Mars Climate Orbiter Team Finds Likely Cause Of Loss
Andrew Plotkin
> RELEASE 99-113
>
> MARS CLIMATE ORBITER TEAM FINDS LIKELY CAUSE OF LOSS
> [...]
> The peer review preliminary findings indicate that one team
> used English units (e.g., inches, feet and pounds) while the other
> used metric units for a key spacecraft operation. This
> information was critical to the maneuvers required to place the
> spacecraft in the proper Mars orbit.
Several days ago, I wrote:
> In hindsight, *all* errors are preventable.
I think I must retract that attitude. This one is the classic "preventable
in foresight".
I still don't think this is cause for massive firings (what, throw away
the people who *have* experience?) or for public apologies (this is
about getting science data, not caucus votes or Nielsen ratings.)
I think it *is* cause for a lot of people walking around NASA with a sick,
miserable feeling in the pit of their stomachs, much like the feeling you
get when you screwed up, and there's nothing you can do about it except do
better next time.
I think probably a lot of people are.
--Z
"And Aholibamah bare Jeush, and Jaalam, and Korah: these were the
borogoves..."
Edward Lyons
NASA and all of its affiliates and contractors to finally standardise on the
metric system. There is no excuse for this kind of ineptness.
Eddie Lyons
Ron Baalke <baa...@kelvin.jpl.nasa.gov> wrote in message
news:30SEP199...@kelvin.jpl.nasa.gov...
> Douglas Isbell
> Headquarters, Washington, DC Sept. 30, 1999
> (Phone: 202/358-1753)
>
> Mary Hardin
> Jet Propulsion Laboratory, Pasadena, CA
> (Phone: 818/354-5011)
>
> Joan Underwood
> Lockheed Martin Astronautics, Denver, CO
> (Phone: 303/971-7398)
>
> RELEASE 99-113
>
> MARS CLIMATE ORBITER TEAM FINDS LIKELY CAUSE OF LOSS
>
> A failure to recognize and correct an error in a transfer of
> information between the Mars Climate Orbiter spacecraft team in
> Colorado and the mission navigation team in California led to the
> loss of the spacecraft last week, preliminary findings by NASA's
> Jet Propulsion Laboratory internal peer review indicate.
>
> "People sometimes make errors," said Dr. Edward Weiler,
> NASA's Associate Administrator for Space Science. "The problem
> here was not the error, it was the failure of NASA's systems
> engineering, and the checks and balances in our processes to
> detect the error. That's why we lost the spacecraft."
>
> The peer review preliminary findings indicate that one team
> used English units (e.g., inches, feet and pounds) while the other
> used metric units for a key spacecraft operation. This
> information was critical to the maneuvers required to place the
> spacecraft in the proper Mars orbit.
>
> "Our inability to recognize and correct this simple error
> has had major implications," said Dr. Edward Stone, director of
> the Jet Propulsion Laboratory. "We have underway a thorough
> investigation to understand this issue."
>
> Two separate review committees have already been formed to
> investigate the loss of Mars Climate Orbiter: an internal JPL peer
> group and a special review board of JPL and outside experts. An
> independent NASA failure review board will be formed shortly.
>
> "Our clear short-term goal is to maximize the likelihood of a
> successful landing of the Mars Polar Lander on December 3," said
> Weiler. "The lessons from these reviews will be applied across the
> board in the future."
>
> Mars Climate Orbiter was one of a series of missions in a
> long-term program of Mars exploration managed by the Jet
> Propulsion Laboratory for NASA's Office of Space Science,
> Washington, DC. JPL's industrial partner is Lockheed Martin
> Astronautics, Denver, CO. JPL is a division of the California
> Institute of Technology, Pasadena, CA.
>
> - end-
>
Michael L. Belanger
1: The calculations were done on 2 sets of computer software using
different measurement systems (and one of them English!).
2: The calculations were done by hand!
I cannot decide which is more disturbing,
MLB
--
Michael L. Belanger
Ochtauber Interactive
mbel...@deltanet.com
http://users.deltanet.com/~mbelange/
kuhnfucius
spacecraft was built - used different measurement systems, and quality
control failed to notice the discrepancy, said Edward Weiler, NASA's
associate administrator for space science.
LIKE I SAID LAST WEEK!!! Smaller, cheaper, faster and BOOM!! The first
thing to go will be Quality, and the second will be CONTROL. Organizations
sometimes mimic people. Nasa is now proving they are still the same bunch
that produced Hubbles mirror. I was hoping that they learen from the
lesson. Dilbert's boss is now running NASA. This is not a case of
difficulty.
Michael L. Belanger <mbel...@deltanet.com> wrote in message
news:37F3D3...@deltanet.com...
Michael L. Belanger
time. Shouldn't it have appeared after TMI (Trans-Mars Insertion) or
during any number of corrective burns on the way in. Was it a problem
with just part of the software, the part responsible for orbital
capture?
And just because I'm curious (as at a train wreck), which group was
using which system? Who was living life in the 1960's?
kuhnfucius wrote:
>
> Two teams - one at JPL and another at Lockheed Martin in Colorado where the
> spacecraft was built - used different measurement systems, and quality
> control failed to notice the discrepancy, said Edward Weiler, NASA's
> associate administrator for space science.
>
> LIKE I SAID LAST WEEK!!! Smaller, cheaper, faster and BOOM!! The first
> thing to go will be Quality, and the second will be CONTROL. Organizations
> sometimes mimic people. Nasa is now proving they are still the same bunch
> that produced Hubbles mirror. I was hoping that they learen from the
> lesson. Dilbert's boss is now running NASA. This is not a case of
> difficulty.
>
> Michael L. Belanger <mbel...@deltanet.com> wrote in message
> news:37F3D3...@deltanet.com...
> > This presents one of two horrifying scenarios:
> >
> > 1: The calculations were done on 2 sets of computer software using
> > different measurement systems (and one of them English!).
> >
> > 2: The calculations were done by hand!
> >
> > I cannot decide which is more disturbing,
> > MLB
> >
Michael L. Belanger
matt o'
like it....
"Edward Lyons <eddie...@virgin.net> wrote in message
> NASA and all of its affiliates and contractors to finally standardise on
the
> metric system. There is no excuse for this kind of ineptness.
so your saying that some of the brightest people in the world (if not THE
brightest) are inept..... i'm sure if someone went through your job
performance they wouldn't finde any ineptness, ehh?
Michael L. Belanger
world due to an error in which measurement system you are using, your
are guilty of ineptness. It might as well the the dictionary definition
of the word.
--
rk
> When you crash $125 million in taxpayer's money in front of the whole
> world due to an error in which measurement system you are using, your
> are guilty of ineptness. It might as well the the dictionary definition
> of the word.
definition:
inept: 1. not apt or fitting; unsuitable; inappropriate. 2. not sensible;
foolish. 3a. awkward; clumsy. b. incompetent.
i don't know enough about the cause of the failure to say much w.r.t. "inept"
but i do know that anyone can make a mistake. for example, i recently found
this on the internet:
It might as well the the dictionary definition of the word.
now, i admit my spelling and grammar is not all that good but that looks like an
error.
rk
rk
> In article <37F4106F...@NOSPAM.erols.com>,
> rk <stel...@NOSPAM.erols.com> wrote:
> >> When you crash $125 million in taxpayer's money in front of the whole
> >> world due to an error in which measurement system you are using, your
> >> are guilty of ineptness. It might as well the the dictionary definition
> >> of the word.
>
> >definition:
> >inept: 1. not apt or fitting; unsuitable; inappropriate.
>
> JPL has operated spacecraft through many planetary encounters, and
> done an excellent job in the past. The loss of a spacecraft due to
> an un-corrected unit conversion error is unsuitable and inappropriate.
>
> >...2. not sensible; foolish.
>
> Failing to notice and correct a unit conversion error, when a $125 million
> spacecraft is involved, is not sensible and is foolish.
>
> >...3a. awkward; clumsy. b. incompetent.
>
> Failing to notice and correct such an error is awkward, clumsy and
> (possibly) incompetent. There, the loss of the MCO spacecraft seems
> to fit all of your dictionary's definitions of ``inept'' (isn't
> semantics fun...)
>
> Seriously. Someone making a unit conversion error is not inept. It
> happens all the time. But ``inept'' does seem to apply to a program
> operating a $125 million spacecraft, and which fails to notice and correct
> a unit conversion error. At least one that mistake affects something as
> critical as safe spacecraft navigation.
fc,
you usually are careful in your posts (i wish i was 1/10 as careful) but i want
to make sure that your snip doesn't change the meaning of what i said or
intended to say.
i gave the dictionary definition of "inept" to follow up the previous poster
and then added:
"i don't know enough about the cause of the failure to say much w.r.t.
"inept" but i do know that anyone can make a mistake. "
so i make no claim of ineptness. in fact, i consider the navigation guys at
jpl to be *wizards* (and i have met some of them and knew some of them - not
sure who's there now from who [or is it whom?] that i knew). it was apparent a
mistake was made and anyone can make one. much of the time mistakes are made
and nothing bad happens. sometimes it does. they are mistakes. there are, of
course, other cases where things are worse than simple mistakes but that's
another story.
at the conference i was at this week a guy went up and talked about a mistake
in his organization; went through it from top to bottom, with the result of a
complete loss of a science mission. bared his soul. spilled his guts. took
the blame. a very courageous act. they simply missed something and made a
mistake. can happen to anyone. it was a very well received talk. i would not
call his organization nor the responsible engineer "inept" at all; it was
simply an error. it's a risky business.
rk
Andrew Carol
jimpi...@aol.com (Jimpiccard) wrote:
> They do less, with less. But you can't compare any of these to a
> Galileo, a Cassini, or a Hubble.
They do less with a LOT less. These things are doing 75% of what the
expensive missions did, but at 20% of the cost.
--
Andrew Carol xaec...@ix.netcom.com
(Remove leading x from my address to e-mail)
Frank Crary
>my car get's three rods to the ram's head and, damn it, that's the way i
>like it....
You do? That's terrible mileage. What do you drive? A tank? A car ought
to get at least four leagues to the ram's head (or a quarter of a league
per, if you get creative about the definition of a league...)
>> NASA and all of its affiliates and contractors to finally standardise on
>the
>> metric system. There is no excuse for this kind of ineptness.
>so your saying that some of the brightest people in the world (if not THE
>brightest) are inept..... i'm sure if someone went through your job
>performance they wouldn't finde any ineptness, ehh?
That depends. I make lots of similar errors. But I also catch them
before I start relying on the results. One of the quotes in the
press release said more or less the same thing: The error wasn't
a failure to convert English units to metric; it was a failure
to catch and correct this mistake.
Frank Crary
CU Boulder
Frank Crary
Michael L. Belanger <mbel...@deltanet.com> wrote:
>This presents one of two horrifying scenarios:
>1: The calculations were done on 2 sets of computer software using
>different measurement systems (and one of them English!).
>2: The calculations were done by hand!
>I cannot decide which is more disturbing,
If I understand it correctly, it was the former problem. Data from
LockMart, presumably the output from a computer program, was sent
to JPL, and then used as input for navigational calculations (also,
presumably done by a computer.) At some point in communicating the
data, someone forgot to convert the units. (Or, equally likely,
someone assumed someone else had/would do the conversion, or even
did not know that a conversion was required.) The calculations
were definitely not done by hand. First, I know that isn't how
JPL does things. Second, the calculations can't realistically be
done by hand. (Just try calculating a solution to a non-linear
equation by hand, then think about how long it would take you
to solve a few thousand...)
Frank Crary
CU Boulder
Frank Crary
kuhnfucius <kk...@acworld.net> wrote:
>Two teams - one at JPL and another at Lockheed Martin in Colorado where the
>spacecraft was built - used different measurement systems, and quality
>control failed to notice the discrepancy, said Edward Weiler, NASA's
>associate administrator for space science.
>LIKE I SAID LAST WEEK!!! Smaller, cheaper, faster and BOOM!! The first
>thing to go will be Quality, and the second will be CONTROL. Organizations
>sometimes mimic people.
And sometimes not. Lunar Prospector definitely didn't go boom. Not did
it have any problems worth mentioning. It was probably the most
extreme case of ``smaller, cheaper, faster'' in the would Discovery
program. SNOE didn't have any problems, and it may very well be the
smallest, cheapest and fastest mission NASA has funded since the 1960s.
There are good and bad ways to deal with small, cheap, fast missions.
MCO definitely dealt with it in a bad way. I do note that LP was managed
and operated by Ames Research Center, and SNOE by the University of
Colorado, not JPL. This may be significant. It may be easier to start
from scratch (or nearly from scratch) rather than change an approach based
on large, expensive, slow missions.
>...Nasa is now proving they are still the same bunch
>that produced Hubbles mirror.
That doesn't follow: Hubble was not a ``smaller, cheaper, faster'' mission
by any stretch of the imagination. Hubble showed that even big, expensive,
slow spacecraft can have serious problems. Small, cheap, fast ones can also
have problems, but for completely different reasons.
>...I was hoping that they learen from the
>lesson. Dilbert's boss is now running NASA. This is not a case of
>difficulty.
I disagree. Doing more or less the same thing for a tenth the cost
_is_ difficult. It is possible, but it requires very different ways
of managing and operating the mission. Learning what those different
ways are isn't trivial.
Frank Crary
CU Boulder
Frank Crary
>time. Shouldn't it have appeared after TMI (Trans-Mars Insertion) or
>during any number of corrective burns on the way in. Was it a problem
>with just part of the software, the part responsible for orbital
>capture?
Rumor has it (i.e. I got an email from someone with more details, but
it was forwarded and I don't remember what the original source was...)
that the problem probably did appear before TMI (and before TCM-4)
but didn't cause enough of a problem to be noticed/considered
significant/whatever.
>And just because I'm curious (as at a train wreck), which group was
>using which system? Who was living life in the 1960's?
LockMart. Hardware is usually designed and built in English units.
There are moderately good reasons for that. If you go to a machine
shop, you will notice that all the dials on the lathes and milling
machines are in English units. Companies like LockMart have a huge
amount of money invested in the tools to build spacecraft, and have
since the 1960s. Switching to metric would mean replacing them, and
that's _very_ expensive. You can't even phase them out without having
to deal with mixed units for a while. As MCO showed, dealing with
mixed units is a bad idea. In contrast, spacecraft navigation is
done in metric units, with some data in sec. or Hz (which is the
same in English and metric units.) The software is much more recent
than the tools in a machine shop, and sometimes written specifically for
the mission, so it uses metric units.
Two footnotes: First, English versus metric isn't the only possible
source of problems. Even if everyone is using metric, is that cgs
or SI metric? I've never heard of that affecting a spacecraft, but
I have seen it cause more than enough confusion and minor problems
in data analysis. Fortunately the cgs and SI numbers usually differ
by several orders of magnitude, so this is an easier problem to catch.
Second, my boss occasionally complains about the term ``English'' units.
She is English, and notes that the English don't use ``English'' units...
Frank Crary
CU Boulder
Frank Crary
rk <stel...@NOSPAM.erols.com> wrote:
>> When you crash $125 million in taxpayer's money in front of the whole
>> world due to an error in which measurement system you are using, your
>> are guilty of ineptness. It might as well the the dictionary definition
>> of the word.
>definition:
>inept: 1. not apt or fitting; unsuitable; inappropriate.
JPL has operated spacecraft through many planetary encounters, and
done an excellent job in the past. The loss of a spacecraft due to
an un-corrected unit conversion error is unsuitable and inappropriate.
>...2. not sensible; foolish.
Failing to notice and correct a unit conversion error, when a $125 million
spacecraft is involved, is not sensible and is foolish.
>...3a. awkward; clumsy. b. incompetent.
Failing to notice and correct such an error is awkward, clumsy and
(possibly) incompetent. There, the loss of the MCO spacecraft seems
to fit all of your dictionary's definitions of ``inept'' (isn't
semantics fun...)
Seriously. Someone making a unit conversion error is not inept. It
happens all the time. But ``inept'' does seem to apply to a program
operating a $125 million spacecraft, and which fails to notice and correct
a unit conversion error. At least one that mistake affects something as
critical as safe spacecraft navigation.
Frank Crary
CU Boulder
Jimpiccard
>>thing to go will be Quality, and the second will be CONTROL. Organizations
>>sometimes mimic people.
I think that may indeed be true in this case; just as it was found true of the
LockMart corporation in the past two weeks, after their string of failures;
i.e, aggressive cost-cutting leading to real cuts on quality control.
Goldin said after the failure last week that occassional failure like this was
normal; and even to be expected and applauded. He is wrong in this case. The
type of failure he was "extolling" was NOT intended to cover increased
screw-ups due to cost-cutting; but simply increased flight rates, etc. This
sort of failure is actually fairly unprecendented in terms of a single failure
of this type causing this mode of failure. Since it hasn't happened in the
past 35 years of planetary exploration, it seems to point doubly at Goldin's
"cut, and if it fails, it's YOUR fault, not mine" behavior of this decade.
>And sometimes not. Lunar Prospector >definitely didn't go boom.
That's not a valid comparison at all. The moon mission was closer, and simpler;
it did NOT have an atmosphere to run into; and didn't even have to have a
computer on board!
> Doing more or less the same thing for a tenth the cost
>_is_ difficult. It is possible, but it requires very different ways
>of managing and operating the mission.
