If you want the highest accuracy you would go direct contact between the fluid and the instrument simply as there are no heat losses at the boundary faces over and above the conduction of the heat thru the metal probe body, ie heat from motor transferred to oil, oil transfers heat to the outer casing of the direct contact probe, heat migrates thru the probe metal, likely thru a conductive paste inside the probe head to warm the thermocouple tips which has a small electrical voltage, and the resistance changes as the metal probe wires heat up, and you get your variable scale output read as DegC.. With the clamp on config, you go engine oil to filter housing body, thru the metal thickness - there are insulating temp drops depending on the thickness, then out the other side of the filter housing, where you strike an insulator, being the paint of the filter, then you bright metal clamp probe, heats up and transfers the heat to the thermocouple head which is soldered to the clamp ring, and the whole voltage/resistance thing happens as above. So there is some efficiency to be gained with direct contact, however you only get the benefit if you have calibrated and matched instruments of a high standard and your physical operating environment is stable. For automotive applications outside of test rigs, the calibration would be relatively ordinary compared to industrial and certainly miles off lab level. This is why there can be such a wide range of results with even two instruments sitting side by side - calibration variance, quality of components and the physical locations could experience het cold, water/cold air etc.Quote:
Originally Posted by Winnie
Years ago I was chasing accuracy on a steam jacketed heating kettle I had designed and it needed to hot a sauce at a preset temp for a 5 minute period to ensure all the bugs were killed. The product was super sensitive to over temp conditions due to the fat content, and would break down and seperate in a horrible blob of circa 800kg, all lost product. If not heated high enough, the kill step was not achieved so the process was delayed until it was, thus impacting production lines with no product. Anyway, I engaged an instrumentation tech to help me and we used all highly calibrated instruments with very high sensitivity, still could not get the accuracy. Turned out the paddle design had generated an un stirred pocket right over where the probe was and was reading 3 deg out. Took 2 weeks to find that. The instruments were fine, the accuracy was exceptional, but the physical location turned out to be crap due to this quirk within the kettle stirring design. I added a deflector to the stirrer arm and moved the probe 25mm away and it suddenly all worked beautifully. Once again, the best instruments can't compensate for a bad installation. In theory it should have work from the get-go, but this one product snookered me fair and square - another design lesson.
