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

[S1gmoid]

- - - Updated - - -

http://imgur.com/a/FIXKk

CoL isn't displaying properly, i can get it to work randomly by spamming Raise/Lower gear. Read about a CoL update button but could not find.

1.0.5 "Helmbold". Clean install(aside from Blizzy's toolbar), not CKAN, not 64 bit.

If there's a quick fix i would love that, spent half the day losing my mind on Spaceplanes because the CoL was ....ing with me.

Indeed! I second this bug, it's happening for me as well. Either that, or the center of lift calculation is somehow off? When I put a basic first tech node crew capsule on the rocket, I get a HUGE effect on both the FAR profiles, and the COL too. Back in 0.2x FAR, when I put 6 fix on a rocket, it always brought the COL down nice... Not any longer, as long as I have a pod on.

Is this as intended? There used to be separators that meshed nicely with the heatshield on the capsule, but I can't find those parts any longer... but I wonder if that miniscule gap warrants such a gigantic effect.

Also, this might not be a FAR issue, the entire VAB seems awfully wonky... For example the engineer report comes up with absolutely wrong alerts all the time, like no command module (when I do have a command module), no parachute (when I do have a parachute), etc.

Edited November 10, 2015 by S1gmoid

[ferram4]

Yes, I'm aware the CoL is wrong initially, I don't know what causes it. The CoL after the reload is correct though. Considering that it will be automatically taken care of when the vehicle is updated by attaching / removing parts, I'm not too concerned. In addition, as already noted, CoL is less accurate than the graphs; use the damn graphs, that's why they're there.

randman2222, as for you, you need to provide reproduction steps and logs. I haven never been able to cause any issues like that in the current build, which means that something you're doing is different so you have to tell me or else there's nothing I can do. I think you just installed it wrong, tbh.

[paul23]

I've yet to learn to use the stability derivates correctly lol. Anyone got quick pointers for the 18 derivates and how to most easily change a single/couple without making the rest worse?

[capi3101]

I've yet to learn to use the stability derivates correctly lol. Anyone got quick pointers for the 18 derivates and how to most easily change a single/couple without making the rest worse?

Pointers are on the FAR wiki - start here. Basically, anything red on the static tab is bad and needs testing on the dynamic simulation tab for sure. Test the value of the sub-script; a value of 1 is sufficient in most cases. The behavior of the graph tells you things - a slowly increasing amplitude in an oscillating curve is bad, and a curve that shoots off to Kraken-knows-where is even worse. An oscillating graph that stays the same or damps down is good. Note that you should test M_w and L_ß even if you're green across the board. As for testing itself, use 0m/Mach 0.35, 10,000m/Mach 1 and 20,000m/Mach 3 as an initial set of benchmarks, at least until somebody around here gives you a far better set...

That answers half your question. As to how "most easily change a single/couple without making the rest worse", it takes experimentation, and i haven't figured all of the details out there my own self. Let's just say don't change something you can't undo, and sorry for the non-answer there.

[Wanderfound]

I've yet to learn to use the stability derivates correctly lol. Anyone got quick pointers for the 18 derivates and how to most easily change a single/couple without making the rest worse?

You can design perfectly well just by eyeball and flight testing; the flight data isn't absolutely necessary. But it does help if you can use it.

The first screen in Ferram is the static analysis screen. This gives you pretty graphs.

It has two options at top: Sweep AoA and Sweep Mach. There are boxes at the bottom which say Lower, Upper and Mach/AoA. This page can produce two different graphs: if you press the Sweep AoA button, it shows the behaviour of your plane from Angle of Attack values between Lower and Upper, at the speed shown in the Mach/AoA box. If you press the Sweep Mach button, it shows behaviour at speeds between Lower and Upper at the AoA shown in the Mach/AoA box.

The blue line is the Coefficient of Lift. It's good when this is high.

The red line is the Coefficient of Drag. It's good when this is low.

The yellow line is the Coefficient of Manoeuvrability/Instability. You want this to be angling down (like it is here), and it's best that the slope of the line isn't too steep and that its X-intercept (where the yellow line crosses zero) isn't too far above zero.

The green line is lift divided by drag. It's good when this is high.

This picture shows how the plane will act at Angles of Attack between 0° and 25° while travelling at Mach 2.

Sometimes the lines split into two lines. This shows how the plane responds after a stall: you get a sudden loss of lift and increase in drag that lasts until you return your AoA to where the line isn't split any more.

This is the same picture at Mach 0.8. See how the plane can stall at that speed?

