totm may 2023 Elcano IV: Circumnavigate all the things!

[damerell]

For a bit of a change of pace, I'm at least thinking about RSS Earth (by sea, I'm not completely mad). It seems at least fairly feasible to cross the routes of the Suez and Panama canals by land; the terrain is flat apart from the occasional inexplicable pinging-up into the air, but in a brief test those haven't been destructive, just alarming - and I'm definitely willing to use quicksaves to work around random invisible lumps in the ground when on real-Earth I'd be _in a canal_.

The Earth's circumference is more or less 40 thousand km. The Kerbian Sea Monster will do 100 m/s full, 130 m/s half-empty, and 200 m/s dry. This suggests maybe its overall average speed is 140 m/s or so; the ten Goliaths guzzle a total of 5.8 LF/sec, the four big tanks and assorted smaller ones hold 76,200 LF (16000 per big tank), allowing us to go 1,840 km on a full tank. This is a bit odd since this would not have got me around Kerbin, and the real KSM topped out at 206 m/s, but it will do for a thumb-in-air estimate. This would mean a bit over 20 refuellings. Each tank would burn for a bit under 4 hours, making for an 80-hour journey - assuming, for the sake of argument, that refuelling took no time at all.

Endurance obviously could be increased by adding more tanks or removing some engines, at the cost of making the trip still longer.

I tried both ideas. Nine big tanks gives 156,200 LF; almost exactly 7 1/2 hours to burn a tank. This will do circa 80 m/s full, 110 m/s half-full, and 155 m/s dry. 115 m/s average, 3,000 km on a full tank, or circa 13 refuellings in a 100-hour trip.

Four engines gives 60 m/s full, 90 m/s half-full, and 145 m/s dry. 98 m/s average, 9 hours a tank, 3,200 km a tank, again circa 13 refuellings but now we are looking at 112 hours.

With some hydrodynamic adjustments and 16 big tanks - and the addition of one slightly cheeky nuclear thermal jet - I get 72 m/s full, 96 m/s half-full, and 165 m/s dry. 111 m/s average, so still a 100-hour trip, but at over 12 hours a tank there would be "only" around 8 refuellings - and on the NTJ alone I can limp along at maybe 30 m/s, I can't be stranded. The roughly 5,000km range this represents would let me cross the Atlantic from Saint Vincent to Cape Verde, and let me stop at suitable islands in the Pacific - assuming I find I have to land to refuel. If at all possible, I'll come up with a way to refuel at sea.

Edited April 10 by damerell

[damerell]

I unwisely decided to send the resupply mission to Tylo with chemical rocketry. This was tricky since the payload is the best part of a kilotonne. As such, I ended up launching the most ridiculous object I have ever put on the launchpad; 5m payload with 7.5m and 10m stages, and a bottom stage of 32 Space-Y "Freya" boosters - by comparison, the Hangarmoth was got off the ground by six of those.

I've been away for a while because this resupply mission was the most amazing mix of errors on my part and KSP bugs. I reached a stage where every time my planned vector went over Laythe the game would crash; since this came at the end of half-hour burns, I got very frustrated. I had weird bugs that just make the dV for a planned maneuver walk upwards for no reason. But I made it to the point where I was aerobraking at Laythe:

Approaching docking. I'm not as short of dV as this looks, because the forward 5m tank is just resource locked.

And docked up. I'll dock one monoprop can on each side of the Hangarmoth, to let me use monoprop to balance my mass fore-aft.

[damerell]

In orbit around Vall, waiting to drift far enough from the QA to use the Orion engines. With tanks so very nearly full, the "mast" can be left with the QA - all the empty LFO/monoprop space that might be filled by ISRU fits into the Hangarmoth alone.

Not much else to say about this, but it's a nice shot of Jool and Laythe. Our kerbals waited nearly five Kerbal years at Tylo for resupply, but they knew this was a long mission; more important is that twelve kerbals arrived at Tylo and twelve left it alive.

Drilling away on Vall. The time to full tanks looks very encouraging, but that is mostly because what's been landed has less capacity and most of that started full.

I went only a short distance on Vall, but was presented with Joolrise - spectacular, since I came over a ridge and there it was.

Vall seems nice and flat so far, although I'm about to review other circumnavigators' accounts. More worrying is that my computer seems slower even though it is the exact same computer; where previously I was getting a near-steady green clock while roving I'm now running about 80% real-time, and performance with more parts is similarly reduced (making running the QA-Hangarmoth-Resupply combo _agonising_).

