Airplane Range on 100 EC

[18Watt]

20 minutes ago, camacju said:

I was playing around with offset props to try to reduce the torque required to spin them. Instead, I think I found a new glitch

I hate to look like an idiot, - No wait, that's not true!  I have gotten used to it!

I watched your video, but wasn't able to really grasp what the glitch was.  If you can dumb it down, I'm curious what you found.

Speaking of glitches, there used to be a glitch that was referred to as 'the infini-glide' glitch.  I'm not sure if KSP ever fixed that one or not.  I did not forbid that glitch for a few reasons, the biggest being I think that one disappeared at some point.

The other reason was the players most likely to respond to aero-challenges don't need or use that glitch.  You and @OJT are at the top of that list in my book right now.  Anyway, curious if you're familiar with that glitch, and if it's still present?  

Anyway, I'm not sure what glitch you've discovered.  I've seen several glitches relating to the robotics parts (including the rotors), but nothing I'd consider beneficial.

[camacju]

The propeller is completely off, but a phantom torque force is causing my propeller to spin up. Basically autorotation on steroids.

It works because KSP calculates propeller blade lift at a point offset way out from the axis of rotation to compensate for Unity's rotation limit. But lift also comes with a lift-induced drag acting opposite to the motion, which is applied at the base of the propeller.

Normally this is just fine, because the lift induced drag acts on the same side of the propeller as the lift point, causing a negative torque. But when the props are offset inward far enough, the drag acts on the opposite side of the propeller, which actually produces a positive torque.

This means that when the propeller produces thrust, the thrust induces a torque that further accelerates the propeller. So all I need to do is get the prop moving initially and then glide infinitely.

19 minutes ago, 18Watt said:

Anyway, curious if you're familiar with that glitch, and if it's still present?  

If I remember correctly, the original infiniglide glitch centered on producing a phantom force by spamming control surfaces. It's no longer present in the latest version of KSP.

On 1.11, which is the version I play on, there's a version of the infiniglide glitch that involves spamming the flag panels at a precise angle. I'm not using it in my runs, but I did consider it.

[18Watt]

10 minutes ago, camacju said:

The propeller is completely off, but a phantom torque force is causing my propeller to spin up. Basically autorotation on steroids.

I'm actually a professional pilot (scary, I know!).  My initial take on that was what happens to a propeller when it stops receiving torque from the engine (engine quits)- which is there is still torque.  But now the torque is reversed, the propeller is driving the engine.  That torque requires energy, which is coming from the airflow, and is also producing a large amount of drag.  Which is why we 'feather' props if we can when we lose and engine, the drag reduction is very significant.

However, it looks like you have stumbled onto something different, which produces positive torque in extreme edge situations.  Well, that's pretty cool!  I still don't really understand it, but it's late on a Saturday night..

Can't remember if it was you or @OJT that made this comment, but I think we're all hoping a lot of the rover-wheel and aero glitches are sorted out in KSP2.  Honestly, I don't think the aero glitches will ever be fully eliminated.  Aero is quite complicated, and to simulate aero a lot of compromises have to be made.  

Wheel-ground interactions are also complicated, but a lot of commercial video games have done a fantastic job simulating those interactions very well.  Not just racing games, even games like GTA do a pretty good job of it.  Really hoping KSP2 gets wheel simulations right, or at least much better than KSP1.

I do not consider the optimizations you and OJT have found to be glitches at all.  Instead, you have found ways to optimize the aero model of KSP.  Not necessarily due to aero glitches, but rather due to what KSP allows you to fabricate.  Most of your gains seem to be drag reductions.  In real-life, those reductions would not be impossible, but rather extremely expensive and difficult to achieve.  KSP allows some pretty extreme drag reductions, it's just a matter of finding them.

The ability to 'find torque' is pretty interesting though!  If you need an environment to really test that, I suggest Eve.  I did an Elcano at Eve (OJT is rolling his eyes..) using electric props for propulsion.  Whatever happens on Kerbin you can multiply by at least 5 on Eve, so if you think you've found a possible aero exploit Eve would be a good place to validate.

[OJT]

4 hours ago, 18Watt said:

I suspect we are getting close to the limit of the attainable distance.  

Ahhh, about that...

