Elcano II: Destination Mun

[damerell]

After getting around Minmus, Svetlana Kermana and the rest of the engineering department have been hard at work on a Mk II rover design intended to address all the issues with the Mk I (mainly, that everything that could possibly fall off did).

A surviving prototype Mk I can be seen in the background. While the Mk II is considerably longer, it's not a great deal heavier, so it should be straightforward to launch. Unlike the Mk I, the main body of the rover is entirely enclosed in a NF Construction octo-truss, a sturdy piece of equipment. The important parts mounted on the top deck are shielded from side impacts by structural panels. The six medium tracks should provide good traction in all conditions, with skids mounted on outriggers as a first line of anti-roll defence. (There are also underbody skids to prevent the main truss from bottoming out.) A Universal Storage hex-core provides a water purifier in addition to the fuel cell and electrolyser from the Mk I to further extend electrical endurance; however, energy storage is significantly lower than the Mk I - the Kerbal Foundries rover body was, apparently, packed to the brim with batteries. KIS storage is provided for a winch and a variety of attachments for it.

RCS thrusters on the side panels provide further roll prevention. When kOS detects a roll, it will activate RCS and return stowable solar panels and the camera boom to the positions in this picture. Once MechJeb has stabilised the rover, the crew will manually deactivate RCS and raise the camera boom.

The side panels and outriggers will also be able to mount Kerbal Foundries screws to convert the Mk II to an amphibious configuration.

Visibility from the pilot pod atop the vessel is excellent. The same cannot be said of the two inline pods within the strut structure, but they are at least heavily protected and can use the camera boom.

The camera boom is provided to reduce the risk of driving off cliffs. Its high vantage point should be useful to alert Jebediah, the pilot, to terrain irregularities. It can rotate and pitch freely.

As on the Mk I, two external command seats are provided for Sally and Svetlana, as and when they feel the anti-roll precautions are sufficient.

To do:

Test the array of headlights and decide which are necessary. Consider mounting a headlight on the camera boom. Fit underbelly camera.

Move the RTG from the easily accessible prototype position to one inside the struts - not adjacent to any crew module.

Fit MKS inflatable habitat for crew comfort.

Fit kOS and SCANsat BTDT modules; fit launcher with full array of SCANsat modules.

Fit forward-facing engine to provide additional cliff-driving-off protection.

(I'm not generally a big fan of Tweakscale, but I'll make an exception for batteries and OX-STATs where mounting lots of little ones drives up the part count. Why am I posting when I'm not quite done? Partly to get these things out of my head.)

[B-STRK]

The boxier solution would likely be more survivable, though in serious need of a makeover (), but I think the safest part would actually be the track clearances on the front and back. Before I made the hallelujah! conversion to KF, octocores served as wheel bogies for XL-3 skid wheels, which were then clipped into the main fuselage (usually a resource tank), and the exposed parts strutted against the fuselage. I swear by NF Construction (so as long as I can spare the memory for the mod! )

kOS on the Mun will be an interesting proposition. (Disclosure: I've got a challenge vehicle on the Mun myself). On wheels I often kinda hit "ruts" than either break stock wheels when on high speed (back in 0.25, the last time I ever touched stock wheels), or (with KF wheels) may send a craft tumbling from tripping (though then again, I'm coasting at 40m/s on downhills at that rate, so the game might be punishing me for my carelessness). Not sure if tracks experience the same kind of problem, or they behave differently, which may be their advantage?

Safe travels! Hope to see how your journey turns out.

[damerell]

Ah, you may be missing some context. kOS isn't driving the rover, it's just providing assistance with the electrical system and detecting rollover. Sorry about the confusion.

The KF long tracks were fine on Minmus (and I hit 55 m/s on downhills, which was fine except when I drove off a kilometre-high cliff); I experimented with the KF wheels, but suspected the suspension would keep boinging the craft up into the air in low-g. The medium tracks seem ill-behaved, so I've switched to two sets of long tracks.

The Mk II is ugly as sin... but functionality has a beauty of its own, I suppose.

[B-STRK]

Ah, you may be missing some context. kOS isn't driving the rover, it's just providing assistance with the electrical system and detecting rollover. Sorry about the confusion.

