Investigating significant drag and heating changes introduced in 1.8.0

[AggressiveYoghurt]

The 1.8.0 update sneaked in some unannounced (bug perhaps?) significant changes to aerodynamic drag and heating. I created 2 craft and tested them in both 1.7.3 and 1.8.0.

Drag tests: Ignore the prograde re-entry profile, i used this just to make this test as accurate as possible.

https://youtu.be/EwdkbTU7aAM

1.7.3

1000m/s 1400kN @ 1400m

1650m/s 2600kN @ 3300m

1.7.3 re-entry drag

60km 0.290kN @ 2118m/s

50km 1.37kN @ 2155m/s

45km 2.76kN @ 2173m/s

40km 5.9kN @ 2190m/s

35km 15.3kN @ 2203m/s

30km 38.5kN @ 2196m/s

25km 96.2kN @ 2129m/s

Overheated at 22.7km and 2085m/s

 

1.8.0

1000m/s 1165kN @ 1360m

1650m/s 2200kN @ 3300m

1.8.0 re-entry drag

60km 0.225kN @ 2118m/s

50km 1.15kN @ 2155m/s

45km 2.5kN @ 2173m/s

40km 5.21kN @ 2191m/s

35km 12.3kN @ 2204m/s

30km 32.3kN @ 2202m/s

25km 80.3kN @ 2132m/s

Overheated at 24.3km and 2119m/s

Overall i found that drag has been reduced by around 16.4%

 

Re-entry heating tests with a craft that can barely survive a re-entry in 1.7.3 showing the internal temperature of the mk.1 cockpit.

https://youtu.be/vOybtZYYNHE

https://imgur.com/a/NQgnGPZ

https://imgur.com/a/HjztFJe

1.7.3                                                   1.8.0

60km internal 480K                         581K

55km internal 591K                         721K

50km internal 685K                         835K

45km internal 768K                         928K

40km internal 851K                         1013K

35km internal 973K                         1101K/Failed at 35.7km

31km internal 1061K

Craft survived re-entry

Edited October 19, 2019 by AggressiveYoghurt

[Kergarin]

I was just logging in to post the same.

All my "on the edge" designs which barely survived reentry in previous versions, now explode around 30-20km due to overheating.

It seems there is way less drag.

Also adding one or more opened mk1 cargo bays, which were a great early game aerobrake, doesn't provide enough friction to make those designs survive again.

 

I didn't find anything in the changelogs. Maybe it's relatet to the new unity engine?

Edited October 19, 2019 by Kergarin

[Kergarin]

But I guess the difference is too big, to be just the engines calculation differences.

Did you check the atmospheres pressure? Did there change something? 

I'm on the run and can't test right now

[AggressiveYoghurt]

27 minutes ago, Kergarin said:

But I guess the difference is too big, to be just the engines calculation differences.

Did you check the atmospheres pressure? Did there change something? 

I'm on the run and can't test right now

The AeroGUI is visible in the imgur links and the videos, although probably impossible to see on mobile, but it doesnt seem like there are any changes to static or measured dynamic pressures.

Also its apparent that the wings show significant heating effects in 1.8.0 where there is none in 1.7.3. Overall i find it very odd that wings would heat at all in a gentle LKO re-entry.

[panzer1b]

I prefer the new behavior, makes heat shields actually useful (lets face it, unless u hit laythe above 3km/s you can aerobrake almost anything but flipping around in the atmo), and it makes it less of a dragfest.

 

The souposphere is not my idea of realistic or even all that fun, best was around 1.0 release (with truly low drag but dangerously hot atmo), and its gone downhill ever since that update.  I may not be the majority, but i vastly prefer new atmo in 1.8 to the old one (now the challenge is more heat related, less drag/dV related as it should be, given that it was virtually impossible to overheat anything in 1.7.3 with heat set to 100%, and even at 120% you had to intentionally do something to melt your ship unless its interplanetary aerobraking).

