Fastest Juno-powered aircraft

[swjr-swis]

15 minutes ago, neistridlar said:

I'd like to claim my second place on the probe core leader board: 813m/s, using only 7 junos.

That looks like it should be able to get there, but it's still being held back by the tiniest bit of body drag.

Try replacing the two basic fins by Elevon-4's (maybe even just a single one, centered), with Pitch/Yaw/Roll disabled, with just the bare minimum deployment deflection to keep it level when over 800m/s. The reaction wheel can handle attitude control, use the elevon(s) purely for the tiny lift needed to stay perfectly prograde.

It may sound silly to swap to a 'wing' that offers less than half the lift of what you use now, but being able to control the deployment angle 'live', starting with a bit more for take-off and subsonic then slowly tapering off as it accelerates, you can maintain the optimum lift to drag ration all the way up to flame-out.

 

31 minutes ago, neistridlar said:

And my second place for the manned category: 818.4m/s

I think you linked the wrong image. The speed read-out says 807.3 on that screenshot.

[Andetch]

41 minutes ago, neistridlar said:

.@Andetch If you want me to I'll beat you in manned level flight as well .

Nah  I'm pretty bruised up as it is, and you could probably be doing better things with your time! And I still claim mine is better as it is still a functional flying plane rather than a capsule being thrown at speed!

@Klapaucius (and also @neistridlar maybe CrazyJG needs to put a link to this in the OP of the new KEA) here is the link. Got some nice informative pictures for aircraft design.

 

Edited June 20, 2018 by Andetch

[Klapaucius]

29 minutes ago, Andetch said:

 

@Klapaucius (and also @neistridlar maybe CrazyJG needs to put a link to this in the OP of the new KEA) here is the link. Got some nice informative pictures for aircraft design.

 

That's just awesome. The whole thing about the KSP drag model was enlightening.   It explains a few things about why some of my designs, which I thought made aerodynamic sense, did not work well (and why other ones, that I figured would never fly, actually do).  I've noticed with seaplanes that the KSP model is somewhat deficient as well concerning speed and planing surfaces.   Have any of you used Ferram Aerospace?  I'm quite intrigued.

[swjr-swis]

3 hours ago, neistridlar said:

It flies almost straight up, then ditches the junos when they flame out. It reaches 37km altitude and free falls back down. In theory it should be able to reach 850m/s without air resistance from that height, so it might be possible to break the 820m/s limit this way.

850 m/s... that would mean having a velocity of 600 m/s when crossing the 18.5km mark on the way up (and assuming zero drag to get that exact same speed at that point on the way down). But the Juno flame-out speed at that altitude is much lower, closer to 500 m/s, if you manage to even get full thrust up to there and don't flame-out before. That maximum flame-out speed/moment is going to be the same regardless of how sleek the craft is: the limit is set by the Juno, not the terminal velocity.

Theoretically there is a bit of margin - using an experimental flame-out of 590 m/s at 17350m (all values rounded up), you could get a theoretical max in a zero drag environment of 830 m/s (rounded up) when crashing into the ocean. I don't think those 10 m/s are enough to counter the drag losses though.

All that said: it's an interesting approach and I'm curious to see if it pays off.

Edited June 21, 2018 by swjr-swis
saved too soon

[Klapaucius]

7 minutes ago, swjr-swis said:

Theoretically there is a bit of margin - using an experimental flame-out of 590 m/s at 17350m (all values rounded up), you could get a theoretical max in a zero drag environment of 830 m/s (rounded up) when crashing into the ocean. I don't think those 10 m/s are enough to counter the drag losses though.

If either of you have the Mission Builder, it would be interesting to plop the aircraft on Laythe and see how that changes thing...

[swjr-swis]

@neistridlar Some encouraging evidence, I hope:

In trying out your flight profile with a stripped down version of my JunoSpeeder-3c (no gear, elevons or command seat, launched vertically), I managed to capture a splashdown at 821.1 m/s. It was a split second too late and the craft is already destroyed in the screenshot, so it doesn't serve as proof for the leaderboard, but it proves you right: there is opportunity to side-step the Juno flame-out limit.

