HSTW Drone Race -- closed.

[Goddess Bhavani]

Well... try that

I\'d stick to low-altitude only. Low-alt flight profile suits better to this kind of drone.

Definitely. This is subsonic drone Now that I got clearance to modify the engine thrust, the rest is simple. Hopefully we should have some semblance of a realistic slow Predator drone with at least 40km range by the end of my vacation tomorrow.

For testing, your rule to launch the drone immediately after launch does give an accurate range representation.

UPDATE:

Using standard launch procedure (no flight controls until launch), new drone has surpassed 15km range. Altitude under 1500m.

Launch data

Angle 15 deg regulated by SAS (does not matter, drone flies itself now. Improper launch angle like straight up means drone just run out of fuel uselessly)

Speed about 130m/s (since throttle was full all the way)

Altitude about 500m

Drone ascended to about 1200m on its own for more efficient cruise and self terminated at 15.1 km

Next step:

Now I must give drone different flight model for when engine is on, and engine is off (gliding). This is to separate the \'boost\' phase where drone climbs up above parent aircraft (regain lost altitude during launch), from the cruising phase (which must be as level as possible for realism).

Also intend to give exactly just enough fuel for 50km flying range. Will prevent over eager test pilots from trying to launch drone like Space Shuttle,

UPDATE 2

New flight model - now floating past 22km range.

But something triggeed the self destruct autopilot although the fuel tanks were half full - and it dove down like a guided missile into a lake. Had enough ballistic range, I reckon, for the full 50km flight.

[Iskierka]

Currently conducting tests here with minimal design modification. Currently only thrust vector and attach conditions have been modified, the latter allowing an overhead mounting, which results in the drone lifting itself clear of the parent vehicle unassisted by jettison mechanism, allowing much more stable launch.

Initial tests are promising, though visual contact with current best was lost due to pilot/control error roughly two minutes after launch. Radio signals imply vehicle is still airborne after more than five minutes of flight, with an estimated cruise velocity of 120-160 m/s, giving an estimated range of 40 km or more. Report shall be updated once ground impact is confirmed, hoped to occur roughly 45 minutes from now, after burnout.

[Tosh]

Report shall be updated once ground impact is confirmed, hoped to occur roughly 45 minutes from now, after burnout.

As all the KSP parts are equipped with small but effective suicide charge they silently vaporize as soon as they recede at more than 5 000 m from the command pod. Since you\'ve lost the track you\'ll not be getting that impact confirmation.

[Goddess Bhavani]

That and jettisoned objects or stages do not give a position marker. Impact craters would be somewhat nice for future releases actually, especially where there\'s persistence.

[Iskierka]

As all the KSP parts are equipped with small but effective suicide charge they silently vaporize as soon as they recede at more than 5 000 m from the command pod. Since you\'ve lost the track you\'ll not be getting that impact confirmation.

This was noticed shortly after reporting, further tests have been more careful to stay within line of sight.

Also, the plan hasn\'t been for seeing where, it was just to see how long it endured.

I\'ve made some progress on figuring out how to trim the drone to fly far, but implementing the design is tricky. Best results so far are with minor modifications of default settings, but are not self-sustaining (Within the bounds of launch conditions tested). My testing ends here for now, with one drone that flies at only ~50 m/s, and very insistently upwards, but is following a stable path.

[Tosh]

Angle 15 deg regulated by SAS (does not matter, drone flies itself now. Improper launch angle like straight up means drone just run out of fuel uselessly)

Speed about 130m/s (since throttle was full all the way)

Altitude about 500m

Please provide an exact speed reading. Drone seems to be very sensitive to launch speed. I -- one again -- cannot repeat your results))).

Moreover I\'ve found the drone extremely sensitive to decoupler placement. I need to somehow mark a correct spot on the model . Please standby until I release a new one.

And -- I still dunno what\'s to do with that plane-dependent behavior. When I launch the drone (with your latest config) from QF-106 plane it flies a nearly-ballistic trajectory and hits the ground. When I launch it from a standard plane, it climbs up and backwards and then circles endlessly above the KSC.

[Iskierka]

The plane-dependent performance might instead be due to each competitor assembling their drone with marginal differences in positioning of parts.

EDIT: I\'d thought this launch would be a dud, seeing its steep and slow climb, but I decided to let it fly for a bit and it seems to be levelling out around 5 km and 90 m/s. I\'ll have to watch it closer and see how it goes.

[Tosh]

The plane-dependent performance might instead be due to each competitor assembling their drone with marginal differences in positioning of parts.

Not only. When launching from the heavier plane airflow tends to press a drone to plane\'s belly giving a drone some pitching moment while a drone launched from the light one falls right below.

And what\'s concerning placement differences -- I\'m currently marking a spot for ejection rack on the drone\'s back to eliminate one of the variables...

[Iskierka]

Status update: the drone appears to be level at 8.3 km altitude and 155 m/s velocity after 15 mins flight. (I was paused for long periods) At this point the distance travelled is at least 90 km, most likely over 100, but an accurate figure is difficult due to erratic flying from the chase plane. Hopefully the drone doesn\'t outlast the chase and burnout can be observed.

