Tylo Landing Tips?

[boccelounge]

I'm getting ready to do one of the few things I've still never done in KSP: land on, and return from, Tylo.

Anybody got any tips?

-----

A few notes:

- I'm playing 1.22.  I'll be upgrading to 1.4.x in a week or two, and looking forward to it, but I'm having a great time with this latest Career save, and want to wrap up a few more accomplishments before I move to a clean, new re-install (i.e. I won't attempt to migrate any of my old saves).

- I have a few mods, including the "informational" ones, like KER, MJ, BetterOrbits, etc., and I have a bunch of "welding"/.cfg tweaks I've made myself, and a few part-adding mods... but all of this is for aesthetic and/or part-count-reduction purposes.  That is, my game is essentially stockalike, for purposes of buidling this lander (no magic Torch or Fusion drives, e.g.).

- I'll be using (I hope) a re-usable, single-stage LV-N "Nerv" powered lander, designed and used exclusively for Tylo.  It will refuel on orbit, and hit all of the moon's surface biomes, one landing at a time.

- I almost have the refueling infrastructure set up around Jool; still waiting for a few more elements to arrive.  That is to say, I'm not in a hurry; I have quite a few other things to finish up in various places (namely Dres and Duna) before I can really focus on Tylo.

- I don't normally like using external command seats for anything but rovers, but I may consider it for this.  I need (by my own requirements) to land (at least) three Kerbals at a time.

- I'd rather not re-fuel on the surface, but I'm not completely opposed to it.

- Funds are not an issue; we've long since cleared the Tech Tree and the program is steadily profitable; I'm currently sitting on 45 million funds (and growing), with relatively little to spend it on.

- I think I have a handle on dV requirements; I'll be shooting for around 5600 m/s from low orbit (10-20km).  Does that seem way off?

- I normally like a high TWR for airless bodies, but for obvious mass reasons I don't want to over-do it.  Does ~2.5 TWR, fully-fueled, seem right?  I'd really like TWR to be as high as possible practical for this.

And... that's about it.  I'm very interested to see what you folks have to add/suggest on the things I've mentioned, and on anything I may be overlooking.  Your comments gratefully anticipated...

-----

EDIT:

This post is probably already too long, but to sum up my most pressing questions:

1. What dV figure should I shoot for?

2. What TWR should I shoot for?

3. What initial orbital altitude will optimize fuel use vs. landing precision/convenience/safety?  I only need to hit biomes, not specific surface targets (unless I find a bunch of anomalies... but that would probably be a follow-up mission).

4. Am I foolish to think I can do this with a single stage?  I could just design a re-usable Ascent stage, and ship out a bunch of disposable Descent stages (one for each biome).

Edited April 1, 2018 by boccelounge

[Gargamel]

39 minutes ago, boccelounge said:

1. What dV figure should I shoot for?

While my initial reaction is don't, just don't :), I have never landed there myself.   But I can help you with this little bit.  Use the subway map.

I keep a copy handy so I can reference it easily. 

[boccelounge]

Just now, Gargamel said:

But I can help you with this little bit.  Use the subway map.

Thanks-- and I did!

I keep it on my desktop, and it's sitting open next to this window as I type this.  That's where I got the idea that 5600m/s might work for me.  I find I always need more than the Subway Map suggests (for airless bodies).  It tells me I can descend/ascend on Mun for <1200m/s, but I find I need more like 1800-2400 to feel safe (depending on the craft, of course).

So I should have been more specific: what's the "real" (i.e. safe and practical) dV goal for Tylo?

 

And I don't care what you say, man... I'm doing this!  Jeb's getting tired of ferrying tourists to Ike (which is very oddly popular to Kerbals in this save).

[-Velocity-]

4 hours ago, boccelounge said:

Thanks-- and I did!

I keep it on my desktop, and it's sitting open next to this window as I type this.  That's where I got the idea that 5600m/s might work for me.  I find I always need more than the Subway Map suggests (for airless bodies).  It tells me I can descend/ascend on Mun for <1200m/s, but I find I need more like 1800-2400 to feel safe (depending on the craft, of course).

So I should have been more specific: what's the "real" (i.e. safe and practical) dV goal for Tylo?

 

And I don't care what you say, man... I'm doing this!  Jeb's getting tired of ferrying tourists to Ike (which is very oddly popular to Kerbals in this save).