My craft did have the advantage over yours that it still has its engines and some spare fuel on the way down, and there is a brief section of the descent where the Junos do kick in (between 16-13km), which recovers some drag losses up to that point. Below that, free-fall velocity never dips below flame-out limit anymore so they are of no further help.

Still, the vertical flight profile can get to at least 821 (and perhaps even a bit higher). Game on!

Spoiler

Stripped all components that were not necessary to launch the JunoSpeeder-3c vertically (gear, wings, command seat).

Just short of 37km, probably due to steering losses to ensure that it would crash into the ocean (at least 70m extra free-fall distance).

The Junos do engage again briefly between 16 and 13 km, recovering some drag losses. It's worth it to keep at least some engines for the way down.

Snapped it a fraction of a second too late, craft already destroyed. But now we know it can be done: at least 821 m/s is possible.

 

[RealKerbal3x]

7 hours ago, neistridlar said:

I'd like to claim my second place on the probe core leader board: 813m/s, using only 7 junos.

And my second place for the manned category: 818.4m/s This was intended as a way to break the 820m/s limit, though I have not managed to do so yet.

This is what the craft looks like at launch:

It flies almost straight up, then ditches the junos when they flame out. It reaches 37km altitude and free falls back down. In theory it should be able to reach 850m/s without air resistance from that height, so it might be possible to break the 820m/s limit this way. And in case anyone wonders, the shock cone on the back is purely for drag reasons, it does not provide any meaningful amount of air to the engines.

@Andetch If you want me to I'll beat you in manned level flight as well .

Ooh, interesting how you aren’t directly using the junos to get up to speed, instead just getting to altitude and then becoming a gravity propelled bullet.

Your two lovely entries will be put on the leaderboard later since editing posts on mobile is flarping annoying

Edited June 21, 2018 by RealKerbal3x

[Andetch]

7 hours ago, Klapaucius said:

That's just awesome. The whole thing about the KSP drag model was enlightening.   It explains a few things about why some of my designs, which I thought made aerodynamic sense, did not work well (and why other ones, that I figured would never fly, actually do).  I've noticed with seaplanes that the KSP model is somewhat deficient as well concerning speed and planing surfaces.   Have any of you used Ferram Aerospace?  I'm quite intrigued.

FAR is a great mod, and I really must start using it. The reason I didn't keep using it after my initial attempts was that it changes the way control surfaces are programmed and at the time I was obsessed with playing with my Flying Car design (it is a car that handles just as well on the ground as it does in the air!) and although it still worked with FAR, I couldn't work out how to get the control surfaces to do what I wanted so I went back. But pretty much on every challenge involving sustained atmospheric flight someone asks if there is a FAR category -  I am surprised no one has came up with a FAR entry for this to be honest! 

Edited June 21, 2018 by Andetch

[neistridlar]

3 hours ago, swjr-swis said:

My craft did have the advantage over yours that it still has its engines and some spare fuel on the way down, and there is a brief section of the descent where the Junos do kick in (between 16-13km), which recovers some drag losses up to that point.

I did experiment with keeping the junos for that bit, and for my early attempts it did help a little bit. For the design I ended up with in the end though, I found the drag losses in the initial free fall to be bigger than the regain bit. And the 850m/s second figure was based on this formula: v=sqrt(2gh), which assumes conservation of mechanical energy, and also constant g, both of which are wrong of course, But I would think that it is not too far off.

@Andetch @Klapaucius The drag model stuff in Keptins tutorial there is outdated, and has been since 1.0. What swjr-swis has posted here is how the new model works. For those interested in the inner workings of it I have written a little piece, it is mostly aimed at modders, but it might be interesting non the less. https://docs.google.com/document/d/1v41jJEl52_WYnGgx2ct1hxPwB4Co_88EeOPg1xtTx5c/edit?usp=sharing 

[Andetch]

5 minutes ago, neistridlar said:

 

@Andetch @Klapaucius The drag model stuff in Keptins tutorial there is outdated, and has been since 1.0. What swjr-swis has posted here is how the new model works. For those interested in the inner workings of it I have written a little piece, it is mostly aimed at modders, but it might be interesting non the less. https://docs.google.com/document/d/1v41jJEl52_WYnGgx2ct1hxPwB4Co_88EeOPg1xtTx5c/edit?usp=sharing 

I figured the drag model stuff was old, but the pretty pictures of COM/COT/COL are nice and still relevant, as is the wing stuff (I think?) Either way the pictures are nice and pretty.

