Hacker News new | past | comments | ask | show | jobs | submit login

For the UPS, I always err on the side of ones that are PFC-compatible, pure sine wave, line interactive. Partly because some of the gear I rack has PFC PSUs, but also because I don't want to be flirting with edge cases in other gear. I've owned 3 such rackmount UPSes, but it doesn't seem to be the default.



Just had a PFC capable APC UPS fail on me. Not that it necessarily means anything. UPS's do fail, that's all.


The two APC SMX1500RM2U rackmount units that I had at home seemed built very well.

(Currently have two Cyberpower rackmount units, which don't need as much rack depth in my living room, but I kinda expect something industrial-sounding like AMERICAN POWER CONVERSION to be more reliable than cheesy-sounding CYBERpower. :)


I used to swear by APC but they have really gone down hill in the last decade. The first rule of a UPS should be, don't burn down the house. So when buying a new unit this year I went with the brand whose consumer PFC-compatible model didn't have several reviews about it catching fire: Cyberpower.


I had a UPS fail last week, but I have two of them and my servers have dual PSUs plugged into each one. I don’t really need the uptime, but I’m mostly glad I was able to finally justify my unnecessarily high power bill to myself. :)


How does a UPS cause a high power bill? Doesn't it just charge up the batteries once and basically maintain after that?


The high power bill is due to using enterprise rack mount hardware with dual PSUs, instead of consumer hardware.


"Idle" is often measured in the hundreds of watts on enterprise/DC gear.

Newer stuff is sometimes a little better about ramping down at idle, but heck, some server fans will use more power than either of my desktop computers at idle.




Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: