It seems that in one of the last few version the physics have been tweaked to make re-entry and landing much more dangerous.
In particular, parachutes now (1.0.5) seem absolutely useless. I've been able to get by with a little trickery where capsules are concerned, but I recently decided to put parachutes on my stages so they touch down softly instead of exploding, instead of blowing up. I know that in vanilla KSP stages beyond a few km just vanish anyway, but I'd still like to design a stage that in principle could land in one piece.
I decided to test it by building only the individual part of my bottom stage as its own rocket, and trying to launch and then parachute down with it. My rocket was made up of:
- Stayputnik (only for the test, so I can control it, should have negligible effect on mechanics)
- 2x FL-T800 liquid fuel tanks (only 10% full for the test)
- LV-T45 "Swivel" engine
- Variable number of Mk2-R radial parachutes
All these components are stacked like a tower, except the parachutes. With these, I tried several different configurations:
- One parachute on the side of the fuel tanks, above dry CoM (located by looking at CoM with tanks 0% full)
- One parachute on the side just below the CoM, a second parachute above the CoM on the same axis as the first one
- One parachute above the CoM, cluster of 3-5 parachutes below CoM
- I have also tried RealChute Radial Chute from the RealChutes mod (I did all applicable tests with and without the mod, seems to make no difference)
Settings on the parachutes were always default, and they definitely opened way ahead of the critical moment (speed was constant for the last several seconds in each case).
So the idea is I launch this "rocket", get it to about 1 km before fuel runs out, activate parachute and see how well it lands. Since it's difficult to control this, most landings have been on the ground, not water. You should be able to reproduce my experiment if you want to.
The rocket has dry mass of about 3.5 t. According to the calculator, 2 radials should be enough to land with 8 m/s and 5 should be enough for 5 m/s. However, 2 chutes result in a very rough (~18 m/s) crash. 6 chutes result in about 8 m/s and the engine and tank explode on ground contact.
Surely I can just attach 100 chutes to this stage and it will land very gently. But the idea here is that I want expensive stages to land intact to be more economical (again, leave aside how the game handles stage recovery for a second). The point is that this rocket costs about 2.5k and the radial chutes are 300 each. If I need to deploy more than 1-2 chutes, a large fraction of the money starts going to the chutes, which seems silly to me.
Are parachutes supposed to be this useless, or am I doing something wrong? I noticed, for instance, that the chutes overlap when deployed - I never noticed this being important before, but if two chutes overlap does that somehow decrease their effectiveness? Is there some sort of trick to correct chute placement? It seems like I should have more than enough chutes to land safely, but with few chutes I get 10+ m/s, and even 6 chutes only gets me to 8 m/s which still explodes the rocket.