That thing called drag...

[ShadowZone]

First off: this is not me asking help for a specific craft or problem with a build, this is my trying to understand how a certain physics principle is working in this game with the help of this example:

 

Maybe I misunderstand how aerodynamics should work or the game behaves in a way that I have no idea how to comprehend. Either way, thing is this: I have a vehicle with 4100 m/s of dV and enough TWR but cannot reach orbit. The reason for this appears to be drag related. However, I really dont' know why. I was under the impression, parts behind other parts would be occluded and therefore protected from the airstream (i.e. during reentry) and should also not produce drag. However, as you can see from my screenshots, there is apparently A LOT of drag being caused by tiny parts in the back.

Can anybody explain to me, why this drag is occuring? Is there any way to avoid generating drag by putting them behind other parts? Fairings are not an option since this vehicle is supposed to visit many planets and moons during its mission.

 

 

Edited December 7, 2015 by ShadowZone

[panzer1b]

Drag is now oversimplified and the ONLY way to shield parts is to have them stacked one atop the other (as in connected by the nodes).  Anything else regardless of whether its clipped inside, or placed behind another parts suffers FULL drag as if it was exposed to the entire airflow.

 

Right now, the only way to actually kill drag on radially attached crap uis either stuffing them inside a bomb bay or a service bay (those disable all drag forces on anything inside), or alternatively place them in a fairing (which has its problems like the inability to reuse).  Until either squad alters the aero to take into account what is behind a part, or you choose to install FAR mod (which models drag much more realistuically), you are going to have to use as little radial attachments as possible.

[ShadowZone]

10 minutes ago, panzer1b said:

Drag is now oversimplified and the ONLY way to shield parts is to have them stacked one atop the other (as in connected by the nodes).  Anything else regardless of whether its clipped inside, or placed behind another parts suffers FULL drag as if it was exposed to the entire airflow.

 

Right now, the only way to actually kill drag on radially attached crap uis either stuffing them inside a bomb bay or a service bay (those disable all drag forces on anything inside), or alternatively place them in a fairing (which has its problems like the inability to reuse).  Until either squad alters the aero to take into account what is behind a part, or you choose to install FAR mod (which models drag much more realistuically), you are going to have to use as little radial attachments as possible.

well... that's a bummer - und puts quite a problem in my current mission plan.

[KerikBalm]

While drag is unrealistic now, I like that it prevents massive abuse from part clipping.

(though you can still abuse part clipping to make a lot of stuff lift in a little cargo bay/fairing).

Compression heating does work like you were hoping though...

[n.b.z.]

1 hour ago, ShadowZone said:

I was under the impression, parts behind other parts would be occluded and therefore protected from the airstream (i.e. during reentry) and should also not produce drag.

Is my subjective impression correct that this works differently for drag and heating? While I think that panzer1b's explanation is entirely correct for drag, I have a gut feeling that "shielding" things behind other parts may work to protect from heating.

I'm basing this on the fact that I have managed to fly externally mounted Gigantor XL's on spaceplanes, and return said planes from orbit to the KSC without ablating the XL's away. I seem to remember that this required some fiddling with their position.

[Windspren]

1 hour ago, panzer1b said:

Drag is now oversimplified and the ONLY way to shield parts is to have them stacked one atop the other (as in connected by the nodes).  Anything else regardless of whether its clipped inside, or placed behind another parts suffers FULL drag as if it was exposed to the entire airflow.

 

Right now, the only way to actually kill drag on radially attached crap uis either stuffing them inside a bomb bay or a service bay (those disable all drag forces on anything inside), or alternatively place them in a fairing (which has its problems like the inability to reuse).  Until either squad alters the aero to take into account what is behind a part, or you choose to install FAR mod (which models drag much more realistuically), you are going to have to use as little radial attachments as possible.

I didn't know that, so I guess I'll be using FAR when 1.1 comes out...

[KerikBalm]

16 minutes ago, n.b.z. said:

Is my subjective impression correct that this works differently for drag and heating? While I think that panzer1b's explanation is entirely correct for drag, I have a gut feeling that "shielding" things behind other parts may work to protect from heating.

I'm basing this on the fact that I have managed to fly externally mounted Gigantor XL's on spaceplanes, and return said planes from orbit to the KSC without ablating the XL's away. I seem to remember that this required some fiddling with their position.

Yes, it does work differently for heating

[theend3r]

Why not try FAR? Things work like they should in FAR.

[Frozen_Heart]

1 hour ago, panzer1b said:

Drag is now oversimplified and the ONLY way to shield parts is to have them stacked one atop the other (as in connected by the nodes).  Anything else regardless of whether its clipped inside, or placed behind another parts suffers FULL drag as if it was exposed to the entire airflow.