If you can do a reasonable approximation of a suicide burn, 6000 m/s works well as a "safe" value.  "Safe" and "suicide" don't really belong in the same sentence, but it is what it is.  If you don't use quick saves then don't get too attached to the Kerbals you're going to try sending down there.

[Human Person]

High TWR. If you have to start your landing burn very early, you are going to have a harder time getting your landing approach right as you want to arrest your horizontal velocity close to the ground (everything else will cut deep into your delta v budget, especially on Tylo). The final touchdown is also pretty tricky because of the high gravity. you can practice that by doing powered landings on Kerbin.

Edited April 1, 2018 by Human Person

[boccelounge]

2 hours ago, -Velocity- said:

If you can do a reasonable approximation of a suicide burn, 6000 m/s works well as a "safe" value.

 

25 minutes ago, Human Person said:

High TWR.

 

Thanks, @-Velocity- and thanks, @Human Person.

But you both get to the crux of my question without really answering it: how much does Tylo favor TWR over dV versus dV over TWR?

More specifically, what is a good TWR for descent and ascent on Tylo?

What does "high TWR" mean to you?  2.5?  5?  10?

 

I'm currently roughing out a design in the VAB.  It would have about this performance (these are very rough figures; I haven't added all the science and attitude-control parts):

- Fully fueled for Descent from low orbit: 6500dV / 2.4TWR

Assuming 3000m/s dV for landing (the Subway Map suggests 2270), it would have:

- On Ascent, 3700dV . 3.5 TWR

I -think- this basic design will work, but I know it won't be quite this efficient; there is more mass, more piloting errors corrections, more cross-range/inclination changes, etc. to work into the design.  And I always increase my "safety" (read: fuel) margins when I have Kerbals on-board.

Does that sound like enough (or close to enough) TWR to effectively use the nominal dV?

-----

Thanks for the help, past and future.  This really is a great game that is only exceeded by its amazing community of fellow players.

[Kryxal]

This sounds safe, your TWR means you can always have a good 85% of your delta-v directed to changing horizontal velocity.  I'd even say it might be more than you need, and the throttle's going to be tricky on landing.  If you have 3+ engines, consider setting a toggle to turn some off for a "landing mode"  For that matter, since you're using nukes, how's ground clearance?

[ElWanderer]

I wrote an Excel spreadsheet to calculate how TWR would affect delta-v and burn distance, when performing a zero-descent burn. My designs with nuclear engines tend to end up with such poor TWR that the extra delta-v and distance required shoots up.

https://github.com/ElWanderer/kOS_scripts/tree/v1.1_draft_changes/misc

Note that the spreadsheet doesn't include calculation of a deorbit burn (50m/s roughly for Tylo) or the final suicide/landing burn (allow 250m/s, I guess, though the more the merrier if doing it manually).

My kOS lander script has made it down on target using about 2500m/s, but it isn't consistent and sometimes wants 3200m/s+

[blakemw]

One of my hobbies is making Tylo Hoppers: A ship capable of landing on Tylo, refueling using onboard ISRU, taking off again and landing again. In other words the tanks have enough fuel to do an ascent and descent without refueling in orbit. This is borderline possible with chemical fuel for a rocket consisting of a vector engine and large amounts of fuel, so if you don't need single-stage-to-orbit-and-back capabilities you actually have a fair amount of margin and you're not completely crazy wanting SSTOAB capabilities.

For my SSTOAB designs I do not use LV-N's, I mainly use Vectors and just lots of LF/Ox, maybe an Aerospike or Poodle for when most the fuel mass has been burned off. Since you need pretty extreme deltaV the key is performing a good constant altitude burn. You don't need much TWR when landing, about 1.8 is enough. With my designs, they take about 2600m/s to get into orbit when fully laden, and about 2400m/s to land from orbit, when most the fuel has been burned off. If I were to use a mostly LV-N powered design I'd probably still have LF/Ox engines for the landing and takeoff to minimize the weight of LV-Ns, they can still perform most the of the ascent and descent burns (maybe TWR of 0.4-0.5 on LV-N's, about 2.0 for the LF/Ox engines) but more deltaV would definitely be required since the drymass is much, much higher when using LV-N's. Actually LV-N's are pretty horrendous when you need TWR so I'd be tempted to not use them at all, and just ramp up the fuel haulage to tylo orbit, I assume it's coming from Pol or Bop and big tankers can be landed there.

