Glitch messing up perfect Ike landing

[Donkey32]

Forgive me if this is in the wrong place, but I'm looking for advice on how to proceed.
I am currently in the middle of a big (maybe too big/elaborate, but that's a different topic) mission to the Duna system. Part of that mission involves a landing on Ike.
I like to build rovers with rockets instead of landers, it builds in a LOT of fault tolerance, like having little to no problem with a bit of down-range drift on landing, etc.

Anyhow, I've got my rover all lined up for landing on Ike. I get down to and begin the suicide burn and about halfway through the burn, the craft suddenly spins out of control as if being hit by something. I've now tried the landing several different ways with the same result... Is there some Kraken like bug that I'm not aware of?? Have I done something that should be obvious?? How can I deal with this or work around it to finish the mission??
Any ideas??

Thanks in advance!
D32

[AlamoVampire]

With out seeing your craft id guess maybe fuel imbalance or sas issue?

[Donkey32]

I managed to finally land it.. I started the suicide burn too high, so when the sudden spin happened, the rover was moving less than 5 m/s.. I hit the ground on my back, just not hard enough to do damage.

What's interesting is that I "simulated" several landings while building this craft and putting together the mission. With some tweeking it worked just fine in my simulations (using Hyperedit to place the craft in orbit around Ike and try landing/takeoff, etc.). The thing looks a bit goofy, but I assure you, it is well balanced with zero engine torque, full tanks and empty.

As to SAS issues.. I don't really know anything about that. I've only been playing for a few months and this game is SO deep!! What kinds of things should I watch out for?


Aaand images,
Here she is, in her current state


I folded up the solar panels, so as not to block view of the craft.


This next image is rather dark. It shows the fuel/engine arrangement. 2 little doughnut tanks per engine, 2 engine arrangements on the craft, one each fore and aft.


and a 3/4 top view.. Everything is centered on that docking port. It flies arrow straight to that... When it's not suddenly rolling wildly out of control for no easily apparent reason..

[Razorforce7]

What could be the case is that you have SAS on with autopilot set to pro/retrograde with you navball velocity marker set to "orbit" instead of "surface" while trying to land.
You can switch between both modes by clicking on the speed indicator in meter per second on the Navball.

In orbit mode you ignore the surface rotation speed of the planet. So if you aproach Ike and slow down enough with the Navball in orbit mode it will read 0 meter per second and flips the marker around while your still moving over the surface. You want to land in surface mode while having sas autopilot on. And I personally switch it of alltogether and land completely manually.

Edited April 5, 2017 by Razorforce7

[Donkey32]

Quote

What could be the case is that you have SAS on with autopilot set to pro/retrograde with you navball velocity marker set to "orbit" instead of "surface" while trying to land.
You can switch between both modes by clicking on the speed indicator in meter per second on the Navball.

In orbit mode you neglect the surface rotation speed of the planet. So if you aproach Ike and slow down enough with the Navball in orbit mode it will read 0 meter per second and flips the marker around while your still moving over the surface. You want to land in surface mode while having sas autopilot on. And I personally switch it of alltogether and land completely manually

Oh... Thanks for that. I'll look into it.
I didn't check what mode the navball was in, this is true. If memory serves, autopilot was set to stability only this last go... Not totally sure about that.
Looks like I've got "simulations" to perform to work this out.

[Foxster]

Looks like you landed at night...Dumb question: You didn't run out of power?

[Donkey32]

I landed right at dawn.  Started down with panels folded closed and full batteries.

 

[Donkey32]

I've learned NOT to try landing with the panels unfolded. There are now solar panels strewn across the face of Mun and Minmus, as learning the hard way seems to be my little way.
Anyway, this landing was just after the crack of dawn.. I wanted a full day, in case...

So far, no power problems. The batteries barely dipped at all during descent and are now completely full, with just a few minuets rest on the surface, panels open.

[KerikBalm]

Looks like the fuel tanks aren't inline with the CoM, given the sci jr, wheels, and structural pieces behind the landercan. I cant see the engine, but i assume its off center and has no alternator.

Maybe the sas is too weak to counter the torque as fuel is burned, or maybe it sucks up all the power and you lose control when power is gone?

[Donkey32]

Just now, KerikBalm said:

Looks like the fuel tanks aren't inline with the CoM, given the sci jr, wheels, and structural pieces behind the landercan. I cant see the engine, but i assume its off center and has no alternator.

