[skimoose]

If you are only losing communications for a half second occasionally, it's probably found in a poor connection between the dLink and cRIO. Check or replace your CAT5 wiring.

If it was a loss of power or an under voltage issue, the dLink takes about 40-45 seconds to reboot and re-establish comms. This can be the result of loose connections in your wiring or more likely a poor connection at the dLink barrel power connector. These power jacks are pretty poor quality and we've started duct taping our power connector into the dLink to insure proper connectivity. The other reasons for the dLink losing power are the voltage converter failing or not correctly wiring the voltage converter to the regulated power port and using any WAGO power port through a snap action breaker. These radios seems to be very susceptible to voltage sags. If the battery is getting low or the voltage regulation isn't doing its job, a voltage drop will knock the dLink off line.

We've had one dLink actually fail at Championships. After our radio began knocking off line mid match and getting progressively worse each match, we ran through every possibility I mentioned above. We finally swapped in a loaner radio from another team and the issue went away. That trouble shooting cost us three matches of going dead and sitting there for 45 seconds hopelessly waiting to re-establish comms.

When our team returned home, I opened the dLink to search for the cause of our problems. Under magnification I found a surface mount capacitor with apparent cold solder joints. One side had broken free from the pcb and when I barely touch the component it fell off. These units are manufactured in China and the board does not appear to be wave soldered but hand soldered. Looking at the components under magnification many are not lying flat on the board and are not uniformly placed on their pads. So while the cRIO will handle a 50G shock load I don't think these radios are quite as robust.

We had another radio on our practice robot begin to have similar issues after one of our students accidentally drove at full speed into a tower base. I opened that radio to look for similar bad solder joints but didn't find any. There was a distinct smell of "magic smoke", but the radio was clean with no foreign matter inside the unit which could have contributed to a short.

I hope my problems are isolated issues. As far as I know, there haven't been many reports of failed dLinks, but if others are having problems, we should probably let Kate know up at FIRST HQ.