RE:
I agree that "sign wave" sounds promising. But I don't think we have seen anything specific from Prusa regarding wave forms, small spikes or whatever it is that trips up the Core One L's power monitoring. And I don't think I have seen any report of success here on the forum (or elsewhere) where someone found a good UPS that actually works with he Core One L. Maybe the unsuccessful attempts all used cheap stepped-approximation supplies; maybe something more subtle is getting in the way?
Obviously the core one I never had a problem with from that perspective… the H2D draws a crazy amount of current when it’s heating the bed, and I’d be rather surprised if the core one L did not do the same thing… and that along with that rather significant draw, it would stress out any sort of lesser unit. I was actually surprised that I could get the APC to work on overload for about 30 seconds, but that was not a core one L. it screamed the entire time but the SINE wave was totally clean… I suspect eventually I would’ve burned something out internal to the UPS.
that’s why it ended up with the C 2000 GEN two. once the printer is actually operationally past it’s warming phase for the heat bed - It will run the printer for a significant amount of time… And you can actually add additional batteries should you find the need. realistically, that kind of solution is about the only thing that you’re gonna find to really do a good job… unless you can find some sort of conditioner that can clean up a crappy waveform. I have used some basic ones in the past that have rectifier in them and will strip out any sort of riding waveform on top of the SINE wave. problem is they’re all made relatively cheaply nowadays as opposed to 30 years ago when they would be built fairly well.
RE: UPS for CoreOne L
I think all APC units are sine wave, and their larger products are designed for commercial data centre use (especially the three phase units) so they all have very high quality output. I have checked in the past with an oscilloscope and the waveform they output is a better quality sine wave than I get through the wall socket.
But I suspect the problem with the Core One L is always going to be the switchover delay as any standby UPS takes a few moments to switch the inverter on so there is always a small interruption to the supply. The delay is very short, only a few AC cycles, but if you plug a light into a standby UPS it's enough that you can see the briefest of flickers when you switch off the wall socket. For most appliances it's no problem, but it sounds like the Core One L monitors the power quality very closely so it seems the firmware would have to be designed to handle short interruptions like this for any standby UPS to work, regardless of brand or quality.
I don't know, but I suspect, the quality of the waveform is not so important (the stair stepped one might be fine) since power waveform monitoring hardware is typically quite expensive, and resistive loads like heaters don't care about the waveform quality (and will typically run just fine off DC as long as the voltage is the same). But if they are running the bed directly from mains AC then any interruption would risk a change to the bed temperature so I imagine that's what they are most concerned about. I am just guessing here though, as I don't know how the heater is run and how they are switching the power to control the temperature. I don't think the switching process can be PWM or similar as that would result in a terrible power factor and IIRC there are rules in the EU about keeping your power factor high.
APC have very good overload protection in them. Like any good inverter they have a surge rating which allows them to provide more than their rated power for a short while, e.g. to handle the momentary inrush current when switching on say a fridge compressor. After a little while though they will shut off the output before anything internal gets damaged.
This is because there is no fixed point at which an electrical circuit is overloaded. All electrical current generates heat as it moves though a circuit, and typically the air cools the circuit so the heat never builds up beyond a certain point. But as you push more current through a circuit, there's a point at which the heat cannot escape quickly enough, and the circuit continues to heat up until eventually something will melt and you get a failure. But that is not an instant process, it takes time for all the metal to heat up, so you can always run things above their ratings for short periods (the further above the rating the shorter the time, since more power means things heat up faster).
This is why fuses and circuit breakers don't magically trip the instant you reach their rated current, but even with a small overload they can still take hours, sometimes days before they blow/trip. It's because the extra current just isn't generating enough heat for it to be a problem, especially if the outside temperature is cold enough that the additional heat can be removed.
It's also why certain cheap home appliances have a duty cycle - so many minutes of operation followed by some time to allow it to cool down. For things like coffee grinders that are typically run for only a few minutes at a time, it allows the manufacturer to get away with a smaller cheaper motor and overload it, because they are betting you'll have finished grinding your coffee before the overloaded motor gets dangerously hot, and then it will have plenty of time to cool down again before you next need to use it.
RE: UPS for CoreOne L
Duty cycle would be the key phrase there... The current setup I'm using is a 120v APC Transfer switch with one leg in the wall and the other into the battery pack. You can run the unit full time off the battery pack, I just tend to not want to stress that Li-Ion unit out by running in continuously. The transfer switch has bunch of beefy capacitors in it, and that keep the AC at line voltage without any dips while the power cuts over to battery pack. What I haven't checked is if the output of the battery pack is exactly the same amplitude as the wall outlet in every facet.
RE: UPS for CoreOne L
Capacitors can only store DC voltage so at best you'll get one downstroke of the AC waveform out of them, but if the transfer happens at a low point in the AC cycle the capacitors will be empty and won't do anything. You might be seeing filter capacitors intended to clean the AC waveform rather than to ride out any interruptions during the switchover delay.
A transfer switch can switch faster than a UPS because it's moving from one live source to another. If the two sources are in sync then you might only see an outage for a fraction of a cycle so the filter capacitors might be able to smooth that out, but if it's on a rising part of the waveform you'll still get a little flat stair-step there where the voltage stops rising for a moment.
But a UPS is slower than a transfer switch because it also has to get the inverter running first, and because it takes more than one AC cycle to do that, you can't ride it out with capacitors. That's why dual conversion UPSes exist, which run the load off the inverter even when mains power is available, using mains power as a DC source to run that inverter and keep the batteries charged. Similar by the sounds of it to your battery pack, except the charger is beefy enough that it can run everything while keeping the batteries fully charged.