[1.3.1] Ferram Aerospace Research: v0.15.9.1 "Liepmann" 4/2/18

[jrandom]

Yeah, I know the tricks. I was less asking, "How do I tell?" and more, "I have designs that I know don't work, because I've tested them, but they used to, so I'm looking for new ideas because I'm obviously in a design rut if others aren't having this problem."

I appreciate the advice, though.

Ah! Well then, pictures! Including flipped upside-down with CoM/CoL markers visible.

[Traches]

Is anyone else having some trouble making stable reentry vehicles?

You can try putting small fins (if you have tweakscale, basic fins at 50% seem about right) on top of the command pod so it ends up looking sorta like an old school nuclear bomb. They have to be pretty small so they don't destabilize your rocket when you launch (or you could cover them with a fairing)

[Jovus]

YAFIYGI

Javascript is disabled. View full album

[Wanderfound]

Mass at the bottom, drag at the top. A few high-up goo pods often work well as reentry fins.

But: why bring it back at all? Go EVA, collect the data, reenter with just the pod.

[jrandom]

YAFIYGI

Yep, your CoL is on the wrong side of the CoM and that's why your craft is flipping on reentry. Either move mass towards the heat-shield end, or drag fins on the opposite end to get the CoL on the correct side of the CoM.

[Jovus]

Mass at the bottom, drag at the top. A few high-up goo pods often work well as reentry fins.

But: why bring it back at all? Go EVA, collect the data, reenter with just the pod.

Yeah, I get how it works. But how do I move from expertise to execution?

As you can see, mass is already at the bottom. (Heat shield and payload bay.) And I can't put parts like goo cans up top on the capsule, because they explode without heat shielding. Does anyone have actual examples of working re-entry vehicles to share? I'm having a hard time believing I'm the only person who has a problem.

[jrandom]

Yeah, I get how it works. But how do I move from expertise to execution?

As you can see, mass is already at the bottom. (Heat shield and payload bay.) And I can't put parts like goo cans up top on the capsule, because they explode without heat shielding. Does anyone have actual examples of working re-entry vehicles to share? I'm having a hard time believing I'm the only person who has a problem.

Unfortunately, you can't argue with physics. You have too much mass high up, simple as that. Maybe add a wide, flat procedural tank just above the heat shield and add lead ballast until the CoM moves to the other side of the CoL. Short of that, I don't really have a solution for you that hasn't already been given here in this thread.

[Jovus]

So, do others just not bother trying to return Science Jr.'s?

[jrandom]

So, do others just not bother trying to return Science Jr.'s?

OR IMPLEMENT THE VARIOUS SOLUTIONS GIVEN IN THIS THREAD. Holy cow. It's an engineering problem.

Edit: A science jr. under the Mk1 + DRE heat shield shows that it should be reenterable:

Javascript is disabled. View full album

If you don't have DRE, the shield mass might be to low, which would put the CoM on the wrong side of the CoL, hence my recommendation of using Procedural Parts to put a lead ballast tank just above the heat shield to manually place your CoM where it should be.

Edited June 4, 2015 by jrandom

[Jovus]

OR IMPLEMENT THE VARIOUS SOLUTIONS GIVEN IN THIS THREAD. Holy cow. It's an engineering problem.

...whoa. It was an honest question, not a snippy one. I'm curious, given what Wanderfound said and the apparent lack of experience in common with mine.

[jrandom]

...whoa. It was an honest question, not a snippy one. I'm curious, given what Wanderfound said and the apparent lack of experience in common with mine.

Apologies. It looked snippy in light of the rest of the thread. Darned text don't carry tone very well. I'm going to be attempting a reentry of the above pod in a moment to verify.

[Jovus]

I understand where' you're coming from on the text/tone issue, so no harm done. I'm beginning to suspect a bum install or something similar on my end, so let me know how your re-entry goes and I'll go digging around in my logs and debug panel to see if I can't find something odd. 'cause long poles shouldn't be aerodynamically stable when set at 40 degrees to the airflow...

