- #1
Skaperen
- 90
- 1
I'm looking for information on how hot swap PSUs as used in computers (used typically for data center servers) are designed. In particular, I want to focus on the common aspects between the 2 or more PSU subunits that can be swapped. What kinds of connections are made between the individual swappable PSU module, and the common power circuitry? Is there any chance the any mains AC voltage would go across this connection, or by any other means, even if it is intended as a low current or current limited connection? What I am wondering about is whether this could be exposed to a phase-to-phase higher voltage (e.g. 416 volts) if/when the individual PSU modules are powered by separate phases (in a 240 volt phase to neutral system).
Hopefully, no hot swap PSU systems even exist where they allow AC mains voltages from separate PSU modules to be applied together anywhere, or if they do, are specifically designed to handle the worst case (e.g. at least 480 volts differential for PSUs that can operate at 100-240 volts). Common non-swappable PSUs for PCs only have DC and control lines that I have ever seen. But I can't get enough info on the hot swap type to rule out any phase-to-phase risks.
I am wanting to wire up servers which have 1+1 PSU redundancy (e.g. can run on 1 PSU being powered) to separate UPSes (small scale units, 1 or 2 per rack, not a giant room scale UPS). These UPSes would be wired to different phases in sequence. If I lose power to one phase only, and time passes by that exhausts the UPS battery, the phase diversity should still keep everything running (as long as the two remaining phases have that much capacity). Phase diversity UPSes would be nice, too, but I just don't see that on the market in rack size scales (3 to 6 kVA).
But in any case, what I am exploring at the moment is any possible cross phase risks, where phase to phase exceeds 240 volts, in the 1+1 redundant hot swap PSU system designs.
Hopefully, no hot swap PSU systems even exist where they allow AC mains voltages from separate PSU modules to be applied together anywhere, or if they do, are specifically designed to handle the worst case (e.g. at least 480 volts differential for PSUs that can operate at 100-240 volts). Common non-swappable PSUs for PCs only have DC and control lines that I have ever seen. But I can't get enough info on the hot swap type to rule out any phase-to-phase risks.
I am wanting to wire up servers which have 1+1 PSU redundancy (e.g. can run on 1 PSU being powered) to separate UPSes (small scale units, 1 or 2 per rack, not a giant room scale UPS). These UPSes would be wired to different phases in sequence. If I lose power to one phase only, and time passes by that exhausts the UPS battery, the phase diversity should still keep everything running (as long as the two remaining phases have that much capacity). Phase diversity UPSes would be nice, too, but I just don't see that on the market in rack size scales (3 to 6 kVA).
But in any case, what I am exploring at the moment is any possible cross phase risks, where phase to phase exceeds 240 volts, in the 1+1 redundant hot swap PSU system designs.