Groups | Search | Server Info | Keyboard shortcuts | Login | Register [http] [https] [nntp] [nntps]

Groups > comp.arch > #3594

Re: Solving the Floating-Point Goldilocks Problem!

Cross-posted to 2 groups.

Show all headers | View raw

sidd wrote:
> On Thursday 08 September 2011 10:10, jmfbahciv wrote:
>
>> How did you find the disappearing shorts?
>
> Depended on the specific fault. First, log the fault. Then find,
> isolate and monitor and immediately set one or more recorders to
> monitor them. Back in those days, this might be a chart recorder, or a
> grad student.Then you consulted the logbook, pored over heavily
> modified schematic and illegible scrawlings. You had, of course,
> measured resistances and  impedances of all legs of each circuit at
> room temperature, when the cryostat was open, and you had
> periodically measured many things as the cooldown proceeded. You had
> data on the composition and behavior of the various bits of your
> wires. From these and the currently measured values, you could often
> make an intelligent guess at the location. Before you warmed up the
> cryostat you could do further testing...if for example if i push a
> current through the short, which bits of the cryostat warms up ? or if
> a coax line, you might break out the storage scope and polaroid, and
> do TDR off the line to get an idea of distance to the fault.

Distance?!  Were there loops?

>
>
> After you were done with any lo-T diagnosis, you warmed up the
> cryostat watching the fault all the way to room-T. You have
> thermometers and sensors all thru the cryostat, you might pause
> along the route if you saw sumpn interesting...eg circuit
> parameter changed when this bit of the circuit warmed through
> superconducting transition temperature...and such. Along the
> warmup, you take advantage of the hysteresis, in that the innardz
> of the cryostat cool the slowest and warm the slowest, so you are
> applying different thermal (and thus mechanical) gradients to the
> cryostat and you know which parts are being stressed in which order;
> this order is different during cool down and warmup.
>
> Disappearing faults disappear differently.The manner of their
> vanishing often betrays whence they came.

Ah!  That would be useful.

> When all else fails,
> you rip out and replace, an operation that (in some cases) can
> take months.

Thanks.  No wonder physicists love their gadget (not to be
construed as a slight).

how does one learn all those tricks such as "disappearing
differently"?  Wouldn't the behaviour be unique to each
cryostat?  Or are they constructed so alike that working
on one trains you to be able to do the same diagnoses
on another.

/BAH

Back to comp.arch | Previous | Next — Previous in thread | Next in thread | Find similar | Unroll thread

Thread

Re: Solving the Floating-Point Goldilocks Problem! jmfbahciv <See.above@aol.com> - 2011-09-08 14:10 +0000
  Re: Solving the Floating-Point Goldilocks Problem! sidd <sidd@situ.com> - 2011-09-08 14:08 -0400
    Re: Solving the Floating-Point Goldilocks Problem! jmfbahciv <See.above@aol.com> - 2011-09-09 15:21 +0000
      Re: Solving the Floating-Point Goldilocks Problem! "Rod Speed" <rod.speed.aaa@gmail.com> - 2011-09-10 06:00 +1000

csiph-web