 

Right now, the only way to actually kill drag on radially attached crap uis either stuffing them inside a bomb bay or a service bay (those disable all drag forces on anything inside), or alternatively place them in a fairing (which has its problems like the inability to reuse).  Until either squad alters the aero to take into account what is behind a part, or you choose to install FAR mod (which models drag much more realistuically), you are going to have to use as little radial attachments as possible.

Well its still so much better than the old souposphere...

[Red Iron Crown]

The pictured craft would generate far less drag if it used cylindrical tanks instead of Mk2 adapters. The Mk2 parts generate a lot of lift (and associated drag), which isn't helping the type of craft shown.

[n.b.z.]

25 minutes ago, KerikBalm said:

Yes, it does work differently for heating

Thanks!

 

23 minutes ago, theend3r said:

Why not try FAR?

For some reason, I tend to stay very close to stock. I don't even know why, which is a bit alarming considering that this has kept me from using even KER!

[problemecium]

3 hours ago, panzer1b said:

Drag is now oversimplified and the ONLY way to shield parts is to have them stacked one atop the other (as in connected by the nodes).  Anything else regardless of whether its clipped inside, or placed behind another parts suffers FULL drag as if it was exposed to the entire airflow.

This isn't entirely true. I've found that parts clipped inside other parts end up being shielded (not that I've ever exploited this - it feels so very cheaty), and parts near the back of a vessel seem to generally be affected less by re-entry heating, so I presume they'll make less drag as well.
Still, all the above suggestions are valid. Also try redesigning it so those side pods are all facing straight forward - fuel tanks (and most other parts) pick up a lot of extra drag when oriented at an angle or sideways compared to straight forward.

[Red Iron Crown]

Reentry heating and drag use different occlusion rules. Drag is like panzer1b says, stack occlusion and containment within a bay, compartment, or fairing are the only way to hide something from drag, clipping has no effect. Reentry heating determines leading edges and makes a conical sort of shape behind them where heating is greatly reduced.

[ShadowZone]

3 hours ago, Red Iron Crown said:

The pictured craft would generate far less drag if it used cylindrical tanks instead of Mk2 adapters. The Mk2 parts generate a lot of lift (and associated drag), which isn't helping the type of craft shown.

Probably, but this looks better

26 minutes ago, Red Iron Crown said:

Reentry heating and drag use different occlusion rules. Drag is like panzer1b says, stack occlusion and containment within a bay, compartment, or fairing are the only way to hide something from drag, clipping has no effect. Reentry heating determines leading edges and makes a conical sort of shape behind them where heating is greatly reduced.

I have no idea how drag would behave in the real world, but I would imagine there would (should?) be more occlusion than in this case. This completely negates the possibility to avoid drag other than using bays or fairings. Which is actually quite annoying the more I think about it

[samstarman5]

Drag not only increases in higher air pressure, but also to the faster you are going. In lower altitudes it pays to take it easy and not go full throttle.

[Motokid600]

3 hours ago, n.b.z. said:

Thanks!

 

For some reason, I tend to stay very close to stock. I don't even know why, which is a bit alarming considering that this has kept me from using even KER!

Treat yourself... A few quality of life mods won't jeopardize that "stockalike" thing people rave about for some reason or other.

[Chaos_Klaus]

I think you guys are going about this the wrong way. You see a large number of red arrows and assume that there is a lot of drag. Acutally, many of these arrows are quite short. Even with an unaerodynamic rocket that has 4100m/s of delta v, you should be able to reach orbit no matter what. I guess the problem is with ascent profile?

 

6 hours ago, panzer1b said:

Drag is now oversimplified

 

What do you mean ... "now"? Before 1.0 there was no occlusion at all. Every part was subject to drag as if it was flying around alone and the size of the part was estimated by its mass. A full fueltank would cause more drag then an empty one. That's what I call oversimplification.

 

Instead what you see in the picture is perfectly sane physics. The outer tanks are angled. And the sides of these stacks are exposed to the air stream at an angle. Naturally, this creates drag.

Drag also occurs with parts that are exactly behind other parts in a stack. That is because drag is induced by all faces that are exposed to moving air. Obviously that applies to faces that are pinted into the airstram, but it also is true for all the faces at the rear of the vehicle. Actually most of the drag we mean when whe say "drag" is actually caused by the low air pressure *behind* the vehicle. Air moving over a surface causes skin drag. That even applies to faces that are perfectly parallel to the airstream.

Occlusion only happens when two faces are ontop of another. The top face of a tank won't cause drag if it is stacked behind another part of the same size. All the other face will however still create drag.

It's more complicated then you think and KSP's aero model does a relatively good job. Remember that in real physics we make many seemingly stupid simplifications. Like talking about point masses. But if it works, it works.

In cases where you clip parts into one another and build things that are not quite so rocket shaped, the model struggles. If you want something more sophisticated, use FAR. But it's more sophisticated and that's the price you pay.

 

Edited December 6, 2015 by Chaos_Klaus

[n.b.z.]

