Martin Brown wrote:
On 29/10/2020 09:24, Dave Liquorice wrote:
On 29 Oct 2020 00:44:08 GMT, Bob Eager wrote:

Despite the ring circuit being correctly wired and being quite short,
and there being little load on it...when I switch on the laser printer,
the UPS shifts taps to compensate for the voltage drop - it is clearly
audible.

That sort of surprises me, there will be an inrush as the filter
capacitor and PSU capacitors charge but wouldn't expect that to be
greatly different to a PC. Think I'd be looking to see if the UPS is
a bit "sensitive" or what the volt drop drop is on that ring when
loaded. May be indicating a slightly iffy joint somewhere.

Laser printers are a brutal load to get the fuser element that melts the
toner onto the paper up to temperature quickly is a brief but very high
surge current from cold. Easily enough to trip a UPS out.

Inkjet printers are fine though.


Laser printers are separated by generation.

It was the first generation that took 1kW when
bringing the fuser up to standby temp.

Modern toners (at least the color toners) melt at
a lower temperature, which means the fuser design
now has a lower mass, and also a slightly different
implementation (quartz tube with heater inside).

The number now should be less than the "old 1kW" number.

Example: Samsung ML-1670 (Samsung Laser division now owned by HP...)

Power consumption (printing) 270 W
Power consumption (standby) 40 W
Power save mode 2.8 W
Power off mode 0.45 W
AC input voltage 220 - 240 V
AC input frequency 50/60 Hz

Since nice cheap gutless UPS are available,
there will still be a temptation to run too
high a load on them. You should never buy a
UPS from the "first tier" which is the
cheap and cheerful tier. We tried that at work...
The failure rate from new, was around 10% on those.
None of the failures were destructive (of ~100 units tested).
The quality issues should have been sorted at
their factory.

*******

A Kill-O-Watt meter measures Watts, VA, and PF.
And has taken all the guesswork out of picking a UPS.

Connect all the loads to the Kill-O-Watt. Then use
the three numbers that result, to pick the UPS.

On the computer, run Prime95, as a means of getting
a representative load. At the same time, send a print
to your Samsung laser printer. Now, look at the
Kill-O-Watt display, to see what your "estimated
worst case" is.

Before the Kill-O-Watt meter existed, it was a
lot harder for ordinary people to size these things.

Even incandescent light bulbs can be deceiving.
When cold, a 60W lightbulb draws 120W at startup.
If you had 5 incandescent light bulbs, each switched
on individually, you'd need 4*60W + 120W or 360W max
(at the moment you're switching on the fifth lightbulb).

If all bulbs were switched via a common lightswitch,
the power (worst case) would be 5*120W or 600W max.
It depends on whether a UPS has a "hair trigger",
as to whether it drops out the instant a light bulb load
appears like that (whether a 600W load for 5-10 mains
cycles would be enough to trip it).

ATX power supplies (desktop PC) have terrible
characteristics. They can draw 40A for a few mains
cycles. But only occasionally does this cause
a problem. Anandtech evaluated one ATX PSU, a
device rated for 1kW DC output, and every time
they turned it on at the back, it tripped their
lab breaker. Which is surely an achievement
(as to trip a breaker instantly, takes around 5x
the breaker rating).

The transient behavior of loads, and the cheesiness
of some UPS designs, are confounding issues.

And on electric motors, the stall current or
locked rotor current rating, can be 10x the operational
value. A 2HP motor could draw 20HP at startup (15kW).
And that causes indigestion for a lot of solutions.
A 2HP motor is the largest motor I've seen in a
domestic situation (YMMV). I'm sure many here will
have bigger kit than that. That's why even fractional
HP motors aren't a good fit for UPS.

*******

For the comedic value...

https://teslamotorsclub.com/tmc/thre...l.99078/page-9

"I have three Powerwall 2's and a 2-stage A/C which consumes about
2kW and 4kW depending on first and second stage. The blower motor
uses another 200W. My stack of three PW2's cannot start this
particular A/C unit. When they installed the Powerwalls, they checked
the inrush current of the A/C with a clamp ammeter. I would not expect
to run an A/C of that size with one Powerwall at all. I am trying to
get Tesla to come out and install a soft start unit on my compressor,
which should reduce the inrush current from over 150 amps to something
reasonable like 60 amps.

At the moment, when the A/C starts up, the Powerwalls do not enter a
fault state, but my house blacks out for a quarter second (the lights
literally go off) and all the electronics reboot.
"

An easy way to do reboots.

Paul