Straight from Goldin's publicist; it's also wrong. Contrary to the propaganda,
these mini-missions do NOT do "more or less the same thing at a tenth of the
cost". They do less, with less. But you can't compare any of these to a
matt o'
> so i make no claim of ineptness. in fact, i consider the navigation guys
at
> jpl to be *wizards* (and i have met some of them and knew some of them -
not
> sure who's there now from who [or is it whom?] that i knew). it was
apparent a
> mistake was made and anyone can make one. much of the time mistakes are
made
> and nothing bad happens. sometimes it does. they are mistakes. there
are, of
> course, other cases where things are worse than simple mistakes but that's
> another story.
>
> at the conference i was at this week a guy went up and talked about a
mistake
> in his organization; went through it from top to bottom, with the result
of a
> complete loss of a science mission. bared his soul. spilled his guts.
took
> the blame. a very courageous act. they simply missed something and made
a
> mistake. can happen to anyone. it was a very well received talk. i
would not
> call his organization nor the responsible engineer "inept" at all; it was
> simply an error. it's a risky business.
>
my thoughts EXACTLY! it seems like allot of people in this newsgroup look at
every engineering mistake as the death throws of the space program.... hey
it's not like their engineering a can of soup
matt o'
> You do? That's terrible mileage. What do you drive? A tank? A car ought
> to get at least four leagues to the ram's head (or a quarter of a league
> per, if you get creative about the definition of a league...)
>
depends on what type of ram your talking about... is it a white hoofed north
african or a squiggly tailed european ram?
> That depends. I make lots of similar errors. But I also catch them
> before I start relying on the results. One of the quotes in the
> press release said more or less the same thing: The error wasn't
> a failure to convert English units to metric; it was a failure
> to catch and correct this mistake.
>
but because of this would you consider the engineers at JPL inept? i
certainly wouldn't.... i took umbridge to the origianl poster's use of the
word inept. the metric/imperial thing was a mistake, but hey that's life....
if you happen to find the perfect person , tell me, i'll vote for him or
her.
_______________
"part of the fun of being alive is knowing that you're annoying the hell out
of someone else"-- matt groening
matt o'connell
www.personal.psu.edu/mjo139
ma...@psu.edu
Dan Hartung
once cried out:
>> Galileo, a Cassini, or a Hubble.
>
>expensive missions did, but at 20% of the cost.
Not to mention 400% of the redundancy.
--
Dan Hartung | "I believe we can fly
dhartung (at) wwa (dot) com | on the wings that we create"
http://www.wwa.com/~dhartung/ | -- M. E.
Lake Effect Weblog: http://www.wwa.com/~dhartung/weblog/
Ben
<stel...@NOSPAM.erols.com> wrote:
>
> at the conference i was at this week a guy went up and talked about a mistake
> in his organization; went through it from top to bottom, with the result of a
> complete loss of a science mission. bared his soul. spilled his guts. took
> the blame. a very courageous act. they simply missed something and made a
> mistake. can happen to anyone. it was a very well received talk. i
would not
> call his organization nor the responsible engineer "inept" at all; it was
> simply an error. it's a risky business.
Which mission?
-Ben
Sean Ellis
>definition:
>
>inept: 1. not apt or fitting; unsuitable; inappropriate.
> 2. not sensible; foolish.
> 3a. awkward; clumsy.
Well, that's the Imperial unit system in a nutshell. Despite ample
evidence to the contrary, I still find it difficult to believe that
scientists and engineers in the US still use imperial units.
> 3b. incompetent.
On the other hand, it is easy to criticise. I admit to making several
errors a day, but I am glad that the worst thing that happens is that
the lighting on someone's 3D model looks a bit odd.
Sean
But remember, I'm not a rocket scientist. I just play one on TV.
Fight spam! Join EuroCAUCE: http://www.euro.cauce.org/
CLVANCIL
>Seriously. Someone making a unit conversion error is not inept. It
>happens all the time. But ``inept'' does seem to apply to a program
>operating a $125 million spacecraft, and which fails to notice and correct
>a unit conversion error. At least one that mistake affects something as
>critical as safe spacecraft navigation.
I find this conversion error business really hard to fathom. How many people
actually check this kind of thing for just this sort of error? Something more
must be added to make sense of this whole affair. I mean the a similar program
must have been used for Pathfinder yet it didn't plow into the ground!
Chris Vancil
Member NSS and Mars Society
http://hometown.aol.com/CLVANCIL/
Gene Nygaard
fcr...@rintintin.colorado.edu (Frank Crary) wrote:
> Second, my boss occasionally complains about the term ``English''
units.
> She is English, and notes that the English don't use ``English''
units...
But they do. Just look at their speed limit signs. Or why do
we still hear all to often that someone weighs so many
stone; that's a unit only used by those weird people who think that
hundred is written in digits as "112."
Furthermore, it is a matter of history, not of usage. Nobody in Rome
uses Roman feet or paces any more either.
Gene Nygaard
http://ourworld.compuserve.com/homepages/Gene_Nygaard/
Sent via Deja.com http://www.deja.com/
Before you buy.
Tom Faber
> must be added to make sense of this whole affair. I mean the a similar program
> must have been used for Pathfinder yet it didn't plow into the ground!
>
Well... Actually, it did! :)
--
Tom Faber
John Schilling
>>And just because I'm curious (as at a train wreck), which group was
>>using which system? Who was living life in the 1960's?
>LockMart. Hardware is usually designed and built in English units.
>There are moderately good reasons for that. If you go to a machine
>shop, you will notice that all the dials on the lathes and milling
>machines are in English units. Companies like LockMart have a huge
>amount of money invested in the tools to build spacecraft, and have
>since the 1960s. Switching to metric would mean replacing them, and
>that's _very_ expensive. You can't even phase them out without having
>to deal with mixed units for a while. As MCO showed, dealing with
>mixed units is a bad idea.
As MCO also showed, continuing to buy English-measure machine tools
*doesn't* stop you from having to deal with mixed units. And I imagine
the cost of one MCO would pay for an awful lot of machine-tool retrofits.
There's a false economy at work here.
--
*John Schilling * "Anything worth doing, *
*Member:AIAA,NRA,ACLU,SAS,LP * is worth doing for money" *
*Chief Scientist & General Partner * -13th Rule of Acquisition *
*White Elephant Research, LLC * "There is no substitute *
*schi...@spock.usc.edu * for success" *
*661-951-9107 or 661-275-6795 * -58th Rule of Acquisition *
Steinn Sigurdsson
press release said more or less the same thing: The error wasn't
a failure to convert English units to metric; it was a failure
to catch and correct this mistake.
Nah, THE error was for L-Mart to be using Imperial units in the
first place. A secondary error was for NASA not to specify
explicitly that contractors should forward all data in SI units
or approved auxilliary units. A tertiary error was for JPL not
to check that the L-Mart units anyway.
It is really hard to believe the L-Mart people were that stupid
to begin with... did they ever mention a root cause for all those
launchers they lost? :-)
Gene Nygaard
>
> >And just because I'm curious (as at a train wreck), which group was
> >using which system? Who was living life in the 1960's?
>
> LockMart. Hardware is usually designed and built in English units.
> There are moderately good reasons for that. If you go to a machine
> shop, you will notice that all the dials on the lathes and milling
> machines are in English units. Companies like LockMart have a huge
> amount of money invested in the tools to build spacecraft, and have
> since the 1960s. Switching to metric would mean replacing them, and
> that's _very_ expensive.
I keep getting a picture of them sawing off a few pounds force (or was
it poundals) to add to these rockets.
George Herbert
>fcr...@rintintin.colorado.edu (Frank Crary) writes:
> press release said more or less the same thing: The error wasn't
> a failure to convert English units to metric; it was a failure
> to catch and correct this mistake.
>
>Nah, THE error was for L-Mart to be using Imperial units in the
>first place. A secondary error was for NASA not to specify
>explicitly that contractors should forward all data in SI units
>or approved auxilliary units. A tertiary error was for JPL not
>to check that the L-Mart units anyway.
If you can find me a US manufacturer who can fab space hardware
in metric without doing internal conversions, I might consider
that a useful comment. As is, reality indicates that specifying
the use of metric for designs is a mistake.
Now, given that, there's also no good reason for not giving
all performance and dynamics data in metric during handover.
I tend to design hardware in inches and do trajectories and
dynamics in meters/km and will probably keep doing so until
materials stress and strength in Pascals make intuitive sense.
-george william herbert
gher...@crl.com
Steinn Sigurdsson
Steinn Sigurdsson <ste...@muon.astro.psu.edu> wrote:
>fcr...@rintintin.colorado.edu (Frank Crary) writes:
> press release said more or less the same thing: The error wasn't
> a failure to convert English units to metric; it was a failure
> to catch and correct this mistake.
>Nah, THE error was for L-Mart to be using Imperial units in the
>first place. A secondary error was for NASA not to specify
>explicitly that contractors should forward all data in SI units
>or approved auxilliary units. A tertiary error was for JPL not
>to check that the L-Mart units anyway.
If you can find me a US manufacturer who can fab space hardware
in metric without doing internal conversions, I might consider
that a useful comment. As is, reality indicates that specifying
the use of metric for designs is a mistake.
Naively one might think that the customer should be able to specify
the design parameters, up to and including the metric.
If L-Mart or other corporations are incapable of doing this, they
can refrain from bidding.
Note that nothing precludes anyone from working internally in
any units they please, but they then have the obligation to
do any necessary conversion to meet the customers requirements,
including any additional quality assurance to make sure the
conversions are actually done.
Now, given that, there's also no good reason for not giving
all performance and dynamics data in metric during handover.
I tend to design hardware in inches and do trajectories and
dynamics in meters/km and will probably keep doing so until
materials stress and strength in Pascals make intuitive sense.
If you actually used SI units they would make intuitive sense,
there is nothing intuitive in the slightest about Imperial units.
The continued use of Imperial units is a combination of institutional
inertia and intellectual laziness. To actually use them in scientific
work seems completely mad.
And I say this coming from a discipline that imposes more auxillary
units on SI than any others, for metric and historical reasons.
But not even astronomy is mad enough to work in poundals or inches.
rk
> rk <stel...@NOSPAM.erols.com> wrote:
>
> >definition:
> >
> >inept: 1. not apt or fitting; unsuitable; inappropriate.
> > 2. not sensible; foolish.
> > 3a. awkward; clumsy.
>
> Well, that's the Imperial unit system in a nutshell. Despite ample
> evidence to the contrary, I still find it difficult to believe that
> scientists and engineers in the US still use imperial units.
hmmm ... at my day job i am required to use SI units ... and i must check
a box on the signed release form that i have done so.
btw, many times it is simply not practical to use metric units. for
instance, if you purchase integrated circuits, many of the packages have
lead pitches of 0.25" or 0.050" or 0.100". if i were to specify the lead
pitch to the guy designing the printed circuit board in metric units,
when his libraries are in english, i will drive up the budget and
schedule (BS) as well as the error rate. it will take a while to
switch. note that the newer packages are going metric.
rk
rk
small explorer wire
rk
Dr John Stockton
15:30:42 in news:sci.space.policy, Gene Nygaard <gnyg...@nccray.com>
wrote:
>In article <7t15oe$8...@peabody.colorado.edu>,
> fcr...@rintintin.colorado.edu (Frank Crary) wrote:
>
>units.
>> She is English, and notes that the English don't use ``English''
>units...
>
>But they do. Just look at their speed limit signs. Or why do
>we still hear all to often that someone weighs so many
>stone; that's a unit only used by those weird people who think that
>hundred is written in digits as "112."
About the only things that Imperial units are still much *used* for here
on British soil are things that the uneducated more or less understand -
measurement of road distances and speeds, beer (but not milk), human
weights, carpet and clothes sizes. The technology behind these will be
as metric as practicable.
Verbal descriptions remain commonly Imperial; I asked today for a pound
of sausages, but they were dual-priced (lb/kg); I expect builders still
talk about six-inch nails, but the size on the packet is metric.
IIRC, it was some while ago that British Rail dropped the 4' 8.5" gauge
in favour of a metric one (not necessarily nominally identical - 1432
mm??).
The Munros will no doubt remain Imperial; but the Ordnance Survey
mapping, as sold to the public, went metric a generation ago.
The Hundredweight is no longer significantly used - sacks of about that
size are 50kg now - but note that it is a convenient fraction of a ton.
There is precedent for generosity in hundreds : the Hundred Years' War
(1347-1453).
--
© John Stockton, Surrey, UK. j...@merlyn.demon.co.uk Turnpike v4.00 MIME. ©
Web <URL: http://www.merlyn.demon.co.uk/> - FAQish topics, acronyms, & links.
Correct 4-line sig. separator is as above, a line precisely "-- " (SoRFC1036)
Do not Mail News to me. Before a reply, quote with ">" or "> " (SoRFC1036)
George Herbert
}gher...@crl3.crl.com (George Herbert) writes:
} If you can find me a US manufacturer who can fab space hardware
} in metric without doing internal conversions, I might consider
} that a useful comment. As is, reality indicates that specifying
} the use of metric for designs is a mistake.
}
}Naively one might think that the customer should be able to specify
}the design parameters, up to and including the metric.
}If L-Mart or other corporations are incapable of doing this, they
}can refrain from bidding.
}Note that nothing precludes anyone from working internally in
}any units they please, but they then have the obligation to
}do any necessary conversion to meet the customers requirements,
}including any additional quality assurance to make sure the
}conversions are actually done.
As a general rule, you design things in even increments and fractions
of units (whole cm or mm for many measurements). Converting between
them after a design is done in one set of units results in really odd
fractions and decimals everywhere.
If you try and specify in metric, and want document handoff in metric,
you're asking for getting a godawful mess of data which never comes out
evenly or rounds. Plus they'll charge more.
} Now, given that, there's also no good reason for not giving
} all performance and dynamics data in metric during handover.
} I tend to design hardware in inches and do trajectories and
} dynamics in meters/km and will probably keep doing so until
} materials stress and strength in Pascals make intuitive sense.
}
}If you actually used SI units they would make intuitive sense,
}there is nothing intuitive in the slightest about Imperial units.
}The continued use of Imperial units is a combination of institutional
}inertia and intellectual laziness. To actually use them in scientific
}work seems completely mad.
}And I say this coming from a discipline that imposes more auxillary
}units on SI than any others, for metric and historical reasons.
}But not even astronomy is mad enough to work in poundals or inches.
I actually use SI units. They make perfect intuitive sense in terms
of understanding what they mean. The value ranges don't. When I'm
thinking steel strength, aluminum fatigue, pressures it only makes
sense in PSI. If I had been taught SI consistently it might be different.
If I could get materials data sheets from US manufacturers that had
SI units (even SI along with Imperial) it might be different.
As it is, I have to pull teeth out of suppliers to get metric data
half of the time, or just convert it myself (with all the conversion
error risks that implies).
-george william herbert
gher...@crl.com
rk
> Steinn Sigurdsson <ste...@muon.astro.psu.edu> wrote:
> }gher...@crl3.crl.com (George Herbert) writes:
< snip >
> }If you actually used SI units they would make intuitive sense,
> }there is nothing intuitive in the slightest about Imperial units.
> }The continued use of Imperial units is a combination of institutional
> }inertia and intellectual laziness. To actually use them in scientific
> }work seems completely mad.
> }And I say this coming from a discipline that imposes more auxillary
> }units on SI than any others, for metric and historical reasons.
> }But not even astronomy is mad enough to work in poundals or inches.
>
> I actually use SI units. They make perfect intuitive sense in terms
> of understanding what they mean. The value ranges don't. When I'm
> thinking steel strength, aluminum fatigue, pressures it only makes
> sense in PSI. If I had been taught SI consistently it might be different.
> If I could get materials data sheets from US manufacturers that had
> SI units (even SI along with Imperial) it might be different.
> As it is, I have to pull teeth out of suppliers to get metric data
> half of the time, or just convert it myself (with all the conversion
> error risks that implies).
going back to what george said earlier, and from an engineering perspective,
it's sometimes hard/impossible to get components using metric. this has to do
with an infrastructure that has been around a long period of time. even if
you could get them in metric tolerances, then you would have to convert them
to be compatible with other tools. tools that are designed to work with
industry standards.
all of this moving numbers around will be error prone and take a lot of
checking to ensure that nothing bad happens. it also negates the use of
already proven, well-checked out libraries.
many new components are being designed with metric tolerances.
i grew up with english units and think of many things that way. some things
are more natural to think of in metric. e.g., temperature testing hardware is
naturally done using celcius (i don't remember of that's SI or if K is or if
both is acceptable) but temperature outside i think of in F. for h/w test, i
have to convert to F for this, not the other way around. however, because of
previous practice and how i learned, large physical dimensions i use english.
at day job we had some projects that were forced to do everything in metric.
a real mess. and we spent a lot of time converting things to english to see
what they mean. for small units, metric is standard, i.e., microns or
angstroms.
so,
i would agree that institutional inertia is a factor. but existing
infrastructure will result in an evolution, not permit a revolution. btw, i
wonder if jimmy carter is building his houses in metric.
i would disagree that not using metric is "intellectual laziness." as i put
down in a previous post, lead pitches for many electronic components are
typically 0.050" or 0.025" or 0.100". while the conversions to metric are
trivial and have nothing to do with laziness of any sort, they will be extra
work, it will be error prone, and existing infrastructure will have to be
modified. in this case, the "fix" of metric is worse than living with the
inferiority of english units.
to use english units in scientific work may be mad; to use english units in
engineering work is often practical and produces the least errors and is
minimum cost.
i think for a while it will be like my toolbox for working on cars. there
will be two sets of numbers and with new designs most likely being done in
metric as sales are now more global then ever.
i also note that there has been a retreat from the metric system down here.
for a while, i think it was in virginia and another county, perhaps maryland,
they required that all construction plans be done in metric. because of a
bunch of reasons, some of which are similar to those mentioned above, there
was problems and it created hassles. as a result, these requirements for
metric are being dropped.
btw, think of the high tech pc you're working on (ok, you're probably on a sun
workstation running unix): what dimensions are specified for the monitor, hard
drive, floppy, and cd-rom? if you take off the cover, you'll see that the
older packages for the ic's are in english units. and breaking out a pci
specification for the backplane, you have a choice on the drawing, for the
recommended connector layout, of a pitch of 0.050" or 1.27 mm.
what a mess!
have a good evening,
rk
p.s. a lot of things stay around a long time; i note that many measurements we
use every day are in bases that i believe go back to babylonia, before ancient
egypt. defninitely not in base 10.