[Tosh]

Status update: the drone appears to be level at 8.3 km altitude and 155 m/s velocity after 15 mins flight. (I was paused for long periods) At this point the distance travelled is at least 90 km, most likely over 100, but an accurate figure is difficult due to erratic flying from the chase plane. Hopefully the drone doesn\'t outlast the chase and burnout can be observed.

HOW????////??

I want your config and launch data))

[Iskierka]

Launch conditions were not accurately taken, due to it not being expected this configuration would work so well, but to estimate:

80 m/s, 400 m altitude, less than 10 degrees above horizon.

As for config, I\'ll look at cleaning up the mess I made once this flight is completed, but the changes I made were:

Reconfigured drone to be overhead mount, which removed problem of drone hitting launch vehicle. Created attachment nodes for forward wings for consistent placement. Tailfins were split into left/right pieces and given very slight inclination, and placed so the leading edge was touching the front of the flat mounting area for them. Throttle was reduced to 2, and I think that should be all changes. No craft editing, though constant roll tendencies make me want to look at the craft for the aircraft itself ...

[Tosh]

I\'d like to review an entire pack and try to launch it myself. If it can travel to 90 km... that is really awesome. Would you mind PMing it to me?

[Tosh]

What a pain to travel 34 km (with good chances to travel whole 50!) just to collide mid-air with that f***ing drone! 8)

And --

Warning! Drone v2.5 released. Attachment spot for ejection rack is now marked on the back of the fuselage and is visible through at highest zoom level. Those who hang the drone up below the plane should use this attachment spot.

Config values are the same as in v2.4.

I\'ve also uploaded updated .craft file to the opening post.

[Iskierka]

Damn, unlucky with that collision.

I\'ll PM you the configs for the drone, craft for the aircraft, and my C7 configs (I don\'t use stock and I doubt this plane could match the drone\'s pace otherwise), though it might not be today, as I should leave soon-ish and would have to tidy up a little. As a further update on the drone\'s progress, it appears to be oscillating between just over 8.5 km and 8.4 km, and is still going after 45 mins (5 mins remain) at a strong ~191 m/s. Ground distance covered by the chase plane is currently around 450 km. Going by the roleplay political map, we\'re just crossing the north coast of Pedestrania, and will probably burnout not far from Forseti. Predicted impact zone is somewhere near the border of these countries.

[Tosh]

-snip- As a further update on the drone\'s progress, it appears to be oscillating between just over 8.5 km and 8.4 km, and is still going after 45 mins (5 mins remain) at a strong ~191 m/s. Ground distance covered by the chase plane is currently around 450 km. -snip-

Holy cow...

[Iskierka]

12 seconds to burnout.

Just in case anyone was doubting.

Unfortunately, the drone seems to be much less aerodynamic than the plane, and didn\'t waste time dropping behind after burnout. This unfortunately means impact point or distance cannot be determined. We can offer, however, a final shot of the drone being left behind, an image of an estimated impact point based on last known trajectory, and an estimate of 500-520 km range launched from low altitude.

On a more useful note; I think this set-up for the drone should be quite balanced, and will probably have minimal dependence on launch conditions. I expect it to be possible to adjust final flight conditions by varying throttle, though I don\'t know how much, beyond that a thrust of 7 from low altitude will begin cartwheeling uselessly. (I think this one might be more useful launched at altitude and speed)

[Tosh]

12 seconds to burnout.

I need at least a screenshot of drone launching to be able to add you to the leaderboard . And since you\'re competing as an engineer configs are also necessary.

[Iskierka]

Fair enough, I would\'ve had a launch photo but as stated, didn\'t expect success.

On a retry, launch:

Hopefully the runway counts.

As far as has been flown, the behaviour is the same. Should be approaching 8 km shortly after the 10 minute mark, and stabilising around 15 minutes.

I\'m attaching a zip of the drone with adjusted configs (though I haven\'t cleaned them up, so there\'s some mess from trying to create attachment nodes. Because there are now left and right tail pieces, and the drone has been altered for mounting above the plane, testing in a different design will require complete (Yes, complete, don\'t try get away with half) reconstruction. I suspect the launch aircraft would be able to keep up with the drone, but it may suffer from lack-of-fuel problems prior to drone burnout in default C7 configs. It\'s also a work in progress and joystick problems make it difficult to tune out issues.

Requires: attached drone, C7 (original and remake), tiberdyne shuttle launch system.

[Tosh]

I\'m attaching a zip of the drone with adjusted configs (though I haven\'t cleaned them up, so there\'s some mess from trying to create attachment nodes. Because there are now left and right tail pieces, and the drone has been altered for mounting above the plane, testing in a different design will require complete (Yes, complete, don\'t try get away with half) reconstruction. I suspect the launch aircraft would be able to keep up with the drone, but it may suffer from lack-of-fuel problems prior to drone burnout in default C7 configs. It\'s also a work in progress and joystick problems make it difficult to tune out issues.

Thanks!

I\'m confirming your results. At the first try the drone just easily flew 50 000 meters in 10 minutes -- and is not showing any sign of losing stability. That\'s far better than all I\'ve seen before.

You\'re currently the Number One on the Leaderboard

P.S. Quite a stable craft you\'ve built. I thought I never can fly a plane without an ASAS (I don\'t have a joystick) -- but your MA-03E is completely tractable to even a dumbass like me

UPD. And now I need some time to sort it out just how did you manage to acquire such a result, and can I reproduce it with some other kind of a plane. After that I\'ll message you to discuss a prize you\'re likely to receive. Of course, if noone succeeds in beating your achievement))

UPD2. Holy smokes! 20 minutes and 165 km -- and this crazy robot still flies! 8)

[Iskierka]

I was mainly trying some things learned from my first semester at uni, regarding aircraft stability and control. In short, what\'s stabilising the drone is the incidence I\'ve given the tails - they\'re angled slightly down now, so as alpha alters, the centre of lift moves back and forth, which means that there is a particular angle of attack the drone wants to settle at. I don\'t think this is strictly necessary, but even a slight change in alpha causes quite a significant restoring moment, making it a very effective correction factor.

EDIT: Actually, thinking about this more, I think I\'ve gotten confused, I should\'ve used opposite incidence to what I did in order to stabilise the drone. I\'m not sure why the lift doesn\'t completely destabilise because of this, though following points still apply.

In addition to this, wings are very near zero drag (when at zero alpha in KSP, they -are- zero drag). This means the high-mounted tails provide almost no drag to cause a pitch-back, but do add to the centre of gravity, moving it above the main body. Since the main drag is then from the main body, drag causes a pitch-down moment, which increases with velocity. With the way the engine is set up, the thrust is almost exactly opposite to drag, but with no proportionality to velocity, meaning there is a balanced velocity where D = T. Since drag is also proportional to air density, it goes faster at higher altitudes.

So, after some complex stability interactions, what happens is high lift causes it to ascend to a point where air density is such that, when D=T, lift = weight, thus resulting in a stable vehicle.

I think what some designs also suffered from was a problem of thrust being greater than weight, which enabled them to pitch all the way over and loop. With a mass just over 0.6, and a thrust of only 2, this thing only has a TWR of around 0.3, meaning instead of looping, it decides to power into the air, looking for its balance altitude.

[Tosh]

Thanks a lot! I was just about to ask you to explain that trick with fins incidence \'cause it\'s the only significant difference from an original config (not counting engine thrust and attachment rules).

I think an information you provided is enough for me to reconfigure the drone to fly any way I like it to.

P.S. Where are you studying at?

[Iskierka]

Aerospace Engineering at University of Manchester. First semester was flight, including a topic covering stability and control, so yeah.

As noted in my addition to that post, the incidence is probably wrong for perfect stability - though it might take quite some figuring out to re-balance other parameters to make it fly well again, so it\'s probably best to leave it as-is, since it\'s stable and flies like a (rocket propelled) arrow already.

[Tosh]

EDIT: Actually, thinking about this more, I think I\'ve gotten confused, I should\'ve used opposite incidence to what I did in order to stabilise the drone. I\'m not sure why the lift doesn\'t completely destabilise because of this, though following points still apply.

Did you mean to make the drone pitch up as long as there\'s enough lift (i.e. enough speed) to do that, and to pitch down otherwise? It works exactly as intended. I even increased incidence 10 times to stabilize it even further.

Moreover I reconfigured the drone to be launched from below the plane, and it still works just fine. Now I\'m trying to make it fly faster (keeping 30 m/s at low altitude makes a difficult task for a chase plane ) and to stabilize at lower altitude...

Aerospace Engineering at University of Manchester. First semester was flight, including a topic covering stability and control, so yeah..

Ah. No surprise you could deal with this robot)))

[Iskierka]

Did you mean to make the drone pitch up as long as there\'s enough lift (i.e. enough speed) to do that, and to pitch down otherwise? It works exactly as intended. I even increased incidence 10 times to stabilize it even further.

Moreover I reconfigured the drone to be launched from below the plane, and it still works just fine. Now I\'m trying to make it fly faster (keeping 30 m/s at low altitude makes a difficult task for a chase plane ) and to stabilize at lower altitude...

Ah yes, ofc. I was thinking the other way; if the plane pitches up with no other changes, the lift will encourage it down. I wasn\'t considering if velocity and nothing else changed. I\'m not sure what\'s providing the former stabilisation mode now, since it\'s necessary to avoid losing control completely, but I\'d probably be able to find something if I tried to construct stability equations for it.

Ah. No surprise you could deal with this robot)))

Only first year and first semester so far

[Volt]

Hmm, I\'m using the latest version posted above with a slight adaptation of the MA-03. I\'m observing an uncontrollable pitch-up in the drone when launched. I don\'t know if that\'s what Iskierka is talking about, but I can confirm it...