Parts & Vessels Massive drag bug with basically every inline auxiliary part (tests, results and pictures)

[FlazeTheDragon]

So heres the thing, basically every inline mounted auxiliary part such as decouplers, batterys, reaction wheels and cargo bays have bugged drag calculations(including those rounded rear hatches but i didnt feel like making a custom plane just to add it to my testing).

Any 1 of these parts will effectively double your drag.

These bugs are not limited to only the parts ive tested, as far as i can tell, literally every size and shape of said parts is affected.

Ive run a couple of tests as you can see below. Every plane configuration is exactly the same other then parts that are in question added to the plane (original plane parts are white and test parts are orange).

First test with base line results:

Nr1. Base line: ~360m/s with drag being ~55Kn.

Base line v.max ~1000m/s with drag being ~110Kn.

Nr2. Same plane configuration, but with added reaction wheel. Plane is unable to accelerate past 350m/s with its drag being ~160Kn.

Nr3. Same plane configuration, but with added 1 inline battery. Plane is unable to accelerate past ~340m/s with its drag being ~160Kn.

Nr4.Same plane configuration, but with added inline decoupler. Plane is unable to accelerate past ~330m/s with its drag being ~160Kn.

Nr5.Same plane configuration, but with added inline cargo bay. Plane is unable to accelerate past ~355m/s with its drag being ~160Kn.

Nr6. Ive sized up the plane to see if the amount of drag scales with the size of the craft, and it looks like it basically does:

Nr6.1 Base line ~750m/s, drag is ~600Kn.

Nr6.2 Base line v.max = ~1100m/s, drag is ~1100Kn.

Nr6.3 Same plane with inline decoupler. Plane speed is ~750m/s with drag being ~1050Kn. Plane is unable to accelerate further.

All of these drag errors are effectively making it impossible to make xeon ssto's, put any sort of cargo capacity on planes or just have auxilary power or reaction wheels, not to mention the numerous implications it has on rockets.... tho i guess there not affected as much, it just dumsters there efficiency, as every vertical stage is going to effectively double your rockets drag.

 

Hopefully this gets noticed and is fixed relatively soon. :)

 

[Thomot512]

Interesting discovery.

Note also that at lower Mach number, your planes appear to need more Lift in order to fly in level flight.

[FlazeTheDragon]

11 minutes ago, Thomot512 said:

Interesting discovery.

Note also that at lower Mach number, your planes appear to need more Lift in order to fly in level flight.

Yup, i really love both the building process and flight characteristics of ksp2 planes, its easily the biggest upgrade over ksp1. Tho these bugs, along with rapier thrust bug are kinda cramping my ssto vibes.

[Duke-49th]

I don't think they played their own game for even just 5 minutes. The amount of bugs are just incredible.

[Gordon Fecyk]

When IG releases official modding support I'll be looking for Ferram Aerospace Research 2. @dkavolis @ferram4 any thoughts?

[FlazeTheDragon]

11 hours ago, Gordon Fecyk said:

When IG releases official modding support I'll be looking for Ferram Aerospace Research 2. @dkavolis @ferram4 any thoughts?

yup, im really looking forward to it as well

[FlazeTheDragon]

UPDATE

Ive discovered a way to still have inline decouplers with out doubling your drag, tho its a bit jank and will likely require you to have modified joint rigidity.

What you do is, instead of using a decoupler, you add a sacrificial ant engine, and once you want stage separation, you jetison the specific ant engine. Engines dont seem to be affected by the drag bug.

The downside is that you cant do it over the staging sequence, but for something thats drag critical, like an eve ascend module, its basically the only viable workaround.

[Kaa253]

Confirmed. My SSTO works fine until I add a cargobay and then it can't even reach the upper atmosphere.

[FlazeTheDragon]

39 minutes ago, Kaa253 said:

Confirmed. My SSTO works fine until I add a cargobay and then it can't even reach the upper atmosphere.

If you do a single cargobay piece thats at least 1 size down from the shuttle size parts, and cap each end off the cargo bay internally with the tiny nose cone, you can actually make cargoplane ssto's. They will still suffer from increased drag and anything inside the bay wont be shielded from drag, but its at least sort of functional then.

[stackocakes]

This explains so much why I was struggling last night when I added my cargo bay...

[oro]

Thank you for discovering all this. I too felt that drag is a little bit too high and now i know the reason. Hofully this bug can be easly fixed by the dev in the upcomings patches

[kkaja123]

What happens when you flip around the inline part under test? E.g. upside-down for a rocket, front facing backwards for a plane.

When I tried that out on my rocket using the small inline probe command module, the drag returned to expected levels (about a magnitude of 10 smaller than the default orientation).

