Do part adapters serve a purpose?

[guitarxe]

Other than aesthetics, do part adapters (those things that connect from small to large sizes) serve any real purpose?

[Plusck]

Aerodynamically, yes. Without the adaptors you have blunt objects facing the airflow. Not too sure about using the short adapters at the back - e.g. to put a 1.25m rocket on a 2.5m stack: I think I read that blunt retrograde surfaces add drag too but I haven't tested this in any way.

Also, when using a 2.5m fairing for a 1.25m probe, it helps to put the ultra-short adapter on the fairing first, so you can strut the payload effectively with minimum added weight and minimum added height to the fairing.

[Blaarkies]

@Plusck That is certainly true for FAR, but does it apply to stock as well? Last i heard, stock calculates drag by the "open" nodes towards the front of the vessel(so im guessing a 2.5m front section, covered by the 0.625m nose-tip would be effective)

[Plusck]

1 hour ago, Blaarkies said:

@Plusck That is certainly true for FAR, but does it apply to stock as well? Last i heard, stock calculates drag by the "open" nodes towards the front of the vessel(so im guessing a 2.5m front section, covered by the 0.625m nose-tip would be effective)

Easy enough to test... but that's OK I'll do it for you

Set-up: one advanced nose cone, one remote guidance unit, one 1.25 ore tank (empty), one 2.5m fuel tank (full), one SRB.

Tested with both the standard adapter and the short one (that looks like it's meant for engines), then without the engine-side adapter, then without either.

I put the ore tank on because I was going to use it to compensate for weight differences, but it was a bad choice since the ore graduations are not fine enough to cover the 0.08t of engine-side adapter and the 0.1t of top-side adapter.

Therefore, the weight of the tests without the adapters was slightly lower, and total dv was higher (457 m/s with no adapters compared to 450 m/s with both). The results are therefore even more compelling since a properly balanced test would have given even greater differences.

As it was, with adapters the rocket reached 2861m altitude. Without the engine adapter: 2818m. Without either: 2786m.

So yes, the adapters make a difference both in the direction of the airflow and on the blunt rear of the craft.

Incidentally, I deliberately made the set-up heavy to avoid excessive TWR. The rocket never attained any great speed - and so if you like to go really fast in the atmosphere you should experience a much greater difference.

 

Edited April 1, 2016 by Plusck

[Blaarkies]

15 minutes ago, Plusck said:

As it was, with adapters the rocket reached 2861m altitude. Without the engine adapter: 2818m. Without either: 2786m.

So yes, the adapters make a difference both in the direction of the airflow and on the blunt rear of the craft.

Incidentally, I deliberately made the set-up heavy to avoid excessive TWR. The rocket never attained any great speed - and so if you like to go really fast in the atmosphere you should experience a much greater difference.

 

Kuddos! Really well done test, thanks for taking time to do it 
I am quite surprised by the +- 50m difference that a little adapter made(especially on such low altitude). The game probably finds "draggy" section the same way that it occludes part with re-entry, so shape will definitely matter. I wish adapters had a bit of fuel though 

[Plusck]

Just now, Blaarkies said:

I am quite surprised by the +- 50m difference that a little adapter made(especially on such low altitude). The game probably finds "draggy" section the same way that it occludes part with re-entry, so shape will definitely matter. I wish adapters had a bit of fuel though 

Yes, I was surprised too. I didn't try swapping them around to see if that makes a difference (i.e. if being rearwards has a "spoiler" effect which would not be equivalent to the "streamlined" effect if the stubby adapter was in front). But now I know what I'll be putting on more rockets in future...

[pizzaoverhead]

8 hours ago, Plusck said:

I think I read that blunt retrograde surfaces add drag too but I haven't tested this in any way.

 

2 hours ago, Blaarkies said:

@Plusck That is certainly true for FAR, but does it apply to stock as well? Last i heard, stock calculates drag by the "open" nodes towards the front of the vessel(so im guessing a 2.5m front section, covered by the 0.625m nose-tip would be effective)

I've done some testing with this. Blunt retrograde surfaces, such as those on radial drop tanks or 2.5m tanks stacked on 1.25m lifters, cause high drag, therefore tailcones and a gently varying silhouette are effective. The stock system uses drag cubes (generated images) rather than the open node count old versions of FAR incorporated, so the 0.625m nose cone on a 2.5m tank shouldn't be effective, though I haven't tested this particular case.

