Bug, excessive drag, or what?

[RizzoTheRat]

Solid boosters cut out, half second pause, decouplers fire and the Reliant engine starts up.  KER reckons it's got a TWR  of just over 2 (I did check the stage can lift from the pad just in case KER's reporting dodgy numbers)...but the ship slows down and the now unpowered boosters overtake it!

I know it's not a particularly streamlined design (not got farings yet) but surely it can't have that much drag can it?

 

 

Edited April 24, 2018 by RizzoTheRat

[Johnny Wishbone]

Doesnt F12 toggle the aerodynamic overlays? Turn it on and see for yourself what parts are producing drag, lift, etc.

Also, a 20-25 degree AoA at 10km seems awful shallow for a rocket. Still in the “soup” and burning a lot of fuel fighting air.

Edited April 24, 2018 by Johnny Wishbone

[Mattasmack]

I can't see your rocket too clearly in that screenshot but I'd say, yes, it's possible you have that much drag.  Radially-attached parts add drag.  Size changes between stack-mounted parts (e.g. a .625m engine or decoupler on a 1.25m tank) add drag.  It looks like you have lots of both of those.

And SRB casings, even after burn-out, are pretty heavy for their size (high ballistic coefficient) so they probably aren't decelerating much in that picture.

Johnny Wishbone's advice to try F12 is good too -- it's also possible that you're getting excessive drag from a bug; the aerodynamic overlay would show you that if a whole lot of drag is coming from one unexpected place.

(And BTW, you're likely to get better/more answers in the 'gameplay questions and tutorials' subforum.)

[JAFO]

I concur with these points below.

8 hours ago, Johnny Wishbone said:

Also, a 20-25 degree AoA at 10km seems awful shallow for a rocket. Still in the “soup” and burning a lot of fuel fighting air.

8 hours ago, Mattasmack said:

I can't see your rocket too clearly in that screenshot but I'd say, yes, it's possible you have that much drag.  Radially-attached parts add drag.  Size changes between stack-mounted parts (e.g. a .625m engine or decoupler on a 1.25m tank) add drag.  It looks like you have lots of both of those.

And SRB casings, even after burn-out, are pretty heavy for their size (high ballistic coefficient) so they probably aren't decelerating much in that picture.

Johnny Wishbone's advice to try F12 is good too -- it's also possible that you're getting excessive drag from a bug; the aerodynamic overlay would show you that if a whole lot of drag is coming from one unexpected place.

And speaking of bugs, it's only just today been discovered and confirmed that there's a drag bug in some stock-DLC parts, with variations of up to 7% depending on which skin you apply to them. Might well be related.

 

[RizzoTheRat]

It's a ridiculously draggy design (4 commsats with 1.25 bodies but 0.625 tanks with radial batteries on them), which is why it's so shallow as I gave it the usual initial turn but it's accelerating slower.  All stock parts apart from the kOS processors.

I'd hadn't though F12 would show me anything useful as the drag would all be on the centreline, but thinking about it would show if the batteries are draggy despite being tucked under the 1.25 parts.

I'm astounded that it can cause enough drag to decelerate with TWR>2 though.

I don't have farings yet but I guess another approach is needed as it looks like attempting to power through the drag won't work.  Batteries in service bays and Terriers instead of sparks with be less draggy but a lot heavier.

[Daveroski]

54 minutes ago, RizzoTheRat said:

It's a ridiculously draggy design (4 commsats with 1.25 bodies but 0.625 tanks with radial batteries on them), which is why it's so shallow as I gave it the usual initial turn but it's accelerating slower.  All stock parts apart from the kOS processors.

I'd hadn't though F12 would show me anything useful as the drag would all be on the centreline, but thinking about it would show if the batteries are draggy despite being tucked under the 1.25 parts.

I'm astounded that it can cause enough drag to decelerate with TWR>2 though.

I don't have farings yet but I guess another approach is needed as it looks like attempting to power through the drag won't work.  Batteries in service bays and Terriers instead of sparks with be less draggy but a lot heavier.

I suggest you try the same rocket but not start your gravity turn until it reaches 10000m

[The Aziz]

8 hours ago, Daveroski said:

I suggest you try the same rocket but not start your gravity turn until it reaches 10000m

Which is not a good idea either (better though). At 10km it should be around 45 degree.

