On Sun, 12 Mar 2006 10:21:38 -0800, Jim Stewart wrote:



Based on an incident that happened several years
ago, I've avoided drugstore "mineral oil" for any
sort of technical application.

We had a manifold with several high pressure
transducers mounted on it. We were gathering
data for temperature compensation tables. We
would apply a specific pressure with a deadweight
tester then vary the temperature of the transducers
in a thermal chamber while logging and plotting
the outputs. We were using "mineral oil" as
the working fluid.

I got a call at 9pm from the project leader. He
said all the transducers had died at -30 degrees
C. As we were far behind on this project and it
was a make-or-break project for the company, I
came in immediately.

It didn't take me long to find that there wasn't
anything wrong with the transducers, but that at
-30 degrees C, the mineral oil had changed consistency
to something more like a cross between slush and
Vaseline. Rinsing and refilling the system with
Mobil1 5W10 (the best thing I could find nearby at
that time of night) fixed the problem.

So what's in mineral oil? Water? I don't know.
I just don't like the stuff any more.




Probably you just reached the wax point of the oil. the 5W10 had a lower wax
point but would probably have gone the same way eventually.



RAMBLE
We had an oil and transducer problem back in the 80's. A customer claimed that
the transducers we were using to measure boiler feedwater flow during a steam
turbine performance test were reading 1.5% low.

We were using KDG differential transducers, which used lvdt's to sense
diaphragm movement in pairs of dashpots with the back sides of the pots full
of silicone oil and connected via a balance pipe. They were using Rosemount
transducers which had a single diaphragm between two isolating diaphragms with
the enclosed spaces also filled with silicone oil. The Rosemount system
measured the change in capacitance between the diaphragm and the two halves of
the chamber.

Several weeks of testing showed that our transducers had lousy temperature
coefficients (which we knew about and always used the transducers in
temperature controlled boxes). The Rosemounts had excellent temperature
coefficients. However all the Rosemounts consistently read 1.5% high when
subjected to a 2300psi static pressure. This was the source of the difference,
since the measurement in dispute was on a flow meter at boiler inlet pressure.

The cause of the error was simply that silicone oil is compressible and as it
compresses, its dielectric content changes.

The investigation paid for itself because the error was worth $2m per 650MW
turbine in 1984 money.
/RAMBLE


Mark Rand
RTFM