[John FX]

18 minutes ago, pizzaoverhead said:

 

I've done some testing with this. Blunt retrograde surfaces, such as those on radial drop tanks or 2.5m tanks stacked on 1.25m lifters, cause high drag, therefore tailcones and a gently varying silhouette are effective. The stock system uses drag cubes (generated images) rather than the open node count old versions of FAR incorporated, so the 0.625m nose cone on a 2.5m tank shouldn't be effective, though I haven't tested this particular case.

The newer versions of FAR do a proper voxel shape of your craft so all the things affect your drag in the way they should.

[Plusck]

I decided that the earlier tests needed to be redone properly, with a higher TWR and a set mass so that all results would be strictly comparable.

The set-up: small nose cone, oscar tank, NCS adapter, remote guidance unit, [optional 2.5m adapter #1,] size 16 rockomaxx tank, [optional 2.5m adapter #2,] thumper SRB.

Spoiler

The oscar tank and ncs adapter tank were intended to let me get an exact weight no matter what. I plumped for a TWR of 2.00 which gave a mass of 11,567 kg.

I tested the two "empty" 2.5-to-1.25m adapters, and also the long adapter tank. The heavier adapter tank took a bit more work to get the mass right.

Spoiler

I only tried the adapter tank in standard configurations. They won, significantly, which wasn't all that surprising but nice to know all the same.

What was much more surprising came from the exhaustive mix of the short and standard structural adapters.

It turns out that the greatest efficiency doesn't come from putting the adapter on the front, but on the back. The difference is very small, but it was consistent: putting your most streamlined adapter on any blunt rear is actually better than putting it on a blunt front of your rocket.

Results:

Spoiler

I finally tried a number of other options like a docking port just above the probe core, struts and fins. I didn't include the docking port or strut tests here because the results were pretty obvious: not great for aerodynamics.

What was surprising, though were the fins (hence that last image above). First of all - they made SAS stability assist go off prograde every time. Even the control fins ended up going slightly further off the vertical than not having fins at all, oddly.

Most oddly, though, the smallest "basic fin" actually ended up giving less drag than if they hadn't been there at all. I triple checked that I was getting the mass right and couldn't find any error. Is this a bug or is there some process that I'm missing that would explain this (like better airflow over the SRB? Pretty sure that can't be modelled in KSP).

So the real, real winner is: adapter tanks between 1.25 and 2.5m parts, and basic fins on the bottom.

[Norcalplanner]

9 hours ago, Blaarkies said:

I wish adapters had a bit of fuel though 

Fuel tanks plus by necrobones has conical fuel tanks and fueled nose cones in every diameter.

[N_Danger]

 There are fuel tanks in stock that are conical and can be used as adapters.

[guitarxe]

Thanks, Plusck!

[Newpossum]

Plusck, this was hugely educational for me. Thank you very much for taking the time to do all these tests, and document your results so meticulously.

[sardia]

On 4/1/2016 at 8:22 PM, Plusck said:

Aerodynamically, yes. Without the adaptors you have blunt objects facing the airflow. Not too sure about using the short adapters at the back - e.g. to put a 1.25m rocket on a 2.5m stack: I think I read that blunt retrograde surfaces add drag too but I haven't tested this in any way.

Also, when using a 2.5m fairing for a 1.25m probe, it helps to put the ultra-short adapter on the fairing first, so you can strut the payload effectively with minimum added weight and minimum added height to the fairing.

Is it worth it to put tail cones (both top and rear cones)embedded into our SRB boosters? I have a rocket with a 12 pack of them. 

Edited April 11, 2016 by sardia

[Plusck]

1 minute ago, sardia said:

Is it worth it to put tail cones embedded into our SRB boosters? I have a rocket with a 12 pack of them. 

I really couldn't say. I don't do much clipping other than the obvious things.

It is a question that I've asked myself though: if a rocket is firing, are drag effects even counted (since IRL the exhaust cone will necessarily eliminate drag on the open end of the nozzle, leaving just the junction between the rocket body and the engine itself to cause drag)?

So I'd suggest trying it (SRB with minimal fuel, probe core, small ballast tank, with and without clipped tail cone) and seeing what happens.

 

[ibanix]

This is some of the best !!SCIENCE!! I've seen in ages. Guess I'm putting adpaters all over my stages now....