If I can't find an easy fix I guess I'll reinstall Better Time Warp to let me use tiny increments of physwarp when roving, and just suck it up when in space.

[damerell]

I found out the maximum physics time per frame setting was set to its minimum. I don't know how that happened, but increasing it cured the yellow clock entirely. What a relief!

After some time on Vall I saw this. As always, I wondered if it was close, or far off and absolutely huge? But I knew from other reports that there is a mountain at 90W, so I expected it was huge.

Yep, it's huge. And why does the landscape have stretch marks?

Looking at the peak out the side of the cab, I am really glad I did not feel the need to attempt it.

Here I hit a chunk of very bland landscape - so little local detail I had almost no sensation of motion, which is awkward in terms of assessing if one is going too fast.

And here, I confess, after going 1/6 of the way around Vall wondering why performance uphill was so bad, I remembered this rover might have been left in a lower gear.

Now, however, with the ability to go full speed uphill, Vall's low gravity, and running at 100% of real time, it all feels a bit like the Dukes of Hazzard, only with a less unfortunate paint job.

[damerell]

Inevitably, the mention of the Dukes of Hazzard led to this incident. One nice thing about playing with headphones on is if one set of wheels lifts - or both - it's immediately obvious.

Mountains in the distance again. One thing I'm really liking about Vall is how the rough terrain seems so often visible from a great distance.

I plan to rove straight up this ridge.

Again, a very clear view of the pass through the mountains.

Crossing the date line.

Finally, I plan to steer for this cleft in the distance.

Edited April 11 by damerell

[damerell]

Heading up to another pass.

Another mountain being roved around, not into.

This one I will not rove around so well.

I regret I didn't capture what happened next, but I found myself heading down the mountain too fast to brake and also on too much of a slant to steer upslope. I could easily have come a serious cropper but by luck and the aid of MechJeb Stability control, I ended up down on my wheels.

More hills ahead.

Here I am up close to them.

Is this a long way off, or absolutely huge?

It is huge!

And here I am passing it to one side.

Vall remains a pleasure to rove on. There are these vast mountains, but they're all visible from a distance; and you can rove slightly faster than the dawn, so you can always rove in daylight.

[damerell]

Another mountain looms. I'll pass to the right of it.

Going past it.

This vast (horizontally, not vertically, although it's pretty big vertically too) range appeared. I went southwest down the edge of it for a while but then checked the ScanSAT maps; it would be a huge detour, so I resolved to cross it.

Here I am crossing the mountains.

After that, that mottled patch in the centre of the map at bottom left is very easy terrain. Not much to write about that, but these moons appeared - it's Laythe close by, and Tylo a long way off, coincidentally in apparent proximity.

The end is in sight!

Home on Vall.

Edited April 19 by damerell

[18Watt]

On 4/9/2023 at 5:37 PM, damerell said:

it all feels a bit like the Dukes of Hazzard, only with a less unfortunate paint job.

Every time I go over a jump with a rover that Waylon Jennings song plays in my head.  "Just two good old boys..."  Every single time.

[damerell]

To Laythe! Taking off from Vall. I'm afraid I really like these great-big-Jool screenshots, so you're going to be seeing a few of them.

Here's another one, having arrived at Laythe. Once again I am cursing my blithe assumption I could balance the thrust axis with pulse units alone (and indeed I'm now doing it with other resources); you cannot balance the ship with a resource that is very full or very empty. I'm also (again) cursing the redesign of the Hangarmoth which put two big Ore tanks at the back and two big MetallicOre tanks at the front - if each resource's pair was diagonally opposite each other on the thrust axis, they'd be far more useful for balancing. It takes over a month to use a full load of MetallicOre, a few days to use a full load of Ore, so I'm constantly wanting to take off and manuever with mass all-forward.

However, having realised I could DockRotate so one of the Fertilizer tanks was aft, one forward, I could sort out a lot of this; and also that I could stop converting Ore and fill up the Metal tank on the Orion spinal truss that's towards the rear. But did that help? I was consuming more pulse units by having a heavier vessel full of Ore in order not to have to jettison MetallicOre which would make a tiny number of pulse units.