I've been conducting lots of experiments today with various plane designs and various flight profiles and this is what I found out:

1) Fan shrouds actually do have a purpose: they have very low skin drag. Hell, the aero debug menu shows 0 OccA and 0 WDrg in both Y directions if both nodes are occluded. Putting the engine inside the shroud makes the whole plane more aerodynamic

2) And continuing to that: props in fan shrouds work even if the fan shroud is completely covered! So, for example if you attach the fan shroud to the end of a plane, put engine in it and then place, say, a nose cone on the rear node, the props will still generate thrust. This gives extra opportunities for drag optimizations

The most important discovery however doesn't have to do anything with plane design. It is the flight plan: I've completely revised it

All of my (and seemingly camacju's aswell) attempts followed this profile: slowly get to high altitude, play around with variables (engine power, prop angle etc.) during ascent, try to spend as much time high up as possible while maintaining speed, slowly glide down when power runs out.

I've figured out completely different approach: pick the engine RPM, Torque and prop angle very carefully, take off and then fly very low and very slow, like about 50-70 m/s. If you pick your variables very carefully and maintain perfectly level flight, the electricity usage will be extraordinarily low and even despite the low speed, you can reach flight distances of over 800 km! Yes, you've read that correctly

This bring along good news and bad news. Bad news is, the stuff named above assume a single-rotor plane, which makes the "maintain perfectly level flight" part of the equation pretty much impossible in stock KSP because you will need to constantly correct the heading with trims and whatnot. And considering the estimate distances and velocities, this flight can take up almost 4 hours. And I don't think there are people crazy enough to attempt to fly this with the limitations that I've mentioned

What's the good news? I've actually found an autopilot mod that can do exactly what we need! It is called AtmosphereAutopilot (I think there's actually a forum topic in the Mod Showcase section, I found it on SpaceDock) and it has very sophisticated cruise flight options. In fact, I am flying my mission at this very moment that I am writing this. Unfortunately, phys time warp is screwing up the said Autopilot, so I have to fly it in real time, but at least I don't need to actively participate in the flying: I just need to occasionally check the progress, maybe correct the heading from time to time to avoid islands and whatnot, but other than that it is entirely self sustained. I just need to hope that the game doesn't crash midflight 

This brings in different problem though. You didn't seem to discourage any so to speak "flight assistance" mods in the OP, but this Autopilot mod is so overpowered that it makes the challenge absurdly trivial, because before we needed to build AND fly our creations, but now we basically just need to build and let the Autopilot fly it for couple of hours and claim the leaderboard spot, and when it comes to building you eventually reach a certain point where no more optimizations can be applied (which I feel like I've reached now with my concepts)

What do you think about all this? Will you allow my "Autopilot" submission (that is flying at the moment) if I post it? Or will it not be in the spirit of this challenge? Let me know what you think @18Watt

[18Watt]

1 hour ago, OJT said:

And I don't think there are people crazy enough to attempt to fly this with the limitations that I've mentioned

I run the Elcano challenge.  4 hours is a drop in the bucket..

1 hour ago, OJT said:

I've actually found an autopilot mod that can do exactly what we need!

Awesome!  I don't even use mods, but it does bother me that the common autopilot mods I've seen don't really manage all the inputs we need in a satisfactory way.  As @camacju commented, PID controllers take a lot of dedication to get working well.

1 hour ago, OJT said:

Unfortunately, phys time warp is screwing up the said Autopilot

Yeah,  I believe you commented that during glide, just going to 2X warp caused the plane to pitch down.  I've seen that too, and there's not a perfect way to compensate for it.  Yet another thing that we hope will work better in KSP2.  We're putting a lot of pressure on the KSP2 team.

1 hour ago, OJT said:

You didn't seem to discourage any so to speak "flight assistance" mods in the OP

No, I didn't.  And I'm not going to discourage it now.  I personally play stock.  For challenges like this, I like to specify 'Stock Parts' because otherwise it's pretty easy to come up with a modded engine that produces a lot of power with no actual EC drain.  So stock parts to keep everybody on the same playing field.

However, as I pilot, I recognize that maximum efficiency happens when the airplane is maintaining course and attitude to a high degree of accuracy.  This is easy to see in KSP when hand-flying a plane badly.  Altitude and heading excursions rapidly cut into efficiency.  The computer is awful at making rapid maneuvers, but it can fly a straight line better than I can.  Compensating for torque is a unique problem, the torque available in KSP rotors far exceeds anything I've seen in similar motors in the real world.  It's such a wide range of possible torque values that I do understand making a PID controller to compensate for all possible values is difficult if not impossible.

Anyway, for the purposes of this challenge, autopilot mods are specifically allowed. (I'll update the rules..)  If you use such a mod, please make a note of which mod you used.  In some cases, autopilot mods may require a non-stock part.  That's ok, as long as the rest of the plane uses stock parts, and no physics are changes.  The basic idea is:  Theoretically, if you could react as fast as a computer, you could fly the non-autopilot plane just as well and get equal performance.