[linuxgurugamer]

5 minutes ago, The Aziz said:

Which is not a good idea either (better though). At 10km it should be around 45 degree.

I disagree, thays a bit too much.  I normally am at about 22.5 degrees at 10km, and 45degrees at 25km

[Daveroski]

7 minutes ago, The Aziz said:

Which is not a good idea either (better though). At 10km it should be around 45 degree.

What do I know.. must be a noob compared to you.

[pandaman]

15 minutes ago, Daveroski said:

What do I know.. must be a noob compared to you.

Most of us can be considered a noob compared to @linuxgurugamer  ...

Apart from Scott Manley of course.

[The Aziz]

Okay before we get to who's better (I am no expert, come on, I still struggle to land on Jool' moons) I'll say it depends on rocket design and its capabilities. BUT, my method works for me every time, so I only can recommend that, and everyone has the same right to share their own way.

[pandaman]

6 minutes ago, The Aziz said:

Okay before we get to who's better (I am no expert, come on, I still struggle to land on Jool' moons) I'll say it depends on rocket design and its capabilities. BUT, my method works for me every time, so I only can recommend that, and everyone has the same right to share their own way.

Absolutely.

We each have our own way of doing things.  I've been playing ages and never have developed a consistent 'personal method'.  Just an 'ideal' in my head that i usually fail to achieve by varying degrees.

[RizzoTheRat]

I upped my turn speed so it climbed a bit steeper, and when it separated it just managed to overtake the boosters by enough that they took out the main engine , so it does look like it's just generating way too much drag.

Redesigned it to be all 1.25m with just the solar panels sticking out and it behaved a lot better, but doesn't have enough dV to get to orbit yet.

 

As for debate about turn angles, I start my turn at 50m/s, which is about 250m for my usual 1.6ish TWR, and if it's a streamlined design gravity losses are way worse than aerodynamic losses so it's best to turn as early and as hard as you can.  For a decently streamlined rocket starting my turn at 50m/s used about 100m/s less dV than starting it at 100m/s, and for a given start speed, the most efficient turn angle was less than half a degree short of the angle at which it wouldn't make it to space.  The colours are for different TWRs all starting the turn at 50m/s, but my turn angle algorithm only uses TWR so far, at some point I need to do some more experiments and include drag in it.  Graph was generated in 1.3 but I don't think he aerodynamics have changed since.

ETA:  Forgot to say that gives me about 45-50 degrees above the horizon at 10km.

Edited April 25, 2018 by RizzoTheRat

[Mattasmack]

11 hours ago, RizzoTheRat said:

...

I'd hadn't though F12 would show me anything useful as the drag would all be on the centreline, but thinking about it would show if the batteries are draggy despite being tucked under the 1.25 parts.

...

Careful here, it sounds like you're applying real-life logic where you need KSP-logic instead.  Radially attached parts produce drag, period**.  All that matters drag-wise is how they're angled to the vessel's direction of travel.  Location does not matter; you can tuck them under another part for esthetics, but KSP doesn't know that that ought to keep them out of the airstream.  In fact, if tucking them under means they present a larger cross-section in the direction of travel, drag will actually be higher.  (As an aside, this often comes up when putting solar panels flat on a cone-shaped part like a 1.25m -> 2.5m adapter.  In real life, the panels would add almost no drag if laid flat on the underlying part.  In KSP-logic, such panels are seen to be angled to the rocket's direction of travel, and they produce significant of drag and lift.  The lowest-drag orientation is to make the panels vertical, regardless of the shape of the part they're attached to.)

** (unless contained within a closed service bay, cargo bay, or fairing, of course.)

1 hour ago, The Aziz said:

Okay before we get to who's better (I am no expert, come on, I still struggle to land on Jool' moons) I'll say it depends on rocket design and its capabilities. BUT, my method works for me every time, so I only can recommend that, and everyone has the same right to share their own way.

Back in the pre-1.0 souposphere days, aiming for 45 degrees at 10km really was universal, along with an optimal TWR of 2.  With the current aerodynamics, the design of the rocket matters much more.  If you design your rockets fairly consistently, you might be able to find a single launch profile that works well for all your designs.  But others will have different experiences.

