Thinking about making the switch to FAR.

[Van Disaster]

Move some of the rockets. The plane itself is asymmetric around the horizontal, no reason the rockets have to be symmetrical too. I've been developing something today where none of the things are aligned and it manages happily. I should probably align the orbital engines better at least, I guess .

[capi3101]

Alright - moving the rockets around helped. The torque is still there but the onboard SAS can handle it now.

Now I need to re-balance the fuel, looks like. I had it set up with the FL-T800/Nacelle equivalent; I've wound up with not enough liquid fuel to make orbit with the current setup, though the plane tasted space.

[capi3101]

Gah. Alright - the Death Trap either is experiencing more parasitic drag from those 24-77s than I thought and it's time for me to admit defeat, or I'm flying the damn thing wrong. I've only been able to get it to Mach 3.5 or thereabouts before the jet flames out at 34k; I panic, kick in the rockets and run out of fuel before making orbit. The plus side of this is that I'm getting good practice with emergency landing procedure; I've only crashed the one time. Some of the successful landings have been on grades as high as 9 degrees...

I'ma thinking I'm ascending too fast with the design - as fast as I can to 10k, then level out. My ascent has been moderately fast (sometimes up to 100 m/s) in the 10-20k regime but I've been keeping it around 20 m/s above that as much as possible. Should I be trying to keep that down to 20 degrees in the lower regime as well? I'm generally just crossing the Mach 2 barrier when I've reached 20k, Mach 3 has been up around 26k or so. Plane has 0.68 TWR on the runway, single Turbojet, 440 units of LF and 440 of Oxidizer.

[eddiew]

Fast ascent does sound like a probable problem I would totally expect the jet to flameout by 34k... Personally I'd have been looking to be at mach 4.5-5 by then though, really milk that jet engine for all it's worth. Most of my designs level out at 25km and then just build up speed until they stop accelerating - then turn the nose up.

Bear in mind you can take out a little oxidiser too - maybe 10% - to save a bit of weight. Might not be a lot, but sometimes that extra 0.03 TWR makes more difference to the flight characteristics than you'd expect

Oh, also, remember you can use FAR to check the drag of each component! In the KSC screen, hit the FAR button, and find the option to enable it on the right-click menu. I've found a few surprising drag-heavy parts, including the radial engine bodies, and B9's S2-2.5m adapter. When I pulled these off a struggling plane, suddenly it started going way faster way sooner...

[capi3101]

Alright - I'm throwing the towel in on the current design - going to pull the 24-77s off, and a couple of side-mounted rockets with fuel tanks. Did three ascents last night; never did do better than Mach 3.7 before reaching 34k, and that was with an excruciatingly slow ascent (around 20 m/s from 10k on up). Meantime I think I'll toss it over to y'all - here's the craft file. Maybe see if there's anything about the design I overlooked. Definitely would like some pointers on the tail surfaces - if I have them set up correctly or not. Two more successful emergency landings last night, one failed because I had reached the night side of the planet and couldn't tell that I was over the water ("Shores" = water, apparently); got over land but with a very high grade with only 45 m/s of horizontal speed...

[Van Disaster]

If you think 20m/s vertical is "excruciating" then you won't want to try something really big... I call that "quite fast, actually".

Two jets and a nerva is a nice combo for a smallish craft.

[eddiew]

(around 20 m/s from 10k on up)

Bear in mind that the atmospheric density at 10km is really quite high, so drag is still very significant. There's not a lot of advantage in a shallow ascent until you're above 20, or even 25, by which time you've got a lot less drag to fight. I like to think of 25km as my launchpad (higher is better, but also harder); how fast and steep a plane flies up to it doesn't matter a jot, as long as it's stable when it gets there (i.e. at its minimum speed) and you can get the velocity vector shallow again before you're up even higher and out of the stability window.

For reference, my ascents tend to be 60-100m/s, until I've reached cruising altitude, at which point I'll try to hold that down until the jets just won't go any faster. If that speed turns out to be below mach 4, then it's time to do a right-click drag analysis during flight and figure out which part is the problem.

And for consolation, remember you're building a really small spaceplane there, which does make life harder ^^; I don't think I've ventured to lift cargo with anything below ~18 tons...

