Mk16 vs Mk2-R parachutes

[peewee69]

Hi fellow kerbanaughts!

A curious phenomena has occurred when comparing the early career parachutes;

The Mk16 has an effective diameter of 20.7 and the Mk2-R has 41.3.

Yet when I tested each of them, I found that the Mk16 slows the descent of my test craft to about 6 - 6.5 m/s and the Mk2-R to about 7 - 7.5 m/s.

Surely a larger effective diameter should lead to a slower descent?

[Plusck]

One of the two descriptions (probably the Mk2-R Mk16 both, or all of them) is just wrong.

If you place both parachutes on a craft and look at them once they're fully deployed, the Mk16 is visibly larger than the Mk2-R.

 

edit: took another look - the MK16 looks to be about 2/3 the diameter of the Mk16 XL. So if the Mk16-XL is 50, the Mk16 is about 35 and the Mk2-R about 28...

edit 2: looking again (again) at a couple of screenshots (and obviously, subject to distortion of the image due to camera angle) the actual deployed diameter of the Mk2-B looks to be 8m, the Mk16 10m, and the Mk16-XL 15m. I don't know what "effective diameter" is supposed to mean. Simplifying the parachutes to assume they're hemispherical, that would mean (Pi x r²): Mk2-B 50m², Mk16 75m² and Mk16-XL 175m²...

Edited October 14, 2016 by Plusck

[peewee69]

Thanks!

[Plusck]

I found an equation to calculate the required diameter of a parachute to fall at a given rate here. "Diameter" in this context appears to mean the constructed diameter, not its deployed diameter.

Plugging in 6.5 m/s into that equation, assuming a perfect parachute and a 0.9t load, gives a parachute diameter of 17m (which corresponds roughly to the Mk16, once you take its highly curved form into account).

For 7.5 m/s, that parachute needs to be just under 15m (which seems a bit much for the Mk2-R compared to its visible size, but maybe looks are deceiving).

So, in any event, the numbers given in the VAB are certainly not in m (but may be in feet, who knows?), don't correspond to the visible size of the parachutes, and aren't coherent between the different parachute sizes and effects. The visible on-screen size seems much more coherent.

 

edit: actually, most of the above is irrelevant. Something odd is going on and the visible sizes don't really correspond to their effectiveness.

Edited October 14, 2016 by Plusck

[Tyko]

what would be really helpful is a "rule of thumb" that gives a maximum safe weight for each parachute type - i.e. "each Mk16 can safely land up to XXXXkg", "each Mk2-R can safely land up to XXXXkg"

 I realize that "safe" is negotiable, but I usually shoot for about 6m/s.

Edited October 14, 2016 by tjt

[Kelderek]

I'm sure this is completely outdated, but I suspect this parachute calculator  would still give you a good starting point to use for figuring out what parachutes to use and how many:

http://ksp.freeiz.com/

[bewing]

To simplify it a lot: In previous versions, one MK2-R had exactly the same drag as one MK16. Now it has about 80% as much drag. Yes, the "effective diameter" numbers in the part descriptions probably got swapped -- so you should ignore them for now.

[Plusck]

6 hours ago, peewee69 said:

...

Yet when I tested each of them, I found that the Mk16 slows the descent of my test craft to about 6 - 6.5 m/s and the Mk2-R to about 7 - 7.5 m/s.

...

Having looked even further into this, it's actually even odder than it first appeared...

1 hour ago, Kelderek said:

I'm sure this is completely outdated, but I suspect this parachute calculator  would still give you a good starting point to use for figuring out what parachutes to use and how many:

http://ksp.freeiz.com/

Good link. I wanted to see if it was still valid-ish, and that's when things stopped making sense.

In fact it is relatively accurate, still, as a rough guide.

I tried a few ships. One standard command pod, then a 21-ton thing, then a lander can.

On the command pod, the Mk16 was most effective. This is contrary to what the wiki and the calculator say. However, if the Mk16 was offset slightly to one side, it was less effective than the Mk2-R.

Album of images here: Parachute weirdness album on imgur

For the 21t monster, the calculator told me that 12 small parachutes or 6 large ones would bring me down to about 7 m/s, but that the Mk2-Rs would be better.

Correct!

Spoiler

However, here the Mk2-Rs were far, far more effective than the Mk16s. 5.2 m/s compared to 7.8 m/s.