[futrtrubl]

As you can see, mass is already at the bottom. (Heat shield and payload bay.)

There's the problem, those aren't massive, especially compared to the pod, that is why your CoM is closer to the pod than the shield. IE mass is not at the bottom.

[jrandom]

I understand where' you're coming from on the text/tone issue, so no harm done. I'm beginning to suspect a bum install or something similar on my end, so let me know how your re-entry goes and I'll go digging around in my logs and debug panel to see if I can't find something odd. 'cause long poles shouldn't be aerodynamically stable when set at 40 degrees to the airflow...

Alright, it wants to flip around pointy-end forward, but that's what I'd expect from a re-entry vehicle of that shape. The CoL, while being on the correct side of the CoM, is too close to the CoM and the lighter science jr. at the bottom is what's doing it. It's still re-enterable so long as you keep aimed as retrograde as possible (eg. using SAS -- and I've severely nerfed the Mk1 SAS on my install because I feel the stock value is too strong to be even remotely realistic). In short, this design is just marginally workable. Use drag fins on the pointy end, or a larger heat shield. In the below image, take note of the where the CoL is. That's the sort of positioning you want in a reentry vehicle.

Javascript is disabled. View full album

[Jovus]

Thanks for your help, Jrandom. One further thing I'm curious about: when you flew your test pod (without the larger heat shield) did it orient itself directly prograde when you left it alone, or did it form some angle with the prograde vector?

[jrandom]

Thanks for your help, Jrandom. One further thing I'm curious about: when you flew your test pod (without the larger heat shield) did it orient itself directly prograde when you left it alone, or did it form some angle with the prograde vector?

Mostly prograde. I don't remember if it slewed in at an angle.

[Gryphon]

I've been contemplating a crazy solution to have fins for reentry (but not during launch): With KIS, I could have them on my upper stage, oriented for launch (Or just stored). on orbit, my engineer would EVA and move them to the reentry-oriented position. Of course, placement may not be entirely precise...

Also, PAGE 1000!

[Nerd1000]

I've been contemplating a crazy solution to have fins for reentry (but not during launch): With KIS, I could have them on my upper stage, oriented for launch (Or just stored). on orbit, my engineer would EVA and move them to the reentry-oriented position. Of course, placement may not be entirely precise...

Also, PAGE 1000!

Use infernal robotics- attach your fins to hinges and clip them into the ship so that you can hide them inside the ship for launch and extend them later. placement problem solved.

[Wanderfound]

I've been contemplating a crazy solution to have fins for reentry (but not during launch): With KIS, I could have them on my upper stage, oriented for launch (Or just stored). on orbit, my engineer would EVA and move them to the reentry-oriented position. Of course, placement may not be entirely precise...

Also workable: fins that act as fins on the way up, but flip sideways (max-deflection spoilers, fins placed in pairs so the deflections oppose) to act as airbrakes on the way down. That usually requires a nose-first reentry, but an inline cockpit fronted with a heatshield/decoupler/nosecone (nosecone for ascent, heatshield for descent) works well for that.

[Jovus]

So, it turns out, for those watching at home, that even with just FAR (and Hyperedit) installed, a Mk1 pod, parachute, 3 Science Jr's, a payload bay and a heat shield is stable at about 40 degrees (!) to the airflow. It doesn't look like an obvious bug (no voxelization errors, nothing in the debug log). It's just a very surprising aerodynamic feature.

[DKnight54]

Hi Ferram,

first off, thanks for all the hard work you've put into FAR. That being said, I've noticed a problem with using Procedural Parts for cones with FAR, and haven't notice anyone mentioning it before

I've noticed it in the last few Dev Builds of FAR, using the latest version of Procedural Parts.

You can see it using voxel debugging in the editor.

I tested this on a clean install of 64-bit KSP 1.0.2 on Linux with only modularflightintegrator, most recent FAR dev build, procedural parts and ModuleManager 2.6.5 and was able to recreate the issue.

