OT: 30th anniversary of the Challenger

Manned spaceflight, besides being useless, is dangerous.

On Thu, 28 Jan 2016 14:47:53 -0800, John Larkin Gave us:

John Larkin, besides being an abject idiot, is an uniformed ineducable twit as well.

It bummed me out so much that I stopped out of college for a term. I ended up traveling, meeting my wife, moving across the continent, and finishing up at a different school.

I'm not sure if ALL of that was because of the Challenger, but I'm sure that some of it was.

Not entirely useless, definitely dangerous, and even more dangerous when "engineers" and management convinced themselves that an O-ring burnthrough process occuring in the SRBs that wasn't specified in the design, and wasn't well-understood, meant they had a "2/3rds safety factor.

I have a lot of respect for Feynman, but he and the rest of the Rogers Commission got it mostly wrong. Everybody including NASA and Thiokol knew about the O-rings getting harder at low temperatures, so that water glass stunt of his was entirely beside the point. Rogers & Co. didn't have time to understand the very complex NASA process and vocabulary. The House Commission didn't get the publicity, but it did mostly get the right answer.

I highly recommend Diane Vaughan's "The Challenger Launch Decision"--normally I don't have much time for sociologists, but she's a shining exception to the general rule. It's the fruit of a good three years' immersion in the primary documents and many interviews with the dramatis personae. (Charles Perrow's "Normal Accidents" is another good sociology book.)

Cheers

Phil Hobbs

I became a member of the fault analysis team... found a power supply redundancy, when one power supply fails it takes down them all.

Was the first person to get a HexFET space-qualified to solve the problem. ...Jim Thompson

That sounds more like a lack of redundancy. ;-)

joe

For an amusing take on modern sociology in general, consult the great Tom Lehrer:

It's not quite as good as his earlier stuff (e.g. "Poisoning Pigeons in the Park") but very amusing.

Cheers

Phil Hobbs

On Fri, 29 Jan 2016 04:51:46 +0000, Joe Hey Gave us:

These days a switcher can detect the overcurrent and shut down without failing.

Those were likely linear supplies and they are not so tolerant.

But I do not know what was used.

I do know I have a friend who had a triple redundancy supply set that would do the same thing on a critical BAE ATC displays on US ships, and they implemented his little detection, cut-out circuit which took care of it fleet wide.

Oh wait, I think I know what he meant: Although there was redundancy built in, if one of the redundant supplies would fail, it would do in such a manner that the others would still go down anyway as a result of the way in which it failed.

Was that it? :)

joe

On Fri, 29 Jan 2016 10:52:09 +0000, Joe Hey Gave us:

I think you have a major literacy problem.

One supply fails and that failure takes the other one down with it. A five year old kid could get it.

}snip{

I think you got that wrong.

Yes I got it now, thanks.

BTW I'm not a 5 yr old kid. :)

joe

On Fri, 29 Jan 2016 13:25:47 +0000, Joe Hey Gave us:

OK Hoe Jey... whatever you say.

DecadentLinuxUserNumeroUno Wrote in message:

I think the Space Shuttle used a combination of linear and switching supplies, system dependent. I know many of the supplies for the display units were linear and very heavy.

The schematics for the Apollo guidance computer are online including power supply. I'm not sure what the redundancy situation for the power supplies/computers was exactly, but there are of course plenty of places in the one supply I see on the schematic where if that little thingy goes open-circuit it sure looks like the whole system is down for the count.

I don't remember the details now, 30 years ago I was contracting at Sperry/Honeywell Space, but I remember adding the HexFET's as disconnects. ...Jim Thompson

That's called a single-thread failure mechanism, or used to be. Redundancy refers to a usually unneeded capability that is switched in to take over, fully or partially, the function of a detected failed subsystem.

Sure you were.

That's a lot verbiage to explain away a SHIT design that was later changed. Clevis and tang garbage.

I don't think the Apollo had any redundancy built into the computer/guidance system. But, they had TWO of them on most missions. One in the command module, one in the LEM. On Apollo 13, they used the computer in the LEM to do a bunch of stuff. The whole mission design was really smart and had a lot of fail-safe stuff in it. So, for instance once the SM burn for injection toward the moon happened, they would sail once around the moon and then, without another burn to reduce enrgy and hold them in a lunar orbit, they'd loop right back toward earth. So, by doing nothing at the moon, they would get returned to earth, but with some delay. Then, they had the engine in the LEM which had enough energy to get them down from earth orbit. So, as long as only one system went kaput, they had, at least orbitally, a way out.

Jon

The idea of using multiple identical units for redundancy such as diesel generators in NPPs relies on the assumption that faults are _independent_.

In reality, contaminated fuel or in the Fukushima case all diesel generators became wet due to the tsunami.

For a truly redundant system, each system should use a different technology.