[Klapaucius]

@neistridlar

Thanks for that. So, if I read that correctly, it recalculates drag somewhat if a part is attached?  Eg, I attach an intake to an engine, theoretically the attached engine face should not count for any drag. Did I get that right?  But that also means the game won't account for any parts sunk into others, say an engine sunk into a wing section is still going to have a full drag calculation.  Am I at all close here?

[neistridlar]

5 minutes ago, Klapaucius said:

@neistridlar

Thanks for that. So, if I read that correctly, it recalculates drag somewhat if a part is attached?  Eg, I attach an intake to an engine, theoretically the attached engine face should not count for any drag. Did I get that right?  But that also means the game won't account for any parts sunk into others, say an engine sunk into a wing section is still going to have a full drag calculation.  Am I at all close here?

Pretty much spot on. If you stack attach the intake to the engine,  the front face of the engine receives no drag, assuming it is the same size (or bigger) as the intake, and the rear face of the intake also receives no drag, with the same assumptions.

10 minutes ago, Andetch said:

I figured the drag model stuff was old, but the pretty pictures of COM/COT/COL are nice and still relevant, as is the wing stuff (I think?) Either way the pictures are nice and pretty.

Yes, the other stuff is pretty much good. There are some minor nuances that I am not entirely sure is absolutely correct, but the big picture at lest is very much relevant.

[TheFlyingKerman]

It seems one needs a lot of mass/cross section area to take advantage of gravity. Also I am a little skeptical of those bundles of Junos. I managed 820m/s with a Juno rocket, using one engine.

Edited June 22, 2018 by TheFlyingKerman

[swjr-swis]

4 hours ago, TheFlyingKerman said:

I managed 820m/s with a Juno rocket, using one engine.

 

Post that single-engine rocket! It deserves the nr 1 spot for using minimal engines, even if at the same speed. Less engines should definitely get a bonus.

I know the craft I entered can do it with less engines, it is over-powered for the purpose - it just made a visually pleasing arrangement.

[swjr-swis]

5 hours ago, neistridlar said:

I found the drag losses in the initial free fall to be bigger than the regain bit. And the 850m/s second figure was based on this formula: v=sqrt(2gh), which assumes conservation of mechanical energy, and also constant g, both of which are wrong of course, But I would think that it is not too far off.

My calculation was based on conservation of energy too. Since you dump the engines at flame-out on the way up and the rest of the trajectory is unpowered, the speed at that point is the absolute maximum you can hope to have when crossing that same altitude on the way down, assuming zero drag losses during the arc and perfect transformations from kinetic to potential and back again. We don't even need to know how high you can get from there, it's already set by the flame-out point.

Then it's just a matter of calculating how much gravity can add to the speed from the flame-out altitude to sea level , again assuming zero drag loss (v-final squared = v-initial squared times 2g times altitude). That is the limit.

Unless I got my physics messed up. I was trying to do math at 4am again, so it's possible.

[Val]

Manned level flight with 5 Juno's = 777 m/s

Craft

[neistridlar]

6 hours ago, swjr-swis said:

My calculation was based on conservation of energy too. Since you dump the engines at flame-out on the way up and the rest of the trajectory is unpowered, the speed at that point is the absolute maximum you can hope to have when crossing that same altitude on the way down, assuming zero drag losses during the arc and perfect transformations from kinetic to potential and back again. We don't even need to know how high you can get from there, it's already set by the flame-out point.

Then it's just a matter of calculating how much gravity can add to the speed from the flame-out altitude to sea level , again assuming zero drag loss (v-final squared = v-initial squared times 2g times altitude). That is the limit.

Unless I got my physics messed up. I was trying to do math at 4am again, so it's possible.

I ran your numbers as well, the calculation is correct. It is just that somehow I have gotten more energy out of the climb than you assumed in your calculation. Your assumption puts the max altitude at 35km. 

I reran the craft, and it seems I ditch the engines at 15km already, and I got just above 37km altitude. I think that is easier to measure, since the moment does not fly past so fast.