[Wanderfound]

As others are saying: climb steep until 20,000m or so, then flatten off and climb as slow as necessary (usually <20m/s) to get to maximum speed before the jets choke.

I normally do a bounce; steep climb until Kerbal Flight Data starts showing the intake air dropping (20-25,000m, depending on the ship), then pull the nose down to 10° and hold it there. The plane will usually go into a shallow dive that takes me up to Mach 4; the increasing speed and thickening air will gradually flatten out the dive without me touching the controls, and I'll usually start climbing again around 18,000m or so. On the second climb, I keep the nose up and start shutting down jets and activating rockets as the air runs low.

I've only ever seen those tiny radial engines work well on very small Mk1 ships.

[Van Disaster]

Small Mk1 ships might as well use a Rapier - I did a bunch of experiments when 0.90 turned up & Rapier won easily.

I climb subsonic until about 16km & then start flattening out unless I'm superheavy ( those beasts won't climb steeply at all, just cruise up at 30m/s vert until then ), that way I'm not at 20km with a huge AoA & wasting fuel on drag. Given we're all flying FAR, just put the speed display to EAS & keep that roughly constant.

[capi3101]

I've had more success with RAPIERs so far my own self - the idea with the Death Trap is to get used to Turbojet/rocket combos, though, so I can go back to my career game. Haven't got RAPIERS unlocked on account of the R&D building only at Level 2, see.

"Right-click drag analysis in flight" - something I think has been mentioned before but I'm not quite sure to what it's referring. How does that work exactly?

So last night I did do a redesign of the craft again, this after re-installing LoadOnDemand and getting rid of some excess mods that I a) no longer needed and didn't really use that much, like KVV (being able tell which button in the ModuleManager does what again is very nice; my screenies from here on out shouldn't look like dreck either). The version of the craft on my Dropbox is pre-revision, so I could go revisit it again if I wanted to (I'd just download it from myself). The redesign saw the removal of the structural intakes and 24-77s and the removal of the fuel from the tail, and the addition of side fuel/intake/nacelle pods with two side-mounted LV-909 engines. FAR greenlit the flight except for L_ß at 30/4 (it thought I needed a tailplane). 0.46 TWR on takeoff. Don't know if I would've made it or not; what I do know is that at Mach 3.9 and 36k the jet flamed-out (I guess those 24-77s really were putting on the drag), I hit the button to light the rockets......and only the portside one came on. Well, long story short there is that after I recovered from the resultant spin and dive I dumped out all the oxidizer and did an RTLS abort, despite being 600 klicks downrange of KSC at the time. Damn plane almost got back up to as fast as it had been going during the slog back. Landing went okay for the most part; did get the Q up to 10kPa at one point trying to slow down, but nothing much to report there otherwise; did S-turns and applied the spoiler to slow down. I was a little low and left of center when NavUtilities called the middle marker, probably should've gone around but it was getting late and I needed to get to bed, so I went for it. Blew the jet off on touchdown (was coming down around 8 m/s - too fast - when I touched down) but otherwise a successful landing.

Not really much I could say about that flight - the starboard rocket failed to light due to a classic SPH action group bug, which I made sure to fix right before I shut the game down for the night. I won't know if that alone will ensure a successful flight until I get a chance to fly again. Meantime y'all might fill me in on whether the 909 is a good choice or not...I really wish KSP had an inline engine in the 100-120 kN thrust range. I mean, yeah I could tune down a pair of LV-T45s, but that's an extra two tonnes...

Anyways, I'll try it y'all's way next chance I get - 20k as fast as possible without losing control, then flatten it out and build up speed in that regime. I should be able to handle that.

One more thing - how do I know what my takeoff speed is again? Is it when the AoA at ground level with the flaps in the right setting goes below 5 degrees, or have I got that wrong?

Edited April 15, 2015 by capi3101

[capi3101]

Alright - changed the engine setup to two turbojets and one LV-T45 centerline. I wanted to get to orbit, dammit. Good thing I made it...

Not a great design, though. The monoprop is acting as ballast, and there is some pretty wicked engine torque indicated (hence the number of reaction wheels). The nose fuel tank is also full and acting as ballast, so there's actually less working fuel there than what it appears. I doubt very much that the design could handle any substantial payload; I suppose I could (and should) give it a shot. Thrust would be okay - it's the delta-V I'd be worried about.

