Boeing software glitch sends Starliner crew capsule into the wrong orbit

originally posted by: firerescue
a reply to: charlyv

Big problems for software engineering at Boeing, especially after all of the fallout with the Max.

That's what you get when pay bunch of coolies in India $9/hr to write flight software

Get what you pay for ………….

'Coolies' that currently have a probe in orbit around both the moon and Mars as well as a profitable satellite launching system? Those 'coolies'?

a reply to: charlyv

This makes me skeptical of their modeling. Their modeling and simulation check out for a hypothetical in-flight abort test and this is their justification for not doing an actual in-flight abort test.

That said, SpaceX is doing their in-flight abort no earlier than mid-January. Hopefully I can be there for it.

I should get to post them in the next couple days, but I got pics of her going by Las Cruces on the way to White Sands.

originally posted by: jrod
a reply to: charlyv

This makes me skeptical of their modeling. Their modeling and simulation check out for a hypothetical in-flight abort test and this is their justification for not doing an actual in-flight abort test.

That said, SpaceX is doing their in-flight abort no earlier than mid-January. Hopefully I can be there for it.

A vector placing the capsule in the wrong orbit is a big glitch. Something like that should have been caught in a simulation or the modeling software made the same assumptions that the real-time flight software did.

Orbital mechanics is something that should certainly be well understood, so cost cutting corners comes to mind.....

I know this stuff ain't easy, but a company like this cannot afford this kind of failure, especially being a prime defense contractor as well.

a reply to: charlyv


From another website - United Launch Alliance (ULA) ATLAS V services guide

www.ulalaunch.com...

Chapter 5 gives the interfaces between launch vehicle and payloads

Signal from Launch vehicle should have triggered the clock in CST Starliner

a reply to: Zaphod58

Starliner made successful landing at White Sands this morning

www.youtube.com...

Came down before dawn local time

originally posted by: firerescue
a reply to: Zaphod58

Starliner made successful landing at White Sands this morning

www.youtube.com...

Came down before dawn local time

I watched it live on YouTube. The mission, while not a complete success, certainly was not a failure. This is how it's done folks. Build, test, repeat until an astronaut feels confident enough to crawl in and take that ride. That's the real test.

a reply to: LogicalGraphitti

It was cool as hell to see it going over in to White Sands.

originally posted by: machineintelligence

Yes they took a beating today it seems.

A 1% decline in share prices is a 'beating'?

originally posted by: jrod
a reply to: charlyv

This makes me skeptical of their modeling. Their modeling and simulation check out for a hypothetical in-flight abort test and this is their justification for not doing an actual in-flight abort test.

That said, SpaceX is doing their in-flight abort no earlier than mid-January. Hopefully I can be there for it.

Hence a glitch, it doesn’t say the coding was incorrect just yet, what if there was a failure within the FCS which meant an instruction wasn’t followed?

Recovering the craft is a start to find out what kind of a glitch it was.

At least Starliner survived reentry and the parachutes and airing deployed. Still no official word if astronauts will be flying on the next mission.

Embarrassing, but I don’t think it’s a big deal. If astronauts had been piloting the Starliner they would have made it to the ISS or any LEO with no problem. Everything else seemed to work OK, although it had to return early for lack of fuel I believe. I heard it made a good landing in New Mexico. It was its first flight; space flight is hard and something almost always goes wrong on a first flight. Look at the trouble SpaceX had with the test of its emergency launch escape system, and the Dragon is a lot more mature than the Starliner. Nevertheless, I’d fire whoever was responsible for the timing error on the spacecraft, because Boeing is getting too much bad publicity for mistakes and sloppy work. I noticed the stock market went up 108 points just because the Boeing CEO resigned. It was about time!

a reply to: Scapegrace
Astronauts are pretty much along for the ride. There’s no more re-entry vehicle that needs piloting, most everything is automated, humans are too inaccurate with our motor skills to be wasting limited fuel with mistakes, Maybe at close distance docking with the ISS but still.

a reply to: Scapegrace

The timing error was between the first and second stage. The Atlas sets the timing for the Centaur second stage.

It may be that there is no issue with the Boeing software.

Software systems can get quite complex and testing can be challenging. That is why live testing happens. Have to give them a couple of failures in a row to start really blaming their work as subpar or a failure.

the Boeing CEO resigned

I saw news that said he was fired.

a reply to: roadgravel

It's being reported both ways.

originally posted by: roadgravel
It may be that there is no issue with the Boeing software.

Software systems can get quite complex and testing can be challenging. That is why live testing happens. Have to give them a couple of failures in a row to start really blaming their work as subpar or a failure.

Indeed - I write code, baby level stuff, but enough to know that no matter how many error traps you put in someone will do something you haven't thought of. It would only take a tiny delay in a mechanical system to throw out an entire chain of electronic events.

a reply to: OneBigMonkeyToo

will do something you haven't thought of

That is what unit tests, static code analysis, fuzzing are for. It is not hard. It just takes time. But time is money. And nobody is going to pay for that...

originally posted by: moebius
a reply to: OneBigMonkeyToo

will do something you haven't thought of

That is what unit tests, static code analysis, fuzzing are for. It is not hard. It just takes time. But time is money. And nobody is going to pay for that...

After 40 years of programming, I would say it is not that simple. Sometimes it is not even an error, it is just incorrect data which cannot be known as incorrect. Software worked but not in the manner expected.