Around the World in 80... Minutes (ORIGINAL - DONE)

[Lt_Duckweed]

5 hours ago, Mars-Bound Hokie said:

Exactly, so depending solely on rockets to go around the world below the 35-km line won't do any good. What's worse, any depleted rocket boosters you have will be dead weight.

The rules explicitly state that you must take off and land in one piece, so no detaching rocket boosters when you're done with them (as tempting as it is to add SRBs). You're free to add rockets to your plane, but everything you take off with must also come back safely.

It takes very, very little liquid fuel to sustain velocity at 35km with a Nerv or two.  In Laie's earlier post, his craft had less than 10kn of drag and had enough fuel to fly on Nerv's for nearly 3 hours.

[nelso]

Not an official run (no command seat), just experimented with using nuclear engines and RAPIER.  57 minutes roughly.  I used the nuclear engine to jump out of thicker atmosphere repeatedly (up to 35k), thinking it would allow higher speeds.  The nuclear also helped quicken the initial ascent, as well as some spare oxidizer.  I'd turn off the nuclear when the RAPIERS turned back on, and flew it all manually.  The nuclear is tilted downward slightly, the theory being the plane could point forward and the nuclear would keep it up, but my assumption is probably wrong there.  More pics: https://imgur.com/a/ODKV0IH

Great job everyone else.

 

 

[Laie]

5 hours ago, nelso said:

The nuclear is tilted downward slightly, the theory being the plane could point forward and the nuclear would keep it up

It *does* help, if the thrust vector lines up with the CoM.Just see to it that the offset thrust doesn't create much torque: otherwise, the control forces needed to keep your plane pointed into the right direction will eat up all the benefits.

5 hours ago, nelso said:

I used the nuclear engine to jump out of thicker atmosphere repeatedly (up to 35k), thinking it would allow higher speeds. 

For that, you're rather slow. Are you afraid of the heat?

[Klapaucius]

Arg!!! Landed and then hit the wrong key and flipped the plane. But, I did touch down at 51:55--highest altitude 30,837, but mostly stayed between 15 and 20,000.  6 Kerbals had a fast trip around the planet before perishing nobly.

 

[nelso]

14 hours ago, Laie said:

It *does* help, if the thrust vector lines up with the CoM.Just see to it that the offset thrust doesn't create much torque: otherwise, the control forces needed to keep your plane pointed into the right direction will eat up all the benefits.

For that, you're rather slow. Are you afraid of the heat?

The heat was a problem, but mainly the craft was too heavy to keep the height up.  With a small redesign, I was able to get into "orbit" with help from rocket mode from the rapier engines.  This orbit was ~2166 m/s surface speed at ~34k, so to go any faster it would have to thrust downward to keep from going into space.  At this point I could use very little thrust to maintain speed.  Just based on kepler's laws, it should be a significant amount faster than kerbin orbit in space).

I did manage an actual run in about 35 minutes, though I got my altitude too high at some points, I overshot the runway by a lot, then sorta landed at 38 minutes, but this was mainly a test.  I got decent results by using aircraft autopilot, hold vertical speed at 0 and hold speed at 2166, and manually adjusting occasionally to keep under 35k altitude.  It also said I had about 3 hr - 1 day worth of fuel at this thrust level, so it can clearly get away with a lot less fuel.  The majority of the fuel usage was getting to to orbit in the first 5 minutes, exactly as logic would say - this is like launching and maintaining the orbit of a low satellite.

You can see me overshoot here

 

[vyznev]

I just wanted to leave this here. More detailed flight report coming later. Maybe. Sleep first.

Spoiler

 

Statistics summary:

• Launch mass: 8.485 tonnes (+ 1 kerbal in command seat, optional)
• Part count: 16
• Cost: 23,880
• Passengers: 1
• Flight duration: 35 minutes 38 seconds
• Highest altitude reached: 34.66 km ASL
• Fuel left in tank: 0.61 units

Basically, I thought my old 5 part LF only SSTO might be a good starting point for this challenge, so I tweaked it until I got something that worked and met the challenge requirements. The idea is to minimize drag and heat issues by having just a single stack of 1.25 m parts.

The craft climbs up to just under 35 km using the whiplash jet engine, then turns around and uses the nuke to build up and sustain orbital velocity at that altitude. The initial version had no aero surfaces at all, but I ended up including two advanced canards for more flight control and to stabilize the craft when flying backwards using the nuke (which this craft does for most of the flight; only the initial climb up to 30 km and the final glide back to the runway are done in the "forward" orientation). Takeoff and landing are both vertical (using parachutes for the latter) because, while this thing actually glides quite well and is surprisingly agile in the air, it has no landing gears.

