Hello James,
Am 07.08.2025 um 19:42 schrieb James H. H. Lampert via MIDRANGE-L <midrange-l@xxxxxxxxxxxxxxxxxx>:
I just powered up the "MERCURY" 170 with an "amp clamp" (and an adapter allowing me to use an "amp clamp" on a corded device), and I'm in the middle of doing so with the E4A (with one PSU unplugged).
Interesting readings. But simply doing the math is not giving you correct power values for AC. Something you need for UPSes — and what you eventually pay for per kWh.
There was some discussion about this very topic not too long ago. In short: Amps times Volts is only valid if tension and current are in perfect sync for phase as well as amplitude. This is the case with e. g. a classic electric stove or other pure resistance based devices. Once you have nonlinear devices such as caps, coils, or diodes, this is no longer the case, and you most often get higher values with simple math compared to what's really drawn.
The key is the term "reactive power". See
https://en.wikipedia.org/wiki/AC_power#Reactive_power — I admit that the article is overly mathematic.
I suggest to buy a proper power meter, the household type "in between plug" type. The cheapest ones are not very precise, especially when measuring low power values (standby). Can be reused many times to get a better idea what device uses how much energy.
Obviously, a "fat" 170, like MARS, would draw quite a bit more, since the second PSU is *not* for redundancy's sake.
Now I understand what you mean with "fat". You're referring to the expansion box I usually name "sidecar". :-)
Running, but quiet, on both PSUs: 1.5-1.7A (180-204VA) per PSU.
This is surprisingly close to my 400 W I remember to have measured.
So if the box is only in intermittent use, and not usually left powered up unsupervised, and you plan on doing an immediate PWRDWNSYS at the first sign of flaky AC, it's probably not necessary to use that big of a UPS.
Rule of thumb: Buy an UPS which gives you enough runtime for a clean shutdown, and some extra headroom. I admit, I have no experience with IBM i and how it decides when it's time to shut down. Maybe immediately after an "on battery" is signaled? But then, this happens once a week with the usual UPS self test. Which is alt least signaled as short outage via SNMP with APC devices. I assume it's also via serial port.
:wq! PoC
midrange.com
