When responding to a network crash, don't just reboot the system and consider the problem solved when it comes up again. Network crashes can be caused by a variety of factors, and they aren't all IT problems.
Some electrically-related problems to look for include:
- Data loss points. A network analyzer can help you see where and why data loss is occurring. The data loss point may be an active device (e.g., a switch) or passive device (e.g., connector).
- Poor cabling connections. Contrary to common assumptions, this affects wireless networks in addition to wired ones. In a "wireless" network, the network is wired all the way to the "access point" devices that provide the "wireless" part of the network. Use a network cable tester appropriate to the level of the cabling system (e.g., Cat. 5).
- Grounding and bonding issues. Review the definitions of these terms in Art. 100 of the NEC. Grounding is probably not where your problem lies. Review Part V of Art. 250, and then inspect the network for conformance.
- Power anomalies. Because industrial network devices tend to have high tolerance for voltage deviations, power issues may seem unlikely. However, if the MOVs typically used for surge protection are zapped, the circuits will continue to operate. They just don't have MOV protection anymore. Check the power monitor logs for voltage spikes. If such spikes show up, they may have caused the crashes. Replace defective or suspect MOVs, and then determine if your facility's surge protection scheme follows best practices.