Maybe the sas is too weak to counter the torque as fuel is burned, or maybe it sucks up all the power and you lose control when power is gone?

There are two sets of tanks and engine, aligned so the craft is centered and balanced on the top docking port.
This craft has done quite a bit of time on Minmus without trouble. Lots of up and down, to orbit and back.. Lots of hopping around after refuel..
Which is what has me confused with this particular landing.

[Donkey32]

It's possible that I messed something up assembling the new mission though. That would make sense.
Though there was no problem during the entire descent, right until the last second, when it suddenly flipped wildly.
 

[wibou7]

Can we see the COM and COT with full tank full and empty? Maybe your COM shift too far out of your COT when your fuel goes low?

Also, we do not see the engine on those picture... Do you have one engine or several? Maybe the fuel flow is messed up causing one engine to flame-out, resulting in asymmetric trust.

[Donkey32]

Ok.. I've now done some simulations and I think I've learned a couple of things.

Firstly (and slightly off topic but important), RCS Build Aid and Kerbal Engineer do NOT agree about engine thrust balance!! When in doubt, believe Engineer, as RCS Build Aid is a little off.
I discovered that my craft is a tiny bit off kilter. It generates about .33Kn of torque, which isn't really noticeable with SAS on. According to RCS Build Aid it's dead on, which is what I was trusting when I built the craft. Not ideal, slightly annoying, but it's not generating my "flip out of control" problem.

Check it out, this (different version) of the craft now produces zero torque when firing the engines. Notice the disagreement in Torque values in the readouts.

Apparently, I somehow built a glitch into this rover, when assembling the Ike mission.. None of the other copies of this craft has the problem, they land just fine regardless of which mode the NavBall is in (orbit/surface), I tried it on different bodies (Minmus, Ike and Mun) and ONLY had the anomalous flip out when using the rover that I used in the mission. I had to pull the rover out of the full mission package and use that to experience the glitch at all, and then it does it everywhere, landing on any body.
So it's only the craft, coming from the larger assembly that is broken. The same craft, pulled from any other source so far works without a problem.

[Donkey32]

One more piece.. In doing the simulations with this craft, I experienced a different kind of error. Sometimes, coming out of warp, the craft will (only sometimes) suddenly spin wildly. Once I get it under control, parts are often shifted off of the craft. The parts that shift are always the same ones too, they are the Cubic Octagonal Struts in the rear of the craft, attached to the wheels.
This does not happen with other versions of this craft either, so my current theory is that I must have done something to break physics in this lander and only in the version that went to Ike.
So, what does that? How do I avoid it in the future?

Edited April 5, 2017 by Donkey32
Specificity

[Snark]

Moving to Gameplay Questions.

Possibly dumb question, but... are the fuel tanks' flow priorities matching?  If they don't, it would cause one set of tanks to drain before the other one, which could cause lopsided mass distribution as it burns.  I'm guessing your testing has already covered this scenario, just wanted to check.

 

[Donkey32]

Good question.. They are identical.

But wouldn't I notice that earlier on?? The craft wouldn't go suddenly whack if that were the problem.. It would move gradually towards whack, as fuel is used..

[Razorforce7]

What if you put a small probe core just under the dockingport jr, make that the active probe core (that from which you control) and then retest it.
It can also be a form of the kraken. Clipped parts or certain parts attached or in the vicinity of certain other parts are sometimes recipe for a kraken soup.
This is often depicted by parts shifting wildly from your craft and sudden uncontrolled spins.

You could play detective and rearrange your craft until you "may" find a culprit part that may be the leading cause of this.
But with kraken craft it usually is best to rebuild the whole thing and hope it doesn't happen again.

[Lunar Sea]

How does the kerbal see out of the pod? 

[bewing]

1 hour ago, Lunar Sea said:

How does the kerbal see out of the pod? 

He drives around with the hatch propped open, o'course.

 

[Lunar Sea]

4 minutes ago, bewing said:

He drives around with the hatch propped open, o'course.

 

Of course. On one hand I very much want to sanctimoniously criticize damn near every aspect of this design. On the other I have to admit it is very much in the kerbal spirit and wish nothing but the best of luck (you'll need it). 

[Donkey32]

10 hours ago, Razorforce7 said:

What if you put a small probe core just under the dockingport jr, make that the active probe core (that from which you control) and then retest it.

