An engineer proves once again that the best diagnostic tool is the one
that finds the source of the problem, not the newest or prettiest one
in the bag
About a hundred engineering years ago, I worked for a company that made
sunlight readable (high bright) LCDs. They used thin CCFLs (cold
cathode fluorescent lights) and optical films to do this; the CCFLs
used high-voltage inverters and worked very well. I had released a new
design to production and things went along fine for a while.
One day I was called out to the production floor – brightness was
way down. The usual swaps fixed nothing. The voltage across the tubes
was way down, but disconnected (open-circuit) it was normal.
We had just gone through a 6 Sigma training and certification
session and the engineering manager was anxious to show that 6 Sigma
was the universal cure-all. I had the normal engineer’s love of
paperwork and ceremonial procedure – I wanted to find the cause, fix
it, document it and go back to shipping product. I understand 6 Sigma’s
power in high volume production, but this was a 20-unit-per-month
problem, not millions-of-ballpoint-pens-per-day problem. While the
wizards pondered, I grabbed a failing unit and headed back to the lab.
I had been bit by this high voltage too many times before during
design and debug and wanted to think more with my brains than my
fingers. What looks (electrically) like a digit without the bloodshed
and puncture burns? I got a Q-tip, dipped one end in water, clipped a
lead to it, grounded the other end of the wire to circuit ground and
held onto the dry end. I then traced out the HV wire paths from
inverter through the chassis up to the backlights. We used flexible
silicone wire between the inverter and the lamps. As I ran my "finger"
across a tie wrap fastening HV wires to the chassis, the lights went
from dim to OFF. I shut off power and looked at the tie wrap. It was a
soft metal pair of ears embedded in a nylon tab backed with PSA. The
metal tabs were anchored in the nylon with a wide metal tab – think of
a stamp with two flexible arms.
These metal tabs wrapped around the HV wires created a capacitor to
ground and, since the blocking cap from the HV secondary of the
transformer was on the order of a few pF, the tabs created a voltage
dividing cap and my probe was shorting out this cap. The tie-down was a
good manufacturing part, easy to place anywhere, cheap and adjustable.
It just wasn’t good for tying down HV wires. Manufacturing changed to a
nonmetallic part and my problem was solved without another puncture
Management hailed this as another case of "smarter by the numbers."
I smiled and nodded and went back to the next design problem.
About author John Linstrom: "I'm a degreed EE with 30 years experience
in analog/low level/low power instrumentation in
medical/communications/industrial environments. I currently work for a
contract manufacturer in NW Montana and for myself at SUNSPOT."