2 hours ago, Motokid600 said:

Treat yourself... A few quality of life mods won't jeopardize that "stockalike" thing

No question about it - I could not live without Hyperedit, and have a KOS script that can fly some of my Rapier-based planes from standstill to circularization. (The only thing left for me to do during ascent is to grab another coffee. Again.)

To get myself back on topic... I think it's worth pointing out that this statement...

9 hours ago, panzer1b said:

the ONLY way to shield parts is to have them stacked one atop the other (as in connected by the nodes)

...means that it is hugely important to avoid exposing open nodes to the air, unless that can't be avoided for some good reason.

[KerikBalm]

I try to stay "stock"... and I use KER... I just got it for 1.05 2 days ago.

Its so much easier than calculating dV manually. Before KER, I'd look in the map view, note my mass, look in the resources tab - take my remaining oxidizer (assuming an excess of LF), multiply by 2/1.1, multiply by 0.005, subtract that mass from my total mass... take the reciprical (1/that amount), multiply by total mass, take the natural log, * 9.81 * Isp... and I had my dV (assuming a single stage, like a spaceplan... it also works for tugs).

Multi-stage designs had me doing much more math in the VAB. KER doesn't make my design not work on another player's computer... I don't know of any challenge that won't allow use of KER.

Hyperedit... well, i got that too, its way cheaty... but only if you actually use it, its fun for testing things on other bodies... and before I got it, I was manually editing the orbit and reference body parameters by text.... so hyperedit also doesn't really add anything that you can't do in the basic stock game, it just makes it easier.

[ShadowZone]

9 hours ago, Chaos_Klaus said:

I think you guys are going about this the wrong way. You see a large number of red arrows and assume that there is a lot of drag. Acutally, many of these arrows are quite short. Even with an unaerodynamic rocket that has 4100m/s of delta v, you should be able to reach orbit no matter what. I guess the problem is with ascent profile?

I tried:

• Full throttle (TWR at 10km is about 2.5 and > 5.0 at 50km), start tipping at around 1500m, 45° at 9000m, 30° at 25000m, 20° at 35000m --> max Ap around 65km
• Full throttle 90° until 20km, then tip for gravity turn --> Ap above 70km, about 300m/s short for circulization burn
• Kept TWR around 1.75 at 90° until 20km, then tip for gravity turn --> Ap above 70km, about 300m/s short for circulization burn
• Full throttle until 6km, then TWR below 1.85 at 90° until 15km, slowly tip for gravity turn --> Ap about 70km, about 250m/s short for circulization burn

Maybe I am missing something? I was able to get a 400 ton ring station into orbit without any aerodynamics to speak of by keeping TWR low and the ascent to 90° until 20km and then starting my "turn". So imagine my frustration not being able to get this little thing into orbit

 

[Farex]

Could you please provide us with the craftfile so we can try it out? This way we may actually have an idea to help.

 

Edited December 7, 2015 by Farex

[Chaos_Klaus]

Hm. Then maybe the problem really is with the angled mk2 parts providing lift and a ton of drag. 

[Valerian]

1 hour ago, ShadowZone said:

I tried:

• Full throttle (TWR at 10km is about 2.5 and > 5.0 at 50km), start tipping at around 1500m, 45° at 9000m, 30° at 25000m, 20° at 35000m --> max Ap around 65km
• Full throttle 90° until 20km, then tip for gravity turn --> Ap above 70km, about 300m/s short for circulization burn
• Kept TWR around 1.75 at 90° until 20km, then tip for gravity turn --> Ap above 70km, about 300m/s short for circulization burn
• Full throttle until 6km, then TWR below 1.85 at 90° until 15km, slowly tip for gravity turn --> Ap about 70km, about 250m/s short for circulization burn

Maybe I am missing something? I was able to get a 400 ton ring station into orbit without any aerodynamics to speak of by keeping TWR low and the ascent to 90° until 20km and then starting my "turn". So imagine my frustration not being able to get this little thing into orbit

 

Maybe you are going too fast! For a vessel with a lot of drag I try not to go faster than 100m/s at 1000m, 300m/s at 8000m, and 600m/s at 12000m, beyond that I go full throttle.

[sal_vager]

Moved to Gameplay Questions.

As has been said, surface attached parts will not be occluded from drag by the parts they clip into, so if you have open nodes you need to cap them, there are nosecones for this.

Also, any blunt rear parts will create drag, so it pays to cap any such parts with more nosecones.

We can't see what parts are producing the most drag, or determine how much drag is actually being produced, the aero overlay is only telling you that there is drag and a very rough indication of how much, to see just how draggy the specific parts are you need to check the drag values in the action menu, enabled via the debug window.

In addition, drag is greatly influenced by air density and speed, not just the draggyness of the part, as Valerian says above, you need to modify your flight path to cope.

If you upload the vessel other people will be able to test it and help you out.