque had been balanced for the inboard jets, not the rockets. I'd picked the wrong damn pair of engines to analyze...

So yeah, I'll be revisiting the design again; if nothing else I need to add some RCS ports so I can start using that forward docking port for something other than decoration. Pointers? The tailplane looks big to me - and I think I'll try replacing it with a fixed surface/elevator setup. How many control surfaces does that part need anyway? I've also been shooting for a tailplane area that's 40% of the wing area; is that more than I should be shooting for? Should I not be building cruciform tails?

Edited April 23, 2015 by capi3101

[Van Disaster]

You have to think about structural issues too - something long and thin like that tail is going to bend, in KSP's case at the joint, but joint strength for these wings is determined by shape as well as overall size ( actually it's directly related to something aerodynamic for these wings but I forget the formula right now ). Shorter span & longer chord for all the tail parts. Your underslung jet pods are going to provide a big pitch-down moment when you're in rocket mode & still in atmosphere and they're just giving drag. Wing looks fine, you could sweep it more if you wanted. Your CoL is way too aft of CoM though.

I didn't check if FAR includes leading & trailing edges in it's wing area calcs.

[capi3101]

It does - I've seen it jump up a fair bit after adding the tail surfaces.

Curiously enough, I almost never wind up with the wing area where I've calculated it - case in point. I shoot for 0.3 tonnes/square meter of wing loading; I've got a 42 tonne plane there, so I want 140 square meters of wing area, right? Then it's a 17 and change meter long fuselage; root chord 50% of that (8.5 meters) and tip chord 10% of that (0.85 meters, since I'm using B9 procedural wings I can only adjust in measures of one-eighth of a meter so I set it to 0.825 meters). [root + tip] / 2 * span = area, so 9.325/2 * span = 140, thus span = 30.0268, and each wing needs to be half of that...

Dammit, I screwed up the math again. Well, considering last night's pattern, I'm at least being consistent.

I think part of my problems at this point is the cruciform tail - I try to build the tailplane so it has x amount of total area (40% of the main wing), but I'm limited on the length of the tail's root chord by the width of the fin at the point where I mount it, and to compensate I've been increasing the span. I might have better luck if I'd just mount the damn things to the fuselage and leave the fin be. That has gotten me into trouble with engines in the past, though.

I did take a look at my drag-generating parts last night; I was amazed how much drag I was getting on those Mk1 tanks...that just doesn't seem.......fair. Hopefully that's something that'll be fixed come Monday.

I do have an issue with CoM/CoL, admittedly; the plane lacks counterweight up by the nose, so there's a big CoM shift aft once it goes into rocket mode. I have the CoL set as far back as I do so that the plane stay stable as the fuel drains. It probably wouldn't be as pronounced of an issue if I had more up front.

[eddiew]

If you trust yourself not to press spacebar, you could mount the engines on decouplers to prevent direct fuel flow, then run a pipe from the engine to the front tank. That should make the fuel drain from back to front ^^

[Hodo]

Ok I am looking at your SSTO in the last pics... I see several minor issues that will cause bigger issues.

First, the CoM and CoT are not lined up, they are slightly off due to the engines, intakes and liquid fuel tanks being below the wings and bringing the CoM down below the CoT. This with the mass of the wings being below the center line of the craft this would further this shift in mass.

The fix for this is simple. You can shift your turbojets up into the wing by using the movement tool in the SPH. With this you can even remove the pylons they are on which will help with some of the additional mass. Downside to this is your landing gear are going to have be moved.

Second, your wings are attached to the 1.25m tanks which are then attached to the hull of the craft, which means you have a pretty weak flex point that is between your wings and the actual body of the craft. I prefer to attach my wings to the craft then rotate the fuel tanks through them so they clip through the wings. This will keep some things attached at their strongest points. I would also reduce the thickness of the wings to the minimum to place struts from the wings to other parts of the body to increase strength. Then once the struts are placed I would return the wings to the thickness you want, this will hide them inside the wings.

Last but not least the I would remove that MASSIVE horizontal section of the tail. You should not need anything like that if you shift your mass forward just a bit or your CoL back.

[Van Disaster]

I'm not sure why you're insistent on the tail being 40% of the wing, especially for a delta which doesn't have a tail to provide any lift. You just want enough tail to control the craft, never mind any generic formula ( because there isn't one, it's dependent on the actual craft ).

