On the continuing theme of contingency planning, I am puzzled by some of the information circulating about so-called mini UPS for routers. My router (Fritz Box) says only to use the supplied power supply. Can I forget that for a start?
Descriptions refer to 9V and 12V (with a suggestion in one that it does not matter which). It's not clear to me what voltage my router uses (despite going through pages of instructions). Is there an easy way to find out? Does the UPS stay in place permanently, or is it intended to be plugged in during a power cut? If it is left plugged in, would it consume much power?
The output voltage of the power supply will be marked on it. Usually in almost illegible laser etched fine print. This is a requirement for regulatory approval. The information, that is, not the illegibility. Many routers will work over a wide voltage range as they have switching power supplies internally. However, avoid going over 16V as that is a commonly used electrolytic capacitor rating in such power supplies.
It probably voids the warrantee if use it with a non approved PSU.
Polarity wrong, too much voltage and/or not enough current and the magic smoke might come out. Too little and it probably won't work at all.
There should be a ratings plate hidden on the supplied PSU block which might be in embossed 2pt lettering black on black along with centre positive or centre negative indications. I need a magnifying glass to read them on most modern kit. It used to be 6pt font used.
If you provide too much voltage and the device has a linear regulator in it then you will waste more energy as heat in the router.
You need to examine the power supply it came with much more closely.
Many routers do say on them what voltage and polarity they expect.
A UPS for a router shouldn't consume much at all unless it is in the process of recharging after a power cut. It is much easier though to have the entire setup PC, router and fibre modem on the same mains UPS.
I would have thought that most routers wouldn't be too fussy about the supply voltage (eg, say marked supply voltage +/- a volt or two). If so, a suitable battery, a couple of diodes and a resistor (plus, if the existing wallwart power supply can't manage a few more milliamps to trickle-charge the battery, a slightly beefier one) would suffice.
As long as you use an alternative that can supply adequate current at the required voltage, then the router will not know the difference.
It will almost certainty be on the PSU in the small print. They will typically tell you the input voltage and current and the output. You need the output.
Failing that, measure it with a multimeter.
It stays in place permanently - the "Uninterruptible" bit of UPS requires that it is able to jump into action *very* quickly - so quick that the router does not notice the power interruption.
Not really - once its battery has charged, you are just running the router and its normal consumption.
For a UPS to do any good in providing an *uninterrupted* supply (as opposed to a backup supply that needs manual intervention), it must be plugged into the wall all the time, with the protected appliances plugged into the UPS.
I've always wondered:
- Do all UPSes do an equally good job of providing an uninterrupted supply without glitches that can cause computer equipment (PC, router etc) to reboot? Or are some "more uninterrupted" than others?
- What is the standby power consumption of a typical UPS - ie the additional power needed by the UPS itself, over and above the power used by the protected devices?
We only very rarely have power cuts that last more than a minute, but in late summer and early autumn we always (for the three years we've lived here) get lots of brief one-second power cuts - just long enough to make equipment reboot, and sometimes we get several in rapid succession which are liable to make a Windows PC run a forcible "chkdsk /f" before the PC will boot up again. Our power provider offers various reasons when we raise formal complaints: sometimes it is "overhanging branches" (in which case, why don't they patrol lines so they can trim branches before they get close enough that the power even needs to be turned off while the branches are trimmed, let alone getting close enough that they cause shorts to earth?); other times it is "livestock in fields using wooden poles as scratching poles" (hilarious, but it might even be true, though it doesn't say much for how firmly the poles are fixed into the ground, and how often they are monitored for starting to get wobbly).
So a UPS for me only needs to "paper over" an outage of a couple of seconds; and if it lasts for a few minutes, so much the better.
The difficulty is getting power from one UPS to the various important devices around the house. Getting a mains cable from one side of the living room to the other (router and associated Ethernet cables on one side; "master" wifi mesh node on other side where it has best line of site to other rooms) is achievable if I chase out a channel in the concrete floor. But getting protected power to corresponding nodes in other parts of the house is impossible, so realistically I'd need two more UPSes - grrrr. The problem is that the "slave" mesh nodes don't always reconnect to the master when the power goes off, so if they can be kept permanently on, no human intervention is necessary.
Yep. The best of them power the stuff plugged into them all the time and charge the batterys continuously, so there is no glitch at all on mains failure because there is no switching involved at all.
Line interactive ones essentially pass the actual mains through, until it fails; then start-up the inverter section.
The better ones use double-conversion, so effectively they're always charging the battery with one section and providing inverter power from the battery with the other section ... they have a higher vampire power loss but have no glitches.
So do the line interactive ones tend to cause glitches that can reboot computers and computer-like equipment? I've heard that the main advantage of double-conversion ones is that they guard against brown-outs or surges (voltage varying outside the acceptable range) and frequency variations. I'd tended to assume that the switchover of line-interactive ones was fast enough for most PSUs to ignore any glitch.
A UPS that glitches is fine for lighting etc, but useless for anything that will reboot on a glitch. Something to check very carefully before deciding which type of UPS to buy...