The KF long tracks were fine on Minmus (and I hit 55 m/s on downhills, which was fine except when I drove off a kilometre-high cliff); I experimented with the KF wheels, but suspected the suspension would keep boinging the craft up into the air in low-g. The medium tracks seem ill-behaved, so I've switched to two sets of long tracks.

The Mk II is ugly as sin... but functionality has a beauty of its own, I suppose.

So far I've not gotten much boinginess from the wheels, even on landing--in fact, they sometimes seem to depress/"hunker down" at speed, then afterwards raise the craft to its suspension height--probably in response to throttle input, I guess?. I did tweak dampening and spring strength to what I felt was appropriate to prevent any boinginess (up a couple notches, to 0.04 or 0.004, I forgot how many zeroes on the dampening seemed to do the trick). In fact, I feel like I should have increased spring strength a bit, as I'm kinda afraid of a nose strike being more possible when the rover hunkers down at speed. BTW, I also use the plugin with the stock KSP wheels, and they also exhibit this behavior. It's probably going to be less pronounced on the tracks, if it is even observable. A predecessor half-track design I've tested on Kerbin, employing both the wheel and track, showed the hunkering behavior with the wheel; I couldn't notice any appreciable difference in the tracks.

[damerell]

BTW, I also use the plugin with the stock KSP wheels, and they also exhibit this behavior.

I don't suppose you have a Module Manager rune to hand for that? I'd be most grateful.

[B-STRK]

I don't suppose you have a Module Manager rune to hand for that? I'd be most grateful.

They're here, lo-fi's/Gaalidas' configs, actually. To quote lo-fi,

These are the wheel configs you're looking for Just slap them somewhere inside GameData and you're good to go. There was a small bug with the small wheels, but I think I fixed that the other day.

No overspeed or impact damage is currently modelled with KF, and you'll have no problems with the wheels sounds mod. You'll just need what's in the plugins and sounds directory (IIRC) and can remove all the directories containing parts.

Have fun!

[Geschosskopf]

I like rugged, functional vehicles so I think your Mk II is quite appealing. In fact, I sometimes go out of my way to make my ships as "ugly" as possible

On the issue of lighting, the B.Dynamics mod includes a rotatable searchlight....

[damerell]

On the issue of lighting, the B.Dynamics mod includes a rotatable searchlight....

Hm. I'm still fiddling with the launcher, so I could have a look at that (although the mod thread is quite coy about its exact contents, sigh).

[damerell]

Another day, another learning experience. Here we are just after takeoff:

Although the rover itself is quite a bit heavier than the Mk I and a more awkward shape (hence the huge fairing with adapters above it), the launcher itself weighs almost exactly the same while still bringing a considerable dV surplus... or, given the shenanigans I had to go through to get it landed on the Mun, perhaps it should be thought of as a dV reserve.

One interesting feature of this design is that the SRBs have LFO nosecones on top with fuel lines linking them to the central engine. This lets the SRBs be dialled down a bit - the main engine keeps TWR reasonable at launch and is shut off when the nosecones are dry - but in the meantime TWR has risen because weight has been burned off. The SRBs burn longer and don't spend so much time shoving us upwards at terminal velocity.

The plan of installing SCANsat modules on the launcher proper is no good. The huge mass of the launcher has to be got into first a high orbit for the high-res altimetry sensor, then a medium one for the multispectral sensor; you have to wait around burning up life support supplies (yeah, but every little helps) and unless you're going to make multiple passes, you don't even get anything but a narrow strip around the equator. Instead, the launcher should carry and deploy a probe for each scan type.

kOS power management works well but should be decoupled from roll detection. In orbit, you don't want kOS folding up the solar panels all the time.

NERVAs are fine for a variety of purposes, but probably not landings. When they heat up on a suicide burn, there's no shutting them off and doing the rest of the burn later. I had to circularise just above the Mun's mountains and cool off before doing the landing... and at that, it was a close-run thing, hitting about 90% of maximum engine temperature.

Here we are coming in for a landing before I think better of it and abort. Just as well I had all that spare juice.

And here we are landed. Not one of my better landings, either, more "engine nozzles touch, craft falls over, rotate to land tracks-down" - but nothing actually fell off, so no foul.