I do wonder if taking out the Mk2 Rocket tank centerline was a good idea. It shortened the fuselage and saved on some mass, yes, but the CoM is most certainly aft of the cargo bay. Not as much ability to fiddle around with fuel tank placement, see.

The "bouncing" technique works pretty well, I must say. Been a bit since I've actually hit Mach 4 above 30k...

I've had this crazy notion this morning of sticking the wing directly onto the fuselage and mounting the engine pods below the wing, perhaps attached to a Hardpoint. Something I've noticed - I get less torque with the engines mounted low, but with the wing mounted to the engine pods, that has a tendency to put the CoL low as well. Torque generally doesn't appear until I put the fin on. Something to try with another design, perhaps.

Edited April 16, 2015 by capi3101

[eddiew]

"Right-click drag analysis in flight" - something I think has been mentioned before but I'm not quite sure to what it's referring. How does that work exactly?

KSC screen - FAR options. One of the options in there relates to your right-click menu for parts in-flight; just turn on the drag display. From there on, bringing up the right-click menu for a part will also display the kN of drag that it is creating

The in-flight visualisations for drag are also handy. Alter the threshold a bit and you'll soon fight your highest drag parts are glaring out ^^

- - - Updated - - -

And yeah, getting the weight forward in a drone is challenging; check out

for example! Pretty much 2/3 of the drone is actually tail; but it does balance and work

Moving the Deathtrap to a nacelle design will help you with this, since you can swap out the MP tanks for LFO, shorten the nacelles, and have the engine weight forward in a similar fashion - if you can handle the aesthetics of it xD If you flipped the cargo bay into a bomb bay, you can lift the main wings so they clear the engines, and cause the weight of the drone to hang underneath the CoL, which should make it resist flipping. Alternatively, a low wing (like a NASA shuttle) would give some counterweight to the high tail, and reduce torque overall.

- - - Updated - - -

One more thing - how do I know what my takeoff speed is again? Is it when the AoA at ground level with the flaps in the right setting goes below 5 degrees, or have I got that wrong?

Depends on the plane Most can lean back about 10 degrees on the runway; some more so, if they have a high or short tail. I've also never found a need for flaps ^^;

That said, the smaller the AoA at low speed, the easier it'll be to lift off. I have some that don't hit 10 degrees until mach 0.4, but they're high-TWR so they can get away with it. For low-TWR birds, you need more wing to enable lower lift off speeds. The flip side is that if it takes off at low low AoA and low speed, it'll be a bugger to get back onto the ground. 5 degrees at mach 0.3 would be hard to bring down on it's back wheels; probably end up having to nose-dive

Edited April 16, 2015 by eddiew

[Wanderfound]

1) Get the Stock Bug Fix mod. No more symmetry bugs.

2) Fix engine torque by slight movements of the lateral tanks/engines up or down, and slight movement or angling of the centre engine. It's not that hard to zero it out if you have RCS Build Aid.

3) Fine tune engine torque by shifting the tailplane up and down and adjusting the mass tweakable on the tailfin. You could probably also afford to shrink the size of the fin considerably.

4) Dump the monoprop for fuel/cargo/passengers, strip down to one SAS unit and you'll have plenty of room in the bay for a small satellite.

(PS: as with the 24-77's, the LV-909's have insufficient punch for all but the smallest spaceplanes)

Edited April 16, 2015 by Wanderfound

[Van Disaster]

Use a tail or canards, not both - with pWings there's no reason to be splitting things up like that, all it'll do is make pitching a wonky experience. I'd be tempted to use a V tail for that particular one.

As Wanderfound said it's an option from the space centre button - it might be masquerading as a debug option.

[capi3101]

Did I over-engineer this one?

Plane ran out of gas before it made orbit - which I chalk up to my own damn fault. I'd only put 440 Oxidizer each of the two fuselage tanks (880 total). By my own guidelines, I should've had about twice that. The torque you're seeing there is being generated solely by the Skipper in the back; it occurred to me this morning that I could try to fix that by replacing the aft fuselage fuel tank with a slanted piece instead and either raise or lower the position of the Skipper and see what that does to the torque. I imagine it needs to be raised. The jets are well balanced - and each has its own Mk1 fuel tank.

