View Single Post
jim hardy
jim hardy is offline
Feb13-12, 01:24 PM
Sci Advisor
jim hardy's Avatar
P: 3,139
i'm no industry apologist, just an old instrument guy.

If i recall these sensors are thermocouples.
If thermocouples have one weakness it's this - moisture in their connecting wires.

A thermocouple is just two wires of dissimilar metals joined together at the end.
The dissimilar metals when heated produce tens of microvolts per degree , and one reads that miniscule voltage to infer temperature at the business end.

So in effect you have a microvolt source that's a function of(temperature) in series with the resistance of the wires . Most thermocouple wire has resistance that is substantially higher than copper. Make a thevenin equivalent out of that and measure it with instrument of your choice.

Now should the insulation of those wires get damaged and allow moisture to ingress,
you no longer have just a thermocouple, you also have dissimilar metals in an electrolyte which make a galvanic cell.
That places a galvanic cell electrically in parallel with your thermocouple junction.
Galvanic cells make thousands of microvolts not tens.
If there's enough wetted area on the moist part of the wires the galvanic cell will deliver enough current to overwhelm the thermocouple's meager voltage.
If you're lucky the polarity of the galvanic cell will cause a voltage indicating temperatures below freezing so you know immediately it's haywire.
If you're unlucky it makes a voltage that indicates high temperature and everybody wrings their hands until reading becomes impossibly high.

SO-- my point is this - they changed injection flow and one thermocouple departed from its neighbors. It bears watching but is suspect. I think they have a water leak above that thermocouple.

In my plant i checked for such things with a simple analog multimeter, which you dont ordinarily use on thermocouples.
The way you tell is read resistance, switch the polarity and read again. If the reading changes by more than about 1/4 inch needle deflection it's likely wet. (Modern digital meters dont have that reversing switch you have to swap the leads and infer from the numbers, which is much less intuitive)
Then you read the current it'll deliver into your multimeter on current scale and if it's more than a few microamps you know there's a substantial galvanic cell out there. If the thermocouple is the ungrounded type a simple resistance to earth confirms insulation damage..

It is sometimes difficult to explain these details to people who aren't mildly autistic like me.
And you understand why those details aren't in the press release - they infer uncertainty which is not acceptable public image.

And while a lowly instrument technician like me might firmly believe in his results, there's no such thing as absolute certainty.
I'm sure they have better technnicians than me over there and that's why they are saying they think the sensor has failed. It'll probably come back when it dries out. I saw a report of resistance measurement "1.5X normal" and would love to know what it read with polarity reversed - 2/3 normal would be a telling answer. (Just as would "1.5X normal with no sign of wetness".)

Dont know why i rambled so - just want to help those guys over there but cannot.

Anyhow dont bet a lot of money or invest much worry on a lone thermocouple that's suffered the abuse those have. But keep an eye on it.

please advise if this post is out of line.