 

So yeah, might not be all that popular an opinion, but i say keep it as is in 1.8, much more realistic and enjoyable (not to mention heat is finally somewhat of a thing even if its still way too low imo, especially lack of heat gen from NTR engines which was such a neat mechanic back in the 1.0 days).

[Kergarin]

I would be ok both ways. I like challenges.

On the other hand ascending now became even more easy with this less drag. Eve ssto again possible? 

 

Do we know by now, if this change is intended?

Don't want to spend weeks to change all my designs if this is just temporary

[AggressiveYoghurt]

14 minutes ago, panzer1b said:

I prefer the new behavior, makes heat shields actually useful (lets face it, unless u hit laythe above 3km/s you can aerobrake almost anything but flipping around in the atmo), and it makes it less of a dragfest.

 

The souposphere is not my idea of realistic or even all that fun, best was around 1.0 release (with truly low drag but dangerously hot atmo), and its gone downhill ever since that update.  I may not be the majority, but i vastly prefer new atmo in 1.8 to the old one (now the challenge is more heat related, less drag/dV related as it should be, given that it was virtually impossible to overheat anything in 1.7.3 with heat set to 100%, and even at 120% you had to intentionally do something to melt your ship unless its interplanetary aerobraking).

 

So yeah, might not be all that popular an opinion, but i say keep it as is in 1.8, much more realistic and enjoyable (not to mention heat is finally somewhat of a thing even if its still way too low imo, especially lack of heat gen from NTR engines which was such a neat mechanic back in the 1.0 days).

Im not completely against this, however seeing as stock KSP lacks the option to add heat-shielding to existing parts like wings and tanks to create a re-enterable structure i fear it restricts all interplanetary designs to be some kind of Mk.3 spaceplane. Also im not a huge fan of making the game easier in terms of SSTO drag design, i find that drag is the biggest factor in how efficient your spaceplanes really are and making it easier just seems kinda weak to me.

[Kergarin]

4 minutes ago, AggressiveYoghurt said:

Im not completely against this, however seeing as stock KSP lacks the option to add heat-shielding to existing parts like wings and tanks to create a re-enterable structure i fear it restricts all interplanetary designs to be some kind of Mk.3 spaceplane. Also im not a huge fan of making the game easier in terms of SSTO drag design, i find that drag is the biggest factor in how efficient your spaceplanes really are and making it easier just seems kinda weak to me.

Yes, getting to orbit and launching SSTOs was already way to easy.

A solution to make both - reentry and ascend - harder and more realistic, would be a bigger kerbin, leading to higher orbital speeds. Something around 4.000m/s to stay in orbit just above the atmosphere would fit ksp parts best. It feels verry real in terms of burntime, reentry and SSTOs are barely possible to LKO.

But that's on another paper and would change to much.

 

Back to the topic.

This behaviour feels weird. There is almost no heating and drag now in the upper atmosphere, and then at 20km everything burns up.

If this is intended to make reentry harder, why not just set higher heating in the entire atmosphere? This would feel much more realistic, and wouldn't make the easy ascend even easier. In fact this way it would make also ascending harder. The optimum flat gravity turn ascend path actually covers rockets in plasma. Didn't see this irl by now on ascend.

[T1mo98]

1 hour ago, panzer1b said:

I prefer the new behavior, makes heat shields actually useful (lets face it, unless u hit laythe above 3km/s you can aerobrake almost anything but flipping around in the atmo), and it makes it less of a dragfest.

 

The souposphere is not my idea of realistic or even all that fun, best was around 1.0 release (with truly low drag but dangerously hot atmo), and its gone downhill ever since that update.  I may not be the majority, but i vastly prefer new atmo in 1.8 to the old one (now the challenge is more heat related, less drag/dV related as it should be, given that it was virtually impossible to overheat anything in 1.7.3 with heat set to 100%, and even at 120% you had to intentionally do something to melt your ship unless its interplanetary aerobraking).