Chuck Stewart
It's 5.25" steel going into a 2 1/2" HW die in an 80-ton press...
He needs to make 32,000 lbs and the tolerance on parts to spec is 1/2
mil... The kitty arms and activates the press...)
Allen Thomson wrote:
> Sean Ellis <sel...@geocities.com> wrote:
> > I still find it difficult to believe that
> >scientists and engineers in the US still use imperial units.
>
> By and large, US scientists don't, except at the grocery store.
It's just that the machinery to make things is gauged in inches, mils,
pounds and tons.
--
Chuck Stewart
"Anime-style catgirls: Threat? Menace? Or just studying algebra?"
Michael L. Belanger
MLB
rk wrote:
> p.s. a lot of things stay around a long time; i note that many measurements we
> use every day are in bases that i believe go back to babylonia, before ancient
> egypt. defninitely not in base 10.
--
Michael L. Belanger
Ochtauber Interactive
mbel...@deltanet.com
http://users.deltanet.com/~mbelange/
Allen Thomson
rk
mathematics class? i was curious if anyone would remember this. sorry about the
rush post, it's in an interesting topic to talk about, but i only have a minute - or
should i say, 60 seconds. :-)
rk
p.s. my minor league brain also remembers that once learned, arithmetic was more
efficient done the way the babylonians did it as opposed to the way we do it today,
if all things are equal. then again, the babylonians would probably still beat
people today as they wouldn't have to wait for windoze to boot.
=======================
Alexey Goldin
>
> As a general rule, you design things in even increments and fractions
> of units (whole cm or mm for many measurements). Converting between
> them after a design is done in one set of units results in really odd
> fractions and decimals everywhere.
>
> If you try and specify in metric, and want document handoff in metric,
> you're asking for getting a godawful mess of data which never comes out
> evenly or rounds. Plus they'll charge more.
This is exactly the kind of a trouble that Tupolev met when ordered by
Stalin to copy B-29 bomber. Most the design effort was spent on
assuring that conversion was done right (right sheet metal thickness,
as USSR had no supplies of 1/16'' aluminum, for example). The
experience demonstrated even to Stalin that it is more rewarding to
design aircraft from the scratch :-)
Gene Nygaard
Aha! Now it's becoming clear what JPL is up to. They are letting the
Chinese steal all our secrets, because that'll just tie up all their
scientists making conversions from the English units!
Gene Nygaard
Except that even in the grocery store, enough of them have been
brainwashed by some introductory physics classes or whatever so that
they don't know what a pound measures in a pound of beans or whatever.
Altogether too many of them have some screwball notion that pounds are
only or primarily units of force, not units of mass, and the equally
wrong idea that the "weight" in the net weight on their food labels
means something different from mass.
Gene Nygaard
http://ourworld.compuserve.com/homepages/Gene_Nygaard/weight.htm
Dr John Stockton
18:30:02 in news:sci.space.policy, rk <stel...@NOSPAM.erols.com> wrote:
>Sean Ellis wrote:
>hmmm ... at my day job i am required to use SI units ... and i must check
>a box on the signed release form that i have done so.
>
>btw, many times it is simply not practical to use metric units. for
>instance, if you purchase integrated circuits, many of the packages have
>lead pitches of 0.25" or 0.050" or 0.100". if i were to specify the lead
>pitch to the guy designing the printed circuit board in metric units,
>when his libraries are in english, i will drive up the budget and
>schedule (BS) as well as the error rate. it will take a while to
>switch. note that the newer packages are going metric.
Perhaps it would take you a while to learn. On Mainland Europe, which
has used Metric for about two centuries, they appear to have been able
to do this (and similar) ever since the introduction of 0.1" pitch for
components and standard Eurocard sizes. We in the UK have adapted, too.
Really, it only requires a suitable arrangement in the software used for
the design.
ISTM that a significant factor in generating the MCO error must have
been sloppy practice, in transferring numerical values from one system
to another without stating and checking the units.
If you were to learn to operate your shift key, as so many others have,
then your writing would be significantly easier to read. Please do so
as a courtesy to your possible readers : you are, after all, no e.e.
cummings; and, if you are lacking in manual dexterity, there is nowadays
software to assist.
--
© John Stockton, Surrey, UK. j...@merlyn.demon.co.uk / JR.St...@physics.org ©
rk
> JRS: In article <37F535EA...@NOSPAM.erols.com> of Fri, 1 Oct 1999
> 18:30:02 in news:sci.space.policy, rk <stel...@NOSPAM.erols.com> wrote:
> >Sean Ellis wrote:
>
> >hmmm ... at my day job i am required to use SI units ... and i must check
> >a box on the signed release form that i have done so.
> >
> >btw, many times it is simply not practical to use metric units. for
> >instance, if you purchase integrated circuits, many of the packages have
> >lead pitches of 0.25" or 0.050" or 0.100". if i were to specify the lead
> >pitch to the guy designing the printed circuit board in metric units,
> >when his libraries are in english, i will drive up the budget and
> >schedule (BS) as well as the error rate. it will take a while to
> >switch. note that the newer packages are going metric.
>
> Perhaps it would take you a while to learn.
perhaps.
let's see: 0.1 = 2.54
0.050 = 1.27
0.025 = 0.635
in this case i find english easier to remember and to measure for checking purposes
with my calipers. and i don't have to worry about round-off errors for devices with
large lead counts on a side.
===============================================
> On Mainland Europe, which
> has used Metric for about two centuries, they appear to have been able
> to do this (and similar) ever since the introduction of 0.1" pitch for
> components and standard Eurocard sizes. We in the UK have adapted, too.
> Really, it only requires a suitable arrangement in the software used for
> the design.
i know the one time we did specify a metric connector the part had to be built by
hand for the layout software. of course, it is expensive and error prone to make
the software mods for the libraries. perhaps there are other libraries for the
components that could be purchased or installed. i'll check with some of my
european friends and see how they do it. library creation, checking, and
maintenance is expensive and error prone. personally, if we get dimensions from
vendors in english units, the software has english units and is checked out and
verified and proven in english units, it's hard to see why one should switch. even
simple things can get you. on one board that i saw, the french assumed numbering of
a connector one way and the americans the other. i know the american numbering
scheme was "standard" - at least for a noo yawker. it made for an "oops" and a
quick set of mods for correction. i'd hate to introduce risk and errors where there
is no benefit. and unfortunately i don't see the benefit for using metric
dimensions for parts and software that naturally works well in english units.
==================================================
> ISTM that a significant factor in generating the MCO error must have
> been sloppy practice, in transferring numerical values from one system
> to another without stating and checking the units.
>
> If you were to learn to operate your shift key, as so many others have,
> then your writing would be significantly easier to read.
ok. i just tried it. pressed it a few times. nothing happenned. perhaps it's in
the netscape help file.
===================================================
> Please do so
> as a courtesy to your possible readers : you are, after all, no e.e.
> cummings;
who? i've never seen him/her post here or the other newsgroups that i pop in and
out of. but i'll keep my eyes open for him/her.
====================================================
> and, if you are lacking in manual dexterity, there is nowadays
> software to assist.
actually, while my manually dexterity is not what it was 20 years ago, it's still
quite adequate for most things, although my pool game isn't what it was, more
probably a result of older eyes.
perhaps you could be so kind as to recommend some software; i tried pressing this
shift key a few more times as you suggested but nothing happens. i sort of like the
netscape newsreader for the most part, but i would be amenable to trying something
else.
then again, perhaps the keyboard is simply defective. it is rather old, i bought it
in 1986, and i know the design is flawed. the key-to-key spacing is exactly 3/4 of
an inch. oops! i did it again. let me try once more:
the key-to-ley spacing is 19.05 mm. please note that i deleted the word "exactly."
rk
> © John Stockton, Surrey, UK. j...@merlyn.demon.co.uk / JR.St...@physics.org ©
> Web <URL: http://www.merlyn.demon.co.uk/> - FAQish topics, acronyms, & links.