I bought a UPS because my router kept being affected by overnight brown outs. This was years ago pre-ADSL2. I never had any further problems
I now have 3 separate UPSs. One for the incoming router, One for my computer deak and a third to cover security cameras. Separate because of separate rooms.
IME, no. I have three APC smartUPSes they are line-interactive, and numerous at customers ... you get a big "chonk" as the relay cuts in, but power supplies in PCs, network kit etc easily survive the glitch, I think they claim a 6ms transfer time, so not not even half a mains cycle.
Mine spends a few hours a day "bucking" the voltage, the supply here switches taps in a substation somewhere at least a couple of times a day when demand fluctuates,usually 8am or midnight, the UPS kicks in for a few seconds, then goes back.
There is a picture of Line Interactive here, and it has an AVR in the forward path.
formatting link
If you remove the AVR from the forward path, the block diagram is for an SPS or standby power supply style of UPS.
One of the UPS companies used to have a PDF, with about five flavours of UPS defined in the document. Double conversion you already know.
Another way to (manually) regulate mains voltage, is with an autotransformer. Autotransformers are a favorite of teachers in school, and are used for various purposes for doing demonstrations of stuff. After you leave school, you might never see one again after that.
formatting link
"Autotransformers can be used as a method of soft starting induction motors."
That sounds like a traditional backup UPS - switches to inverter on mains failure, but beyond some passive surge suppression do nothing of interest when running on mains.
The line interactive are not full online double conversion ones, but can run the inverter and trim or boost the mains when still running on mains power.
Most computer and network equipment will happily ride out gaps of
20ms (one mains cycle). They generally have power supplies that work down to 85 or 90V, so the input storage capacitor can discharge a long way before anything gets upset. This means there is plenty of time for a relay to operate. Part of the testing for regulatory approval involves subjecting products to varying length and depth gaps in the supply as well as quite nasty voltage spikes. John
"The ATX specification sets the minimum hold-up time to 17 ms with the maximum continuous output load. In many cases, manufacturers use smaller capacitors in the APFC converter, resulting in a measurement of less than 17 ms. Manufacturers do this mostly to cut production costs, as these capacitors are expensive. The smaller bulk capacitors also improve efficiency by a little bit. "
Before Active Power Factor Correction, there was hardly any cheating on holdup time. Now, you may run into ATX supplies that don't meet the spec. Normally, an SPS-flavoured UPS switches in time, before the "tank runs dry" on the ATX supply.
When the ATX supply doesn't have a load (just USB devices on the PC), then the holdup capacitor can last for 60 seconds. At minimal load, the capacitor shows an impressive time constant.
PCs and routers can take quite dirty mains and remain working.
There are different levels of UPS uninterruption.
The crudest notice the mains drop and step in within a half cycle or so
- very cheapest and nasty ones probably take a bit longer than that. Their waveform is usually nothing like a sine wave (but good enough).
Next level up line interactive try to keep the mains voltage presented to the PC at closer tolerance and better approximate a sine wave. The converter still only runs when power is lost (and may only be able to survive at maximum load for the period of time that its battery lasts).
The final up market ones do mains to DC then DC to mains with guaranteed frequency and voltage output - they are always running flat out so no gaps at all. They are expensive, least efficient and most secure.
The options are summarised here:
formatting link
Somewhere between 5-10% of the load that it is supporting plus a few W for housekeeping.
It is true after a fashion there are plenty of poles near me at rakish angles and badly thinned at cattle neck height from them scratching. I reckon poles in fields with beast in regularly last about 5 decades or so before they are ready to fall over spectacularly and all at once.
That was exactly what happened in storm Arwen. Fell down like dominoes.
Just about any cheap UPS will do that but go too cheap and nasty and you won't be able to replace the batteries.
Power cuts where I am include a few short ones but have recently become a lot longer so many hours or days rather than minutes.
Focus on which bits you absolutely need to keep running - the more load you put onto a UPS the shorter time its battery lasts.
You could probably DIY a suitable standby battery supply for your remote routers trickle charged from the mains by the existing PSU and with two Schottky diodes so that the battery takes over if the mains goes down. Be sure to put an automotive low voltage fuse in series with the SLA.
A PC should not be all that glitch sensitive. It typically takes 20ms or longer before a normal office PC will be affected. There is enough stored energy in the main PSU capacitors and the on board local capacitors to bridge the gap between mains fail and inverter starts.
If you are running a server farm then you may want to be sure that everything stays operation no matter what happens. You pay dearly for that level of over engineering since the inverter has to work 24/7 at full power all the time. A UPS has a maximum run time determined by its battery. A few of the up market ones you can double the battery capacity with an approved add-on battery in matching case.
The cheapest ones are rather close to their maximum operating temperature by the time they run out of battery at maximum rated load.
AC mains is crossing zero volts every 10ms.
The PC PSU can easily handle input voltage being under 80v for fractions of a second (although it pulls ever more current to do so). The spec used to be should survive 20ms total gap in mains with the PSU loaded to
90% of rated output (longer if more lightly loaded).
Gaming machines might be at risk from shorter glitches but relatively conservative domestic or office PCs should survive short glitches OK.
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.