If you click the Sweep Mach button, you instead get a look at a bunch of different speeds with Angle of Attack held constant. This shows Mach 0-6 with a 3° AoA. The bumpiness on the left shows the effect of breaking through the sound barrier.

The second page of Ferram is data and stability derivatives. This produces scary looking numbers.

To get those numbers, you need to put in values for altitude and speed.

All of the confusing letters that appear when you hover your mouse over the output numbers relate to this picture here:

x is forwards, y is sideways, z is down. P is roll, Q is pitch, R is yaw. Don't worry about the Greek for now.

If you hover your mouse over any of the numbers, it'll pop up a tooltip explaining what it refers to. Mostly, however, all you want to do is make as many as possible of the numbers green and as few as possible red.

The one other useful thing on this screen is the "level flight" stuff up top right. If you set the analysis for zero altitude and the speed for whatever you think you can reach on the runway, you can find out how much AoA you need to take off (the "level flight" value). Try to keep that number below ten for easy takeoffs.

Transonic drag and Area Ruling

Go to transonic design and toggle the area curves. The green line shows the cross-sectional area of the aircraft as you move from nose to tail; the yellow line is a measure of how smoothly the green line is changing. Mach 1 Wave-Drag Area is the thing that you're aiming to minimise; if you manage to get the wiggles in the yellow line to the smallest amplitude possible, that is what you'll achieve. The lower the wave-drag, the less power you'll need to get up to speed.

With all of these analyses, you should have your gear toggled up. See how much they affect the outcome here:

How to Apply FAR Aero Analyses

The pretty graphs: use these to check for excessive drag and regions of instability. AoA graph is the more useful of the two.

Numbers: use the Level Flight figure to work on your takeoff speed and flap settings. For the stability figures, check takeoff (0 altitude and .35 or so speed), low altitude (5km and .8 speed), Mach 1 (speed obvious, 10km) and edge of rocketry (25km, Mach 4.5). You want green everywhere. You won't get it, at least to start with.

Hover your mouse over the red ones and use the picture above to work out what the tooltips mean. Once you decode the x's and y's, usually they just mean something like "has a tendency to roll when pitching up" or similar. Sometimes the solutions are obvious (e.g. too much yaw slippage, add a rudder), sometimes they take a great deal of trial and error to sort out.

With the transonic drag stuff, just wiggle parts about and see what happens. To achieve the lowest drag, "fill in" dips in the green line and "shave down" peaks.

TLDR version: if you've got pitch problems, shift lift rearwards and CoM forwards. If you've got yaw problems, beef up your vertical stabiliser and shift CoM forwards. If you've got roll problems, increase wingspan or add some dihedral. If you're just falling out of the sky, lose some weight and add some lift.

Edited November 10, 2015 by Wanderfound

[S1gmoid]

Indeed! I second this bug, it's happening for me as well. Either that, or the center of lift calculation is somehow off? When I put a basic first tech node crew capsule on the rocket, I get a HUGE effect on both the FAR profiles, and the COL too. Back in 0.2x FAR, when I put 6 fix on a rocket, it always brought the COL down nice... Not any longer, as long as I have a pod on.

Is this as intended? There used to be separators that meshed nicely with the heatshield on the capsule, but I can't find those parts any longer... but I wonder if that miniscule gap warrants such a gigantic effect.

Screenshot here. This seems totally different from what I remember from 0.2x... And it doesn't even look correct, but please correct me if I'm wrong.

[Kagame]

CoL is working badly on mine too.

[ferram4]

@S1gmoid: I don't see anything particularly wrong with that, though perhaps the smoothing function is too generous near the back of the rocket. I see that you haven't activated the CoL though, so I don't know what you're complaining about; everything looks fine.

@Kagame: Thank you for repeating what has already been acknowledged. I supposed you have some extra info on the issue that you'd like to share?

[Van Disaster]

Derivatives: if you've got something plane-shaped with enough control authority most of them will be green anyway. The vital one is Mw, because it's an indicator of where your CoL is relative to CoM ( actually it's if your plane has a natural pitch down moment or not at this speed/altitude, but that's the result ). Usually other problems will turn up at high AoA & be yaw/sideslip-related & solvable with surfaces with a vertical component - where & what is pretty much design dependent. Zu will go red occasionally around 1.2M ( harmless ), I don't think I've ever seen any of the pitch-rate or roll-rate ones turn red.

[ferram4]

Okay, so good news: I know what causes the CoL issue, and it's just an inconsistency with wings. It is accurate for anything without wings, with wings it underestimates their effects. The graphs are still correct though, so make use of that if you can.

