Short post before bed… I am building tomorrow, so I need my beauty sleep.

But before falling asleep, I need to write some notes. I have just wasted most of an evening seaching for a problem with an RS232 stream. I have *FINALLY* found the problem but it was as much luck as the years of experience that allowed me to track this one down. It was a heisenburg bug. Whenever I searched for it I would not find it since the bug would go away. When I snooped on the datastream things were fine.

When I snooped on one direction of the datastream things were fine. And the other direction. But if I snooped on nothing it would not work at all. I could not work it out. Then I had a thought… What if it was a grounding issue. It should not be that since the laptop doing the snooping is really not connected to anything. So I made up the worlds strangest RS-232 cable… It only had the ground connected. When it was plugged into the snooping cable things would work. When it was open circuit things would not work.

Then I tried connecting this to the RS232/Network interface box (Terminal Server), and when I touched the ground to the case things worked. Seems like there is a gound loop or something in operation. In the next few days I will connect the ground to the chasis inside the box, probably via a 100 ohm resistor. That should fix things.