Craft file here.

Flying it is quite tricky, not least due to flip-over maneuver needed to switch from jet to nuke propulsion. Said maneuver is really only practical above 30 km, and of course the challenge forbids going above 35 km. Also, the whiplash jet starts losing thrust above 15–20 km and flames out around 25 km, and you really want to have both the correct apoapsis altitude and a decent fraction of sideways orbital velocity built up by then. Oh, and while turning around, don't forget to reverse the control authority slider on the canards or the thing will be quite unstable indeed. And remember to do it again when turning back for the final glide.

During the cruise portion, you only need about 1.5% thrust on the nuke to counteract drag. I recommend dragging the thrust limiter on the nuke way down for finer control. You'll basically want to set SAS to retrograde hold (since you're flying backwards, of course) and control your orbit using the throttle. Only use the W/S keys if your climb rate is way off and needs to be corrected ASAP. A useful rule of thumb is to try to keep your apoapsis just under 35 km and about 1 minute ahead of your current position. Although once your orbit gets circular enough, the time to apoapsis kind of stops being meaningful.

Hotkey 1 toggles the jet engines and 2 toggles the nuke. Hotkey 3 opens and closes the cargo bay; it makes a decent airbrake when open, and you also must open it before deploying the chutes. Just don't try to slow down even faster by turning sideways at high speed with the cargo bay open, or this can happen:

Spoiler

So, yeah, it turns out that kerbals explode if they overheat. Who'd have known.

This particular flight was hardly optimal. In fact, it was originally meant to be just a practice flight, but it turns out I got a pretty good climb and ended up in a nice cruise orbit. Still, I could probably shave at least a minute off each end. The real challenge, I feel, is getting below 30 minutes. The orbital period around Kerbin at 34.5 km is about 28 minutes, so it's technically possible even without any downward thrust tricks. But there's not a lot of margin left for takeoff and landing.

Edited December 21, 2019 by vyznev

[nelso]

31 minutes ago, vyznev said:

Flight duration: 35 minutes 38 seconds

Nice!  That is a nice optimization to put everything inline and put the nuke backwards.

53 minutes ago, vyznev said:

The real challenge, I feel, is getting below 30 minutes

There is another potential time save, by going into retrograde orbit instead of prograde, so KSC gets closer to you while in flight instead of farther.  It means going ~340m/s faster relative to the air, and a harder ascent, but by rough calculations it would reduce time by 3 minutes.  It would use more fuel but that is hardly a concern since while in orbit, the only loss is drag and mass doesn't matter very much.

[RealKerbal3x]

@Mars-Bound Hokie I think I’ve got a good design for this challenge, but I’ve got a question.

My design uses two LFO tanks which have been switched to holding only LF by the Simple Fuel Switch mod. It uses no mod parts, so would this be within the rules or not?

If not, I can easily build a 100% stock version, it’s not a problem.

[nelso]

For anyone wondering how to do nuclear engines for this and approach the 33 minute range, here's a video of an ascent and getting auto throttle control.  The craft is kinda cheaty, with the nuclear engine inside the main fuel tank, and was very unstable when the fuel ran out.  The mechjeb autopilot can work for speed, but takes some babysitting to manage altitude.

 

[Nantares]

challenge accepted...here my attempt to run this mission (I will upload my plane to KerbalX if you want)

NA-59 SICRA (the Supersonic Intercontinental Reconnaissance Aircraft) inspired from Lockheed martin's X-59

not the fastest but it fun to fly

craft spec.

Mass : 30.621 ton

parts : 57

Height : 4.7 m

Width : 12.3 m

Length : 19.3 m

Cost : 86,554 BTH

Get the craft here

a mission video...

 

[Laie]

On 12/21/2019 at 4:12 AM, vyznev said:

light duration: 35 minutes 38 seconds

Whoopee! Congratulations!

On 12/21/2019 at 5:07 AM, nelso said:

There is another potential time save, by going into retrograde orbit instead of prograde, so KSC gets closer to you while in flight instead of farther.  It means going ~340m/s faster relative to the air, and a harder ascent, but by rough calculations it would reduce time by 3 minutes.

Interesting thought. You're still flying the same distance in atmosphere, so airspeed has to be the same in either direction. However, you lose that much orbital velocity and can go faster before you have to actively steer downwards. And on landing, you come in from the sea and don't have to worry about hills & mountains - that alone would be a selling point, as far as I'm concerned.

