A reader writes:

For various reasons, we still use wired landline phones at our house. I have an old 80's phone (first generation touch-tone) that belonged to my grandmother that I would like to continue to use. The only thing really wrong with it is the ringer.

It has an actual bell and clapper ringer. But the plastic posts that hold the bell in place inside the body of the phone have broken, so instead of ringing properly when it gets an incoming call, it just makes a feeble rattling sound. Since I'm not all that fond of the harsh sound of an old-style phone bell, what I would like to do is rip out that bell and replace it with the kind of electronic warble circuit that modern phones have.

First, what on earth are those modern ringer circuits called, and can one buy one or do I have to get a cheap phone and rip it apart? And second, is this a drop-in replacement, just solder the modern ringer in place of the bell and it will work, or is it more complicated than that?

And finally, are there instructions out there on the net for doing something like this and I just haven't found the magic search term that will get Google to take me there?

I tried googling this question a few times, but the signal to noise ratio of people talking about unrelated phone topics was too bad and I always came up empty. And then I realized, Hey, if I send Dan a few bucks I'm pretty sure he knows the answer and I will have saved myself a lot of time and frustration.


(And he did, indeed, send me a few bucks!)

The basic parts of old-fashioned phones - from the Bakelite era - should all be interchangeable, because they're just passive and electromechanical components powered by, and dependent upon, the phone line to operate.

So if you were working only with gear of that era, before the pushbutton era, you could probably build a telephone using the earpiece from one phone, the mic from another, the ringer from another, and the pulse-dial mechanism from yet another.

I think the basic electrical characteristics of the Plain Old Telephone Service, and its modern equivalent which is full of high-tech networking gear that pretends to be the POTS as far as phones are concerned, are about the same around the world.

The phone line constantly provides about 48 volts at low current to run telephone systems. (It's sometimes called minus 48 volts, because the phone system is "positive ground", to reduce corrosion of underground components. The power supply wires are at about -48V relative to the power returns and ground.)

The phone line tells a phone to ring by superimposing about 90 volts at about 20Hz AC on the normal 48V DC. When you pick the phone up and it goes "off hook", the line voltage drops to single-digit volts, and your phone should draw double-digit milliamps of current, at most.

(There's a long-lived strain of Magical Free Energy Machine that claims to allow you to tap useful power from the phone company's battery banks, taking advantage of that 48 volts that's sitting there all the time. Actually, the continuous current consumption of a phone is supposed to be in the microamps and the entire subscriber loop isn't expected to draw more than ten watts. The phone company will notice and may get angry with you if your house draws even a few watts from the phone line. You can power a little LED reading lamp or similar low-current device from the phone line without much trouble; the free-power scammers usually promise that you can substantially reduce your electricity bill, though, which is not true.)

Modern phones with transistors in 'em react to these different voltage conditions by doing what the old phones did, only fancier. This presents a problem for you, because when everything's controlled by a circuit board that activates different things in response to different input voltage characteristics, there's not necessarily a simple "ringer" that can be removed and put in another phone.

I wouldn't be at all surprised, though, if you could take the circuit board of a modern phone, with ringer attached but no handset any more, and just stuff it inside the casing of an old phone that still has everything but a ringer (just removing the half-broken bell mechanism would do the job there). Wire the newer phone's board in parallel with the old phone's parts and it ought to work. You may be violating a law or two by connecting an unapproved home-made device to the public telephone system, but if it's electrically the same as the original two phones plugged into a Y-adapter, I don't see any potential for real harm.

I'm not certain about this, though, and welcome input from commenters who've monkeyed with phones more than I have.

(Oh, and if you just want to reconstruct the broken plastic parts inside the phone, that's a perfect job for polycaprolactone, the nylon-like plastic that softens in hot water or a barely-warm oven and can be used, and reused, to make plastic parts of any shape. The main problem with hastily-squished-together polycaprolactone parts is that they tend to look blobby and weird, or spookily organic, but that doesn't matter for parts that're inside a casing that nobody will see.)