Spoiler

Test Rocket Parts List

Aero stats of the test rocket (part in normal orientation) at about 2 km. Drag = ~170 kN.

Aero stats of the test rocket with the part in the upside-down orientation at about 2 km. Drag = ~18 kN, even though the velocity is higher at the same altitude.

Note that in the first test, the probe failed to keep the rocket pointed straight up due to the massive drag. In the second test, the probe had relatively no difficulty in keeping the rocket pointed upwards.

[FlazeTheDragon]

On 3/16/2023 at 5:15 AM, kkaja123 said:

What happens when you flip around the inline part under test? E.g. upside-down for a rocket, front facing backwards for a plane.

When I tried that out on my rocket using the small inline probe command module, the drag returned to expected levels (about a magnitude of 10 smaller than the default orientation).

  Reveal hidden contents

Test Rocket Parts List

Aero stats of the test rocket (part in normal orientation) at about 2 km. Drag = ~170 kN.

Aero stats of the test rocket with the part in the upside-down orientation at about 2 km. Drag = ~18 kN, even though the velocity is higher at the same altitude.

Note that in the first test, the probe failed to keep the rocket pointed straight up due to the massive drag. In the second test, the probe had relatively no difficulty in keeping the rocket pointed upwards.

lmao, will need to test this out, could be huge.

Edit: So i ran some tests and was not able to re-create the same effects with any of the parts.

Im pretty sure the discrepancy in drag in your pics is down to the fact that the craft in the upper picture has substantially more aoa (17.5o vs 1.5o) which would increase its drag considerably.

Edited March 17, 2023 by FlazeTheDragon

[Dakota]

Just popping my head in here to say we're aware of this report and QA is working on reproducing right now. Thanks for all the details you've shared thus far!

[Buzz313th]

On 2/28/2023 at 1:15 PM, Thomot512 said:

Interesting discovery.

Note also that at lower Mach number, your planes appear to need more Lift in order to fly in level flight.

Sarcasm I hope?

[Nate Simpson]

Just adding to what Dakota said above: we've got a good local repro and we're hunting this down right now. The documentation at the top of this thread was fantastic and really helped us focus in on the problem. Thanks @FlazeTheDragon!

 

[FlazeTheDragon]

@Dakota@Nate Simpson Thanks for the update, this specific bug has been bothering me a lot haha.

Glad i could help

[Thomot512]

On 3/17/2023 at 11:14 PM, Buzz313th said:

Sarcasm I hope?

Nope see: 

 

[Joker58th]

Its not just inline parts.  The radial air intake seems to have excessive drag as well.  I reported to support but thought I would mention here as well for more visibility.  Notice the speed difference in the two shots.  These are maximum speeds with afterburner for both examples.

[Buzz313th]

On 3/21/2023 at 1:40 PM, Thomot512 said:

Nope see: 

 

It is normal to need more Pitch authority with a higher alpha to maintain level flight at lower airspeeds.   Regarding you not able to maintain level flight in a 250 m/s glide..   Where was your COP in relation to your COM on that build and was your thrust vectoring turned off?     I would be willing to take a guess and say your COP was far too aft of the COM with that build and your thrust vectoring was on, both would explain  the lack of positive pitch authority in power off flight.   I asked if you were being sarcastic because a lower relative mach number usually means a lower airspeed which means a higher AOA.

Edited March 23, 2023 by Buzz313th

[Thomot512]

On 3/23/2023 at 3:18 AM, Buzz313th said:

t is normal to need more Pitch authority with a higher alpha to maintain level flight at lower airspeeds.   Regarding you not able to maintain level flight in a 250 m/s glide..   Where was your COP in relation to your COM on that build and was your thrust vectoring turned off?     I would be willing to take a guess and say your COP was far too aft of the COM with that build and your thrust vectoring was on, both would explain  the lack of positive pitch authority in power off flight.   I asked if you were being sarcastic because a lower relative mach number usually means a lower airspeed which means a higher AOA.

It seems I did not explain the problem clearly.
I'm not speaking about pitch authority of a specific design. I'm merely observing that for a given airplane with a given mass, it appear that the amount of lift needed to maintain level flight (load factor of 1) varies in function of the mach number.
This is not something that has any logical explanation. Indeed in order to be in a maneuver with a load factor of 1, typically level flight, the lift force must be equal to the G-Force. In the measured case in order to maintain level flight the lift needed to be around twice the weight.

I'm aware that an airplane can be too stable when CoP is too far aft of the CoM, and that in those case it is difficult to steer the airplane, but if the lift is higher than the weight it should still enter a maneuver corresponding to the load factor.

 

Edit: typo corrected.

Edited March 24, 2023 by Thomot512

[Dakota]

This has been fixed as of v0.1.3.