Space X CR6 First stage landing

Both the failing field joints ("O-rings") and detaching external tank insulation weren't quite bad enough to kill a Shuttle, were all but ignored, until of course they couldn't be ignored.

Stopping a first stage rocket traveling back to an ocean going barge from supersonic speeds with substantial lateral velocity is hard, no matter what engine is performing its suicide burn while coordinating with RCS thrusters for attitude keeping.

If you deliver your primary objectives well and keep improving at everything else in the process, your customers will keep on loving you for it.

Look at a couple of the issues they've faced and managed to deftly conquer. Take vehicle design, for example. If you design a purely expendable launcher you tend to want to minimize engine count and concentrate on 2nd stage performance, which leads to a design with an expensive 2nd stage and a cheaper 1st stage with very little throttle authority (due to using a small number of engines). That makes reuse very difficult and not very worthwhile because you'd need to either try to reuse the 2nd stage (a much more challenging task) or to heavily modify the 1st stage, which would then only save the cheapest component of the launcher. So if you want to concentrate on reuse you need a vehicle like the F9, with lots of 1st stage engines, and most of the cost in the 1st stage.

But even then you need to figure out how to run your business. It would be easier to envision the launcher with reuse an integral part of the design from day 1, which would lead you towards testing the reusability aspects of the design prior to the first launch. But that would be tremendously expensive and would make it more difficult to keep a company running while that was going on. SpaceX has been able to provide numerous orbital launches with their rocket and earn a tremendous amount of revenue before it is capable of actually being operated in a reusable fashion. A common engineering dictum is "you ain't gonna need it", to avoid introducing unnecessary design elements that aren't strictly immediately necessary. But without designing the Falcon 9 with reuse in mind from day 1 they would have had a much more difficult time in developing the reusable aspects. Note that SpaceX has made use of hundreds of millions of dollars of flight hardware in their various landing tests, all of which they not only didn't have to pay for out of pocket but actually earned a profit for. That's smart business. Meanwhile, for all of the launches where they haven't attempted landings they've added operational experience and real-world testing of their overall vehicle design and components.

Big, bureaucratic companies that are bringing in billions a year in revenues (like ULA or Arianespace) don't tend to engage in behavior like this, because they don't have the hunger, they don't have the boldness, they don't have the coherent foresight. What SpaceX has done is something that a lot of other companies are going to replicate, likely fairly successfully, once it's been proven. But getting to that point has required not just engineering acumen but enormous risk management skills and business savvy. They'll have a significant market advantage due to their head start for quite some time, and hopefully they will retain enough boldness and savvy to continue moving ahead even as others attempt to catch up.

But the economics is probably more important. The incumbents have such sky-high development costs that reusability seemed like an impossible task to do economically. Last time anyone tried they came up with STS, which was highly capable but ultimately dangerous and ruinously expensive. A new reusable system seemed like it would take half a trillion dollars in development, and would probably never have the flight rate to make it worthwhile. Elon thought that he could do it better. NASA believes designing a reusable system would be too expensive for them. They're probably both right.

Since valves cannot open instantaneously, the control system has to know how long the valve takes to open/close, and start that process a little bit early so the valve is open when it needs to be open. If something goes wrong with the valve, it doesn't open as quickly, and so the computer model thinks it should be open and isn't, which could easily go off into an unstable state.

The "lag" is a delay between your steering signal and the state of the system changing in response. In this case, because of static friction, valves controlling fuel flow reacted with delay (I'm not sure if they're talking about Merlin's valves or attitude control engines though) - so the rocket did not react immediately after being told to do so.

There's a whole math field around that topic, known as "control theory". There are ways to determine the limits of inputs and outputs that will allow your control system to keep everything in the desired state. In this case however, the input lag wasn't accounted for and it pushed the system out of the space of stable states, beyond the ranges where the control unit could keep up with the changes, so it ended up oversteering. Fortunately, in this case it's mostly a software problem, i.e. they could tweak the controller to deal with the lag.

The best way to imagine it is to recall your first experiences with a bike or ice skates - trying to balance yourself on them is exactly the kind of thing the rocket was doing, probably with similar results.

My control theory is a bit rusty nowadays, but I hope this clears up some things.

[0] - if you ever hear about "PID controller", this is exactly where the name comes from - proportional, integral, derivative.

The Falcon 9 cannot hover. The thrust of the single landing engine, even when throttled down to minimum, will provide significant lift to the empty rocket. That's part of what makes this maneuver so difficult and so awesome: it's called the hoverslam, where you achieve 0m altitude and 0m/s velocity at time = 0.

For reference, though, the Grasshopper/F9R-dev1, could hover due to it only having 3 engines, not 9, and because they could ballast it with additional fuel.

Also, falling over almost certainly also leads to a writeoff, as otherwise they'd just fly the rocket into a big pit of fire retardant foam blocks.

I'd imagine that the lateral stresses involved in a crash landing are more than enough to write off the rest of the rocket, and so the explosion would just make the barge slightly harder to clean up.

They probably are spending more effort into not letting it fall over.

Some of the parts from the first missions were re used on the last. The SRBs weren't disposable, but they were essentially refurbished after each launch.