On the pulverisation of potatoes

I should write things here. I used to write things here all the time. I stopped to write a book, and haven't written that yet either. Sorry about that.

Apropos of what I am currently doing: Potatoes!

Because I am a human and am sane, I like mashed potatoes. I didn't make mash very often, though, because I am also lazy, and all the peeling got me down.

Because I did not wish to use a drill, a brush and a bucket, I bought one of these contraptions...

Crank peeler

...which actually works pretty well for peeling vegetables in bulk. (It can also be used to make curly fries, and the long spirals of peel it creates when peeling potatoes make surprisingly good biodegradable cat toys. The crank-peeler I have has a suction cup that holds it onto the kitchen bench well enough; there are also versions with a clamp.)

I also bought a potato "ricer", like this one...

Potato ricer

...which works like a gargantuan garlic crusher, and turns any chunk of cooked potato (or other vegetable) small enough to fit into it into instant extruded mash, with one squish of the handle. (There are many other models, some quite large. I don't know if the suspiciously cheap metal eBay ones are any good.)

But this was all still too much of a production. The ricer's a bit annoying to clean (one of the deadly flaws of many kitchen gadgets), and you can't effectively mix additives (butter, egg, garlic, whatever) with the spud in the ricer. So you end up having to mash "manually" anyway, to mix the additives in.

But now, in what I think we can agree is one of the greatest examples of laziness improving the human condition ever witnessed, and which I'm sure we can also agree is a brand new idea that could have come from nowhere but my own genius, I have developed a technique which solves all of your pesky potato problems. (You won't believe this one weird trick! Chefs hate me!)

What you do is, you don't peel the potatoes.

Buy washed potatoes so you don't have to wash them either. Cut them into similar-sized lumps. Lumps with skin on will, I think, cook a bit slower on the skin side than lumps without. Adjust lump size and shape accordingly.

Remove any really egregious eyes or gashes or other spudular injuries. Don't worry about this too much, though. You're going to be pulverising these things; little imperfections will vanish.

Continue to dress and cut spuds and dump them into your large pot until you think a suitable mountain of spud has been achieved. Put pot in sink, run it full of water, slosh some out again, let it flow through a bit, to wash the lumps clean of whatever stuff still clings to supermarket washed potatoes.

(My house has one of those "instant" hot water systems where the water flows through a gas heater rather than sitting in a big tank, so our hot water doesn't taste funny. I therefore do this washing stage with hot water, to make the cooking a bit faster.)

Transfer spuds to stove. Cook until when you poke a chunk with a knife or skewer there is no longer any perceptible crunch.

Drain spuds and mash, with ordinary or fancy handheld masher.

To prevent your masher from getting clogged with skin, pull a knife through the spud-chunks to slice the skin-bits up. (My potato ricer is completely useless on spuds with skin. The skin blocks its holes immediately, and then what it manages to mash squirts out vertically or seeps around the blockage in a quite faithful simulation of the phenomenon of "paradoxical diarrhoea". Don't all thank me at once for this information.)

Add butter and egg and whatever else you like. To make mixing butter with spud easier, melt the butter in the microwave. (Cover the butter if you do this; it tends to explosively splatter before all of it is melted.) If you've somehow managed to make the spud too dry - by, for instance, not adding all the butter that exists in the world - add some milk too. If the spud is too soggy because you've by some terrible wicked miracle managed to add too much butter, add egg; the egg cooks in the hot potato and stiffens the mix.

And you're finished. There's surprisingly little difference between mash with and without skin; when it's thoroughly mashed the skin can be quite hard to see. And leaving the skin in lets you pretend you're being virtuous and getting more vitamins and stuff.

I invite suggestions of ways to streamline my mashed-potato workflow even further!

