Trekker wrote:
On Mon, 30 Jun 2008 15:09:28 +0000, Derek Lyons wrote:

They spent a lot of time scratching their heads mostly because the
problem was intermittent. Anyone with actual technichal experience
knows that intermittents are stone cold bitch to troubleshoot under the
best of conditions - and filling an ET in order to attempt to induce the
fault is just about as far from the best of conditions as one can get.

I have to wonder how much time and effort was spent on this problem
(prior to the last time of actually finding and fixing it.)

Quite a bit.

Sure, intermittent problems are probably the most difficult to
troubleshoot but what circuitry was involved in the tank itself? Was it
just some sensors and wiring going out to the connector on the tank? If
that's all that was involved, I think they should have found the cause
earlier. Connections would be the first thing I would check with an
intermittent problem, especially one where a great degree of temperature
changes are involved.

The problem is that in order to induce that temperature change, you have
to put the tank through a fill/drain cycle. You get one of those "for
free" every time you launch, but for troubleshooting, you'd want to
insert instrumentation into the loop - instrumentation you most likely
can't launch with. That means a dedicated tanking test, which is a lot
of money and puts an additional thermal cycle on the tank, which makes
the tank more vulnerable to foam shedding when you finally do launch it.

That's why I don't think a whole lot of effort was put into figuring this
issue out until it finally because more than a nuisance and started
causing flight delays. Someone finally said "Enough is enough" and
actually troubleshot the problem instead of shrugging it off.

No, it was causing flight delays as far back as STS-114 and people were
saying enough is enough back then. The problem is that there were really
two separate root causes. Last time around, they discovered a batch of
faulty sensors. Which was a real problem, but it wasn't the *whole* problem.