At 800 km I think we've gone far beyond what would be possible in 2022 on planet Earth.  But that's not the point of this challenge.  The point is what can you do with the stock parts in KSP.  (again, autopilots are fine..)

1 hour ago, OJT said:

Will you allow my "Autopilot" submission (that is flying at the moment) if I post it?

Yes, as long as it doesn't alter the functionality of stock parts or alter physics- autopilots are fine.

[camacju]

2 hours ago, 18Watt said:

Whatever happens on Kerbin you can multiply by at least 5 on Eve, so if you think you've found a possible aero exploit Eve would be a good place to validate.

I talked to Lt_Duckweed about this and he confirmed that this is actually a real effect and he's used it before.

I'm considering carefully balancing the prop so the phantom torque cancels out lift-induced drag and prop drag as much as possible, so I can massively reduce the torque limits on my motor and still fly at high speed. However that might be a bit too cheaty for me  

1 hour ago, OJT said:

I've figured out completely different approach: pick the engine RPM, Torque and prop angle very carefully, take off and then fly very low and very slow, like about 50-70 m/s. If you pick your variables very carefully and maintain perfectly level flight, the electricity usage will be extraordinarily low and even despite the low speed, you can reach flight distances of over 800 km! Yes, you've read that correctly

I suspected something like this would be better, as you get a superior lift to drag ratio at lower speeds. I think that in our recent submissions, my craft was more drag optimized, but range was similar, and I'm pretty sure it was because I was flying faster and getting a lift/drag penalty. So you'd want the RPM limit to be as high as possible but set torque limit incredibly low and play with prop angle until you can just barely take off. The slower the better probably?

I tried to maximize speed because I'm impatient and wanted the mission to be over with quickly  

Note that for jet endurance missions this flips on its head because reaction engines have a specific impulse so their specific energy increases with increasing speed, but props drop off sharply as speed approaches Mach 1.

Another possible approach that I investigated (but couldn't get to work reliably) was a propeller made from ailerons or wing panels. I've found that in some cases these can perform better than the DLC prop blades, but they also seem to be pretty power hungry.

1 hour ago, OJT said:

Fan shrouds actually do have a purpose: they have very low skin drag. Hell, the aero debug menu shows 0 OccA and 0 WDrg in both Y directions if both nodes are occluded. Putting the engine inside the shroud makes the whole plane more aerodynamic

KSP's drag cube model means that any part can have zero WDrg values corresponding to a fully occluded node - I think the fan shroud just increases mass and drag with no real benefit. Ideally you'd put the prop motor entirely inside a fairing and keep the propeller blades outside the fairing.

[18Watt]

Just now, camacju said:

I'm considering carefully balancing the prop so the phantom torque cancels out lift-induced drag and prop drag as much as possible, so I can massively reduce the torque limits on my motor and still fly at high speed. However that might be a bit too cheaty for me

Cheaty?  Not sure.  The point of this challenge is to see how far KSP will allow you to stretch 100EC.  I think everyone would appreciate you specifying any physics quirks you take advantage of.  But if stock KSP allows you to do it, this challenge does not prohibit it.  However, I'm glad you commented that the 'infiniglide' thing doesn't function anymore, that's not what I was looking for..

3 minutes ago, camacju said:

I tried to maximize speed because I'm impatient and wanted the mission to be over with quickly  

Hah!  I thought 100EC would be a limit that would make this challenge a short one to enter.  Boy was I wrong!!

 

[OJT]

13 hours ago, OJT said:

If you pick your variables very carefully and maintain perfectly level flight, the electricity usage will be extraordinarily low and even despite the low speed, you can reach flight distances of over 800 km!

13 hours ago, OJT said:

reach flight distances of over 800 km!

Oh boy, was I wrong with my estimates. But let's not get ahead of ourselves

Remember I talked about a mission that I was flying and asking if it would be an eligible entry to this challenge? 