[paul23]

TLDR version: if you've got pitch problems, shift lift rearwards and CoM forwards. If you've got yaw problems, beef up your vertical stabiliser and shift CoM forwards. If you've got roll problems, increase wingspan or add some dihedral. If you're just falling out of the sky, lose some weight and add some lift.

You can design perfectly well just by eyeball and flight testing; the flight data isn't absolutely necessary. But it does help if you can use it.

The first screen in Ferram is the static analysis screen. This gives you pretty graphs.

It has two options at top: Sweep AoA and Sweep Mach. There are boxes at the bottom which say Lower, Upper and Mach/AoA. This page can produce two different graphs: if you press the Sweep AoA button, it shows the behaviour of your plane from Angle of Attack values between Lower and Upper, at the speed shown in the Mach/AoA box. If you press the Sweep Mach button, it shows behaviour at speeds between Lower and Upper at the AoA shown in the Mach/AoA box.

The blue line is the Coefficient of Lift. It's good when this is high.

The red line is the Coefficient of Drag. It's good when this is low.

The yellow line is the Coefficient of Manoeuvrability/Instability. You want this to be angling down (like it is here), and it's best that the slope of the line isn't too steep and that its X-intercept (where the yellow line crosses zero) isn't too far above zero.

The green line is lift divided by drag. It's good when this is high.

http://i.imgur.com/rPK07yh.png

This picture shows how the plane will act at Angles of Attack between 0° and 25° while travelling at Mach 2.

Sometimes the lines split into two lines. This shows how the plane responds after a stall: you get a sudden loss of lift and increase in drag that lasts until you return your AoA to where the line isn't split any more.

This is the same picture at Mach 0.8. See how the plane can stall at that speed?

http://i.imgur.com/M6CfRoR.png

If you click the Sweep Mach button, you instead get a look at a bunch of different speeds with Angle of Attack held constant. This shows Mach 0-6 with a 3° AoA. The bumpiness on the left shows the effect of breaking through the sound barrier.

http://i.imgur.com/xLkx98L.png

The second page of Ferram is data and stability derivatives. This produces scary looking numbers.

http://i.imgur.com/JmXc3vw.png

To get those numbers, you need to put in values for altitude and speed.

All of the confusing letters that appear when you hover your mouse over the output numbers relate to this picture here:

http://www.mathworks.com/help/aerotbx/ug/bodycs.gif

x is forwards, y is sideways, z is down. P is roll, Q is pitch, R is yaw. Don't worry about the Greek for now.

If you hover your mouse over any of the numbers, it'll pop up a tooltip explaining what it refers to. Mostly, however, all you want to do is make as many as possible of the numbers green and as few as possible red.

The one other useful thing on this screen is the "level flight" stuff up top right. If you set the analysis for zero altitude and the speed for whatever you think you can reach on the runway, you can find out how much AoA you need to take off (the "level flight" value). Try to keep that number below ten for easy takeoffs.

Transonic drag and Area Ruling

http://i.imgur.com/QMJY7km.png

Go to transonic design and toggle the area curves. The green line shows the cross-sectional area of the aircraft as you move from nose to tail; the yellow line is a measure of how smoothly the green line is changing. Mach 1 Wave-Drag Area is the thing that you're aiming to minimise; if you manage to get the wiggles in the yellow line to the smallest amplitude possible, that is what you'll achieve. The lower the wave-drag, the less power you'll need to get up to speed.

With all of these analyses, you should have your gear toggled up. See how much they affect the outcome here:

http://i.imgur.com/xtV7a77.png

How to Apply FAR Aero Analyses

The pretty graphs: use these to check for excessive drag and regions of instability. AoA graph is the more useful of the two.

Numbers: use the Level Flight figure to work on your takeoff speed and flap settings. For the stability figures, check takeoff (0 altitude and .35 or so speed), low altitude (5km and .8 speed), Mach 1 (speed obvious, 10km) and edge of rocketry (25km, Mach 4.5). You want green everywhere. You won't get it, at least to start with.

Hover your mouse over the red ones and use the picture above to work out what the tooltips mean. Once you decode the x's and y's, usually they just mean something like "has a tendency to roll when pitching up" or similar. Sometimes the solutions are obvious (e.g. too much yaw slippage, add a rudder), sometimes they take a great deal of trial and error to sort out.

With the transonic drag stuff, just wiggle parts about and see what happens. To achieve the lowest drag, "fill in" dips in the green line and "shave down" peaks.

Well for me it's Zw and Xw that are unstable (around 1 positive)

Edited November 11, 2015 by paul23

[Stop_signGEORGE]

It seems to me like the stock aerodynamics are still in use. If I put an MK16 parachute at the bottom of the craft, the CoL shifts down. And it doesn't where I put my wings, it will always move the lift marker the same amount.