Pitch authority on this plane is pretty lousy - I have the wing control surfaces set up (from outboard to inboard) as aileron-spoileron-flaperator-elevator. The canards proved to be necessary - first flight without them, and elevators set to 30 deflection, this plane went into one of those damn unrecoverable dives. Load-tested the plane in the process; when it finally crashed into the ocean (and nothing 'sploded in the meantime), the Q meter reading was just over 115 kPa. For reference, surface air pressure of the Earth in the standard atmospheric model is 101.3 kPa...

Ultimately I widened the control surfaces, set the deflection on the elevator to full (40), and added the canards. Second flight the plane got up over Mach 4 below 30k and I'm convinced it would've made it - had the plane been loaded with a proper amount of fuel. The emergency landing did not go well; I wound up slow over highlands. The crapper of it is, though, that that was one of those situations where it was getting late in the evening again - the plane's turbojets still had plenty of fuel and I probably could've flown back to KSC despite being over a thousand klicks out at that point...

Going to try again with this plane tonight after adjusting the Skipper and putting in a proper fuel load. I still don't know exactly why I built this plane - the idea was to try and have something I could switch over to my career game, where I don't have the Mk3 parts...

Really interested to see how heavy of a paywad I could load onto this thing. I also could stand to get some RCS thrusters on the thing - otherwise I might as well replace that forward docking port with a nosecone.

Edited April 17, 2015 by capi3101

[Wanderfound]

Lower the wings until the Skipper's torque is zeroed. It doesn't matter if you have a bit of pitch-up on the turbos; it'll actually help to ease the load on your control surfaces. But the vacuum thruster needs to be balanced.

[eddiew]

Little tip; consider putting the probe core/battery inside the cargo bay for smoother aesthetics on the outside

Also is that another monoprop-ballast design? xD

[capi3101]

Also is that another monoprop-ballast design? xD

Yeah, it is. What can I say besides "3.4 tonnes"? Haven't found another solution yet that doesn't involve adding Kerbals...

Will try lowering the wings to see what happens. I assume I'm going to want the Skipper in-line with the DCoM, and when I go to test things I should zero out the thrust on the jets?

[bigyihsuan]

learned that MechJeb and FAR do not get along at all. I was better off manually flying the rocket than I was letting MJ do the ascent, and I messed with the settings like crazy.

There's a mod to MechJeb that let's it work nice with FAR and NEAR.

I tried out NEAR, but I ended up uninstalling it from frustration. I would recommend it if you're very patient with rockets flipping.

[Wanderfound]

Yeah, it is. What can I say besides "3.4 tonnes"? Haven't found another solution yet that doesn't involve adding Kerbals...

More cargo bay. Passenger cabin with no Kerbals in. Inline docking port. More fuel. Anything except useless monoprop tanks.

Will try lowering the wings to see what happens. I assume I'm going to want the Skipper in-line with the DCoM, and when I go to test things I should zero out the thrust on the jets?

Just put the Skipper and Turbos in separate stages. RCS Build Aid only looks at the first stage.

[capi3101]

Alright - I adjusted the wing position downwards and that killed the rocket torque. Now the plane has lousy pitch authority - this with the elevator and canards set to maximum deflection, and with the "flaperator" set to 50%. Last night's test flight I had to hold down the S-key the entire time; plane never wanted to pitch above 10 degrees. That's pitch angle - not AoA.

Trying to figure out what I might do about the pitch authority at this at this point. Maybe move all four jets below the wing? Except I don't know what that would do to the rocket's torque...

[capi3101]

...it did the trick. I put all four engine nacelles below the wing, and that required me to raise the wing up to kill the torque from the Skipper. Was able to move the CoL closer to the DCoM in the process. The below-the-wing position of the nacelles produced a pitch-up torque from the jets - which is what Wanderfound suggested I try earlier - on the order of about 300 kNm and it definitely gave my pitch surfaces some working room; I was able to tone down the canards and though I didn't adjust the elevators to less than maximum deflection I think I probably should for the next flight. No problems with the test flight of note. FAR gave me a red lights on the main parameters in the 30k/Mach4 regime but I went anyway and I was able to hold it steady in the crucial region. Wound up with the plane in a stable 82x79 orbit, 382 oxidizer remaining, 641 LF remaining with 88.23 units of LF remaining in each of the nacelles (so 288.08 left for the Skipper - it started with 1680 units of Oxidizer/about 1374 units of LF). I have a screenie of the plane in orbit but imgur's having a hiccup this evening; I'll try to post it up later.