13 hours ago, OJT said:

In fact, I am flying my mission at this very moment that I am writing this. Unfortunately, phys time warp is screwing up the said Autopilot, so I have to fly it in real time, but at least I don't need to actively participate in the flying: I just need to occasionally check the progress, maybe correct the heading from time to time to avoid islands and whatnot, but other than that it is entirely self sustained

Well, I forgot to adjust one time and the plane crashed into the island, the very same one that me and camacju finished our best entries so far. To say I was annoyed would be an understatement, however I decided to use this opportunity to optimize the design and see just how much I can squeeze out of 100 EC units

I did a LOT of test flights, tested lots of part combinations and different settings on engines and props and in the end, I ended up with this. Behold! And ignore the name, this was far from 6th iteration but I couldn't be bothered to make lots of craft saves for every single minor adjustment

Firstly, I reduced the weight. From my tests the weight didn't influence the battery consumption much and it more came down to drag, but more on that later. However, tests showed that lower takeoff weight allows for easier acceleration on the runway which can preserve every tiny bit of battery charge. I replaced or removed some parts, most notably I swapped 4 hardpoints for two decouplers. Google search showed that even though jettisoned decouplers leave small mount points on the part it detaches from, these points have zero drag and zero mass, so the functional shape wouldn't be influenced. I also went for deltawing configuration similar to camacju's designs to get rid of rear winglets which also saved a bit of mass

Spoiler

Most important adjustments came from aerodynamic optimizations. I played around with fairing design to get the most optimal shape I could manage (both in terms of streamlining and weight). Size of control surfaces has been reduced and positions of said surfaces has been carefully adjusted to make the plane cruise as stably as possible. Reason being I wanted to eradicate any unnecessary corrections that Autopilot might do to hold its course, which is why, for example, I am still utilizing two small control surfaces as yaw stabilizers: single Basic Fin  made the flight a bit twitchy and Autopilot used a lot of adjustments to hold course, reducing prograde velocity. Two small control surfaces might be heavier, but they make the plane much more stable. Every minuscule detail that I could consider I did. I even swapped out one pilot for this mission

Both pilots for this mission will be females. Why? Because during tests I actually noticed, to my own surprise, that even though all kerbals weigh the same, female kerbals are actually a bit shorter than male kerbals. This allowed me to make the fairing a bit smaller than I otherwise would have to. Picture of kerbals below for demonstration. Unfortunately, it does mean that Jebediah has been sidelined and has been replaced by an equally capable female pilot. Valentina still stays, repping The Original Quartet in this mission

We know the craft, we know the pilots. Let's proceed with the mission!

Proof of battery charge

Spoiler

Because the plane was extremely optimized for level flight, takeoff proved to be very difficult. I utilized the entire length of the runway to accelerate, then as I left ground I dropped the gear and quickly pitched up before hitting the ground and then I engaged the autopilot to maintain the level flight. Takeoff was quite extreme for the relatively low speeds that we are dealing here, I almost crashed the plane right as I left the runway, which is why I didn't take any screenshots of the takeoff  

After the plane stabilized it slowly started to pick up speed until it reached the cruise velocity. I used 150 RPM limit and 7% throttle at first with prop angle of 50 degrees

Spoiler

Then a bit later, after playing a bit with the settings I stopped at 140 RPM limit and 49 degrees prop angle (it says 48.5ish but the value I inputted was 49). I maintained roughly the same speed, but EC usage reduced by 0.5 microunits per second, therefore making the flight more efficient

Spoiler

I also noticed that the plane was veeeery slowly climbing by itself, so I set the Autopilot to maintain the altitude at 114 meters. No particular reason, it was just the height the plane was flying at when I engaged the altitude hold.

As you can see, I fly very low which is why it was important to maintain flight above water. First I've reached the island that my Personal Best attempt finished at

Spoiler

After that, I simply roughly followed the geodesic line that is used for Kerbin Marine circumnavigations. This ensured that I was flying above water and perfectly straight, reducing the unnecessary corrections

Sun was slowly going down, and with it my cruise velocity. KSP simulates temperature changes depending on time of day (and latitude if I remember it correctly), and temperature is an important factor in aerodynamic equations. At least this is my assumption on why the plane slowed down a bit. It still maintained the level flight however and my EC charge actually reduced a bit aswell due to reduction of engine's RPMs, so I didn't lose out on theoretical range (from my calculations at least)

Spoiler

Sun comes up, I reached another set of islands and I was almost running out of charge

Spoiler

Out of charge, finally. Before the charge depleted I prepared my autopilot to maintain slow descent rate to get those extra meters through gliding. As the charge went out, I engaged the descent rate hold and retracted the props

Spoiler

Splashdown! And final location

But what is the final travel distance? As you can see I travelled quite far north which makes the distance calculation purely through longitude change senseless. So, after a bit of quick googling, I found this website. It can calculate the distances between multiple waypoints based on coordinates, which makes it very useful for this challenge

http://kerbalspace.agency/Tracker/distance.htm

I used three waypoints for calculations in total

1. Position of the flag at the runway (-0.23 lat, -74.43 long)
2. Position of the shore near the island that I flew past roughly halfway into the journey (-4.26, -40.47)
3. Position of the splashdown (27.45, 19.43)