> Correct 4-line sig. separator is as above, a line precisely "-- " (SoRFC1036)
> Do not Mail News to me. Before a reply, quote with ">" or "> " (SoRFC1036)
(c) rk - am i supposed to do that, too?
rk
> In article <37F42337...@NOSPAM.erols.com>,
> rk <stel...@NOSPAM.erols.com> wrote:
> >> Seriously. Someone making a unit conversion error is not inept. It
> >> happens all the time. But ``inept'' does seem to apply to a program
> >> operating a $125 million spacecraft, and which fails to notice and correct
> >> a unit conversion error. At least one that mistake affects something as
> >> critical as safe spacecraft navigation.
>
> >you usually are careful in your posts (i wish i was 1/10 as careful) but i want
> >to make sure that your snip doesn't change the meaning of what i said or
> >intended to say.
> >i gave the dictionary definition of "inept" to follow up the previous poster
> >and then added:
> > "i don't know enough about the cause of the failure to say much w.r.t.
> > "inept" but i do know that anyone can make a mistake. "
> >so i make no claim of ineptness. in fact, i consider the navigation guys at
> >jpl to be *wizards*...
>
> I guess I should have been more specific. Given past performance, I think
> the loss of MCO due to a unit conversion error is inept. But that does
> not mean I consider the navigation people involved to be inept.
ok. i just wanted to make sure that someone reading a snipped post didn't think, by
my posting the definition of inept, that i was calling anyone or any system inept.
that's one problem with usenet in general and a way many flame wars or personal wars
get started. i haven't really come to much of an opinion on this thing and won't
until all the facts are out. and i do note that there are many people who read this
group that are "read-only", many of whom i know. and i wish to avoid someone
reading one post out of context and, well, getting the wrong idea and having net
post creep into real life in a negatory fashion. i get into enough trouble by
myself! :-)
from what i've read so far, i can't think of any good reason why this happenned.
but i'll hold off claims of ineptness and "off with their heads", as i've seen in
other posts, until the facts are in.
personally i consider, as i've said before, the navigation folks as wizards who've
pulled off some amazing things. and, from the people who i've met who do this sort
of thing, they are quite, quite bright. something clearly went wrong and we'll have
to wait a bit longer to find out the details and the whole story. there's always
time with "off with their heads" calls later on, so i would suggest that the usenet
mob wait a bit :-)
=======================================================
> There
> are two distinctions here. First, which noun does ``inept'' modify?
> If it an inept mistake or inept people?
good question. the original poster called the "program" inept. is the program an
entity that follows set rules and procedures established by some higher level such
as iso9001 <rk inserts cheap shot for iso9001 - as iso9001 implementations that i
have seen promote just this sort of failure>. or is the program the sum of the
humans who constitute it, with the policies, spirit, and "behaviour" of the program
a reflection on the individuals involved? most likely, a combination of both,
although i tend to think more of the latter.
-------------------
> As you say, we don't have all
> the facts, but my guess that this is a matter of very ept (if that is
> a word) people making an inept mistake.
that seems a reasonable statement, althought the fattest dictionary i can find in
the house doesn't have an entry for ept. btw, i just delivered my paper on "logic
design pathology and space flight electronics" where we (me and two co-authors) went
through a large amount of logic design errors in flight electronics from the last
several years. virtually all of them resulted from low level things, often things
that were not in direct view. many were made by very "ept" designers who just
missed things - often from working at a higher level of abstraction from the actual
hardware or circuit implementation. some were made by inept designers. but a
reasonable number were simply concentrating on the hard things and missing the
smaller details, not taking the time to look under the hood and see how the engine
works. does this make the designer inept or an ept one who made a mistake? or an
ept one in too much of a hurry?
that's why i'll wait for more information before calling a program "inept" because
of this failure. there tends to be a complexity and a masking of details when you
have a number of organizations and bureaucracies involved in doing technical work as
a team - and a team that is not co-located. ideally, your co-worker should be able
to swivel around in his/her chair and bop you over the head to get your attention to
ask a question or resolve something [see skunk works by ben rich]. so, it might be
a mistake that a number of people looked at and just didn't see - perhaps they were
really scrutinizing things well and the low level detail got by them. or perhaps
they had bozo's on the job (which i doubt). in the paper i mentioned above (which i
should be working on the final text on now) there was one case where they had four
organizations looking at the same circuit, they all misdiagnosed it. it was a poor
design which they thought was ok [blamed it on a defective part which wasn't, it was
well within specification] since they missed an error in the timing equations and
didn't understand the whole derivation of the equations or the tool that did the
arithmetic for them. as a result, they missed the systematic error and other
related circuits that are at risk of failure. are these people inept? did they all
just make the same mistake? that's a hard call, i still can't make it. now, on
this program things were worse than they had to be since they loaded the older,
worse code into a component by accident. that does sound inept. i'm not sure about
the previous ones. i can go either way. some of them are quite smart (from people
i know who know them who speak quite highly of them). i certainly don't think the
original circuits were very good and were at risk; they are designable and
verifiable in principle - but the steps needed to maximize the probability of
success were not done and the verification was done wrong, giving a false-positive
answer. and they misdiagnosed the failures in the hardware.
--------------------
> Second, are we talking about
> the individuals involved or the project as a whole? It is quite possible
> for very competent and hard working people to, as a group, make an
> inept mistake. That gets into communication between the people involved
> (a unit error like this is clearly a communication problem), decisions
> about how much effort needs to be spent on double checking numbers,
> etc.
i think one problem here is that inept is a bad word and the dictionary definition
that i posted didn't help much, perhaps hurts. i think the popular usage of the
word inept is worse then the "legal definition" of the word inept. i believe what
you say here is more or less consistent with what i mentioned above but i would take
this discussion one step further. now, perhaps someone knowledgeable in navigation
(perhaps yourself) would know if there was the opportunity or whether there should
have been made the opportunity to ensure that the burns made the spacecraft head to
where it was supposed to be going? that is, could it have been verfied that the
spacecraft was on course and if not, should that have triggerred [spell] an alarm
that there was a problem?
====================================================
> >...It was apparent a
> >mistake was made and anyone can make one. much of the time mistakes are made
> >and nothing bad happens. sometimes it does. they are mistakes.
>
> Yes, but if this was the only thing involved, very few spacecraft would
> reach their destinations. The odds of one person making such a mistake
> are high enough, as are the risks resulting from a mistake, that you
> can't ignore it. Some numbers need to be double checked; if the
> spacecraft is doing something slightly odd, you need to find out why
> and correct the problem (even if it seems to be a fairly minor problem,
> it might be a symptom of a more significant error.) As I said in another
> post, I don't think the unit conversion error was the real problem.
> That, as you point out, is a fairly easy mistake to make. The problem
> is that the error was not noticed and corrected.
yeah. normally everything in a space program is checked. all designs are subject
to design reviews. all fabrication steps are inspected by an inspector. functional
tests are judged pass and fail by test equipment. environmental tests show whether
things will work over the expected environment + some margin. these results are
then reviewed in a pre-ship review. as i mentioned a paragraph up, could the
results of the burn be checked afterwards to ensure things were going correctly and
things were on track? or was this open loop? i might have missed a press release or
two as i've been buried for most of the last week so this question might have
already been answered.
if the spacecraft is doing something "slightly odd" that should generate a problem
report of some kind, the name of which depends on the organization and perhaps the
project. the initiation of this report forces an analysis to be done, an
investigation of some sort, and signoff and review by multiple parties. perhaps i
missed that part, was the s/c doing something slightly odd? was a problem/failure
report written? if so, was it closed?
i note that i see more and more design reviews either not being done or being done
in such a cursory manner they are there simply to check a box on the bureaucrat's
chart, problem/failure reports are either not being written or are being closed by
arm waving, etc., etc. why is this so? that could be an interesting thread which i
will not start right now but will get sucked into if someone does. also, i note
that failures in space seem to be a quite common occurence lately. is this simply
"bad luck" that the problems are proving fatal to mission effectiveness? i note
that most spacecraft do have a lot of problems but that they don't destroy the
effectiveness of the mission. or is there something else going on, deeper, which
has been discussed, that compressed schedules, lower cost (lower people), etc., etc.
are contributing to the high level of failures? is it morale? <fill in reason
here, our business is in trouble>
good night,
rk
rk
rk wrote:
< major snippage >
> i think one problem here is that inept is a bad word and the dictionary definition
> that i posted didn't help much, perhaps hurts. i think the popular usage of the
> word inept is worse then the "legal definition" of the word inept. i believe what
> you say here is more or less consistent with what i mentioned above but i would take
> this discussion one step further. now, perhaps someone knowledgeable in navigation
> (perhaps yourself) would know if there was the opportunity or whether there should
> have been made the opportunity to ensure that the burns made the spacecraft head to
> where it was supposed to be going? that is, could it have been verfied that the
> spacecraft was on course and if not, should that have triggerred [spell] an alarm
> that there was a problem?
from cnn www site, jpl is asking a similar question:
"We're going to look at how was the data transferred," Gavin said. "How
did it originally get into system in English units? How was it transferred?
When we were doing navigation and Doppler (distance and speed) checks,
how come we didn't find it?"