The calculator also told me that 32 Mk25s would have the same effect.
They didn't.

Spoiler

And all of the above was confirmed when I tried a lander can.

Centrally placed, the Mk16 is better than the Mk2-R.

However, radially placed the Mk2-Rs are significantly better than the Mk16s.

 

So, there you go.
To work properly the Mk16 has to be placed singly and absolutely centred above the CoM. I have no idea why that would be.
Mk2-Rs play well together. Drogues don't.
 

Edited October 14, 2016 by Plusck

[swjr-swis]

Note that parachutes now come with an 'Spread Angle', and this angle - an advanced tweakable that can be edited live even when the chute is already open - affects the efficiency considerably. Iow. you can slow/speed up the fall by whole m/s by changing the angle on already open chutes.

So in your tests, you need to take into account the spread angle. How... no idea; I'm not sure that's been specified anywhere.

Video demonstration: https://www.dropbox.com/s/29imu79itgnayvr/KSP120-ChuteSpreadAngleDrag.mp4?dl=0

Edited October 14, 2016 by swjr-swis
video link

[Plusck]

7 minutes ago, swjr-swis said:

Note that parachutes now come with an 'Spread Angle', and this angle - an advanced tweakable that can be edited live even when the chute is already open - affects the efficiency considerably. Iow. you can slow/speed up the fall by whole m/s by changing the angle on already open chutes.

So in your tests, you need to take into account the spread angle. How... no idea; I'm not sure that's been specified anywhere.

Ah, interesting.

My tests were all done in a clean default sandbox on a completely fresh install, so advanced tweakables were off. I was less trying to give a definitive answer and more trying to see what you should expect from parachute behaviour.
That "spread angle" thing might also explain why an offset Mk16 does so much worse than an aligned one, if it's trying to "spread" due to CoM misalignment.

[Alshain]

@swjr-swis  Do that test again without the stacked parachutes.  I'm wondering if it is smarter than we realize and the lower parachutes are preventing the top parachute from reaching full efficiency.  In the real world you have to have air under a parachute, but if another parachute is blocking it, it's not going to work as well. EDIT: Or more accurately, the combination of chutes won't work as well.  It's a matter of surface area.

In fact I would make the test craft wide enough that at no point do the parachutes overlap regardless of spread angle and then see if the phenomenon occurs.  I bet it doesn't.

Edited October 14, 2016 by Alshain

[bewing]

When Claw put in that little mod to make the chutes prettier by spreading them out so they don't clip, I don't think there was any more logic put into it than that.

I'm sure the loss of efficiency is just a cosine loss. The parachutes create a drag force along their axis, and if that axis is not vertical, you get a cosine factor. The more chutes, the bigger the angle, and the bigger the cosine loss. Which is fairly realistic, so I don't think anyone will find it upsetting.

[swjr-swis]

59 minutes ago, Alshain said:

In fact I would make the test craft wide enough that at no point do the parachutes overlap regardless of spread angle and then see if the phenomenon occurs.  I bet it doesn't.

It does happen though, as this video shows: https://www.dropbox.com/s/0sk42kwh8qwo1jm/KSP120-ChuteSpreadAngleDrag2.mp4?dl=0

 

31 minutes ago, bewing said:

I'm sure the loss of efficiency is just a cosine loss. The parachutes create a drag force along their axis, and if that axis is not vertical, you get a cosine factor.

That would've been my explanation, except the effect appears to be exactly opposite of that: the craft falls faster (less drag) when the spread angle is zero, and slower (more drag) as the spread angle increases.

[Alshain]

@swjr-swis Well that blows my theory out of the water but at least we know now.

[Tyko]

I did some testing on this as well and the less spread the two chutes have, the faster the descent velocity. On a craft with a falling rate of ~7m/s, changing both chutes by one step (7 to 8 for example) seems to vary the falling speed by .1(ish) m/s

[Muskrat]

The spread angle maybe refers to the angle of the runners to the chute? So a smaller angle means the chute opens less?

[Tyko]

On 10/23/2016 at 0:35 AM, Muskrat said:

The spread angle maybe refers to the angle of the runners to the chute? So a smaller angle means the chute opens less?

The spread angle setting changes the angle at which the chutes are spread relative to each other. At a 0 angle they are almost completely overlapping while on a setting of 10 the chutes are spread far enough apart that they might not even touch each other.