The Laytheboat-jet combo was in an awkward orbit and I had the Hangarmoth fly to it "to save dV". Why? The combo had to get into a low Laythe orbit anyway, and at that point if it was short on fuel (if the NERVAs have burned some of the LF for the jet) it could be topped up from the Hangarmoth. I wasn't thinking straight.

Here's a decent view of the top half of the jet-boat monstrosity.

There was, it turned out, enough dV anyway - albeit very little to spare; here's the combo getting down into that low orbit.

The next stage in the operation is to dip periapsis to 40km or so, pump all the LF out of the Laytheboat (until now, the fuel balancer has been allowed to put it in there), decouple the Laytheboat, push periapsis on the jet-and-NERVAs left over out of atmosphere, switch back to the Laytheboat, re-enter it, it identifies a nearby island; the jet dips periapsis to 40km, grabs all remaining LF, O, and monoprop, dumps the NERVAs and their tanks, re-enters, and flies over to land on a beach. Uninstall FAR (does weird things to boats), sail boat to beach; the boat is amphibious enough on KF skids to be able to get onto the beach.

What happened is that I landed the Laytheboat smack bang in the middle of the DeGrasse Sea, about as far from any island as it is possible to be on Laythe. I could not have done it with that degree of precision if I'd tried. At that point I said something rude and stopped for the day.

[damerell]

The second time I landed smack bang in the middle of one of the bigger islands. In the dark. If I'd _tried_ to land on an island on Laythe in the dark, with no descent guidance...

Edited April 22 by damerell

[damerell]

On the third attempt things went much better; separation from the jet portion:

Re-entry quite undramatic - it helps that it's mostly made of Mk2 spaceplane parts - and here I am about to splash down, not too far from land at all.

I found I could not re-enter the jet immediately afterwards; it's dusk when it arrives at the island, and a rough landing when you can't really see the terrain is not a happy thing. It'll have to come down half a Laythe-day later or so, to arrive a little after sunrise.

[damerell]

Houston, we have a problem.

I originally planned to sail around Laythe with Scatterer waves. However, it's hard to average more than about 15 m/s under those conditions, and that would take 60 hours. I could tolerate that for Tylo or Eve, where there's some actual scenery, but Laythe would be vast expanses of water and more water; RSS Earth (no waves) might be longer still, but the Kerbian Sea Monster is very stable; I can listen to something or even read a book, only adjusting the steering now and then. I don't want to constantly fight the steering for 60 hours with nothing to look at.

However, the Laytheboat just isn't stable at the speeds it can reach with no waves; with no Mechjeb it slews uncontrollably to one side at around 20 m/s and stops, and with MechJeb KILL ROT it can get to about 100 m/s before slewing uncontrollably to one side and stopping in a number of unusable pieces. Worse yet, at any speed MechJeb is using more EC twiddling the controls than the engine alternator is producing, so I can't throttle down to 80 m/s or so. I'd have to proceed at 80 m/s for a while, crawl along at low power with the controls dead to recharge the batteries, back to 80 m/s...

This calls for more design and simulation work from the KSC. We can pump ballast into the Laytheboat to affect its pitch (currently, it is well up at the bow) or simply to make it grip the water more effectively. More radically if it can limp to a beach, engineer Йеанетте can start removing parts from it, if the KSC has identified which parts to remove (eg perhaps the central hull of the trimaran isn't actually helping?)

Edited April 23 by damerell

[damerell]

Well, I don't have a problem but I certainly am confused. I was just trying with FAR, having ended up trying things out of order. Uninstall FAR; the boat will do about 30 m/s. Stick in Scatterer, turn the waves down a bit... 45 m/s. I don't really understand this, save that sometimes a boat seems to be a bit "stuck" to the water, can only get away with maximum thrust, sometimes it unsticks later.

[damerell]

This screenshot looks pretty, but it was not the right place to drop out of orbit; I ended up flying about 1/3 of the way around Laythe at the end of the glide, and using up about 1/3 of the LF in the drop tanks... but I'm pretty sure there's plenty of that to spare.

And landed. I don't think you can actually land an aircraft on a sand dune by flying level into it and hoping, but that's KSP for you. (I guess if I wanted to do it less absurdly, I'd install parachutes to pop on the final approach so it would come down more or less vertically).

Next will be to uninstall FAR (again) and sail the Laytheboat to the beach; then pick waypoints - the kOS script I wrote for Kerbin by sea will come in handy again.

[damerell]

It turned out some idiot had forgotten to build a ladder onto the Laytheboat. Ooops. It's also very hard to bring boat and jet together because the boat, having only skids, can't brake on land; it can only stop somewhere dead flat or by turning at a right angle to a slope, and it steers like a cow. After a bit of thought I hit on this:

Kerbals can climb things really quite well. The jet doesn't have to stay with wheels in the drink; when the Hangarmoth is overhead I can get a control connection and get it back on the beach.

Bit of a comedy of errors to get here but the Laytheboat is ready to go sailing.

[damerell]

Laythe is, I think, the least interesting world I have done. Sea, sea, and more sea; and the few bits of land I have seen have hardly been distinctive. I didn't fit a ScanSAT BTDT to the Laytheboat because I thought it didn't work on the ocean (it used not to); otherwise, I might hardly have made screenshots at all, other than a small map with a BTDT track.

Here I am enjoying the all too brief period of making nearly 50 m/s.

All too brief because when I reloaded the next day, Scatterer was painting the sea with horrible triangles. I uninstalled it and plugged along at 28 m/s. Joolrise, and you can see Tylo near Kerbol.

I expected things to get dark as Kerbol went behind Jool. They didn't. Apparently Jool is less opaque than one might expect. (Now I think back on it, I seem to remember something similar on Mun and Minmus; only the SOI body blocks sunlight).

So here's Kerbol about to set for real.

Crater Island, which I wanted to see. It is a grey rock, exactly like every other bit of land on Laythe.

The waypoints for the kOS script that recommends a heading were not very precisely selected, and I'm rounding this headland as closely as I can rather than follow the recommended course. It is also a grey rock.

Normally I do all my circumnavigations in real time. I kind of feel things should take as long as they take - and on land, 43 m/s is quite fast enough to be steering a rover at. However, I cracked on Laythe and am using physwarp for the second half of the journey.

Another headland being trimmed, and now up to 4x physwarp; this posed a bit of difficulty in that even tapping the yaw key with fine control on produced a huge slew, but a joystick with a very small dead zone can give very fine yaw control.

The end is in sight, but this has not been my favourite trip.

Edited May 1 by damerell

[18Watt]

19 minutes ago, damerell said:

Laythe is, I think, the least interesting world I have done. Sea, sea, and more sea

Yeah.  The land is just sand.  I do enjoy beaches in real life, but the whole planet is 99% water and 1% beaches.  I do not disagree with you.

 

24 minutes ago, damerell said:

The end is in sight, but this has not been my favourite trip.

Well, just remember, you get the highly coveted Master Mariner badge on completion of Laythe!  

I actually enjoyed the challenge of Laythe, but I was using a specialized vessel for Laythe, which made it a little less painful.

The list of Grand Master Navigators is way too short, and you're so close!  I think you just have Eeloo left?  

[damerell]

20 hours ago, 18Watt said:

I actually enjoyed the challenge of Laythe, but I was using a specialized vessel for Laythe, which made it a little less painful.

Mmm. I think it's very clear I should have paid much more attention to the speed of the Laytheboat; although it wasn't too bad at 4x time acceleration.

Coming through the Eastern islands. I'd found the boat veered badly to starboard at 4x physwarp - not some Coriolis effect, it does it in both hemispheres, and it veered (less badly) to port with no physwarp. I experimented with using MechJeb as a very fine trim control, a job it did quite adequately.

By here, I'd hit on the idea of picking the exact position of waypoints out of the zoom map and feeding them back into the kOS script that recommends a course to steer, letting me shave the headlands more closely. It's getting dark, but there's a vast sea expanse ahead and the zoom map will stop me running into anything.

I had no waypoint near this island - with limited ability to visualise great circle routes in my head, I had no idea which side of it I should aim to pass - but it turned out a route straight to the target would miss it anyway. What luck!

Also, with the target in view on the HUD, I had less need of the kOS navigation assistance - but I kept it running, the distance still being useful to know.

Close to the island I'd landed on; the jet was on the other side and I certainly had no wish to try and sail around the island in the dark. Instead I could stop here and bring the Hangarmoth down into a low Laythe orbit; I very much doubt the Laythejet has the dV to make the rendezvous, the Hangarmoth will have to come to it, but I may as well load the dice in my favour as much as I can.