"People make errors," Gavin said. "The problem here was not the error. It
was the failure of us to look at it end-to-end and find it. It's unfair to rely on
any one person."
rk
George Herbert
>gher...@crl3.crl.com (George Herbert) wrote:
>> I actually use SI units. They make perfect intuitive sense in terms
>> of understanding what they mean. The value ranges don't. When I'm
>> thinking steel strength, aluminum fatigue, pressures it only makes
>> sense in PSI.
>
>kPa, near enough. How much air do you put in your car tyres?
>Approximately 200 kPa (2 atmospheres).
Quick, what's the yield strength of HY-100 steel in MPa?
More importantly, working backwards from an actual load in MPa,
given a specific structural design, what's my safety factor for
the same HY-100 steel? We do crosscheck our own designs, but when
you have to back-convert again to crosscheck, you introduce more
chances to pick the wrong conversion factor or multiply where you
should have divided, etc.
It's not the basic dumb easy metric / SI stuff you learn in school
which is the problem. I can do Pa or PSI, K or C or F, kg or lb,
meter or cm or inch or foot or whatever. The basics have so little
to do with the totality of really understanding the range of engineering
materials properties that it's very painfully Not Funny. Of my varied
engineering texts and references, exactly two are adequately documented
in both unit sets... the Materials Science text, which is in Imperial
units but has SI in parenthesis next to them (except for a few glaring
conversion errors...), and an English book on composite structures and
materials design and engineering, which is in SI but has Imperial units
in parenthesis (again, with a few glaring conversion errors). If my
text/reference books can't get the damn conversions straight consistently,
how do you expect me to do so converting a whole office full of references
and textbooks and designs consistently?
As I keep having to repeat myself, the key conversion problem is when
you put engineers formative years in really starting to understand how
things all work, when they get good enough to understand things intuitively,
in one set of units and then want to change them. Their intuitive sense
of understanding goes out the window for a while until they re-learn all
the things they really fundamentally knew down deep in the new units
system consistently. Plus, all the references I use are still the old
way, and I haven't even found new SI versions of many of them, much less
bought them yet. Hell, the 1992 editions of Sutton and Huzel&Hwang are
both nearly entirely in Imperial units. Has anyone in Europe written
a comperably good rocket engine design text in SI? How do you guys learn,
anyways?
-george william herbert
gher...@crl.com
Frank Crary
rk <stel...@NOSPAM.erols.com> wrote:
>> Seriously. Someone making a unit conversion error is not inept. It
>> happens all the time. But ``inept'' does seem to apply to a program
>> operating a $125 million spacecraft, and which fails to notice and correct
>> a unit conversion error. At least one that mistake affects something as
>> critical as safe spacecraft navigation.
>you usually are careful in your posts (i wish i was 1/10 as careful) but i want
>to make sure that your snip doesn't change the meaning of what i said or
>intended to say.
>i gave the dictionary definition of "inept" to follow up the previous poster
>and then added:
> "i don't know enough about the cause of the failure to say much w.r.t.
> "inept" but i do know that anyone can make a mistake. "
>so i make no claim of ineptness. in fact, i consider the navigation guys at
>jpl to be *wizards*...
I guess I should have been more specific. Given past performance, I think
the loss of MCO due to a unit conversion error is inept. But that does
not mean I consider the navigation people involved to be inept. There
are two distinctions here. First, which noun does ``inept'' modify?
If it an inept mistake or inept people? As you say, we don't have all
the facts, but my guess that this is a matter of very ept (if that is
a word) people making an inept mistake. Second, are we talking about
the individuals involved or the project as a whole? It is quite possible
for very competent and hard working people to, as a group, make an
inept mistake. That gets into communication between the people involved
(a unit error like this is clearly a communication problem), decisions
about how much effort needs to be spent on double checking numbers,
etc.
>...It was apparent a
>mistake was made and anyone can make one. much of the time mistakes are made
>and nothing bad happens. sometimes it does. they are mistakes.
Yes, but if this was the only thing involved, very few spacecraft would
reach their destinations. The odds of one person making such a mistake
are high enough, as are the risks resulting from a mistake, that you
can't ignore it. Some numbers need to be double checked; if the
spacecraft is doing something slightly odd, you need to find out why
and correct the problem (even if it seems to be a fairly minor problem,
it might be a symptom of a more significant error.) As I said in another
post, I don't think the unit conversion error was the real problem.
That, as you point out, is a fairly easy mistake to make. The problem
is that the error was not noticed and corrected.
Frank Crary
CU Boulder
Frank Crary
>> That depends. I make lots of similar errors. But I also catch them
>> before I start relying on the results. One of the quotes in the
>> press release said more or less the same thing: The error wasn't
>> a failure to convert English units to metric; it was a failure
>> to catch and correct this mistake.
>but because of this would you consider the engineers at JPL inept?
Not exactly. I would say the Mars Climate Observer program, as a
group, did something inept. That doesn't mean any particular
engineer involved is inept. Sometimes a group of people, working
together, can make an inept mistake, even though (as individuals)
they are extremely good at their jobs.
Frank Crary
CU Boulder
Bruce Hoult
Herbert) wrote:
> I actually use SI units. They make perfect intuitive sense in terms
> of understanding what they mean. The value ranges don't. When I'm
> thinking steel strength, aluminum fatigue, pressures it only makes
> sense in PSI.
Pressures are easy to get an intuitive sense for! An atmosphere is 100
kPa, near enough. How much air do you put in your car tyres?
Approximately 200 kPa (2 atmospheres).
-- Bruce
Ian Stirling
>Quick, what's the yield strength of HY-100 steel in MPa?
I'd guess around 1Gpa, to within a factor of 10.
(I don't know the figure in PSI, either)
I find the simplest way to visualise Mpa as how many apples a 1mm square wire
can hold up.
Allen Thomson
Dr John Stockton <j...@merlyn.demon.co.uk> wrote:
>About the only things that Imperial units are still much *used* for here
>on British soil are things that the uneducated more or less understand -
>measurement of road distances and speeds, beer (but not milk), human
>weights, carpet and clothes sizes. The technology behind these will be
>as metric as practicable.
Speaking of that, perhaps you can solve one of life's great mysteries:
what is the significance of the stylized "e" on liquid (and other?)
containers in Europe? I've asked this a couple of times when visiting
DRA (now DERA, I think) labs at Farnborough and Malvern, and nobody
had a clue.
George Herbert
>George Herbert <gher...@crl3.crl.com> wrote:
>[...]
>>Quick, what's the yield strength of HY-100 steel in MPa?
>
>I'd guess around 1Gpa, to within a factor of 10.
>(I don't know the figure in PSI, either)
>I find the simplest way to visualise Mpa as how many apples a 1mm square wire
>can hold up.
And you want me to design spacecraft using error margins of 10
and apples? This is supposed to be an improvement over inches and PSI?
-george william herbert
gher...@crl.com
Dr John Stockton
in news:sci.space.policy, George Herbert <gher...@crl3.crl.com> wrote:
>Plus, all the references I use are still the old
>way, and I haven't even found new SI versions of many of them, much less
>bought them yet. Hell, the 1992 editions of Sutton and Huzel&Hwang are
>both nearly entirely in Imperial units. Has anyone in Europe written
>a comperably good rocket engine design text in SI? How do you guys learn,
>anyways?
I don't know of rocket engine texts and other specialist stuff; but for
ordinary engineering ISTR that CUP, OUP, and many other European (&
Canadian?) publishers have outlets or representation in the USA, and
would be happy to supply such books as are used over here. Indeed,
given a hint of a US market for these things, the ****** mainland
continentals would no doubt translate their books into something near
enough the American language and undercut & outsell us of the UK...
--
© John Stockton, Surrey, UK. j...@merlyn.demon.co.uk Turnpike v4.00 MIME. ©
rk
> >...also, i note
> >that failures in space seem to be a quite common occurence lately. is this
> >simply
> >"bad luck" that the problems are proving fatal to mission effectiveness?
>
> It could just be an observational artifact: We are flying more missions
> now than we were ten years ago. So even if the odds of failure were constant,
> failures would be more common. A few people have posted some rough statistics
> that suggest this, but I wouldn't say I'm sure until I've seen a more
> rigorous, statistical analysis.
note i used the word "seem." i have seen a number of articles debate this ...
and reasons for this .. but in any event, we should, all things being equal, be
driving down the rate of failure. if we go to lower-cost, higher-risk missions
than we should not be upset by a slightly higher mission failure rate. of
course, with redundancy cut, we can expect a decrease in mission effectiveness
for a constant failure rate. by personal observation, which is of course a very
limited data set, i see the cut in a lot of areas that is, well, troubling. less
reviews, less thorough testing, riskier designs, cuts in margins, and less
rigorous problem "disposition." the one area where i do see an increase, or at
least the appearance of an increase, is in relative computational power. i see
less of the fight to fit into available memory margins and processor time
intervals, although that seems to be the general case and definitely not the
rule, with some poorly defined architectures sticking out.
it "seems" that failures are up not only in space science missions but in launch
vehicles, too. and it does seem to be a rather odd cluster of failures; the last
time it "seemed" that we had this was in the early '90s, and it seemed that was
after a wave of mega-mergers. heck, i think even surrey satellite lost one!
don't have the number of failues at hand, but the numbers are too low, most
probably, for any rigorous statistical work.
it does seem troubling though.