(Another more serious issue with the Mk II rover is that KSP believes the hatches on the two lower pods are blocked because they're set into hollow trusses. I don't - there's clear air above each pod - but I still can't get anyone out of any pod but the top one. I'm not sure what to do about this.)

[damerell]

I suppose at some point I'd better actually drive somewhere, but first a kerbals-eye view of a flag planted by Svetlana at the landing spot. In theory it should be possible for the rover to reattach to the engine module in an emergency. In practice it is not, because as soon as I undock the engine module is going to roll away downslope and get lost or destroyed; and in any case the rover probably needs legs to elevate its nose for a ground-scraping takeoff.

Here's the in-cockpit view, which I'm using whenever I'm moving. I can't see as much as I would like, but that's the idea.

Another problem arose; the rover has so many parts that KSP runs at about 60% speed, and it's going to take a wee while to drive around the Mun at 18 m/s anyway. However, with the aid of the BetterTimeWarp and KerboKatz PhysicalTimeRatioViewer mods, I can physics-warp up to about 100% speed. This itself cuts down on the driving-off-cliffs problem - it turns out that 18m/s real-time across the Mun's lumpy surface (Minmus has hills, but smooth ones; the Mun is locally quite bumpy) is quite alarming enough and I can resist the urge to let it zoom down hills at silly speeds.

It turns out the Farside Crater is really, really, huge. I landed in a sub-crater, and crested the rim of that (quite a steep slope), but have now moved 5 degrees longitude to the east over a seemingly endless series of small ridges, and I'm still inside the crater proper.

This is just for demonstration purposes, it still being light - but if driving in the dark proves too annoying, plentiful life support supplies mean the kerbals can just stop every night and relax in the inflatable habitation module.

I plan to do more than 1/72 of the trip in each update, but it's gone midnight here and it's taken me a while to work through all the annoying discoveries.

[Geschosskopf]

Why put the camera on the pole if you don't have it on the MFD in the front cockpit?

[damerell]

Er... I do, centre MFD. That's the only camera on the rover. It just looks like the ordinary view out the cockpit because I'm in a relatively boring bit of landscape.

[B-STRK]

Regarding the "trapped" kerbals in the lower two pods, I think you could still transfer them to other free pods within the vehicle, so as long as you can click on their hatches (I guess this means having to zoom well within the truss structure, and click on the cockpit and hope for the best). If that can be done, it's only a matter of asking the main cockpit guy (Jeb, right?) to step out and let the others in so they can step out.

But if the truss' collider has completely blocked even clicking on the inline pod's hatch... ouch. Sorry. I guess even if Nertea had designed that truss to hold a tank inside, it might not technically be hollow like a cargo bay?

The hab section is a nice touch, reminds me of

.

ADDENDUM: If the click-hatch-transfer thing doesn't work, and you really need to get the crew out (e.g., at the end of the journey, so they can go home) if you're not averse to it, you could temporarily install Crew Manifest so you can get them out via the top pod?

Edited June 16, 2015 by B-STRK

[damerell]

Regarding the "trapped" kerbals in the lower two pods, I think you could still transfer them to other free pods within the vehicle, so as long as you can click on their hatches.

Thank you! I don't even have to do that - I can turn off Ship Manifest's Connected Living Space support and teleport them around with it. (I don't really regard this as cheating since blatantly the pods _are_ accessible.)

I guess even if Nertea had designed that truss to hold a tank inside, it might not technically be hollow like a cargo bay?

That's my guess too - internal attachment points, but actually modelled as a solid cylinder.

The hab section is a nice touch, reminds me of

.

That's a nice looking cab on that: I really have to find some better control cabin than the Mk I Cockpit. The hab section is kind of gratuitous, but I figure in sandbox one may as well practice oddball assembly techniques.

That video's also quite familiar (to, I suspect, all Elcano participants) in the way a) some of the rover features that sounded good in the VAB don't make sense once it's actually down on the ground and rollover is catastrophic.

Also, the Kethane drill does suggest the hypothetical amphibious Mk III might include a Karbonite drill and engines to speed up water crossings. Early tests of the KF screw model suggest it might just about do 4 m/s with a following wind, and there's no wind on Kerbin...