This gives me the travelled distance of... 1047.055 kilometers. Yes, I cracked the thousand. And even though the real distance from the flag will probably be a bit lower than that, the fact that I've built a plane that can fly over 1000 km with just 100 EC units is pretty damn good

Now that I see camacju's offset prop experiments, I wonder if I can push it even further 

 

[camacju]

4 minutes ago, OJT said:

female kerbals are actually a bit shorter than male kerbals

I would never have even considered this optimization!

Wow, you've smashed this challenge out of the park. I might have to start getting crazier with my aero optimizations. 

Or maybe start using an autopilot mod, or both

[18Watt]

32 minutes ago, OJT said:

KSP simulates temperature changes depending on time of day (and latitude if I remember it correctly)

I’ve seen this as well, in one circumstance it actually gave me unexpected results.  During a sea circumnavigation I had expected better performance from the jet engine with cooler temperatures (evening, and I was up north a ways).  Instead, performance was reduced- lower speed from the same fuel flow.  Which I can only attribute to increased drag from the denser air.

37 minutes ago, OJT said:

Yes, I cracked the thousand

Crikey!!!

31 minutes ago, camacju said:

Or maybe start using an autopilot mod, or both

I normally just play stock, but if y’all found an autopilot mod that functions fairly well, I’m sold.

Trying to decide wether to modify the distance measurement method (again).  Originally the idea was just as far from the flag as you can get, but since I adjusted to make great circle distance measurements, I think it also makes sense to use multiple points.  Especially since you only really made one major course adjustment at the tip of the peninsula.

For now, let’s call the multi-point method provisionally valid.  If I incorporate that into the rules, I’ll probably set a limit on the number of points.

 

[OJT]

36 minutes ago, 18Watt said:

For now, let’s call the multi-point method provisionally valid.  If I incorporate that into the rules, I’ll probably set a limit on the number of points

Let's assume one would build a plane that could circumnavigate Kerbin and that would fly in a similar way to my record setter (as in, low to water and in straight lines). Quick glance tells that it would require at the very least 5 major corrections, therefore 5 points: one at the edge of peninsula east of KSC and then roughly following the Elcano marine circumnavigation trajectory, making a full circle at the peninsula edge. Photo of that below. My attempt finished halfway between 1st and 2nd point, near two small islands. In addition, the website I linked has a limit of 20 waypoints, so I think, incase someone actually manages to build a 100 EC circumnavigator, 20 points should be more than sufficient if one would like to make the corrections smoother instead of sharp turns because that can save a lot of kilometers

Now will such plane even be built? 2 days ago I would say that a man must be tripping to even consider that, but seeing how stupidly far we managed to travel I can't entirely dismiss such possibility 

Edited May 15, 2022 by OJT

[camacju]

It seems like you've fallen for the map projection - "straight lines" on this projection don't necessarily correspond to great circles.

Here's an example of a geodesic that follows a path very close to sea level:

Taken from one of Stratzenblitz's videos.

On a rectangular map projection this looks like a curved path, but on Kerbin's surface this will just be a straight line.

[OJT]

4 minutes ago, camacju said:

It seems like you've fallen for the map projection - "straight lines" on this projection don't necessarily correspond to great circles.

Here's an example of a geodesic that follows a path very close to sea level:

Taken from one of Stratzenblitz's videos.

On a rectangular map projection this looks like a curved path, but on Kerbin's surface this will just be a straight line.

Yeah, I realized that I done goofed up after posting the message. Gotta firm the L 

[OJT]

So, I was playing around with the craft design and managed to make a huge breakthrough. The new plane is currently flying and, if everything goes smoothly, I will be posting results later in the day 

[camacju]

Oh no

If you continue pushing the bar I'll have to either graciously bow out of this challenge, or begin to abuse aero exploits 

[18Watt]

1 hour ago, OJT said:

So, I was playing around with the craft design and managed to make a huge breakthrough.

 

8 minutes ago, camacju said:

If you continue pushing the bar I'll have to either graciously bow out of this challenge, or begin to abuse aero exploits 

It's like working with two Dr. Evils!

Edited May 17, 2022 by 18Watt

[sevenperforce]

Opens this thread

.

Hmm, I'm sure I can figure out a way to beat that time

.

Reads everything from @camacju

.

Reads everything from @OJT

.

