Guys.
Its looking like a component issue. We're still testing here to get to the bottom of it. In the meantime, can you all please stop running your boards. I know this sucks but it will make things better in the long run. Whatever is going on here is fixable, but will be quicker and easier to fix if the boards are preserved as they are now rather than risking damage by running them while things aren't working as they should.
More soon.
Can you please give us some insight into what you guys are seeing or thinking? While it may be better to stop running the boards, it is hard for us to make the risk assessment without information. The other main issue here is ability to mine with a significant delay. I worry quite a bit about the future returns of boards if there is a delay of a few weeks. At this point I don't think it's too much to ask to be included in your internal discussions.
Based on my calculations, it would actually as profitable to run it for four weeks now then have it die, rather than waiting a month and running for 6 months. This is a very serious risk analysis problem, one that we need proper information on failure rates and what components are at risk.
I'm sure you understand my concerns.