View Single Post
  #16  
Old April 29th 19, 02:01 PM posted to sci.space.policy
David Spain
external usenet poster
 
Posts: 2,901
Default SpaceX Dragon 2 capsule destroyed in abort motor ground test

On 4/22/2019 7:05 AM, Jeff Findley wrote:

This is bad. My guess is at least a year delay for SpaceX commercial
crew. Here is hoping that Boeing gets its act together because we need
something to replace Soyuz for US crew.

Jeff


FWIW the conventional wisdom seems to be focusing on the COPV tanks used
to fuel the SuperDracos. But this is pure speculation (SWAGs if you
will). I have seen ONE still frame taken from a normal speed camera that
appears to show an explosion taking place "around" the capsule where the
hatch window still appears in the frame. In another video you can see at
least two explosions, the first as mentioned where the capsule is
largely still intact and a 2nd to the left (from the viewer viewpoint)
of the first which appears to blow the capsule off the test stand. Which
would tend to indicate multiple possibly cascading explosions.

AFAIK know from what has been published in other forums the capsule is
believed to be a total loss.

Yes this is bad, but the test engineer in me is very happy that this
happened during *testing*. Although a RUD is never a welcome event it is
a learning opportunity with the net result of an improved design. This
is what happened after the Apollo 1 fire. The following block
improvements to the Command Module made the follow-ons very different
from the original article including the wiring and hatch design from
what I have read. It was poor judgement and bad test design that I feel
were the real reason behind the fatalities that should not have happened.

You could say we were "lucky" that this happened while no crew were
on-board, but I *hate* that term. What is proper to say is that an
anomaly was caught in testing, just as it should be. What keeps me up at
night are the anomaly's that I didn't test for. That is why you do
design review after design review and test and test again, and then
alter and add to the testing regimen and test again to prove out design
margins.

Luck as defined simply means you missed a test and found a failure mode
at an opportune time.

David