Edited April 25, 2018 by Mattasmack

[Daveroski]

2 hours ago, pandaman said:

Most of us can be considered a noob compared to @linuxgurugamer  ...

Apart from Scott Manley of course.

One only has to hover over an avatar and one can see the noobs.

[klesh]

3 hours ago, linuxgurugamer said:

I disagree, thays a bit too much.  I normally am at about 22.5 degrees at 10km, and 45degrees at 25km

 

45 @ 20 for me, but that sounds about right.  Been flying that profile since they got rid of the souposphere.

[Snark]

On 4/24/2018 at 1:30 PM, RizzoTheRat said:

I know it's not a particularly streamlined design (not got farings yet) but surely it can't have that much drag can it?

Oh, it can absolutely have that much drag.  I had difficulty interpreting your screenshot (small size, very dark), but with a bit of photoshopping, it becomes a bit clearer:

You've got multiple sub-components stacked up, going repeatedly from 1.25m to 0.625m and back to 1.25m again.

This rocket is "not particularly streamlined" in the same sense that the ocean is "a bit damp" and Greenland "tends away from the tropic".    It's basically the poster child for hideous aerodynamics-- you might as well have mounted those little satellites radially around the center body and then not put nosecones on them.  That's how draggy it is.

For good aero, your ship needs to be a smooth cylinder.  If you don't have fairings, then use 1.25m engines and tanks instead of going down to that 0.625m stack each time.

Or, if you can't fix the problem by fixing the ship, you can mitigate it by adjusting your flight profile.  Normally I'm a fan of high TWR (e.g. 2) and fairly aggressive gravity turns-- if the ship is nicely streamlined.  But if you've got a draggy monstrosity (and hey, we've all been there, every once in a while it's necessary!).... that's often a good time to go for a lower TWR, at least to start with.  By having a lower TWR, your rocket doesn't get going really fast until it has climbed to a higher altitude, which means it's out of the densest part of the soup and drag isn't as much of an issue.

On 4/25/2018 at 12:35 AM, RizzoTheRat said:

it would show if the batteries are draggy despite being tucked under the 1.25 parts.

Be careful how you think about drag-- this isn't how it works, at all.  Positioning doesn't matter.

This seems counterintuitive, because logically (according to actual physical reality), it should.  That's how the world works.  However... KSP isn't IRL, and it uses a very simple aerodynamic model for drag.  The actual physical positioning of parts doesn't matter in the slightest.  This leads to all sorts of common player misunderstandings-- e.g. thinking "I've got part A sheltering behind part B, therefore that mitigates part A's drag."  Nope.  I've also seen people do things like trying to use clipping to "hide" parts aerodynamically (e.g. "I put some stuff inside the nosecone")... which also doesn't work.

So, the radially attached batteries are going to generate exactly the same amount of drag no matter where on the outside of the rocket you put them.  Position doesn't matter.

Moving them inside the service bay (by attaching them to the inside, not by just attaching to the outside and clipping inwards) would, in fact, eliminate their drag.  That's because service bays, cargo bays, etc. have special code so that stuff inside them "doesn't count" for aero when they're closed.

However, my guess would be that although moving the batteries inside the service bays would help a bit with drag, it wouldn't make that much difference.  They have some drag, sure, but I suspect it's not all that much.  I suspect that by far the lion's share of the drag is the fact that you've got that repeated 1.25m/0.625m iteration thing going.  I suspect that turning your rocket into a smooth cylinder would be both necessary and sufficient to solve the problem.

On 4/25/2018 at 11:43 AM, Mattasmack said:

Careful here, it sounds like you're applying real-life logic where you need KSP-logic instead.  Radially attached parts produce drag, period**.  All that matters drag-wise is how they're angled to the vessel's direction of travel.  Location does not matter; you can tuck them under another part for esthetics, but KSP doesn't know that that ought to keep them out of the airstream.

^ This.