[Van Disaster]

Well for me it's Zw and Xw that are unstable (around 1 positive)

At what speed/altitude & consequently what AoA? and at the risk of getting this moved to the other FAR thread, pic?

Xw is signless, if you got that red then you've done something quite spectacular. Zw +ve implies any positive change in downward velocity will result in downward acceleration increasing - which would rather imply your overall drag gets lower as you go faster downwards into thicker atmosphere, which I guess isn't impossible but also might mean something is upsidown, perhaps also might imply I'm reading it wrong.

Edited November 11, 2015 by Van Disaster

[cantab]

So I think wrong Zw means that as your angle of attack increases you're generating less lift instead of more. That suggests you're doing something odd. I advise having your fuselage level when doing FAR analyses, beyond that a picture would help.

[paul23]

Hmm might be then that I'm using the VAB and a vertical position. (I wish to make a gliding decent to base, but otherwise fully rocket-engine based). This however puts a high strain on the stability: since I don't have any fuel left to burn.

[cantab]

Provided everything matches with the default orientation in the VAB you should be fine, but it's even easier to mix the orientations up and confuse FAR. Design the gliding part in the SPH if you can.

[S1gmoid]

@S1gmoid: I don't see anything particularly wrong with that, though perhaps the smoothing function is too generous near the back of the rocket. I see that you haven't activated the CoL though, so I don't know what you're complaining about; everything looks fine.

It might just be a mistake or misunderstanding on my part... yea I forgot the COL, but it seems to roughly correspond to the curvature (on this rocket it's shown to be slightly above center of mass). Now as I remember, back in 0.2x, if I put 4-6 fins on a rocket, it brought the COL right down to the fin. Here, it barely budges, and the curvature of the fins is much smaller than the curvature at the cabin, which might or might not be right. (Is there a good guide for doing transonic design for rockets? In .2x I thought I had it down, but now I seem to be having a harder time.)

[cantab]

Well rockets are normally pretty cylindrical anyway. Just make sure you have smooth lines - use fairings or service/cargo bays where needed, and nosecones on your boosters. Since rockets also have a lot of thrust any transonic drag hump doesn't matter as much either, you can just blast through it. Personally I've virtually never used the FAR analysis tools for a rocket, the design, flight profile, and stability requirements are much simpler than a plane.

[S1gmoid]

Well rockets are normally pretty cylindrical anyway. Just make sure you have smooth lines - use fairings or service/cargo bays where needed, and nosecones on your boosters. Since rockets also have a lot of thrust any transonic drag hump doesn't matter as much either, you can just blast through it. Personally I've virtually never used the FAR analysis tools for a rocket, the design, flight profile, and stability requirements are much simpler than a plane.

Yea, the rule of thumb I got used to was "center of mass up, center of lift down, and it will fly okay".

[tetryds]

There is a better rule: do not make fairings larger than the body of the rocket right bellow it.

If you do, keep it to a minimum.

[paul23]

There is a better rule: do not make fairings larger than the body of the rocket right bellow it.

If you do, keep it to a minimum.

You know that won't when trying to lift huge bases off the ground. Those are way wider than the largest body.

[taniwha]

paul23: why are you lifting huge bases off the ground?. There are better ways.

[Weyth]

I got problem with center of lift this is what it shows with FAR http://www.imagehosting.cz/?v=farlnl.png

and this what it shows stock http://www.imagehosting.cz/?v=stockkfk.png . Why FAR makes stock planes unstable? Why is FAR so different than Stock?

[Van Disaster]

I got problem with center of lift this is what it shows with FAR http://www.imagehosting.cz/?v=farlnl.png

and this what it shows stock http://www.imagehosting.cz/?v=stockkfk.png . Why FAR makes stock planes unstable? Why is FAR so different than Stock?

Use the graphs rather than the CoL ball, it's pretty unreliable. Why is FAR so different? because FAR is a pretty detailed model of actual aerodynamics ( obviously not at the full CFD level ), and stock is a coarse approximation.

@paul23: make narrower bases? make bases out of multiple stacks & plug them together with KAS? and/or as Taniwha said, build on site, or in orbit. You can launch ridiculous stuff in a fairing if you make the rocket broad too, of course.

[tetryds]

You know that won't when trying to lift huge bases off the ground. Those are way wider than the largest body.

You cannot build something inherently unstable and expect an easy solution.

Develop your own techniques to fly that if you want, all that FAR will do is what it was designed for, simulate the aerodynamic effects on your ship as accurately as possible.

It's not like you can't make anything fly, you just have to figure out how.