Index Home About Blog
From: (Henry Spencer)
Subject: Re: Rocket Software (was: Another tank question)
Date: Tue, 23 Feb 1999 15:39:28 GMT

In article <>,
Frank Jenkins  <> wrote:
>> And assuming that you have not made any errors in your assumptions...
>Absolutely true.  However, in the case of the new tank, they were
>dealing with the optimization of an existing structure based on a
>higher-strength material and known statistical flight loads.  Still
>plenty of ways for someone to make a mistake, though...

Or to overlook some novel problem.  Being able to do elaborate calculations
on the well-understood parts of the job does not enhance your ability to
spot things that you've overlooked.  Indeed, it may reduce it, by giving
everyone a false feeling of confidence.

Disastrous problems rarely arise from mistakes in calculation.  They tend
to come out of left field, in areas that were never suspected of being
important and hence were not studied in depth.  That's why good testing is
so crucial.

There is a good analogy here to structural safety factors.  Safety factors
are *not* there to compensate for issues like materials tolerances, stress
concentrations at joints, etc.  The safety factor is applied *after* you
allow for all known problems; it is solely and only a safeguard against
surprises... and a very necessary one.

>> Boeing has large, powerful computers too, but that didn't help the first
>> Delta III.
>IIRC, that was blamed on the guidance software overcorrecting and
>running the gimbal controls out of hydraulic fluid.  A software problem,
>to be sure...

No, a mistaken assumption about which modes of oscillation were important.
To be precise, the assumption was that the modes which were largest at
launch would dominate the others throughout flight.  More or less correct
for the Delta II... but not for the supposedly-straightforward upgrade to
the Delta III.

The software did its assigned job correctly, but unfortunately that wasn't
quite the job which was needed.
The good old days                   |  Henry Spencer
weren't.                            |      (aka

From: (Henry Spencer)
Subject: Re: rocket reliabilty
Date: Thu, 18 Nov 1999 19:45:16 GMT

In article <80sf0n$hq6$>,  <> wrote:
>For example, civil engineers may build things 3 or 5 times stronger
>than needed to take into acount varying loads due to winds and movement
>and variation in manufacture and assembly.
>Aeronautical engineers might build things 1.5 to 3 times as strong as
>they strictly need to be.

Note that this isn't really a matter of the aeronautical engineers taking
greater risks.  To a large extent it's a matter of accepting higher costs
during design and production to get lighter structures.  Even the civil
engineers, for example, reduce the necessary safety factor on a boiler if
you X-ray all the welds after making them, and it's your decision whether
you want to spend the money (and perhaps accept some design constraints to
permit weld X-raying) for the sake of using less metal.

>But, when running this close to the limits, there are variations that
>can occur in manufacturing that can weaken a part.

Yes, and the manufacturing processes are supposed to be tightened up to
the point where the variations are smaller and the excessive ones are
caught.  (Mind you, this doesn't work as well as hoped; the rocket
engineers definitely have gone too far on cutting back the safety
factors.  With some exceptions -- Proton uses 2.0 throughout.)

As others have already noted, structural failures are not a significant
cause of launch failures.

>Since failure rates are also a function of safety factor, how overbuilt
>parts are, its clear that advanced materials technology SHOULD NOT be
>used to DRAMATICALLY reduce structural fraction!  (this is exactly what
>SSTO is trying to do!)

Ah yes, advanced materials technology like what they use in golf clubs?
One should bear in mind that the rocket business is well behind normal
commercial practice in many areas; while it has been guilty of excesses,
it has also been far too conservative in many ways.

Besides, you don't *need* "advanced materials technology" for this, except
perhaps for the added demands of reusability.  Rocket stages with SSTOish
structural fractions were built in the 1960s, some for volume production.
What you need is good designers who are strongly motivated to reduce
weight.  (Hint:  extra stages kill the motivation.)
The space program reminds me        |  Henry Spencer
of a government agency.  -Jim Baen  |      (aka

Index Home About Blog