[capi3101]

To be honest, I'm trying to remember where I saw that 40% figure...reddit, maybe. I don't think I actually calculated that anywhere when I was looking at the proportions of RL supersonic planes. It'd be worth another go-through.

Nobody's said anything about it being a bad figure...

Meantime, tonight I can report success. After taking the time to re-design and re-balance the probe in the VAB, adjust the wing position so that the LV-T45s wouldn't generate torque relative to the DCoM, add RCS ports, push the magic button and re-set up all the action groups, I took the Phoenix 7b to orbit and delivered its payload.

Flight was reasonably smooth - plane did pitch down when I switched to the rockets but after a bit it smoothed out. Still had about a 30% rocket fuel reserve when it made orbit - so the design oughta be able to handle heavier payloads than the 6.1 tonne probe I launched this evening.

Admittedly, landing took a few attempts, the big thing being that I needed to be way gentler on the controls coming in from orbit. After I figured that out, the jets had more than enough juice to fly the craft the rest of the way to KSC. Keeping my airspeed up while on final was a challenge - and once again I could've stood to be more gentle on the controls as well (had one attempt where I pulled up too hard trying to line up on the runway and jerked myself right into a stall because I lost too much airspeed, and couldn't recover afterwards).

[Wanderfound]

Yup; low, slow and smooth is the key to landing. Until you get the hang of it, you're usually better off powering up and going around again rather than trying any last-second hard manoeuvres.

Keeping the jets on at minimal throttle is useful for reducing spool-up time when you need some power; try to get down to landing speed before you reach the runway, then apply a touch of thrust to hold that speed constant.

For reentry, I usually just keep the nose 5-10° above prograde until I'm back to level flight at 20,000m or so.

Edited April 24, 2015 by Wanderfound

[capi3101]

Went back through my notes today - I don't know where that 40% tail-plane figure came from...hell, even the damn glider figures said 25%. There's a few more things I'd like to do with the Phoenix (like maybe give it the ability to turn on RCS thrust - right now all it can do is translate because I used the little single squirt ports instead of blocks, and its inherent turning capacity is crappy because at the heart of the thing is an OKTO2), so I'll shorten the span of the tail plane like y'all have been telling me to do and see what that does. I'll try the 25% figure first, see what FAR has to say about it and let y'all look at it. Probably still going to be an all-moving tail (for now anyways).

I'm also thinking about moving those Mk1 Fuel Tanks from the front of the fuel fuselage to the back (i.e. just ahead of the engine, changing it from Intake - Mk1 - T400 - T400 - Engine to Intake - T400 - T 400 - Mk1 - Engine); this one's from the other day. Would there be an advantage to doing that y'all think, or am I getting drag on those parts simply because of how things look up there (i.e. if I move that part, will I just start seeing drag on the FL-T400s?)

[Van Disaster]

You can't really think in terms of areas like that - a long plane will need less tail than a short one even with the same wing area, or more specifically the further away the tail is from CoL the smaller it can be. What you need is enough control moment, and FAR will tell you if it's not enough - but you have to test it at a bunch of speeds & altitudes because you can fall out of the envelope for a bit; the one that people usually stumble over is going supersonic & suddenly losing the ability to hold the nose up, because CoL movement away from CoM has increased the inherent pitch down moment beyond the tail's pitch up moment at that particular dynamic pressure. Simple answer to that, go supersonic at higher Q. Better answer is to redesign the craft a bit, of course, but at some point you realise that it'd be better just to fly a little differently.

Edited April 24, 2015 by Van Disaster

[Hodo]

Capi,

I just want to say congrats, but now comes the fun part, advancing the design.

May I suggest building a multipart wing so you can kill that long thin tail, unless that is the look you are going for.

You can end up with a wing that looks more like this, and you will still have great pitch authority, supersonic flight control and pretty decent low speed landing characteristics.

That craft can haul 70tons into orbit with ease and it is only 127 parts.

Or a wing like this for a true tailess delta design without canards.

Again that is a SSTO and more than capable of reaching 100km orbit.

If you look at most modern supersonic aircraft, I suggest researching Russian Fighters or US Fighters of the 4th generation, this is closer to the technology you are dealing with in KSP. You will see that the wings change sweep angle at some point along the wing.