I haven't been idle, either:

Flying near orbital velocity requires so little fuel that you don't necessarily have to bring a nuke. However, going even faster comes at an ever higher cost, for pointing downwards. Lesson learned: heat-wise, a sustained 2600m/s are possible. Maybe 2700m/s. However, you also want to accelerate quickly at the beginning, and decelerate decisively near the end.

Check out this flick if you haven't thought about landing yet.

That was done using atmosphere autopilot, which, btw, has a great "fly straight and level" functionality. This time, I'm writing my own lander script in kRPC, which might give me an undue advantage. But it's fun!

[nelso]

My best with actual landing: 34:19.  I wrote a kOS script to maintain a near perfect flight for nuclear engine types, much better than Mechjeb was doing.  It sets throttle based on an estimate of orbital energy, and has a quite stable altitude control.  I only used half the fuel so it may need a redesign.

More in this album https://imgur.com/a/LPmim7w and video of kOS altitude maintenance and landing.  I went into a few intentional spins to slow down to land: 

 

1 hour ago, Laie said:

Flying near orbital velocity requires so little fuel that you don't necessarily have to bring a nuke. However, going even faster comes at an ever higher cost, for pointing downwards. Lesson learned: heat-wise, a sustained 2600m/s are possible

Similar thoughts here, and air speed appears to be the ultimate limit, so to get beyond that, as you said, is fast takeoff and landing.  Good job on 30 minutes.  I can post the kOS script as well if you want.

[vyznev]

4 hours ago, Laie said:

Interesting thought. You're still flying the same distance in atmosphere, so airspeed has to be the same in either direction. However, you lose that much orbital velocity and can go faster before you have to actively steer downwards. And on landing, you come in from the sea and don't have to worry about hills & mountains - that alone would be a selling point, as far as I'm concerned.

My experience with the run I posted above (and the various more or less successful practice landings before it) is actually that I probably started decelerating too soon, and would've saved some time by waiting until I was almost above the KSC before making a pretty steep descent. And I still cleared the mountains with plenty of room to spare. Of course, a perfect 90° turn just above the KSC isn't optimal either (even if your craft could do it), but you really do want to minimize the time spent flying (relatively) slowly through the lower atmosphere.

Anyway, I'll probably try a new flight in the retrograde direction, and with a steeper descent, at some point. My gut feeling is that I might be able to get below 30 minutes that way, but that'll require optimizing the ascent too.

Edited December 23, 2019 by vyznev

[Mars-Bound Hokie]

 ATTENTION ALL FLIERS 

 

First of all, I have to say that I am quite impressed with everyone's entries. I bet everyone got their much-needed fix of speed in this thread - and by speed, I mean going at over 2.5 the speed of sound.

 

Second of all, I decided to set a deadline for the challenge to February 20, 2020 at 11:59 pm Eastern Standard Time. If that seems unreasonable, don't hesitate to say something. 

 

Thirdly, I suppose some of you want to know who's leading in this contest. Here's the leaderboard.

 

CURRENT LEADERBOARD

1. LIGHTNING DIVISION
   1. @nelso - 34 minutes, 19 seconds
   2. @vyznev - 35 minutes, 38 seconds
   3. @Laie - 39 minutes, 47 seconds
   4. @Mars-Bound Hokie (ME) - 53 minutes, 18 seconds
   5. @Nantares - 57 minutes, 47 seconds
2. THUNDER DIVISION
   1.  
3. HEAVY RAIN DIVISION
   1.  

@Klapaucius was disqualified because his kerbals died after landing.

 

Am I missing anything? If I am, please let me know so I can ensure accurate results.

[nelso]

I gave it a test in reverse and got significant savings on fuel, approximately 40% less at cruising for the same surface speed (@ around 2560 m/s surface, circular orbit ~33km).  I was able to test this precisely with my kos script.  It does take stronger engines to get into orbit, since you need to go faster to start feeling weightless, but the fuel consumption is so much less for the cruise, it enables options like using higher power rocket engines for the ascent.

Question for more experienced players - do wings create any drag if they are headed straight in the prograde direction (any helpful threads would be appreciated)?

Below would've been my best run (crash landed at 32:00, and this included doing a 180 turn after takeoff), except I blew up the nose cone on descent.  It used an aerospike instead of nuclear because of better thrust, lower weight, and decent enough ISP.  During cruising I was thrusting about 8 kN.. with the same craft going prograde orbit, it was about 14kn.

[Laie]

5 hours ago, nelso said:

Question for more experienced players - do wings create any drag if they are headed straight in the prograde direction (any helpful threads would be appreciated)?

