One of the most challenging experiences I have had as an engineer was dealing with an aircraft's airflow sensor.
The sensor measured airflow in the aircraft's air conditioning system, also collecting various inputs from other sources, such as the outside ram air pressure and electrical signals from the cockpit. It then translated all of this data into an air pressure signal that controlled the valve that in turn controlled airflow from the engine compressors into the air-conditioning and pressurization system.
This device was not only highly complex, it was unique in that almost everything about it was wrong. The overall system was of questionable design, the technical manual for overhaul of the sensor a total mess, and the overhaul kit contained a bunch of unneeded items while lacking required parts. The final killer was a tricky manufacturing defect. But one problem was especially baffling.
In an apparent attempt to prevent the sensor from sending pressure signals to the valve that varied too quickly, an orifice paralleled by two pop-off valves pointing opposite directions and feeding a small surge tank had been added. The mechanics in the overhaul shop reported that setting the pop-off valves to the required opening pressure was a total crap-shoot. Sometimes they worked and could be set correctly, and sometimes they did not, for no apparent reason.
Examining the parts, I rolled them out on a sheet of paper and was astonished that they seemed to be very slightly tapered. There was nothing to prevent you from installing the taper so it would point one way or another. The manual did not mention the fact. In reality, the small end of taper was to point down into the hole. Problem solved! Or almost.
I got data from the manufacturer and found that the pop-off valve calibration procedure was supposed to be done with the center orifice plugged! Once again, no mention of that requirement was in the manual. I designed a little device that would both plug in the orifice and interface the device to our test equipment. That made a big difference to the results as well.
And then there were only about three or four other things to fix on that sensor...
This entry was submitted by Wayne Eleazer and edited by Rob Spiegel.
Wayne Eleazer graduated with a BS in Mechanical Engineering from the University of South Carolina in 1974, was commissioned a 2nd Lt in the USAF, and served in a number of engineering and program management assignments. He retired from the USAF in 1999 in the rank of Lt Col, his last job being Chief of Advanced Planning at Cape Canaveral AFS. He now works for ACTA as its expert on space launch failures.
Tell us your experience in solving a knotty engineering problem. Send to Rob Spiegel for Sherlock Ohms.