[damerell]

So, I made it to 50 degrees West, but the most interesting development was this:

Pedals and a yoke (also a big throttle and a joystick), all CH Products, which turn out to come with a truly remarkable remapping facility with its own programming language; so now I can drive the rover with this lot rather than with 5 keys on the keyboard I own already... er. Well, it is quite satisfying (and non-linear steering really doesn't hurt, although the way KSP uses a single throttle axis for forwards and reverse on a rover is odd).

As is traditional I drove over one cliff:

but got the situation under control before falling into the really big hole. The Mk III really must have a downward-facing nose camera. I cleverly turned off to the right and less cleverly noticed a bit later that I was now travelling pretty well south, having turned further than I expected.

Since then, however, it's been locally bumpy but otherwise easy going. With any luck, better screenies next time. I was very impressed by someone else's reconnaisance of all the big craters from overhead; but in the grand tradition of "Leadfoot" Kermana, we shall spurn such decadent innovations.

[Geschosskopf]

CH makes good stuff. I took many scalps in online dogfights with it back in the day .

[damerell]

CH makes good stuff. I took many scalps in online dogfights with it back in the day .

I've used CH sticks since the month we broke 3 cheap joysticks playing TIE Fighter. I'd still be using my previous one if Gameport support hadn't gone for a Burton.

[B-STRK]

So, I made it to 50 degrees West, but the most interesting development was this:

http://www.chiark.greenend.org.uk/~damerell/games/ksp/newtoys.jpg

Pedals and a yoke (also a big throttle and a joystick), all CH Products, which turn out to come with a truly remarkable remapping facility with its own programming language; so now I can drive the rover with this lot rather than with 5 keys on the keyboard I own already... er. Well, it is quite satisfying (and non-linear steering really doesn't hurt, although the way KSP uses a single throttle axis for forwards and reverse on a rover is odd).

As is traditional I drove over one cliff:

http://www.chiark.greenend.org.uk/~damerell/games/ksp/mun-7-cliff-small.png

but got the situation under control before falling into the really big hole. The Mk III really must have a downward-facing nose camera. I cleverly turned off to the right and less cleverly noticed a bit later that I was now travelling pretty well south, having turned further than I expected.

Since then, however, it's been locally bumpy but otherwise easy going. With any luck, better screenies next time. I was very impressed by someone else's reconnaisance of all the big craters from overhead; but in the grand tradition of "Leadfoot" Kermana, we shall spurn such decadent innovations.

As a longtime Microsoft Flight Simulator fan (I have got to get a copy of X-Plane, one of these days), may I just say how envious I am that you have a full-set yoke + pedals. If you also had those repeater instrument boxes that go with them, dang, I'd feel the need to rob a bank to get my own . Seriously, however, rep that you got the yoke working with KSP. It certainly is better than having to hold down W (or autokeying it as Fengist had done), moreso when that key is also tied to pitch-down. BTW, can you also control vehicle attitude as well as rover controls with the yoke, or is it really all on MechJeb + kOS? If so, how did you configure the controls?

On a note of trivia, regarding your observation of a single axis for throttle, it might be that the Unity engine's setup for throttle axis could be based off an Xbox 360/PS3-4 bumper-trigger setup, which is what I'm using. The rover throttle axis lines up neatly with using the triggers, which in Xinput are both modeled to a single axis (defaults to 50, LT pulls it to 0, RT to 100, like a joystick rudder-twist), and it also happens that seventh-gen racing games use the triggers as gas and brake/reverse. Also--and this would potentially betray my age --I think the DOS-era MechWarrior games configed the Thrustmaster throttle so that 50% was dead stop, and forwards/reverse on the same axis--just like as observed here.

EDIT/ERRATUM: Oh wait. I think those instrument repeaters I'd rob a bank for are Saitek's.

Edited June 19, 2015 by B-STRK

[damerell]

Seriously, however, rep that you got the yoke working with KSP. It certainly is better than having to hold down W (or autokeying it as Fengist had done), moreso when that key is also tied to pitch-down. BTW, can you also control vehicle attitude as well as rover controls with the yoke, or is it really all on MechJeb + kOS? If so, how did you configure the controls?

