Questions and speculations re fuel sensors

I'd like to ask some questions to anyone who might know (not just think
they know) about the ET fuel sensors. My guess (and it's just a guess)
is that they might be cryogenic temperature sensors. I suppose they
could be low T pressure sensors, but I don't know anything about such
beasts. I do work with cryo T measurements. From the descriptions I've
read, it seems that the signal processor converts the raw info into a
binary (wet/dry) signal, which tends to suggest some sort of comparator
circuit with a little built in hysteresis for noise suppression. The
"simulator" circuit must be some sort of signal relay (solid state maybe
but mechanical? {yuck}) that cuts in a known signal to then test the
complete signal pathway (i.e. removing the sensor itself from the equation).

The question is, if these are T sensors, what kind are they? Remember,
this is a 1970s design. Today, one would probably use a Si diode or a
PT100 RTD as they have become the standard parts, but shuttle is an old
system. So old that it must be a bear to get parts for it. It's the same
vintage as my old Pentax SP500 camera which takes mercury batteries (no
longer available but there is a work-around). Thermocouples can be used
at LH2 temperatures, but they are prone to noise pick up. RTDs and
diodes are much more common now.

The important design issue for a fuel sensor is not accuracy in the
usual cryo measurment sense (you don't need to get it right to within a
mK) but speed of response. The sensor has to give you the right answer
almost instantly. If it is a T sensor, then you get that by doing things
like making the sensor small, or giving it good thermal contact with
what you want to measure. If it's an RTD, and you wanted to measure a T
rise, you might even make use of a normally "bad thing" for cryogenic
measurments, that is self heating of the sensor. The RTD needs a certain
amount of current to measure the resistance. For precision measurements,
this is undesirable and most modern circuits keep the current very small.

But this is an ancient design, not necessarily trying (or wanting) to do
a precision measurment. Is is possible that over the years, engineers
have gradually substituted newer "better" parts with "better" specs that
may have lost sight of some of the original design requirements?

As I said, I don't know the answers. I'm just reading between the lines
and speculating. Anyone know the real story?

--
Chris M. Hall
Dept. of Geological Sciences
University of Michigan

Chris Hall wrote:
mK) but speed of response. The sensor has to give you the right answer
almost instantly. If it is a T sensor, then you get that by doing things
like making the sensor small, or giving it good thermal contact with
what you want to measure.


Ok, so you have a basically unpressurised tank of liquid LH2 and the only
reason the whole thng doesn't evaporate is that there is some insulation that
prevents sufficient heat from penetrating the tank.

Once the liquid level drops to expose a sensor, would the sensor really warm
up instantly ? Wouldn't the ambiant H2 just above the liquid surface still be
extremely cold ?

Or is this more a question of the sensor being heated by current, and when in
liquid, the heat is immediatly absorbed by the liquid which acts as a huge
heat sink, whereas when in gaseous H2, the sensor can heat up more ?


There was mention of similar sensor technology used for the O2 tank. Wouldn't
there be a lot of issues with liquid O2 and whatever metal is being used for
the sensor , and what about any electric current and potential for sparks in
the O2 tank ? Is that a concern ?

John Doe wrote:

Chris Hall wrote:


mK) but speed of response. The sensor has to give you the right answer
almost instantly. If it is a T sensor, then you get that by doing things
like making the sensor small, or giving it good thermal contact with
what you want to measure.


snip

Once the liquid level drops to expose a sensor, would the sensor really warm
up instantly ? Wouldn't the ambiant H2 just above the liquid surface still be
extremely cold ?

Or is this more a question of the sensor being heated by current, and when in
liquid, the heat is immediatly absorbed by the liquid which acts as a huge
heat sink, whereas when in gaseous H2, the sensor can heat up more ?



Yes, it would be a case of the sensor being able to warm up above the
LH2 point once you were in a gas rather than immersed in liquid. The
heat transfer would drop. Then any internal heater (either in the sensor
assembly or the sensor itself via resistive heating) would cause the
sensor T to rise. It won't be hot but it would be high enough resistance
to flip the comparator. If the sensor is a platinum resistor (100 ohms
nominal resistance at 0C, only a few ohms at LH2) and if you
deliberately had a high driving current (several mA instead of
microAmps), that could help to move things along as well. You want the
thing to rise several (or 10's) of degrees in a second or less for an
application like this. But I don't even know if that's how it's being done.


There was mention of similar sensor technology used for the O2 tank. Wouldn't
there be a lot of issues with liquid O2 and whatever metal is being used for
the sensor , and what about any electric current and potential for sparks in
the O2 tank ? Is that a concern ?


I don't know, but most Pt RTDs are encased in glass or ceramic and the
current is tiny. Voltage would be miniscule (probably 1V).

--
Chris M. Hall
Dept. of Geological Sciences
University of Michigan