Okay, yeah, never mind.

[camacju]

Well, I've got this:

Airspeed = 52.30 m/s

Electricity usage = 0.0032 u/s

Electricity remaining = 86.47

This translates to 27022 seconds of powered flight remaining, or 1413 km. And that's not including the initial bit of flight to get here, or the glide down to the ocean.

I am completely unwilling to fly for 7.5 hours even with an autopilot. But this shows that a twin engine design is still quite capable! (I went back to two engines because a motor casing is only 18 kg and this way Smart A.S.S. can actually fly the craft without completely dying).

Edited May 17, 2022 by camacju

[camacju]

On 5/15/2022 at 11:32 AM, 18Watt said:

Instead, performance was reduced- lower speed from the same fuel flow.  Which I can only attribute to increased drag from the denser air.

By the way, increased drag from denser air probably isn't a factor here, as all else equal, higher drag will mean higher lift so your plane will simply reach equilibrium higher in the atmosphere.

I suspect it's due to two things:

-Colder air means lower speed of sound, so you're shifting the Mach curve. If you're using a supersonic jet engine at max speed, this means your max speed is reduced.

-Centripetal force is reduced because you're farther north, so you need more lift to cancel out gravity - which means more drag, but not because the air is denser

[18Watt]

1 hour ago, camacju said:

By the way, increased drag from denser air probably isn't a factor here, as all else equal, higher drag will mean higher lift so your plane will simply reach equilibrium higher in the atmosphere.

That was for a boat, for an Elcano run on Kerbin.  So everything was happening at sea level..

1 hour ago, camacju said:

Colder air means lower speed of sound, so you're shifting the Mach curve. If you're using a supersonic jet engine at max speed, this means your max speed is reduced.

I was using the Goliath (or maybe two or three of them?).  Top speed was around 90 m/s, so transonic effects should not have come into play, especially at sea level.

What I expected at a constant altitude (sea level) with decreasing temperature was increased thrust, accompanied by increased fuel flow and increased drag.  As most of my drag was from the water (it was a boat), I did not expect much if any penalties on speed.  I only expected higher thrust and fuel flow, and higher speed.

Instead, what I got was slightly higher fuel flow, combined with a significantly slower speed.  

Edit:  For boats in KSP, trim is a huge issue.  Not aerodynamic trim, but rather fore-aft weight balance, and also managing any offset thrust- especially if using props which have to be mounted well above the water, and thus almost always above the COM.  Anyway, near as I can tell the boat was 'trimmed' perfectly, so I'm discounting trim issues.  Although they could have been a factor too, who knows?

1 hour ago, camacju said:

Centripetal force is reduced because you're farther north, so you need more lift to cancel out gravity - which means more drag, but not because the air is denser

I don't think that's what's going on either, but the fact that I was farther north made me wonder if KSP's calculation of ground speed ('Surface') was thrown off by being at a higher latitude.  I was basing my performance off of what KSP displayed as my 'Surface' speed.  I did not verify my speed by independent calculation (great circle distance/time), so it's possible KSP was giving me a different speed value at high latitudes.  I don't actually think this was the case, but it popped into my mind.

By the way, the Goliath has an incredibly high ISP (as I'm sure you know).  I was able to circumnavigate Kerbin by sea stopping only once for (ISRU) refueling.  Had I traveled at a slower speed, I'm sure I could have done the circumnavigation without refueling.

1 hour ago, sevenperforce said:

Reads everything from @camacju

Reads everything from @OJT

Okay, yeah, never mind.

I promise, if you make an entry I'll add you to the leaderboard!  While getting the top spot is pretty cool, this challenge is also about what YOU can do.  

I fly planes for a living.  Pretty cool ones.  One thing I've found to be true:  No matter how fast you are, there is always somebody faster.  My point is that you don't need to avoid participating just because you're not the fastest plane in the sky.  Even if you're the slowest plane in the sky, it still beats driving!