Actually, it's easier than you might imagine. CH now have a bit of software they call "Control Manager", to deal with the way older games can't cope with more than one USB axis-and-buttons device. Control Manager creates a virtual device (or devices) with a combination of whatever axes and buttons from the physical devices you want to use... and, thanks to our friend scope creep, does arbitary keyboard macros (not just from buttons, but from axes in a variety of clever ways) and has an embedded programming language (which is terrible, but every rose has its thorn).

That said, while W is rover-forward and pitch-down, KSP does have a separate keybind for "just rover forward" - otherwise, it would be practically impossible to drive on Minmus without doing wheelies all the time - and ones for backwards and steering that also don't pitch/yaw with RCS or reaction wheels. It also (of course) lets you bind axes for proportional control of steering and throttle (but not brake, but Control Manager makes it easy to say "press the brake button if I have pressed the brake pedal at all). Unfortunately, the Mk II rover is too long and thin to pivot on the spot on its tracks, but you can't have everything.

I could also bind the pitch/yaw/roll axes but MechJeb's stability control is in charge of those. There is a kerdog in the cockpit to bite Jebadiah if he tries to fiddle with them.

I don't think the throttle is a straight 50%-neutral setup. I think - could be wrong - it does something funky where as you reverse direction the zero point changes so most of the axis is used to control the direction you're moving in.

[B-STRK]

Actually, it's easier than you might imagine. CH now have a bit of software they call "Control Manager", to deal with the way older games can't cope with more than one USB axis-and-buttons device. Control Manager creates a virtual device (or devices) with a combination of whatever axes and buttons from the physical devices you want to use... and, thanks to our friend scope creep, does arbitary keyboard macros (not just from buttons, but from axes in a variety of clever ways) and has an embedded programming language (which is terrible, but every rose has its thorn).

That said, while W is rover-forward and pitch-down, KSP does have a separate keybind for "just rover forward" - otherwise, it would be practically impossible to drive on Minmus without doing wheelies all the time - and ones for backwards and steering that also don't pitch/yaw with RCS or reaction wheels. It also (of course) lets you bind axes for proportional control of steering and throttle (but not brake, but Control Manager makes it easy to say "press the brake button if I have pressed the brake pedal at all). Unfortunately, the Mk II rover is too long and thin to pivot on the spot on its tracks, but you can't have everything.

I could also bind the pitch/yaw/roll axes but MechJeb's stability control is in charge of those. There is a kerdog in the cockpit to bite Jebadiah if he tries to fiddle with them.

I don't think the throttle is a straight 50%-neutral setup. I think - could be wrong - it does something funky where as you reverse direction the zero point changes so most of the axis is used to control the direction you're moving in.

Oops--what I actually meant by "how" was the keymap/axismap you use. You already answered it with everything under "MechJeb's stability control." (Honestly, though, there is such joy in controlling a given KSP craft with a joystick or the thumbsticks of a gamepad or a yoke that Squad could easily partner with Saitek or someone else to sponsor a KSP-themed HOTAS controller or control panel. )

The fifty-fifty throttle setup was what I remembered in the Mechwarrior manual. Never did have a Thrustmaster, so no personal experience to go by. I guess it depends on how the controller software translates the scaled input from the throttle's physical position to the output fed into the game software.

One thing I've learned from driving ESME on the Mun, though: Wheels certainly can't zero-point-pivot her, or turn her when in dead stop, but even in munar gravity, a lot of SAS does the trick.

[damerell]

Another post, another 10 degrees. I hope I'll get a bit more of a chance to get the miles in at the weekend.

Yet another enormous hole in the landscape. I veered to the right a bit to try and make it out through a less-steep bit, but that was mostly an illusion. Rolling over these very steep edges is tough going - no way to see the other side even with the camera boom, and Stability Control gets a bit confused.

I suppose it was inevitable that something mounted on the back would eventually get knocked off. So much for the habitat (and the docking ports, but I didn't plan to redock to the NERVA cluster anyway), but no serious damage done. And finally out of the Farside Crater. The Highlands are still quite lumpy, but at least they have some flat bits in where I can open up a bit.

40 degrees West. Kerbin-rise, so lovely. A great big hole in the ground, not so welcome. However, I had occasion to use the nose engine for slowing down after driving off a cliff and it worked beautifully. Let's hope the supply of LFO holds out.