Magnificent marble machines

In which I shamelessly knock off a couple of videos from this page on Matthias Wandel's site.

More at Other marble machines, including Matthias' own ones, are here.

On the Domestication of Car Stereos

A reader writes:

I have been rereading some of your stuff concerning speaker building lately and have decided to possibly make it a hobby. A mate and I have a bunch of car audio stuff we pulled out of his car that we would like to turn into a garage/shed powered speaker box. I have been looking on the web and there seems to be about a million different opinions as to various aspects of the build including:

* if we can use the car amp
* if a head deck needs to be involved
* power supply – some say even an old computer PSU can be used some say that the current requirements would be better suited to a car battery with a charger attached (seems a little cumbersome)

So I thought I would send a quick email to great and powerful Tech Master who started me down on this road for some clarification. Help?


Ludicrous car speakers
(Image source: Flickr user Nazly)
Just lever 'em out, and shovel 'em into the house!

Yes, you can make perfectly good household speakers out of car-audio components. Actually, car radio tuners can be better than the home-hi-fi kind, because the extremely variable reception conditions for mobile radio, and the lousy antennas they have to use, mean even quite cheap car stereos often have very good RF sections.

The problem with powering car amplifiers at home is, as you've already discovered, that they can want a lot of current at 12 volts.

First, definitions: You standard "car stereo" has one "head unit" in the dashboard, which combines all of the signal sources - tuner, cassette deck, CD player, line-in socket, memory-card socket for MP3s, whatever - and all of the amplifiers, in one box. These days it probably has four outputs, for two front and two rear speakers. Take such an all-in-one head unit and some speakers and a twelve-volt power supply and hook them all up at home and they'll work the same as they did in the car.

Fancier car audio systems have amplifiers separate from the head unit, and may have multiple head components as well, possibly including a separate satellite-navigation screen, DVD player, et cetera.

A separate car-audio amplifier is basically the same as a separate home hi-fi amplifier: It accepts line-level input, and amplifies it to drive speakers. The input has to come from other components. Fancy car-audio head units may have no amplifiers of their own at all. Again, though, you can run the whole system spread out on a table at home, if you want to and have a suitable power supply.

Standard cheap car-audio amplifiers, like the ones that're built into low-cost one-piece head units, have a maximum output voltage no higher than the 12 volts they run from - 13.8 volts, actually, when the engine's running. Car speakers have a nominal impedance of four ohms - home hi-fi speakers are usually nominally eight ohms - so you can use Ohm's Law to figure out the total possible output current. Ohm's Law says current equals voltage divided by resistance (I = V/R), 13.8 volts divided by four ohms gives 3.45 amps, so that's the most that such an amplifier can output per channel. Each channel is basically its own separate amplifier.

(The "nominal" in "nominal impedance" just means that that's the approximate impedance the speaker presents if you run DC electricity through it. The actual impedance varies quite widely depending on the frequency of the incoming alternating-current music signal, but overall it'll be close enough to the rated nominal impedance for rough-calculation purposes.)

13.8 volts times 3.45 amps is 47.61 watts; an amp with four output channels could therefore output 190 watts, which is more than enough to make the inside of your car very loud, even given the not-so-great efficiency of a lot of car speakers.

(Car-audio equipment of all sorts often has outrageously high "peak music power output" numbers printed on it; this problem may be even worse in car audio than it is in cheap home hi-fi equipment.)

Realistically, to avoid hideous distortion from winding the amp up all the way and to also take into account the rather-less-than-100% efficiency of all amplifiers, a ballpark figure of 25 watts of input power per channel, at the 12 volts you'll probably be running it from at home, is likely to be about the real-world maximum for a basic car-audio head unit. Call it 24 watts for a nice round number of two amps, at 12 volts, per output channel.

