European lander works about as well as European Governance.

[Diogenes]

http://www.space.com/34472-exomars-mars ... hotos.html

[paperburn1]

Stupid decimal places

[krenshala]

A relevant post from a member of the Kerbal Space Program community on Reddit: https://gfycat.com/FatHeavenlyChimpanzee

[Diogenes]

A relevant post from a member of the Kerbal Space Program community on Reddit: https://gfycat.com/FatHeavenlyChimpanzee

That was funny.

[happyjack27]

Code: Select all

  //now do final deceleration burn.
  retroRockets.start();
  Thread.sleep(3000); //30 seconds, in milliseconds
  retroRockets.stop();

I have completed implementation of the landing sequence, as per the supplied specifications. Please test before moving to production.

[paperburn1]

stupid zeros

[Diogenes]

Code: Select all

  //now do final deceleration burn.
  retroRockets.start();
  Thread.sleep(3000); //30 seconds, in milliseconds
  retroRockets.stop();

I have completed implementation of the landing sequence, as per the supplied specifications. Please test before moving to production.

You may have written that as a joke, but it appears to be D@mn close to the truth.

For reasons that are still a mystery, the lander ejected both its heat shield and parachute way ahead of schedule. Schiaparelli then engaged its thrusters for a painfully brief three-second burst—a procedure that was supposed to last for 30 seconds once the lander was just a few feet off the ground. The lander’s onboard computer, it would appear, seems to have thought it was close to the surface. Indeed, Schiaparelli even took the time to switch on some of its instruments, including tools to record the planet’s weather and electrical field.

http://gizmodo.com/software-error-impli ... 1788187542

[happyjack27]

I knew it was 3 vs 30 seconds.

Better written code would have been simply:

Code: Select all

retroRockets.timedFire(30.0);

But there is the issue that the system needs to be responsive to its environment, so it can deal with the unforeseen. For instance, it should be able to adjust its timed fires based on actual speed, mass, acceleration, thrust, and distance from the ground.

Acceleration can be measured with an accelerometer. Speed by difference in ground distance. Mass cannot be measured directly, nor can thrust, but they get cancelled out by measuring acceleration per throttle unit. This can be induced by feedback - fire the engines, measure the deceleration. And this tells you your burn time.

At the end your command will be:

Code: Select all

retroRockets.targetVelocityAtDistance(0.0,0.0);

It's precisely this adjustment, that, from what I read, failed.

There are four systems that could have failed: accelerometer, distance measurement, the rockets, or bad coding.

They should have tested these systems on earth, under abnormal circumstances such as coming in too fast or too slow. An they should test the actual hardware that'll go to space. Way cheaper to test it in earth than in space.

It doesn't look like they did.

[Tom Ligon]

Well, as they say, "Mars is hard."

Although clearly you can dent it a little with a fall of several kilometers.

From what I read, the system was set up to be responsive to the environment. It may have thought it had touched down. Likely a touchdown or "squat" switch was stuck, causing early termination of the burn.

If we had people on Mars, i.e. the bootstrapping proposals, we could send somebody over to check. Otherwise you just guess a lot and try another one. If you guess wrong, that one crashes too.