Edited May 17, 2022 by 18Watt

[OJT]

4 hours ago, OJT said:

So, I was playing around with the craft design and managed to make a huge breakthrough. The new plane is currently flying and, if everything goes smoothly, I will be posting results later in the day 

Well, I nodded off and the plane crashed... will be restarting later in the day, results might take a bit longer than I initially planned

[OJT]

Well, I promised you all a huge breakthrough. Here it is. As usual, the name doesn't actually represent the version number: I've had maybe 8 different variations of this craft before I went ahead with this one

First new thing is the fuselage: it has been carefully optimized and improved upon the previous plane by offset/rotate manipulations. Also, 100 EC battery has been swapped with 1k round battery for drag optimization reasons. In the end not only have I got a less draggier fuselage, I also managed to reduce the dry weight of the fuselage a little bit. Although the weight reduction doesn't really make much of a difference and it was not the main reason for new fuselage anyway, but still, a nice bonus. Engine and prop setup is the same as in previous entry

Wings are also new, with a bit of extra wing area for gliding. This made the plane a bit heavier than my previous entry, after jettisoning the landing gear the plane weighs cool round 1000 kilograms (with both Kerbals without jetpacks and parachutes). Speaking of kerbals, they will again be females

Spoiler

Proof of EC charge. As this is a much bigger battery, I reduced its charge to 100 in SPH

At this point you might be wondering: the plane doesn't look that much different from the previous entry, so what's the breakthrough? The breakthrough is flight profile. Carefully adjusting RPM, engine power and prop angle throughout the entire flight delivers fantastic results. So to do this, I assigned the RCS controls to prop angles, engine RPM and Torque limit

I start at 150 RPM and 7% power and hold these for the takeoff phase. The main adjustments are with prop angle at first: right at the start I set the angle at 85 degrees and then gradually lower it to around 50ish degrees as I accelerate on the runway. I look at the Forward Lift produced by the props and on the runway I keep it in 0.2-0.3 kN range through beforementioned prop adjustments. With this a takeoff speed of about 56 m/s is achieved. Just as the plane goes off the runway, I discard the landing gear and engage the Autopilot. I set the vertical speed of 3 m/s to slowly climb to 700 meters as I continue to accelerate through prop adjustment: at this point I don't look at forward lift generated anymore, I simply monitor the velocity

Spoiler

As I reach 700 meters, I tell the Autopilot to hold the altitude and from then it's all adjustments: I carefully lower the prop angle, reduce the RPM limit and raise the Torque limit until I end up at the following settings: 19.19 degrees of prop angle, 50 engine RPM limit and 10 engine Torque limit. This achieves maximum EC usage of 3.48 microunits per second and cruise speed of 69-75 m/s depending on the air temperature. These settings will be kept for pretty much entire flight

Spoiler

Flight path will be the same as in previous entry: fly eastward towards peninsula and then roughly following the geodesic line. Passing over the peninsula now: I've setup a series of "checkpoint" flags before the mission for ease of navigation. Some of those checkpoints I followed to the T, for the others I simply used them to assess the direction I was going in

Spoiler

Sunset and moon crescent

Spoiler

Passing by the islands where my previous entry splashed down: we're still going strong

Spoiler

Reaching the isthmus at the far north of the map. This is where things got a bit scary: as I was approaching the land I wasn't sure whether my plane would clear the hills, so I tried to climb to 800 meters. By doing so however my engine slowed down, props didn't generate much thrust and I stalled the plane. Thankfully, by quickly readjusting all settings and recovering the plane I saved the mission, but I gotta tell you, I was shook

Spoiler

800 meters proved to be enough with a safe margin, so I continued forward. I even did some planet spotting as it got darker lol

Spoiler

Crossed the isthmus. From then on it was all water and no more dangerous landmasses, so I descended back to 700 meters

Spoiler

Slowly running out of charge. Preparing the plane for landing

Spoiler

Out of charge. Retracting the fan blades and slowly gliding towards the water

Spoiler

Splashdown!

Coordinates of the splashdown

Now is the time to calculate just how far I managed to go. Again I will be using the website that I used in previous entry. These are the coordinates I used for calculation

1. Starting Flag near the Runway (-0.2, -74.43)
2. Checkpoint 1 Flag, on the Peninsula (-2.1, -39.22)
3. Water passage between Two Islands, near Checkpoint 2 (27.94, 16.259)
4. Checkpoint 3 Flag, on the western side of the Isthmus (45, 55.37)
5. Checkpoint 4 Flag, in the middle of the Isthmus (47.35, 69.17)
6. Checkpoint 6 Flag, on the eastern side of the Isthmus (47.30, 89.53)
7. Splashdown Site (20.43, 132.22)

Photo of the checkpoint flags below

Spoiler

All of this ends up in the final distance of......