More powerful car amplifiers step up their input voltage so they can deliver more volts of output, and the sky's the limit for those. But you can run any amp expecting 12-volt power input from any 12-volt source. As long as you keep the volume low enough that the power supply isn't overloaded, you can probably even run some preposterous multi-kilowatt boom-car amp, at very low volume settings, from a one-amp plugpack.

(This is related to the reason why it's not dangerous to touch both terminals of a car battery that can deliver hundreds of amps into a load with a low enough resistance, like the starter motor. Your body has a far higher resistance, so far less current flows. There are also 12V power supplies with very low output ratings, like say the plugpack for an ancient calculator; you might be able to blow one of those up by just turning on a 12V amp connected to it. You also might be able to damage a 12V amp by plugging an old-style heavyweight unregulated linear power supply into it, because those deliver root-two times their rated voltage when they're unloaded, and seventeen volts might be too much for the amp. All care, no responsibility. Et cetera.)

A PC power supply actually is a pretty good 12V source for running car audio gear at home. It'll be able to deliver a decent number of amps at 12 volts (the yellow wires coming out of standard PC PSUs are +12V), and it should also deal elegantly with overload, and just shut down if you ask for too much current. Modern PSUs may have split 12V rails and other complexities, but an old one out of a superannuated Pentium II box should do nicely for most purposes. Since most home-audio listening happens at only a few watts per channel, unless you want Party Volume of want to hear the music over your power tools, this option should be fine.

A battery charger and car battery will let you run a car amp at maximum power - well, until the battery goes flat because the charger can't keep up with the amplifier load, at any rate. It is as you say not a very elegant solution, though, and the charger may get confused when you turn the amp up and it suddenly sees a strangely high load. And the output from cheap car chargers can be really filthy, noise-wise, too. Car amps generally deal very well with distortion in their incoming power, but you may still hear a whine or ticking noise.

You can also, however, plug car speakers into a home hi-fi amplifier. A given volume setting will give you more noise from a four-ohm speaker than from the eight-ohm speakers home amplifiers expect, but pretty much any amp should work fine with four-ohm-nominal speakers; some home speakers actually have six- or four-ohm nominal impedance. Any old mini-system hi-fi amp will do, too, you don't need a fancy expensive one.

Another way car-audio enthusiasts get more noise out of an amplifier is by wiring multiple four-ohm-nominal speaker drivers in parallel. Two 4-ohm drivers in parallel give you a nominal-2-ohm speaker, three in parallel give 1.33 ohms, four in parallel give one ohm, and so on. You have to be careful connecting speakers with very low nominal impedance to most amplifiers; it usually won't cause a problem if keep the volume setting very low, but one-ohm or lower speaker arrays may look like a dead short to any amp that isn't designed, as some car amps are, to drive them.

You can also get small 12V amplifiers intended for use in home and mobile applications, which are descendants of the popular, and distinctive-looking, Sonic Impact T-Amp:

T-amp and speaker drivers
(Image source: Flickr user animakitty)

The modern ones are easy to spot; they all have extruded aluminium cases, often anodised a cheerful colour, and a few chunky controls on the front:

Lepai amplifier
(Image source: Flickr user

These little amps may be sonically superior to cheap car-audio amplifiers. The radio tuner in a car stereo may have to be good, but car amps can be quite noisy and distorted, because that's not very noticeable in the lousy acoustic environment of the average car. These little units all use similar class-D amplifier chips, which are good for maybe 15 clean output watts per channel or 25 watts flat out. For full power you have to run them from something with at least a five-amp power rating - so, a car electrical system or computer power supply, or a laptop-power-supply-style 12V PSU, which last is what they come with if you buy one with PSU included. Again, though, you can hook up a smaller 12V plugpack if you like, and just keep the volume low.

(Some of these little amps have a USB socket on the front panel, too. In the cheap ones, only have the power pins will be connected, so you can use that socket to charge most phones or MP3 players, but you can't play MP3s off a thumb drive.)