 

So yeah, might not be all that popular an opinion, but i say keep it as is in 1.8, much more realistic and enjoyable (not to mention heat is finally somewhat of a thing even if its still way too low imo, especially lack of heat gen from NTR engines which was such a neat mechanic back in the 1.0 days).

I have a Space Shuttle replica that worked perfectly fine.

 

Right now, it's impossible for it survive re-entry. Either the cockpit or the right wing explodes, depending on my pitch. 

Unless we get the ability to properly add heat-shielding to wings, cockpits and the other plane parts, this is not a good change.

[Kergarin]

It's on the bug tracker.

There is a workaround of setting convection factor from 7 to 1. Seems to work so far.

Can anyone tell, if this now is the desired heating, or does 1 make the game easier than it should?

[Wolf Baginski]

9 minutes ago, T1mo98 said:

Unless we get the ability to properly add heat-shielding to wings, cockpits and the other plane parts, this is not a good change.

If you check the .cfg you may see this.

 // heatConductivity = 0.06 // half default

It looks like that was the same in previous versions but it might be worth de-commenting that setting. It is used on the Shuttle Delta Wing Part.

I also see these settings on the Corvus capsule.

maxTemp = 1400
heatConductivity = 0.1 // 5/6ths default
skinMaxTemp = 2400
skinInternalConductionMult = 0.625

maxTemp is lower than normal for wings, but those values are worth testing, I think. skinMaxTemp is already pretty high, about 500 degrees above the usual temperature quoted for the Space Shuttle, but it looks like the skin could be allowed to get much hotter than the interior without triggering explosions. I know, it feels like cheating to modify a stock part but (hand-wave) that could be how you can add a heat-shield layer to a wing.

[Kergarin]

The convection factor workaround does not really solve this.

It definitely looks like heat and drag changes.

I've loaded some savegames parallel in 1.7 and 1.8.. It's significant.

 

Example: a booster reentering from almost orbital speed:

At 30km:

1.7: surface speed 1.872m/s, drag 25,730 kN, convective heat 707.852.000kJ

1.8: surface speed 2060m/s, drag 11,861kN, convective heat 2.040.896.000kJ

 

That's almost 200m/s more. Almost like it didn't loose any speed since reentry at 70km.

Less than half the drag.

Three times the convective heat.

 

Also the booster explodes on touchdown in 1.8 while it stays intact in 1.7.

 

What's going on there? It's nothing in the changelogs related to all this

 

Another funny thing: the booster from my 1.7 savegame nevertheless survived reentry in 1.8.. Building exactly the same in 1.8 from scratch and it explodes.

[Wolf Baginski]

I suspect the drag changes are leading to a need for more parachutes to land a capsule. No-parachute descent speeds look higher, the reefed parachute descent speed looks high, the g-spike when the chute fully opens is way into the red, and the final descent speed looks to have doubled.

[Wolf Baginski]

There's a table in this Wiki page Atmosphere for Terminal Velocity, and it has definitely increased in v1.8.0, roughly 10% higher at 10,000 m. Is that significant? The maths makes my brain hurt, but drag is proportional to v² and density.

I have checked. The .cfg for parachutes haven't changed

Edited October 20, 2019 by Wolf Baginski

[Kergarin]

Please upvote to get this solved

Drag

https://bugs.kerbalspaceprogram.com/issues/23978?next_issue_id=23977&prev_issue_id=23979

Temp

https://bugs.kerbalspaceprogram.com/issues/23966?next_issue_id=23962&prev_issue_id=23977

Have to say I'm a little shocked by so little response. Shouldn't every player notice these issues?

Edited October 20, 2019 by Kergarin

[AggressiveYoghurt]

3 hours ago, Kergarin said:

Have to say I'm a little shocked by so little response. Shouldn't every player notice these issues?

Players playing with only rockets will probably not notice a difference which i suspect is the majority. But i agree this is definitely not a low priority bug as it changes everything for people playing with spaceplanes. Some parachute landing heavy landers will probably fail too if they were on the edge in 1.7.3.