[Aetharan]

Well, I know that I'm not exactly setting any records with this attempt, but here we go:

https://imgur.com/a/n9Bfof5

607 m/s recorded in level flight, with a pilot and two passengers, in a quad-Juno-powered equatorial circumnavigation.

[TheFlyingKerman]

19 hours ago, swjr-swis said:

 

Post that single-engine rocket! It deserves the nr 1 spot for using minimal engines, even if at the same speed. Less engines should definitely get a bonus.

I know the craft I entered can do it with less engines, it is over-powered for the purpose - it just made a visually pleasing arrangement.

I did not save a copy of this craft . It is WYSIWYG: an intake, two Oscar B tanks, a reaction wheel, an OTO2 and the engine. I forgot how much fuel I put into the tanks. The tricky part is the flight profile. You launch it like a rocket, do a gravity turn, and then fly prograde. It seems the optimal AP is about 6000m, otherwise the gravity turn is hit or miss.

[Aetharan]

My second attempt went slightly better, managing a successful landing after circumnavigation under the same self-imposed rules: 2 passengers, no dive-bombing for speed, and complete circumnavigation to demonstrate usefulness as a light transport.

https://imgur.com/a/nhnlD9S

Top speed is 630 m/s, and this time, I have a craft file if anybody wants to fly it.  https://kerbalx.com/aetharan/AAC-630-Leer

[Hotel26]

On 6/19/2018 at 4:12 PM, swjr-swis said:

Note: there is a difference between regular clipping (placing parts as the game normally allows, then offsetting them into each other) and 'part-clipping' (a very confusing term for stack-attaching multiple parts to the same node, which is only possible after enabling a cheat menu option).

I live for the moments when I read stuff like this!!

I've been attaching octa-struts to get additional attachment points.  I'd rather enable this 'cheat', attach more parts, then move them away to where I want them.  (Here's a very recent example: I wanted to embed a Juno with airtake orthogonally into a wing to use for a VTOL design.)

Beautiful, thanks.

[Aetharan]

2 minutes ago, Hotel26 said:

I live for the moments when I read stuff like this!!

I've been attaching octa-struts to get additional attachment points.  I'd rather enable this 'cheat', attach more parts, then move them away to where I want them.  (Here's a very recent example: I wanted to embed a Juno with airtake orthogonally into a wing to use for a VTOL design.)

Beautiful, thanks.

It's worth noting that, while useful, this method does involve using the cheat-menu, and many challenges' authors and participants would consider it dishonest and/or disqualifying to use that type of clipping in an entry.  Personally, even with the "place normally and offset" type of clipping, I strive to minimize it so that functional elements don't overlap.  (For instance, the structural exterior of two tanks could share space, but not what is obviously meant to be the pressurized-tank portions of their models, despite the fact that it makes no aerodynamic difference unless you're cheating them into the same stack.)

[Klapaucius]

10 hours ago, Aetharan said:

My second attempt went slightly better, managing a successful landing after circumnavigation under the same self-imposed rules: 2 passengers, no dive-bombing for speed, and complete circumnavigation to demonstrate usefulness as a light transport.

https://imgur.com/a/nhnlD9S

Top speed is 630 m/s, and this time, I have a craft file if anybody wants to fly it.  https://kerbalx.com/aetharan/AAC-630-Leer

It seems, like in real life record setting or racing, the challenge helped develop a viable mainstream aircraft. You've created a reasonably fast, lightweight, efficient, long-range executive jet.  You may not win the prize, but you will win the mass production contract  

[Hotel26]

12 hours ago, Aetharan said:

"place normally and offset"

Yes, for context, I'm merely an observer in this topic; not a competitor.  (An enormous amount can be learned from these competitions that is useful and practical.)  Secondly, the offset gizmo is indispensable in constructing airplanes for attaching everything to the fuselage to prevent wing root departure.  Thirdly, I generally don't clip at all (much), but need to attach multiple items: my first two years in the VAB were severely handicapped because I only knew how to utilize the standard node placement offered by those editors.

[RealKerbal3x]

10 hours ago, Klapaucius said:

It seems, like in real life record setting or racing, the challenge helped develop a viable mainstream aircraft. You've created a reasonably fast, lightweight, efficient, long-range executive jet.  You may not win the prize, but you will win the mass production contract  

@Aetharan you could probably enter that aircraft in the KEA thread!