On 4/25/2018 at 12:35 AM, RizzoTheRat said:

Terriers instead of sparks with be less draggy but a lot heavier

That's one option, sure.  Here's are various ideas that occur to me:

• You could just eat the aero losses, and launch with a lower TWR and less aggressive gravity turn.  (One way you could do that:  instead of having the Thumper SRBs in one group of four, put them in two groups of two.  One pair is at 100% thrust; the other is stepped down to, say, 70%.  Both activate at liftoff, but the 100-percenters burn out and are jettisoned first.  This lowers your takeoff TWR and also helps to smooth out the ride a bit.)
• You could go with Terriers, as you mention, to keep the stack a uniform 1.25m.  As you say, lots heavier-- the Terriers are kinda overkill for those tiny comsats.
• You could just put no engines at all on the comsats.  Launch a vehicle that goes to the desired orbit, releases a comsat, then uses thrust to go to the next orbit and so forth.  It has the disadvantage that the comsats then can't adjust their orbits afterwards, but it's simple, aerodynamic, and super lightweight.
• You could use small radial engines on the comsats, like Spiders or Twitches.  Not super aerodynamic, since the radial engines do have some drag, but it wouldn't be apocalyptically bad the way it is now.  I suspect it would launch just fine.

 

[StrandedonEarth]

1 hour ago, Snark said:

snip

Did any of that change under 1.4.3? (finally downloaded 1.4.3 this morning, but haven't even unzipped it yet)

Does mounting stuff inside (not clipped inside) a structural tube help at all?

While I got the part about mounting stuff on/under a taper-in  doesn't help, does the taper-in itself improve drag (upside down 1.25-to-2.5 cone)?

I'm about to try MH for the first time: I guess an upside-down closed SM-25 is the best bet for mounting stuff inside, assuming the above question is affirmative?

[klesh]

38 minutes ago, StrandedonEarth said:

Does mounting stuff inside (not clipped inside) a structural tube help at all?

 

 

No.  They do not have any aero shielding capabilities.

[StrandedonEarth]

3 minutes ago, klesh said:

No.  They do not have any aero shielding capabilities.

Hmmm, @JPLRepo: Can we get that for structural tubes in a future update please?

Edited April 28, 2018 by StrandedonEarth

[JAFO]

5 hours ago, StrandedonEarth said:

Hmmm, @JPLRepo: Can we get that for structural tubes in a future update please?

I second this request!

[Snark]

7 hours ago, StrandedonEarth said:

While I got the part about mounting stuff on/under a taper-in  doesn't help, does the taper-in itself improve drag (upside down 1.25-to-2.5 cone)?

Not sure exactly what you mean.  Nosecones help, yes.  If you've got a side-mounted droptank, say, so it's just a cylindrical tank with a flat top and a flat bottom, then yes, it helps to put a nosecone on top, and it also helps to put an upside-down nosecone on the bottom.  What you want to avoid is a sudden "stair-step" change in radius, either going up or going down.

[kBob]

8 hours ago, Snark said:

 

Moving them inside the service bay (by attaching them to the inside, not by just attaching to the outside and clipping inwards) would, in fact, eliminate their drag.  That's because service bays, cargo bays, etc. have special code so that stuff inside them "doesn't count" for aero when they're closed.

 

So this is why opening the doors seems to add a lot of drag?  Everything inside suddenly counts as drag.  So I can put a bunch of useless but lightweight parts in there and cheat my way to extra drag when coming in a little fast.

[StrandedonEarth]

18 minutes ago, Snark said:

Not sure exactly what you mean.

Going from, say, a 2.5m fairing down to a 1.5m tank. From what you just said, it would be better to have the adapter in there rather than a straight step...

Like so:

 

[OHara]

8 hours ago, StrandedonEarth said:

Can we get [aerodynamic shielding] for structural tubes ?

I started to make a Module Manager patch, so that hollow-looking nosecones and adapters would act like always-closed cargo bays.  

I'll ask on the Add-Ons thread how to make it more general, after I remind myself what my questions were.

Spoiler

@PART[noseCone|rocketNoseCone|*FuselageStructural] {
@description = changed
    %MODULE[ModuleColorChanger] {
        toggleInEditor = false
        toggleInFlight = false
    }
    %MODULE[ModuleCargoBay] {
        DeployModuleIndex = 0
        lookupCenter = 0, 0.3, 0
        lookupRadius = 1.0
    }
}