Of course they still have drag, but very little.

Generally, the best resource about drag, it's reduction, the relative drag of parts and whatnot:

Don't be fooled by the title, it goes way beyond Mk2 further down the page.

[nelso]

My best yet: 28:45. 43 tons starting mass. I went retrograde, used 2 nuclear engines and 4 rapier engines which I used for orbital insertion and brief boosts at takeoff, and I found rapiers to be the best for this purpose.  Instead of a nose cone at the front I used a fairing which has higher heat resistance for this critical part.  Mainly it was a combination of a good ascent and descent, and going retrograde which lessens lift needed to keep below 35k, allowing significantly higher speeds for the same thrust.  I only recorded a video, which does have an edit for an autosave towards the end.

 

22 hours ago, Laie said:

Don't be fooled by the title, it goes way beyond Mk2 further down the page.

Thanks, this was very helpful.

[Laie]

14 hours ago, nelso said:

My best yet: 28:45

 I think congratulations are in order.

I'm still working on my script: sometimes it works and sometimes it doesn't, even when saveloading from the same situation. Riddle me this. I'm certain to post another entry someday, hopefully today (this challenge has kept me occupied far more than is sane or healthy, time to get it over with), but I'm not confident that I'll be any faster than you.

 

 

[nelso]

 

11 hours ago, Laie said:

 I think congratulations are in order.

Thanks, this was fairly challenging and took quite some time (many failed approaches in ship design), although after a while I got a sense of how fast I needed to be going early in the run: 2650m/s at 3 minutes was my goal, with about 20 minutes of nuclear fuel remaining.  My goal was less about going faster during the cruise - every 50m/s faster is only about 30 seconds off the final time, but consumes a lot more fuel.

Also note this was 43 tons so in the higher mass category.  It was easier to balance numbers of engines, intakes, fuel/engine weight balance, and aerodynamics, and had a higher mass to drag ratio, so a quicker descent and presumably better aerodynamics in cruise. I think it's possible to shave off another 2 minutes using a powerful rocket for ascent, that's probably the 80+ ton category.

You said you were having trouble with kRPC.  I found kOS language awkward, but no reliability problems, just had to restart the script when loading a save.  Here is my script, it's a "PD" control of orbital energy (factoring in speed and altitude) and another PD control for altitude.  This already has the particular issues worked out like disabling SAS while using it, and giving back control to the player when exiting (by pressing any key).  It's configured to retrograde, but changing the 270 to 90 will allow prograde.

PRINT 1234.
SET gui TO GUI(200).
gui:SHOW().
LOCAL label IS gui:ADDLABEL("Low Orbit Autopilot").
SET label:STYLE:ALIGN TO "CENTER".
SET label:STYLE:HSTRETCH TO True.
gui:ADDLABEL("Altitude").
LOCAL field1 IS gui:ADDTEXTFIELD("33000").
gui:ADDLABEL("Orbital Speed").
LOCAL gui_speed IS gui:ADDTEXTFIELD("2360").


UNTIL SHIP:ALTITUDE > 4000000 {
    //PRINT SHIP:VELOCITY:ORBIT:MAG.
	SAS OFF.
	SET targheight TO field1:TEXT:TONUMBER(33000).
	SET targspeed TO gui_speed:TEXT:TONUMBER(2360).
	SET targthrottle TO (targspeed^2-SHIP:VELOCITY:ORBIT:MAG^2+(targheight-SHIP:ALTITUDE)*20)*.0005.
	LOCK THROTTLE TO targthrottle.
	SET pitch to MAX(-14,MIN(14,(targheight-SHIP:ALTITUDE-ship:VERTICALSPEED*80)*.01)).
	LOCK STEERING TO HEADING(270, pitch).
	//PRINT ship:VERTICALSPEED*50.
	PRINT throttle+" "+ pitch.
	//LOCK STEERING TO HEADING(90, 0).
	
	
	//PRINT SHIP:ALTITUDE.
    // pid:update() is given the input time and input and returns the output. gforce is the input.
	if terminal:input:haschar {
	  break.
	}
    WAIT 0.1.
}
SAS ON.
UNLOCK STEERING.
UNLOCK THROTTLE.
SET SHIP:CONTROL:PILOTMAINTHROTTLE TO 0.

 

 

[έķ νίĻĻάίή]

This is, harder then I expected 

Edited December 27, 2019 by έķ νίĻĻάίή

[Laie]

On 12/27/2019 at 2:55 AM, nelso said:

I found kOS language awkward, but no reliability problems

Nah, that's not the kind of reliability I'm talking about. The problem is that properly lining up with the runway takes time, which I don't have. Or, put differently: I'm employing methods that don't work too well when you're still going 500m/s a mere 5km from the runway.

