My lesson from my personal experiments with Lavarand: you must have more than one lamp, not necessarily for more entropy, but for fail-over and uptime. At appr

My lesson from my personal experiments with Lavarand: you must have more than one lamp, not necessarily for more entropy, but for fail-over and uptime. At approximately ~30 hours, my vintage 70's lamp 'gives up' - the fluid temperature becomes pretty even between the bottom and the top. It's all essentially superheated as far as the wax is concerned and it simply stays in one place as a dome at the bottom, barely moving. This isn't good for creating random data. By using multiple lamps, it's possible to power cycle them. Ideally, every ten hours or so, remaining off for a couple hours.