[damerell]

To 20 degrees West. Only one picture, because really, there's only so many pictures of the view out the cockpit looking at a crater rim one can post. I've been through the Highlands and the Highlands Crater and into the Midlands. Driving has been easier, with relatively long flattish stretches broken up by occasional big crater rims, but - sigh - this lulled me into a false sense of security, and anyone who's read my Minmus journey probably thought I'd roll it at some point and break off the camera boom.

That's the situation after Svetlana scrapped the rest of the boom (except the base part, which I'm blowed if I can right-click on) and the damaged solar panels. I also lost one fuel tank for the nose thruster (but demand for that seems very modest) and RCS thrusters on one side. I don't know what the red line on the SCANsat map is. :-/

Lessons for the Mk III (besides not driving like an idiot)? Maybe encase the whole thing in a girder cage; also, I've maligned kOS, which can activate action groups OK; it's that IR is a bit hesitant to act when it does.

[B-STRK]

To 20 degrees West. Only one picture, because really, there's only so many pictures of the view out the cockpit looking at a crater rim one can post. I've been through the Highlands and the Highlands Crater and into the Midlands. Driving has been easier, with relatively long flattish stretches broken up by occasional big crater rims, but - sigh - this lulled me into a false sense of security, and anyone who's read my Minmus journey probably thought I'd roll it at some point and break off the camera boom.

http://www.chiark.greenend.org.uk/~damerell/games/ksp/mun-11-rolled-small.png

That's the situation after Svetlana scrapped the rest of the boom (except the base part, which I'm blowed if I can right-click on) and the damaged solar panels. I also lost one fuel tank for the nose thruster (but demand for that seems very modest) and RCS thrusters on one side. I don't know what the red line on the SCANsat map is. :-/

Lessons for the Mk III (besides not driving like an idiot)? Maybe encase the whole thing in a girder cage; also, I've maligned kOS, which can activate action groups OK; it's that IR is a bit hesitant to act when it does.

I think the red line's the scan bar. Nothing abnormal, just nice timing on the screenshot--unless that line isn't sweeping up. How many degrees left to your journey?

[damerell]

It was indeed the scan bar, just bad timing - and not being used to seeing it unmoving.

After a lengthy session, I've arrived at 30 degrees East, having started at 65 degrees West - so 265 left to go. Don't hold your breath (after all, one thing that has worked is that there is plenty of oxygen).

I know this is a regular complaint with me, but the Mun is really full of opportunities to drive over harmless-looking ridges to find great big holes to fall into, especially after losing the camera boom:

Here I am crossing the meridian. The meridian is not very interesting, especially since it is getting dark - but at least my electrical system works perfectly.

And, to no-one's surprise, I rolled the rover again, knocking off Jeb's cockpit but not destroying it. Jeb is now steering from one of the external command seats; he looks very happy about this, considering how dismal his prospects for survival have become:

I am now driving via the Jeb's eye view provided by Hullcam VDS. He can't look around while on the external command seat, but at least his field of view is unobstructed and perhaps I'll have a better chance of catching rolls before they happen.

I've skirted the edge of the Northwest Crater by driving roughly East until I go "cor, that's a lumpy bit, I'll head a bit more South":

Here I am at 30 degrees East. It is now properly dark, and I'm wondering if my intrepid kerbonauts should shut down for the Munar night - life support supplies will easily permit that - or press on further:

I think it's now clear rovers must be designed to be rolled, especially on low-G worlds where a corner digging in can easily send you spinning off. Trying not to roll them is all very well, but ultimately they must fail safe.

Also getting a third kerbal to volunteer to serve as a pilot with "Deathstroke" Kermana may prove tricky.

[Geschosskopf]

I think it's now clear rovers must be designed to be rolled, especially on low-G worlds where a corner digging in can easily send you spinning off. Trying not to roll them is all very well, but ultimately they must fail safe.

Yeah, I really come to hate rovering on Mun since it got procedural terrain, and the lighting rarely if ever helps. These days, I only use rovers like to drive from 1 side of a base to another and have great respect for those who take trips of any distance. Mun is littered with abandoned rovers bulit by countless folks who didn't know what they were getting into .