There was a thread some time back about powering ONTs, routers and Wi Fi points during a power cut if fed by fibre to the home.
Soemone pointed out usingf a mains UPS was inefficient as its 220V mains down to lead acid battery voltage, then back up to 220 mains out put, then down to 12V via the wall warts for said devices.
Someone kindly posted a link to a module that was mains powered, charged a 12 volt battery and the ONTY, router, wi fi would ben powered eithert with 12 V DC from mains or 12 V DC from said battery during power failure.
In message <ublmgr$3rf9q$ snipped-for-privacy@dont-email.me, SH snipped-for-privacy@spam.com writes
For anything being powered from an approximately 12V DC mains power supply, I would have thought that a basic two diodes and one resistor would be all that is needed to enable it to be fed from a trickle-charged 12V backup battery.
No but not everything runs on 12v, Many routers seem to be 15v. That is why folk tend to use an inverter to make it mains again, doing away with lots of custom power supplies. Brian
I wonder if over time there will be a move to coerce equipment manufacturers to use one standard voltage and one standard power connector, as there as been with USB power/data connectors.
A mains UPS is so much easier (if less power-efficient) so it can drive as many wall-warts as you need (subject to total power consumption versus life of charge in battery). Mains power distribution blocks are easy to string together, whereas there isn't the equivalent for USB (5V) and equipment (12 or 15V) to fan-out from a smaller number of power sockets.
My experience with UPSes is that they are needed mainly to provide coverage for frequent breaks of no more than a couple of seconds. We get a whole run of those every summer/autumn, probably due to overhanging trees touching HV lines, forcing switchgear to try an alternative HV feed. Northern Powergen need to be a lot more pro-active in cutting back trees *before* they become a problem. It is rare to get a power cut of longer: the 2-hour cut that we had last month was the first long one that we've had in the four years we've live here.
But a break of 1-2 seconds is *plenty* long enough for every computer-based device to reboot. That may involve Windows computers chkdsking the disk (a long process), and will almost always involve turning on mesh network nodes in a specific order so they will all connect. It's a crying shame that our mesh nodes are all-or-nothing for 2.4 GHz: you can't configure them so most only use 5 GHz (where there is plenty of available bandwidth) and only additionally use 2.4 GHz on a specific node which needs to talk to a device that can't speak 5 GHz. You can only turn on 2.4 Ghz on all or none of the nodes: and they spend ages faffing around trying to auto-negotiate channels - because the nodes have to be spaced closely enough for 5 GHz connectivity to the parent node, but this means that almost every node can see every other node at 2.4.
USB power delivery allows negotiation of min/max required voltage and max available current. The present range is 5V to 48V and 0.5A to 5A, quite a wide range ...
EV "chargers" can provide mains voltage AC or high voltage DC on the same plug, Tesla's proprietary connector with fewer pins seems to be winning over the CCS standard in the US at least.
USB-PD works by means of negotiation, but it's a bit awkward if you don't need that. In other words, if you have a 12V power brick and a 12V device, putting a USB-C connector on the end implies the PSU can generate at least
5V and 9V as well, and the device can cope with being given 5V - because inevitably the PSU and the device may become separated over time, and people will expect 'any' USB-C PSU to power the device. That adds extra cost to not much benefit.
OTOH having routers powered from a standard 5V input wouldn't be a terrible idea - 5V 3A should be enough... although again somebody will probably try to plug in a 500mA supply and wonder why it doesn't work. And it makes the power circuitry more complex (eg need to provide 5.1V for a USB host port, but the input is on a crappy USB cable and only getting 4.7V under load, so now we need a boost regulator).
So I think it boils down to the barrel jack we have today is the simplest and cheapest solution. Maybe there could be different notches or something to stop fitting a 19V supply to a 9V equipment (or worse, backwards polarised), but a lot of things are 12V centre positive so that is the de facto standard.
I suppose it would be nice if devices had a buck-boost converter with a wide input tolerance (eg 'Input: 4-20V dc 10W') which would mean you don't have to be too fussy about supply. That would possibly not cost too much extra (although I'm unclear on efficiency)
Tesla do it differently by using the two power pins for both AC and DC, whereas CCS has separate AC and DC pins. That means Tesla have to switch the charging current internally (could be up to 1kA), while CCS avoids that.
(They don't have 3 phase AC chargers in the US, hence only two pins)
Yes, well there is a tendency for many devices to run off usb, which is 5v of course, and I bought one of those higher power Apple ones and shoved it into an el cheapo non powered hub with a switch for each outlet, and that charges my phone and powers and charges several other thing without breaking a sweat, given the right conversion cable. One of the devices though, a line level preamp with volume seems to have its own inverter inside to allow it to run on a much higher voltage. The problem is it generates a lot of RFI and interferes with portable radios.
Many moons ago when I could see I ran many things the other way around. A lawnmower lead acid battery and charger, 12 to13.8v I then made a few simple zenner controlled psus to power the different voltages of 6, 9, and anything lower from those. Of course being DC Series regulators they were not going to be as efficient, since they dumped the excess as heat. Brian
It's definitely easier, to start with a higher DC voltage, then use buck converters for the variable/subtending voltages. A buck converter is a lot easier to do than a boost converter. I could almost design one :-)
If you work with SMPS circuits, you almost need an oscilloscope, to verify they're well centered and working properly. Using linear regulator solutions, like you describe, a multimeter is all you need to verify those, which is a lot cheaper.
Using a 48V battery could make sense. Power over ethernet switches use 48 or sometimes 56V inputs and poe can be used to power WiFi access points, VoIP phones and plenty of other things. A buck converter from 48V to 12V or whatever is likely to be quite efficient.
HomeOwnersHub website is not affiliated with any of the manufacturers or service providers discussed here.
All logos and trade names are the property of their respective owners.