I haven't tried making the orbital refueling version, which is obviously different because you can't consume propellant in quite such volumes, but if SSTOAB is doable SSTSAB should be too.

 

EDIT: Just tried a SSTSAB, using a Vector LF/Ox design:

dV in orbit when fully fueled: 5399m/s w/ Tylo Surface TWR of 1.17

dV on surface: 2851m/s (2548m/s used for descent)

dV in orbit after ascent: 527m/s (2324m/s used for ascent)

I felt it had too much deltaV (my feeling was that 5000m/s would be enough) and was gratified to have quite a bit leftover once in orbit. The trick really is a Vector can do it easily if the fuel fraction is really high.

Edited April 1, 2018 by blakemw

[GoSlash27]

I don't think you're going to successfully operate SSTO on Tylo with LV-Ns. The absolute minimum DV (4,540 m/sec) and initial t/w of 1 means you'll need 2 LV-Ns for each tonne of payload. If you add any T/W or DV (and you really need more of both) it quickly spirals into the "impossible" range.

I'd consider 5800 m/sec the minimum "safe" DV for a Tylo SS lander without ISRU. 4 aerospikes per tonne payload would get the job done at 1.3G minimum, which is pretty weak for a lander. If you start bumping up the Gs, it rapidly becomes impossible for any engine.

Tylo SSTO lander return designs, 1t payload, 5800 m/sec DV

local g/ number of engines/engine type/Mass/fuel mass

1.3/ 4/ Aerospike/ 68.9t/ 63.9t
1.4/ 1/ Vector/ 85.0t/ 81.0t
1.8/ 1/ Rhino/ 138t/ 128t

For higher t/w, you'd have to add Rhinos and fuel. You can see how the fuel demands increase rapidly.

HTHs,
-Slashy

 

[Human Person]

I didn‘t mean to go overboard on the TWR. @GoSlash27 ‚s example is reasonable.

Most important off all: practice. You can use the debug menu to spawn a vessel in Tylo orbit and just go for it. Low TWR landers require more delta v and more piloting skills, that‘s why I said go for more thrust but maybe you like the extra challenge. final touchdown is hard either way, so practice especially that (as I said, you can practise that on Kerbin too).

You don‘t need extremely high TWR for ascent, 1.2 or greater will get the job done (remember it‘s basically the same as a Kerbin ascent, just without atmosphere so any rocket that can launch from Kerbin will be more than enough to ascent from Tylo)

Edited April 1, 2018 by Human Person

[GoSlash27]

1 hour ago, Human Person said:

I didn‘t mean to go overboard on the TWR. @GoSlash27 ‚s example is reasonable.

Human Person,

 Actually, I prefer a higher initial t/w myself. Makes it easier to do a pinpoint landing without the cosine losses. It's just that Tylo makes it very difficult and expensive.

If I were to do a fully reusable Tylo lander, I'd definitely go ISRU.

Best,
-Slashy

[blakemw]

2 hours ago, Human Person said:

You don‘t need extremely high TWR for ascent, 1.2 or greater will get the job done (remember it‘s basically the same as a Kerbin ascent, just without atmosphere so any rocket that can launch from Kerbin will be more than enough to ascent from Tylo)

 

One huge caveat is while it's possible to get by with a TWR of 1.2 on Tylo, you're going to get some pretty horrendous gravity losses on the descent and ascent and you DO NOT want those gravity losses on Tylo!

Now on *Kerbin* we tolerate those gravity losses because we don't have to launch the fuel to Kerbin's surface! The fuel you are burning at the launch pad has been delivered through a grand total deltaV of 0m/s, fuel is cheap at the launch pad. That's why we're even happy to use crappy, heavy low ISP solid fuel, because we didn't have to lift it anywhere.

Now, if on Kerbin we want to build the lightest rocket to launch a given tonnage (rather than the cheapest rocket), we actually go for a much higher TWR, probably at least as high as 2.0 and maybe even higher (and we use streamlining so we can ignore the atmosphere). This eliminates a great deal of gravity losses and the rocket will consume significantly less fuel - maybe half as much. It will be more expensive because it is engine heavy and engines are expensive, but it'll be very much lighter.