There is already a probe core under there.. I like to have the capability to drive these around by remote control sometimes, or stick a scientist in there instead of a pilot.. It's already saved my butt on one mission and it's boosted my science capabilities on most missions.
 

Quote

It can also be a form of the kraken. Clipped parts or certain parts attached or in the vicinity of certain other parts are sometimes recipe for a kraken soup.
This is often depicted by parts shifting wildly from your craft and sudden uncontrolled spins.

Thats what I think now.. Seems the most likely scenario.
At this point, I'm going to do what I can to complete this mission in a good way, then leave that rover to the Kraken as a sacrifice!! I only hope it hasn't somehow infected the other craft in the cluster of craft that went to Ike together..

7 hours ago, Lunar Sea said:

Of course. On one hand I very much want to sanctimoniously criticize damn near every aspect of this design. On the other I have to admit it is very much in the kerbal spirit and wish nothing but the best of luck (you'll need it). 

@Lunar SeaI very much would like that criticism!!
I'm still pretty new to all this, the learning curve on every aspect is pretty steep and any help (however sanctimonious) is appreciated..

Before you do, you should know that this ugly little beast has (up till now) performed very well for me on a couple of missions to Minmus, a mission to Mun and several science sweeps of various places on Kerbin..
I'm not sure how I invited the Kraken in to this iteration or on this mission, but without that evil beast on board, it's been a real champ for me.

Edited April 6, 2017 by Donkey32
Cleanup and iterate

[Donkey32]

5 hours ago, Lunar Sea said:

How does the kerbal see out of the pod? 

She uses the force!!
And really good instrumentation.

[FancyMouse]

15 hours ago, Donkey32 said:

Firstly (and slightly off topic but important), RCS Build Aid and Kerbal Engineer do NOT agree about engine thrust balance!! When in doubt, believe Engineer, as RCS Build Aid is a little off.

What's your vessel mass in RCS build aid? It looks like two mods don't agree on that number. Most likely there's some part that they got a different mass number for some reason.

And I agree with the body rotation argument. That explains pretty well your original problem. Worth double checking if it's really happening

[Donkey32]

6 hours ago, FancyMouse said:

What's your vessel mass in RCS build aid? It looks like two mods don't agree on that number. Most likely there's some part that they got a different mass number for some reason.

3.388t  Engineer Redux has it at 3.592t

It looks like there is a disagreement  over struts (at least). RCS Build Aid reports them as massless, Engineer Redux reports correctly at .05t (actually .051)
In my vessel, the numbers check out (mostly) I put 4 struts underneath to make the frame attachment to the wheels into triangles.. The difference in mass reporting between the two mods is .204, 4 struts account for .2 of the difference.. Not sure where the .004 is coming from, but my guess is it's a rounding error as there are 4 of them.. Or, on second thought, simplest answer is that the actual mass of a strut is .051

In this image, the mods agree

After adding one strut, there's a disagreement

And I agree with the body rotation argument. That explains pretty well your original problem. Worth double checking if it's really happening

Yeah.. I liked that answer best too, until I tested for it. The whack comes, no matter what setting the navball is in and it happens with SAS turned off too.

At this point, I'm convinced that I've got a Kraken on board.. A little mini-me Kraken!!
So basically, I've got to try to complete the ground portion of this mission, survive it by not tempting the Kraken overmuch and get out of this rover as fast as possible.
Hopefully, with a little luck, the rover is the only craft of the cluster (that I sent to Ike) that has Kraken aboard!!

[Donkey32]

14 minutes ago, Donkey32 said:

It looks like there is a disagreement  over struts (at least). RCS Build Aid reports them as massless, Engineer Redux reports correctly at .05t (actually .051)
In my vessel, the numbers check out (mostly) I put 4 struts underneath to make the frame attachment to the wheels into triangles.. The difference in mass reporting between the two mods is .204, 4 struts account for .2 of the difference.. Not sure where the .004 is coming from, but my guess is it's a rounding error as there are 4 of them.. Or, on second thought, simplest answer is that the actual mass of a strut is .051

Ok.. But now, I checked with 3 struts, the difference comes to 150kg, even.. And, looking back I can see I was sloppy by not checking the numbers above. The strut part checks out at .05t..
Which means there is some other tiny little disagreement someplace (in my rover) that adds up to .004t

Coffee is the answer..

Edited April 6, 2017 by Donkey32
clarificarion