Several years ago, I visited a refinery to install a Chromatography Data System (CDS). All was going well and I was ahead of schedule. Well, that was the case until it came to the last instrument. The physical connections to the instrument were a snap to make and the instrument was soon connected and we were collecting data. That part was easy, but getting a final result that was consistent with what the customer expected, was not.
After running a number of standards and controls I could not get the result to be within an order of magnitude of the one the old HP integrator was producing. Of course it just so happened that the test method run on this instrument was critical for product release, so the pressure was on. I soon figured out that the integrator was doing some additional number crunching, at the end of the run, to come up with this important result.
I asked what calculation the integrator was doing and I got a lot of blank stares. “Someone must know. Who wrote the program?” I asked. It turned out that a former employee had set up the integrator many years before but had left no information on what had been done. Integrator manual to the rescue!
After a late evening of pouring over the program listing and the manual, I finally figured out what was going on, really quite simple, but with no description left behind it was the worst kind of black box.
A quick update to the CDS method and I was doing the calculation correctly.
The moral of this story: whenever you perform some coding, configuration or otherwise modify a laboratory informatics system, always leave a trail (paper or electronic ) to inform those that come later.