I used to design agricultural electronics, and one of our products was a grain wagon scale. The axles of the grain cart formed the actual load cells for the scale system so that each axle had strain gauges glued to their surface. All four load cells were wired in series to produce one large Wheatstone bridge.
I began to get field complaints from a new customer who was having a very difficult time troubleshooting the problem. Since the load cells were wired in series, a shorted cell could appear anywhere in the Wheatstone bridge, and the symptoms would vary depending upon which socket of the scale the faulty cell was plugged into.
From the customer's point of view, he had a scale that wouldn't turn on, so he would swap the scale out of the system for a new one. Then, the second scale would turn on, but he wouldn't be able to balance the scale and show a reading.
He would try yet another scale and that one would work, but would be rather unstable. I tried to explain to the customer that the problem might be a faulty load cell, but the customer was convinced the problem was in the electronics. Finally, I convinced him to send me the complete systems for examination.
While examining the failed systems, I found that each of them had a short on one of the axle strain gauges. Using a microscope, I saw that each of the failed gauges had a burn mark on the polyimide backing, shorting the gauge foil to the axle's steel.
I called the customer to tell him of my findings. He was becoming quite agitated because he believed I was covering up a serious quality problem with our electronics. From his point of view, every time he swapped out the electronic scale, he saw a different problem, so clearly every one of the scales must be defective. “Besides,” he said, “why would we begin having these problems with the axles now that we've just improved the wagon by adding shock mounts?”
“You've recently changed your design?” I asked cautiously.
“Yes," he said, "now the whole wagon is mounted on huge shock mounts, so the axles aren't even getting bounced around like they used to.”
It was starting to make sense to me, but I had to be careful. I asked the customer to take an Ohm meter and check for continuity between the axle frame and the wagon body; sure enough, the two were isolated. I explained that the tires would build up a significant static charge while the wagon traveled down the road, and without a connection between the axle and the wagon frame, the high voltage would eventually punch through the strain gauge backing, onto the strain gauge, up to the scale, and then to ground.
The customer reluctantly added a ground strap between the wagon body and the axle frame. The field failures stopped with the addition of the ground strap, but within in a year we still lost the customer because he believed that the ground strap was simply a cover story for some sort of rampant quality problem.
This entry was submitted by Chip Curtis and edited by Jennifer Campbell.
Chip Curtis is a development engineer for Teko, designing small consumer appliances and medical devices.
Tell us your experience in solving a knotty engineering problem. Send to Jennifer Campbell for Sherlock Ohms.
Click here to access the Sherlock archives.