2082.916 kilometers

This is an extraordinarily good result, but there is a downside to it: this is basically the limit of the current concept of the aircraft I've flown. Different RPM/Torque/Prop angle settings bring in diminishing results, different wing configurations didn't bring much either and the fuselage is already as streamlined and light as it could be. Not even adding or reducing amount of fan blades bring any improvements, only worsened the theoretical range. So, unfortunately, my dream of circumnavigating Kerbin with 100 EC are not gonna come to fruition, at the very least with the current aircraft design

The only thing I didn't touch in any of my entries (including this one) is the fan blade positioning: perhaps you can get more range if you carefully offset/rotate the blades so they spin themselves up, but at this point if your fan blades spin themselves up without any EC usage, then what's the point of the challenge anyway? But hey, that's just my opinion

@18Watt As for now, I've basically shut down the whole challenge lmao

 

Edited May 17, 2022 by OJT

[18Watt]

9 minutes ago, OJT said:

I've setup a series of "checkpoint" flags before the mission for ease of navigation.

That's a good idea- it's not easy to accurately fly a straight great circle route by just 'eyeballing' it.

20 minutes ago, OJT said:

Not even adding or reducing amount of fan blades bring any improvements,

I noticed you guys generally are using two blades.  When I try to see how fast a prop will go, I sometimes use 6 - 12 blades per rotor.  Not efficient, but seems to allow more speed.  Just curious if anyone has tried just one blade?  Wouldn't make any sense in the real world, but I'm not sure how much KSP would model the vibration.

One more question about blades:  I generally tie the torque to the main throttle.  I tie the blade angle to two keys, usually translate forward/back (H&N).  However, I get clunky results when adjusting blade angle that way- sometimes I want to change 0.5 degrees, but often I end up with 2 or 3 degree changes.  Curious what method you guys use to adjust blade angle.

27 minutes ago, OJT said:

As for now, I've basically shut down the whole challenge lmao

Heh, yeah I suspect @camacju may not have the patience for a run of that length.  Nice run!!

 

 

[OJT]

9 minutes ago, 18Watt said:

I noticed you guys generally are using two blades.  When I try to see how fast a prop will go, I sometimes use 6 - 12 blades per rotor.  Not efficient, but seems to allow more speed.  Just curious if anyone has tried just one blade?  Wouldn't make any sense in the real world, but I'm not sure how much KSP would model the vibration.

One more question about blades:  I generally tie the torque to the main throttle.  I tie the blade angle to two keys, usually translate forward/back (H&N).  However, I get clunky results when adjusting blade angle that way- sometimes I want to change 0.5 degrees, but often I end up with 2 or 3 degree changes.  Curious what method you guys use to adjust blade angle.

All of my planes had 8 small fan blades per engine. And during the testing for the latest entry I tried various combinations of fans, from 2 biggest ones to 16 smallest ones. At the end I determined during testing that 8 fan blades gives the best results: other fan blade configurations either didn't give any speed improvements or slowed down the plane altogether

As for the controls, I personally felt that it would be easier to control all elements through RCS controls, because this way you have one hand on directional (WASD) controls and the other hand on the RCS. Of course, in this challenge the Autopilot handled the flying, so I had both hands on the RCS. I also engaged Caps Lock for fine adjustments and after that I basically tried to tap the buttons as lightly as I could to get the fan blade angle for example: getting the 19.19 degrees was crucial to get the best combination of speed and average EC consumption. The RPM and Torque settings were adjusted by hand at first parts of the flight, but then I just typed in the necessary values once I was going at final cruise speed

16 minutes ago, 18Watt said:

Heh, yeah I suspect @camacju may not have the patience for a run of that length

I mean, the record flight alone took me over 8 hours, not even counting in all of the testing which in total probably took me longer than the flight itself. I have about 7 pages on my iPad filled with lots of data: plane configurations, RPM/Power/Prop angle settings, EC consumption, average speed, theoretical range from 95 EC charge (the 5 EC is basically as a safety margin for the takeoff, climb and acceleration) and so on. I don't think I did as much research for any other KSP mission lol 

[camacju]

3 hours ago, 18Watt said:

Just curious if anyone has tried just one blade?  Wouldn't make any sense in the real world, but I'm not sure how much KSP would model the vibration.

Interestingly enough, single blade propellers do exist in the real world:

But I suspect they'd be impractical in KSP due to the difficulty of balancing them.

3 hours ago, OJT said:

this is basically the limit of the current concept of the aircraft I've flown. Different RPM/Torque/Prop angle settings bring in diminishing results, different wing configurations didn't bring much either and the fuselage is already as streamlined and light as it could be

Have you tried further reducing motor size? A smaller motor would have less weight and draw less power. You'd need to have a higher torque limit, but that is likely a good thing - it gives you finer control over your overall torque.

Another thing you could do is angle your fairing upwards so you get some body lift. It'll be minimal but it will help reduce drag.