So returning to Tylo, in this case we've had to deliver the lander to *frigging Tylo*, through a deltaV of oh, 6000m/s. We had to deliver the fuel to frigging Tylo orbit, the fuel for ascent further had to be delivered to friggin Tylo's surface. We're going to make that lander as light as we can! And that means reducing fuel tonnage by going engine-rich and largely eliminating gravity losses.

There is more scope for low TWR when launching from Tylo using fuel gathered using ISRU on the surface, sure in that case the fuel isn't being lifted from anywhere, but in particular, when using orbital fuel you'd want a TWR at touchdown/liftoff in the range of 1.8-2.5. This is actually true whether it's a single stage or a separate ascent stage.

Edited April 1, 2018 by blakemw

[boccelounge]

Hey, thanks all for such thoughtful responses.  And Happy Easter, to those that observe it.

I sat down this afternoon to test some designs, and quickly realized nukes weren't going to work.  My late-night estimations were way off-- instead of firing up the game, I used an old spreadsheet (that I made when I got the demo, and had no access to KER).  For reasons I can only describe as stupidity sleep deprivation I assumed LV-Ns had 120 thrust and massed 1 ton... and that is of course not the case.

Anyway, I'm trying out a LFO "stage-and-a-half" design.  I'm actually landing now, in a Sandbox test save.  I'll try to post pictures later tonight.

But even from my playing around in the VAB I'm coming to the realization that a reusable (or partially reusable) design that can hit all the biomes is probably not going to happen.  I could do it, but the logistics and infrastructure needed to place and refuel so many boosters so many times would be, well... not enough fun to be worth the effort.

So I think the mission parameters have changed.  We'll try for 1-3 landings, depending on what the ScanSat/Anomaly orbiters find.  I have a small flotilla of orbital research probes about 40-50 days out from Jool, so in maybe 80 days or so I hope I can identify a small number of interesting landing sites.  I'll let you know when I do this "for real."

Thanks again for the help; you guys really got me to sharpen and refine my goals and methods on this.

Edited April 2, 2018 by boccelounge

[GoSlash27]

Boccelounge,

 Good luck and remember: You *can* still do a fully reusable Tylo lander that's not obnoxiously huge and expensive if you use ISRU.

Best,
-Slashy

[boccelounge]

2 hours ago, GoSlash27 said:

 Good luck and remember: You *can* still do a fully reusable Tylo lander that's not obnoxiously huge and expensive if you use ISRU.

Yeah... I know.  I just have an abiding distaste for hauling drills and converters around.  I like them permanently on the ground, or permanently in orbit.  I'm not against the idea though.

However... I got the second test vehicle to work, and it was (surprisingly) a very simple and easy design.  It's also exceedingly ugly, but this is just a prototype...

In the VAB:

 

Thinking that if I were going for fewer landings, each one would need to count, so I aimed for 0N/0W  from a (IIRC) 70km orbit.  I ended up with a longer braking burn than I'd have liked, with all four engines burning at less than full throttle-- in the production version, I'll toggle the Aerospikes so we can brake/adjust plane with just the Poodles.  That should squeeze a little more fuel efficiency out of it.

 

And on the surface:

 

I did the descent/ascent once manually, then let MechJeb have a try.  We both got back to a 50km orbit easily.  I had just less than 200m/s in the tanks; MJ had over 360.  Show-off.

 

I didn't grab a shot at booster separation (I was rather focused on the dV and fuel levels), but in both landings the (onion-staged) side boosters ran dry just under 1000m from the surface.  MJ didn't even freak out when I manually staged.

-----

So, some conclusions/notes:

- The final version won't look exactly like this.  I like to "build" my own tanks by welding small radial tanks, radiator panels, etc.  But it will have the same mass, form factor, fuel capacity, etc.  Plus Science!, ladders, etc.

- I'll probably add a disposable "nose tank" to help with inclination/altitude changes, and to add a little more margin of error.  I'm thinking 400 LFO ought to do the trick if I turn off the Aerospikes for orbital maneuvers.