With no ladders or brake on the boat, I had no better plan than to rub it on the beach and let the kerbonauts bail out and swim ashore, with the last out turning off the engine (the radiator panels were heating up badly with no airflow, and why not leave an intact boat on Laythe if I can) and making a run for it as the boat slides back into the sea. The jet is some way away, but the first in can take the brakes off and roll it backwards a bit.

And ready for takeoff! But I am done with Laythe, and need to reinstall FAR.

 

[damerell]

More problems.

First of all, it turned out in the flight across Laythe, the jet engine was draining LF from one of the LFO tanks in the main body, as well as the two LF-only drop tanks. This meant I had considerably less fuel than I expected; the plan to get as high and fast as I possibly could in airbreathing mode then switch the RAPIERs to rocket mode simply was not going to work. The second was that the decoupler between the plane body and the rocket body is the wrong way around - if it's allowed to fire it puts the rocket body into an uncontrollable spin, and if it decouples without firing it's a huge chunk of offcentre mass on the rocket body, meaning that can thrust no harder than can be offset with RCS. I did have plenty of monoprop, but what I lacked was time - with a very limited ability to get out of atmosphere on the RAPIERS, I wanted to be able to make my circularisation burn as fast as possible.

I removed as much unnecessary mass as I could - antenna, ladders, the monoprop in the Mk2 Clamp-o-tron - and then fiddled very carefully with ascent profile. Here I am going up in airbreathing mode (the middle engine is much higher efficiency, so was used for crossing Laythe; it gets dropped here when I get above its useful operating height).

Rockets burning - and a careful eye on the fuel supply; as soon as the apoapsis will give me enough time (around 55km) I'll stop the RAPIERS, pump it all back into the 1200-FTP tank, and ditch the plane body.

This is a slightly lower periapsis than I might like, to put it mildly, but it's enough. However, the orbital inclination is about 80 degrees from that of the Hangarmoth, and there isn't enough LFO left for this craft to adjust its plane. (In retrospect, I should have run it dry to improve matters as much as possible).

Rendezvous. However, the circa 2 km/sec maneuver to match planes means the Hangarmoth is now very short of pulse units. It's not impossible I'll have to send another resupply mission to Jool - or to Eeloo to give me enough juice to get home.

ETA: I've got enough dV to go to Eeloo and circularise inside its SOI. I'll have a resupply mission meet me there.

Edited May 2 by damerell

[damerell]

Another lesson: don't, waiting at Jool for a maneuver far in the future, wait in an orbit that crosses that of any of the Joolian moons. Obvious in retrospect, but isn't everything? I adjusted the orbit after this alarming but non-fatal intercept.

The resupply mission docked up. With a smaller payload and a better ascent profile than the last resupply mission, those three tanks have enough LFO left to fill up the Hangarmoth - a welcome bonus since I had burned up almost 3/4 of the supply on board in order to reduce the demand for pulse units on the transfer to Eeloo.

Stopping here for the night IRL, but in a low Eeloo orbit. The last planet awaits.

 

[18Watt]

1 hour ago, damerell said:

Another lesson: don't, waiting at Jool for a maneuver far in the future, wait in an orbit that crosses that of any of the Joolian moons.

  Yeah, that's good advice.  No matter how many times I get burned by that, it's something I still do.  

1 hour ago, damerell said:

The last planet awaits.

[damerell]

When designing the Hangarmoth I fitted these Super Dibamus OMS parts (on either side of the auxiliary LFO engines); as well as ordinary RCS, there's a throttle-controlled port on one side of them. This is the first time I've fired them up, which is odd because they're ideal for this sort of job of pushing off from the Queen Agaster after undocking.

However, if I were doing this again I'd just make the auxiliary engines DockRotate, making them useful for slowing down and large fore-aft translation as well as fine control of forward thrust. Particularly, one of the weak spots of the landing script is trying to control horizontal velocity while keeping the ship basically belly-down before it hits the ground; if the auxiliary engines could rotate to do that, it would really help.

Indeed, the landing script did not cover itself in glory, slapping the nose down into the ground and breaking off one of the extendable radiators (which are stowed for landing). However, I probably don't need all three, especially given we're at the coldest place in the Kerbol system - and, perhaps more importantly, installing a fixed radiator onto the reactor directly has fixed most of the cooling woes (if things heat up overall, reactor power output drops, and if it drops far enough the radiators can't get power, you've got to turn the whole ship off and start again; giving it a dedicated radiator makes this impossible).