Alan Anderson
rk <stel...@NOSPAM.erols.com> wrote:
> .. but in any event, we should, all things being equal, be
> driving down the rate of failure.
If all things were equal, we would be doing the same things over and over
again and never progressing. This is *exploration* we're talking about
here, remember? We should "aim high" and expect a certain high level of
failure -- and I think we're closing in on a good balance of mission risk
and mission quantity right now.
Frank Crary
rk <stel...@NOSPAM.erols.com> wrote:
>...could the
>results of the burn be checked afterwards to ensure things were going
>correctly and
>things were on track? or was this open loop?
It isn't an open loop process, but you don't get good trajectory solutions
instantly. The accuracy and reliability of the solution improves with
time and additional data. Optical navigation, which was not used, would
have improved this situation somewhat. Dealing with the problem (i.e.
sequencing and executing a correction maneuver) also takes some time. It
isn't clear if there was sufficient time/data to identify and deal with
the incorrect trajectory.
>if the spacecraft is doing something "slightly odd" that should generate a
>problem
>report of some kind, the name of which depends on the organization and
>perhaps the
>project. the initiation of this report forces an analysis to be done, an
>investigation of some sort, and signoff and review by multiple parties.
>perhaps i
>missed that part, was the s/c doing something slightly odd?
I just saw a press release which said the doppler and range data did not
agree, but this was (as I understand it) after TCM-4. The project considered
another trajectory correction maneuver, but decided against it, since the
solution didn't appear to be very far off. The fact that two data sets
systematically disagreed is definitely odd, and should have been (in my
opinion) an indication that the estimated trajectory could be completely
wrong. (When you have two data sets, measuring the same thing, which
give different results, odds are the average isn't the correct answer.
This isn't really how the trajectory is calculated, but it's close
enough... A systematic difference should be taken as a sign that
something is wrong, and that the results from combining the data sets
are unreliable.) I've also heard a rumor of odd behavior during a
star calibration of the imaging instrument, and that investigating the
problem was deferred until after orbital insertion. (I assume because
of limited resources and the fact that orbital insertion was more
critical.) Assuming that is true, it may or may not be relevant.
Something like a pointing error in an instrument isn't obviously
connected with a unit conversion error in a rocket's thrust. But it
would imply that anomalies were not resolved as soon as possible, and
that isn't a good sign.
>i note that i see more and more design reviews either not being done or
>being done
>in such a cursory manner they are there simply to check a box on the
>bureaucrat's
>chart, problem/failure reports are either not being written or are being
>closed by
>arm waving, etc., etc. why is this so?
I suspect it would depend on the individual mission, but I suspect it
is related to dealing with significantly lower mission budgets. (E.g.
dealing with missions that have to be perhaps half as capable with a
tenth the funding.) Cutting corners on design reviews and problem/failure
reports/investigations isn't a good way to do that (and we will need
more data before we know if this has any relevance to the MCO event),
but I can see how this sort of thing could happen.
>...also, i note
>that failures in space seem to be a quite common occurence lately. is this
>simply
>"bad luck" that the problems are proving fatal to mission effectiveness?
It could just be an observational artifact: We are flying more missions
now than we were ten years ago. So even if the odds of failure were constant,
failures would be more common. A few people have posted some rough statistics
that suggest this, but I wouldn't say I'm sure until I've seen a more
rigorous, statistical analysis.
Frank Crary
CU Boulder
rk
excuse me, i think you're talking about a different subject matter, out of
context. of course we're talking about "*exploration*".
i think the rate of failure should be going down, "all things being equal" as
both technology improves and our experience level grows. on the technology
side, i'll talk about something i know a bit about. for integrated circuits,
we have higher levels of integration and reliability levels not dreamed about
35 years ago. go back and look at what the engineers were writing about back
then - and how they built their own quad redundant gates for long term
reliability - for a mission of 1 year duration! the high levels of
integration just reduce failures by limiting the number of connections -
about 1/2 of all ic failures are in the i/o, not internal to the chips. for
experience, one would figure that we would learn from mistakes and
operational experience. these things would tend to drive down failure rates,
all things being equal.
i think we should aim high ... but i don't expect a certain high level of
failure and i certainly don't want one. why are we still blowing up launch
vehicles (or putting things in wrong orbits), etc., etc.? and i note that
the cost to orbit hasn't dropped much.
i respect your position that you think we have a good balance of mission risk
and quantity.
my personal opinion is that we're launching a lot of crap and we can and
should do better.
Dr John Stockton
of Sun, 3 Oct 1999 23:13:31 in news:sci.space.policy, Alan Anderson
<aran...@netusa1.net> wrote:
>In article <37F8119A...@NOSPAM.erols.com>,
>rk <stel...@NOSPAM.erols.com> wrote:
>
>> .. but in any event, we should, all things being equal, be
>> driving down the rate of failure.
>
>If all things were equal, we would be doing the same things over and over
>again and never progressing. This is *exploration* we're talking about
>here, remember? We should "aim high" and expect a certain high level of
>failure -- and I think we're closing in on a good balance of mission risk
>and mission quantity right now.
Indeed, there should be an appropriate trade-off between risks taken and
the costs of reducing them; the aim must be to get the maximum result
available in a reasonable time using available and expected resources
(KSC..., $$$...).
In that, planners must (and presumably do) realise that two types of
failure are particularly unacceptable :
Loss of life (especially astronauts on missions or simulations;
ground-grunts, astronauts travelling on earth, etc., count for little
more than ordinary citizens);
Making the same mistake twice.
And the public must learn that of quick, cheap, and reliable, one should
not expect more than one or two.
ISTM, perhaps, that NASA should insist either on an all-metric approach,
or that wherever Imperial is used the SI values should also be given,
and both supplier and user of the values are held fully responsible for
checking that the values agree.
That way, if BoeHeed use Imperial & say "That's five pounds = eleven
kilograms", then NASA can either see the error or replace the man who
didn't, and can demand appropriate penalty from BoeHeed.
--
© John Stockton, Surrey, UK. j...@merlyn.demon.co.uk / JR.St...@physics.org ©
Web <URL: http://www.merlyn.demon.co.uk/> - FAQish topics, acronyms, & links.
Dr John Stockton
1999 18:03:28 in news:sci.space.policy, Allen Thomson
<thom...@netcom.com> wrote:
>Speaking of that, perhaps you can solve one of life's great mysteries:
>what is the significance of the stylized "e" on liquid (and other?)
>containers in Europe? I've asked this a couple of times when visiting
>DRA (now DERA, I think) labs at Farnborough and Malvern, and nobody
>had a clue.
I don't know the definition, but I understand it to mean that the size
is a European-approved one. A quick look in the kitchen suggests that
any size will do, provided that it's written in metric (not necessarily
SI).
Alas, a Web search for "e" is unlikely to help; but you might try
"+Europe* +food +label*" or suchlike.
Or, next time you visit, ask the local Trading Standards people?
--
© John Stockton, Surrey, UK. j...@merlyn.demon.co.uk Turnpike v4.00 MIME. ©
CLVANCIL
>CLVANCIL wrote:
>> must be added to make sense of this whole affair. I mean the a similar
>program
>> must have been used for Pathfinder yet it didn't plow into the ground!
>>
>
>Well... Actually, it did! :)
You must have a very different dictionary than I. Plowing requires subsurface
trenching and Pathfinder certainly did not trench, therefore it did not plow
into the surface of Mars!
Chris Vancil
Member NSS and Mars Society
http://hometown.aol.com/CLVANCIL/
Ian Stirling
Yes, you use one standard apple, from a specific apple farm in france.
Adding a factor of 10 greatly simplifies design, anyway.
...
No of course not.
I was just attempting to point out that different people visualise/remember
stuff in different ways.
I actually have a pretty good grasp of what a newton is, so don't bother
with the apples thing.
The fact that I don't know what the yeild strength of a certain steel alloy
is, isn't particularly relevant, as I wouldn't know in any units,
I similarly wouldn't know how many square chains are in a acre.
Ian Stirling
>Tom Faber wrote:
>>CLVANCIL wrote:
>>> must be added to make sense of this whole affair. I mean the a similar
>>program
>>> must have been used for Pathfinder yet it didn't plow into the ground!
>>>
>>
>>Well... Actually, it did! :)
>You must have a very different dictionary than I. Plowing requires subsurface
>trenching and Pathfinder certainly did not trench, therefore it did not plow
>into the surface of Mars!
Pathfinder may have not, sojourner certainly did.