To recreate, just attach any of the procedural parts and select cone and set the tip to 0 or use the procedural cone directly.

Javascript is disabled. View full album

Log files look clear with no errors, but attached anyway.

https://www.dropbox.com/s/0rdqzdac6zx9r97/KSP.log?dl=0

I don't think it's a problem directly from procedural parts because there was a dev build where the problem went away, but latter came back in a latter build. IIRC, the build without problems should have been the one before the negative scale mirrorRefAxis fix.

Edited June 4, 2015 by DKnight54
additional info

[Kitspace]

Hello!

I am experiencing a strange behaviour with stability graphs.

I build a plane fuselage starting with the root part and run the analysis that gives certain graphs.

Then I attach a symmetric wing part to said fuselage and run the sweep again resulting in a different graph.

If I then remove that wing part and place it again in the very same spot in exactly the same way it was placed in the step two the result of the stability sweep will be considerably different from both steps one and two.

No matter how precisely the wing position is repeated between steps two and three the static stability graphs will be different.

Voxels look just the same at all times.

Did anybody encounter that kind of thing?

I have encountered that several times and those generic reproduction steps seem to work reliably.

I am going to make a full report with pictures.

Thank you.

[Ryds]

Hello,

I posted this in the DRE thread but thought it might help the flipping craft discussion from a couple of pages back.

I don't think its reasonable to expect a craft that's 4 science juniors long to behave itself whilst crashing through the atmosphere at 2km/s. The space shuttle is the longest re-entry vehicle we've launched and that needs ablative shielding all along it and must follow a strict re-entry profile.

The craft flipping is annoying but as someone either in this thread or in FAR said, KSP is about engineering, so find an engineering solution to it.

Here's mine. I've used 4 A.I.R.B.R.A.K.E.S and placed them so that the don't cover the EVA hatch. To be honest I've got 50 mods including B9 so I'm not sure if these are stock or not or where they come in the community tech tree (I used sandbox to find something suitable).

When flat at launch they don't overpower the atmo stage wings, although I found I needed to add an extra pair (6 instead of 4) to make sure the CoL was low enough.

At re-entry I open them as soon as I hit atmo. I'm then able to descend using only SAS at a cost of 0.005 EC instead of retrograde hold @0.7 EC. You need to see it down manually tho as if you deviate from the vector indicator by more than a few degrees you'll go poof.

I'd suggest moving this lower down the tech tree, but I'm not a modder and don't know even how to do that. Maybe a someone with the modding chops could pick it up?

Hope this helps, I'm pleased to finally offer a small solution for once instead of just reporting bugs!

[Kitspace]

I have a few cases of abnormal behaviour to report.

First I confirm that center of lift for some reason moves when the fuel load changes.

Second I am experiencing an issue with asymmetric center of lift without any procedural parts at all.

I believe that these two are visual issues with the marker and do not affect the craft behaviour but not completely sure.

And finally the stability sweep issue described a few hours ago.

See the pictures below for description of steps and what happens.

All three issues are reproduced on the same simlpe test craft.

Javascript is disabled. View full album

Edited June 4, 2015 by Kitspace

[Gryphon]

I have a few cases of abnormal behaviour to report.

First I confirm that center of lift for some reason moves when the fuel load changes.

Second I am experiencing an issue with asymmetric center of lift without any procedural parts at all.

I believe that these two are visual issues with the marker and do not affect the craft behaviour but not completely sure.

And finally the stability sweep issue described a few hours ago.

See the pictures below for description of steps and what happens.

All three issues are reproduced on the same simlpe test craft.

You might wish to include a couple screenshots showing the FAR graph that overlays on your ship, with the changes you are making, instead of the COL ball. As I understand it, you shouldn't put too much credence in the stock COL indicator, since it's no longer intended to be the primary tool to use when designing a ship with FAR. (Ferram has posted several times about it in the past, if you look (or search) back some pages in this thread.