[AHHans]

@nestor already said that they are looking into it in this post (if the link works...).

[djr5899]

I don't have any fancy stats on drag, heating, dynamics, etc.....but....

I started a new career after the MOAR Boosters upgrade, and, in a simple craft, first run of getting to orbit of Kerbin and de-orbitting, the following occurred at a normal re-rentry angle at around 2400m/s and decellerating:

• Re-entry consisted of a heat shield, science junior, two chutes on either side of the science junior, a MK1 command pod, 8 goo containers, 6 pressure sensors, 6 thermometers, and a MK16 chute.
• Re-entry heating exploded all objects except for the MK16 chute and the MK1 command pod.  The heat shield didn't actually explode.  It was attached to the science junior, yet, the science junior still exploded, and the heat shield was aerodynamically stuck to the MK1 chute.  Without that, I believe the MK1 would have likely exploded.

In previous versions, that entire craft should have survived with easy, with nothing being destroyed by heating factors.

I'd say there is a definite problem with heating after the update.

[katateochi]

yeah, finding my spaceplanes are failing to reenter now. LKO reentries are ok, but a return from Mun and even just skimming the atmo (approach Pe at 60km) and wings are burning up....oddly it seems to always the starboard wings, even if I hit the atmo perfectly level.

[OHara]

Many of the coefficients of drag in 'PartDatabase.cfg' , and a few of the part dimensions, seem to be clearly wrong in 1.8.0. 
If we delete that file, KSP regenerates it, but again with the wrong values.   

Those of us with a previous working installation can copy over the old 'PartDatabase.cfg',  delete the line containing "version = ..."
and that solves the 1.8.0 problems with aerodynamics and heat.

Edited October 21, 2019 by OHara
simpler to delete the "version =" line

[Kergarin]

5 hours ago, OHara said:

Many of the coefficients of drag in 'PartDatabase.cfg' , and a few of the part dimensions, seem to be clearly wrong in 1.8.0. 
If we delete that file, KSP regenerates it, but again with the wrong values.   

Those of us with a previous working installation can copy over the old 'PartDatabase.cfg',  delete the line containing "version = ..."
and that solves the 1.8.0 problems with aerodynamics and heat.

Thanks! Do the newly added parts still work then?

[OHara]

54 minutes ago, Kergarin said:

Do the newly added parts still work then?

It is worth trying them.

KSP automatically added entries for the new parts, to the 'PartDatabase.cfg' that I copied from 1.7.3. 
One might expect the new parts to have errors in drag, but the ones I've checked seem fine.

[Kergarin]

16 hours ago, OHara said:

Many of the coefficients of drag in 'PartDatabase.cfg' , and a few of the part dimensions, seem to be clearly wrong in 1.8.0. 
If we delete that file, KSP regenerates it, but again with the wrong values.   

Those of us with a previous working installation can copy over the old 'PartDatabase.cfg',  delete the line containing "version = ..."
and that solves the 1.8.0 problems with aerodynamics and heat.

Had the time to test.

This solves the issue and behaves 100% like 1.7.3.

Thanks!

[ExtremeSquared]

If you fly Eve missions in multiple KSP versions, you'll realize how regularly aero and heating changes don't make it into the changelog. The question should be "Is this better or worse?"

[AggressiveYoghurt]

52 minutes ago, ExtremeSquared said:

If you fly Eve missions in multiple KSP versions, you'll realize how regularly aero and heating changes don't make it into the changelog. The question should be "Is this better or worse?"

I have to say i dont remember a change like this happening in a few years. Maybe i just didnt pay closer attention before or they were patched out before i could notice, 1.4 definitely had lots of hotfixes for hotfixes in a short timespan.

The question then is if they are always done on purpose or by accident. I would think that changing something in aerodynamics would be one of the top-priority things to write in to a changelog as it can dramatically change everything for all players whether they notice it or not.