- Instead of the big radial decouplers, I'll use clamp-o-trons.  That way the central stack can be re-used.  I'm tentatively looking at four landings (one for each of the Big Four), so I'll have to send 4 pairs of boosters out to Jool, along with some little assembly/docking tugs, etc.  That'll be a fun exercise in logistics planning (no sarcasm; that really will be enjoyable).

- A cool (if unintended) finding is that this thing could probably ascend from Laythe, if I add a nose tank with parachutes and de-orbit fuel.  So that's awesome-- I really dislike one-use or disposable vehicles.  Almost every payload I send up these days is meant to be multi-role and multi-use. (Note: I'm talking about the central stack, not the side-boosters.)

I think this is going to work well.  I'm of course open to any further suggestions or refinements.  Thanks again!

Edited April 2, 2018 by boccelounge

[DerekL1963]

2 hours ago, boccelounge said:

A cool (if unintended) finding is that this thing could probably ascend from Laythe, if I add a nose tank with parachutes and de-orbit fuel.  So that's awesome-- I really dislike one-use or disposable vehicles.  Almost every payload I send up these days is meant to be multi-role and multi-use. (Note: I'm talking about the central stack, not the side-boosters.)

When I did my Jool-5 mission, that's more or less exactly what I did.  One lander of essentially the same design for Laythe, a second lander with a design much like yours for everywhere else.  In Tylo mode (it's first use) it had a crasher for the deorbit burn then descended on a core and two radials. On mine however, the radials landed to avoid a disastrous last minute drop in t/w ratio.  They burned out and were jettisoned about thirty seconds into the ascent.  A tanker then re-fueled it for future use.

[boccelounge]

On 4/2/2018 at 2:11 AM, DerekL1963 said:

When I did my Jool-5 mission, that's more or less exactly what I did.  One lander of essentially the same design for Laythe, a second lander with a design much like yours for everywhere else.  In Tylo mode (it's first use) it had a crasher for the deorbit burn then descended on a core and two radials. On mine however, the radials landed to avoid a disastrous last minute drop in t/w ratio.  They burned out and were jettisoned about thirty seconds into the ascent.  A tanker then re-fueled it for future use.

I was worried about the same thing-- "a disastrous last minute drop in t/w ratio."  I avoided it by putting the higher-thrust Aerospikes on the main stage. That, I think, was the decision that made it work for me.  It goes against my normal thinking in rocket design-- low TWR, high isp on the final stage.  This comes of course from many launches from Kerbin's surface.  Tylo forces you to think "backwards," as it were.  And then I started thinking how well Aeros would work for a Laythe ascent...  and I think I'm pretty close to dual-use vehicle that (with orbital refueling) will work for both moons.

Now... I'm starting to ponder Vall.  Looks like this might be a "Jool-5" after all.

[blakemw]

My very simple philosophy for Tylo, Eve and other high gravity worlds is that you're almost always better off with more thrust. 

When dealing with microgravity ditching weight as you go along is very effective at extending deltaV, but this works terribly in a strong gravity field if you're ditching engines as you go along. If you can work out a way to ditch empty tanks, that's fine, but often you're better off keeping the engines to mantain a higher thrust. Actually in practice the best thing to do tends to be to just use a powerful engine, like instead of doing asparagus stage with 5 aerospikes (900kN, decreasing), instead you just have a single Vector and maintain that 1000kN of thrust at all stages. On that evil planet Eve, a single-vector design beats the pants off of aerospike asparagus (if the vessel is streamlined enough to be able to largely ignore the atmosphere), it's like, half or as third as heavy for the same payload to orbit. Even on Kerbin, often it's better to keep engines than to ditch them.

Also I managed to find a pretty concrete answer to how much TWR an ascent stage should have to minimize overall lander weight. The Apollo Lunar Ascent Module had a TWR (local lunar gravity) of 2.12, and I bet that was seriously number-crunched. This number ought to be around an optimal for minimizing gravity drag while not carrying too much weight of engine and it should be reasonably independent of actual gravity strength. It comes pretty close to the middle of my experience guestimate of 1.8 to 2.5. (But fun fact is the deltaV from Lunar surface to Lunar orbit is about the same as from Tylo surface to Tylo orbit: 1870 vs 2400 or so).