That said... I managed 26:25, twice as fast as the flight that spawned this challenge:

(Video is unavailable in Japan, I gather. Let me know if that's a problem, I'll try to arrange something)

Crew once more sits in a Mk2 Lander Can inside the cargo bay. The craft file has the Lander Can removed, as that's the most offensive modded part. Everything else should load in stock KSP. There will be complaints about missing part modules, but it will load nonetheless. Then add your own lander can, and a ladder. If necessary I can provide pictures of how crew gets on and off.

Technical stuff: this one is limited to 3000m/s by aerodynamics. The nose cone settles at ~2520K, so there is room to fly faster still. It also needs 3 minutes to reach 2km/s... Somehow I have an inkling that 25 minutes could actually be possible. But not me, not today. It's essentially still the same craft that was meant to break 30 minutes flying east, I only turned it around. "HR" stands for "Heavy, Reversed".

 

Edited December 28, 2019 by Laie

[έķ νίĻĻάίή]

1 hour ago, Laie said:

Nah, that's not the kind of reliability I'm talking about. The problem is that properly lining up with the runway takes time, which I don't have. Or, put differently: I'm employing methods that don't work too well when you're still going 500m/s a mere 5km from the runway.

That said... I managed 26:25, twice as fast as the flight that spawned this challenge:

(Video is unavailable in Japan, I gather. Let me know if that's a problem, I'll try to arrange something)

Crew once more sits in a Mk2 Lander Can inside the cargo bay. The craft file has the Lander Can removed, as that's the most offensive modded part. Everything else should load in stock KSP. There will be complaints about missing part modules, but it will load nonetheless. Then add your own lander can, and a ladder. If necessary I can provide pictures of how crew gets on and off.

Technical stuff: this one is limited to 3000m/s by aerodynamics. The nose cone settles at ~2520K, so there is room to fly faster still. It also needs 3 minutes to reach 2km/s... Somehow I have an inkling that 25 minutes could actually be possible. But not me, not today. It's essentially still the same craft that was meant to break 30 minutes flying east, I only turned it around. "HR" stands for "Heavy, Reversed".

 

That. Looks. So. Cool.

[nelso]

1 hour ago, Laie said:

I managed 26:25

Very nice and congrats!  I guess you fly upside down because the drag from the tail fin helps rotate down, and optimized drag to maximize the use of the nerv.  Craft looks great too and it looks like you really tweaked it.  

Guess I will have to try some new approaches to get faster than that, but it will certainly be tough.

[Laie]

4 hours ago, nelso said:

you fly upside down because

...because the wings are slanted a little, so simply by turning it on the back I already have some negative lift. With some tweaking, this vessel could certainly reach the highest possible airspeed. However, determining the highest possible airspeed is going to take another day of experiments, and anyway, no amount of tweaking will make this the challenge winner (too slow to accelerate). A whole now design is necessary. I was eager to get something done *soon*, however, so I settled for today's highscore.

[Laie]

WIP:

5° wing slant. To make controls easier, the cockpit is angled the same.

At cruising speed and altitude, the plane has to be flown at about 10° AoA, which makes for a lot of drag from the fuselage. Slanting the wings mitigates this, but too much slant will create trouble during the more ordinary flight in the thicker atmosphere. You will have drag from an angled fuselage no matter what.

Mk2 parts are notoriously draggy at AoA, that's why I went with Mk3 on my last run. Unfortunately, that design didn's scale well to speeds above 3km/s: I had to pack ever more engines and the fuel to drive them and it quickly ballooned to 90 tons. Vertical Mk2 parts, however, have turned out to be very sleek even at AoA.

As a side benefit, Using cargo bays allows me to pack the fuel very tightly, and to finely tweak the amount of fuel and oxidizer on the vessel, so it runs out *just* as it reaches 3200m/s and then can settle into a Nuke-powered cruise. You cannot see the nuke, though, because it has been clipped into a Panther.

The Panther saves some 10-15 seconds in reaching Mach-2. After that, the Rapiers provide more than enough power: there is no measurable difference between using a Panther or a Whiplash for the purpose, other than the Ramjet being heavier.

However, there are serious control issues on the way down, the plane is barely stable in the yaw axis. That's 3x as much tailfin as I had on the Mk3 vessel, yet it is still not large enough. I also have yet to figure out how to decelerate safely, the nosecone goes poof more often than not.