Also going to have to figure out what all that implies as far as the plane's potential maximum potential payload. This was another empty run. Switched over to the rocket at Mach 4.2 and 32k.

EDIT: Here we go.

Edited April 20, 2015 by capi3101

[capi3101]

Quick question - when you're building a delta winged plane, when does it need a tail and when does it not? Or is that another "it depends" type of question?

I ask because over the weekend I built another craft, this time in my career game - I was trying to build a replacement for my Barn Owl 7 design, which was my stock probe delivery craft (the one I'd used to fulfill probe contracts). It was a quad-jet slung under the wing /double LV-T45 twin-fin tailless delta-wing design with Mk2 fuselage parts and it flew like a discount brick; into the drink three times, CoM was too far aft, CoL was balanced to the DCoM which was too far back from the CoM at takeoff, wasn't enough pitch authority so I added canards, and since I only have Level 2 R&D in my career game I couldn't use standard/advanced canards - had to try and approximate them from B9 Procedural all-moving wings (which worked, I suppose). When I finally got it into the air, I lost control at 30k/M4...just like the FAR analysis said would happen. She'd roll onto one side and try to flip onto her back, I'd lose altitude and try as I might I couldn't get the plane to right itself again; when I'd pitch down while inverted, the plane would flip over again. It never really stabilized from that behavior once it began, even as I got further down into the atmosphere again.

I should mention that I did make a bonafide effort not to put any monoprop in the design. By which I mean it did have a pair of monoprop cylinders in a forward utilities bay along with the probe core, SAS units and batteries, but it would've been used for RCS maneuvering instead of just ballast. No screenies, I'm, afraid, but I intend to work out the kinks so I'll get some made eventually. Pretty sure this design will need a tail when everything's said and done.

[Wanderfound]

AV-R8s make perfectly good canards.

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 [1]. But if you want a tailed/tailless delta, you can place the wing appropriately.

Changing to a tailed delta usually involves shifting the main wing forwards slightly, so what you're mostly getting is a slight forward shift in the CoM and CoL. The cost is that your elevator duties are now restricted to the stabilator, instead of being able to divide the load amongst the main wing's elevons. The stabilator can probably do the job, especially if paired with canards, but greater deflection = greater drag and earlier stall.

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?

[1] Yes, there are experimental canard-only things. Trying not to overcomplicate it, okay?

- - - Updated - - -

BTW: a sample Mk2 tailless delta satellite carrier. Nukes here, but with a bit of rebalancing you could swap in LV-T45s on a similar airframe.

Javascript is disabled. View full album

Craft file at https://www.dropbox.com/s/refxbd1gub30oqf/Kerbodyne%20Nenu%203.craft?dl=0

Edited April 20, 2015 by Wanderfound

[Hodo]

Quick question - when you're building a delta winged plane, when does it need a tail and when does it not? Or is that another "it depends" type of question?

I should mention that I did make a bonafide effort not to put any monoprop in the design. By which I mean it did have a pair of monoprop cylinders in a forward utilities bay along with the probe core, SAS units and batteries, but it would've been used for RCS maneuvering instead of just ballast. No screenies, I'm, afraid, but I intend to work out the kinks so I'll get some made eventually. Pretty sure this design will need a tail when everything's said and done.

I dont understand why people are telling you to not have monoprop on your design. I use monoprop tanks and RCS thrusters on a great many of my larger and smaller SSTOs. The thing is they make for great counter balance tanks but they are a bit heavier than most other tanks.

Like this tiny fighter it has mono tanks and RCS thrusters. I dont believe in abusing the SAS reaction wheels to achieve level flight. If the craft is well designed you can fly it without a reaction wheel.

And I am a huge fan of "tailess" delta designs.

Edited April 20, 2015 by Hodo