 

Edited April 3, 2018 by blakemw

[ElWanderer]

33 minutes ago, blakemw said:

The Apollo Lunar Ascent Module had a TWR (local lunar gravity) of 2.12, and I bet that was seriously number-crunched. This number ought to be around an optimal for minimizing gravity drag while not carrying too much weight of engine 

I remember reading threads where people ran some of the numbers on this, and the conclusion generally seemed to be that beyond a TWR of 2.5, the delta-v savings were tiny.

[GoSlash27]

ElWanderer,

 

2 hours ago, ElWanderer said:

I remember reading threads where people ran some of the numbers on this, and the conclusion generally seemed to be that beyond a TWR of 2.5, the delta-v savings were tiny.

Yeah, I was part of that. For the ascent to orbit, the most mass- efficient vehicles operated at surprisingly low TWR. This was due to a) diminishing returns due to increased engine mass and b) piloting error increased with shorter burn times. Ideal t/w for most engines was in the 1.2 to 1.7 range. A lot of the engine specs have changed since then, though.

*However*... this only applied to ascents, not descents. Descents are different because precision in the point of touchdown and surrounding terrain may or may not be a concern, and this gives rise to multiple potential approach modes. It's easy to get away with low t/w on a constant altitude approach, because it's essentially a launch in reverse. But terrain and precision dictate a steeper approach, which requires more t/w. Generally, 2G initial t/w is adequate for a precision landing.

Best,
-Slashy

Edited April 3, 2018 by GoSlash27

[ElWanderer]

2 hours ago, GoSlash27 said:

*However*... this only applied to ascents, not descents.

Yes, I could've been clearer about that (though the post I was quoting did reference the ascent stage).

For descent, you can get away with fairly low thrust to weight ratios, but if you're flying manaully it's a rather risky proposition. It is comforting to have a high TWR in case you need it.

[boccelounge]

8 hours ago, GoSlash27 said:

Ideal t/w for most engines was in the 1.2 to 1.7 range... this only applied to ascents, not descents. Descents are different... Generally, 2G initial t/w is adequate for a precision landing.

This is really the question I started with, and though it took a bit of a meandering path to get to the answer...  thanks!

And it makes me wonder if how much I can improve the design.  In the first picture above, you can see there's 2400 LFO on the central stack, 3200 on the boosters.  I wonder if I could reverse that...  it would mean, what, a reduction of 9 tons per landing.  Nine tons I don't have to ship to Tylo.

Away from (Gaming) computer at the moment; will play around with this later.  I have a feeling (i.e. a back-of-the-envelope calculation) that reduction would drop the available dV below "safe margins."

Instead, I may want to make both the boosters and the core 3200 LFO.  That not only gives me a better margin-of-error, it should give the craft more ability-- reach higher latitudes, e.g.  And the extra fuel shouldn't hurt on Laythe...  Well, this post has devolved into me talking to myself, and making notes for things to look into in-game, later tonight.  Thanks for your indulgence...

 

6 hours ago, ElWanderer said:

For descent, you can get away with fairly low thrust to weight ratios, but if you're flying manaully it's a rather risky proposition. It is comforting to have a high TWR in case you need it.

I second that!

[GoSlash27]

ElWanderer,

 

7 hours ago, ElWanderer said:

Yes, I could've been clearer about that (though the post I was quoting did reference the ascent stage).

Yeah, not contradicting anything you've said I'm just trying to be specific about what I'm saying. I hope it didn't come off the wrong way.

It turns out that the DV to orbit doesn't vary much with t/w, and even very low t/w will still allow a DV to orbit that varies little from the theoretical "perfect" burn. But descents wind up being an entirely different problem due to the aforementioned considerations.

Best,
-Slashy

[GoSlash27]

Boccelounge,

Your design looks pretty solid by my eye. If I were to alter it, I'd look at shifting DV outboard so the ship lands prior to staging, and attach the legs to the descent stage. The high- thrust would be reserved for the descent instead of ascent.

Best,
-Slashy

 

 

Edited April 3, 2018 by GoSlash27

[ElWanderer]

The funny thing with descent is that you can start with fairly poor TWR (even less than 1, for a constant altitude style approach), as long as it rises to something respectable by the time you're primarily fighting gravity rather than bleeding off horizontal velocity. This is where nuclear engines struggle, as they burn off fuel mass too slowly (on top of being weak and heavy)! That is based on my spreadsheet simulation rather than actual experience, mind.