Also, it redeemed itself by, through sheer luck, picking one of the few biomes on Eeloo with all three of Ore, MetallicOre, and Uraninite to mine. Supplies are not short, especially since I have nowhere else to go but home, but this certainly won't hurt.

The new Mk VIII design which met us in Eeloo orbit (no screenshot of that, oops); compared to the Mk VII, the rollcage comes further forward protecting the spotlights, it has a second reaction wheel for stability control, and there are rotatable beams added at the sides which - in theory - can be used to re-right the rover if it turns turtle.

Edited May 5 by damerell

[damerell]

As usual, I've roved diagonally towards the Equator, which I'll mostly aim to follow; but I think I'll try and track down the big ice canyon running southwest from the Equator down towards the landing site - from the looks of things, I landed with it in sight.

120 degrees west, and it's getting dark. Eeloo has been extremely easy to rove on, gentle ridges which just provide enough variety to be more interesting than Tylo. It doesn't make for much in the way of screenshots, though, I'm afraid.

Clearly being outpaced by the sunset, so I shut down here overnight.

At 150W at dawn.

Here I'm afraid my screenshot-transferring broke down and screenshots to 120E are lost. Not the end of the world, since the BTDT is doing a superb job of mapping progress and that patch of Eeloo is an endless series of Midlands - the terrain was so easy I switched the righthand MFD to biome mode, giving me a plain sheet of Midlands - with the lost screenshots just being ridges with no good view of the terrain ahead, but still a pity. However, I think that puts me just short of halfway through, since I set off from 77W.

[damerell]

Of course my screenshot process would break down immediately after being thread of the month (although I really think the main Elcano Challenge thread is much more worthy). I've sought to redeem myself somewhat with a shot of me setting off replacing one of the lost ones. Kerbol is high overhead so we have half a roving day or so left.

You would think after roving around 95% of the system, I would remember to run the kOS script to run the fuelcells, prevent overspeed, etc - or have arranged for kOS to boot it automagically; I think I did the latter on the Mk VII, but this is a Mk VIII. You'd be wrong; the wheels broke down this slope, and so while I don't generally plant flags (ScanSat does such a good job), engineer Arald could not resist the urge while out to fix them all.

Eeloo remains a pleasure to rove on - enough gravity to let you accelerate and brake sensibly, and to actually turn corners, but not so much that hills are a chore, not so little that every bump means two minutes off the ground. This is about the biggest jump I have done this session, and sure, it's big, but not worrying.

Crossing an ice canyon that runs north-south. I'm going to cross this area of high ground on the Equator, not try and follow these ice canyons.

[damerell]

Kerbol was low in the sky, so I stopped for the KSP night. After this I went into the Highlands, some of the most interesting roving on Eeloo - but it remained a pleasure, the terrain was less even but it was never difficult.

This was quite an alarming jump - my shadow is far below me - but in low gravity with a solid rover design, it was not a problem.

Here I hit more "soft ground", like on Pol. Exasperating, so close to the finish. I found that quicksaving and then reloading would cure it for a bit, and I struggled on for about 2 degrees of longitude, but then the ground was firm again.

However, the cure did tend to ping me up into the air, and once I landed upside down. I could have quickloaded again - no objection to doing that in response to KSP bugs - but I took the opportunity to test the self-righting mechanism on the Mk VIII. It performed well, getting me on my side so I could then rotate around the anti-roll arm to put me back on my wheels.

I turned south, partly to try and get into the canyons but also to explore what turned out to be the Craters biome.

I'd been guided by the biome map for some time, but it turned out the cleft in the landscape I was planning to rove down was outside the biome on the map. I steered straight downwards, trusting the terrain.

I'd been seeing an object on the horizon, and it turned out it was, thanks (I think) to Distant Object Enhancement, the Queen Agaster VI.

At this point I'd given up on the unreliable biome map and was just following the topography scan - but now they coincided, and I was in the Ice Canyons.

Turning the corner of the apparent T-junction in the canyons.

I circled around here, as you can see in the top right MFD, because I'd found a spot below "sea" level on a dry world. I didn't know that could happen.

And finally, up the slope out of the ice canyon and home. I shall bring the kerbals home, but at this point their mission was accomplished.