These little amps may or may not sound any better than a cheap car head unit, and all they are is an amplifier - no tuner, CD player or whatever. But they look a lot better indoors than a bare car head unit, and the going rate for one without a power supply is under $US20 delivered.

As I've mentioned before, you can find these amps by searching for "class D" or "class T" (Tripath's trademarked version of class D). You'll find the very cheapest ones if you just search for 12V amps, though. You can get these same chips on little amp-module circuit boards, too, for hobbyists to install in their own enclosures, like a normal sort of amplifier enclosure or powered speakers. The above search is sorted by price, so finds lots of those little modules before it gets to the assembled amplifiers.

Car speakers are a great choice if you want to play with transmission-line speakers, too. The basic transmission-line design has a single widerange driver at one end of a folded tube; here's one under construction:

Building a transmission-line speaker
(Image source: Flickr user Moisturizing Tranquilizers)

A cheap six-by-nine oval car driver with a separate tweeter and/or midrange on a bridge in the middle of it is an excellent low-cost choice for a speaker like this.

Sharp edges are just good aerodynamics

Here's something I didn't know existed.

Behold, the rocket monocopter!

It's a wing on one side, and a rocket pointing horizontally on the other.

Or, in the case of this monstrous one...

...two J engines and a K in succession, as this spinning surfboard wanders the sky looking for heads to lop off.

It doesn't seem very difficult to build a small rocket monocopter, from scratch (a spare R/C helicopter rotor blade is helpful...) or a kit.

It's apparently possible for these things to get themselves into a nice flat autorotation spin on the way down, too, so you don't need a parachute.

Well, that's the theory, anyway.

One-bladed propellers are actually very desirable, essentially because the more blades you have, the more turbulence each blade must cope with from the blades in front of it.

The obvious problem with dropping to one blade is that if it's mounted on a shaft and developing thrust, then that thrust will be off-centre, spinning around the shaft with the blade, and thus ruining the shaft bearings quite quickly. So you only see single-bladed props in oddball applications like these monocopters, and extremely tweaky ultra-speed control-line models, and certain rather expensive ceiling fans.

There are also propeller-driven monocopters, where a motor and prop takes the place of the rocket in the above contraptions. With modern digital control and image-processing systems, monocopters may be useful as unmanned aerial vehicles:

Posted in Hacks, Toys. 2 Comments »

More photons, fewer pixies

A reader writes:

This gravity powered light has been getting a little attention lately. It reminded me of your post on the Gravia light.

So is the GravityLight a sensible design that conforms to the laws of physics, or is it also powered by Pixie Dust?


Gravity Light

There's nothing inherently wrong with the basic idea of converting the energy of a falling mass into electricity. (That's how hydoelectric power stations work, after all.) It may even be possible to do it quite efficiently, and cheaply, on a small scale now.

In the olden days you would have needed to gear up your dynamo a lot from the pulley your falling weight (or flow of stream water) was turning, but today super-powerful magnets are very cheap and so efficient lower-speed dynamos are easier to make. Especially if you're only trying to light one high-intensity LED to a brightness that'll let someone read a book at close range.

The GravityLight people claim thirty minutes of light per lift of the weight, and the tape connecting the weight to the light looks to be, being generous, about 1.5 metres long. They don't say exactly what the weight is, but they do say you can fill the GravityLight weight bag...

GravityLight kit

...with use "anything weighing about 20lbs"; let's again be generous and say you get ten kilos of stuff in there, which is a perfectly liftable weight for a wide range of humans.

So we've got ten kilograms falling 1.5 metres in standard gravitation; that gives us a maximum of 147 joules to play with.

Split over thirty minutes, which is 1800 seconds, that gives us only 0.082 joules per second. That means a power of 0.082 watts, or 82 milliwatts.

A single standard white LED will have spec-sheet power numbers of about 20 milliamps at 3.6 volts, which is 72 milliwatts. It'll work fine - and more efficiently - at rather lower current, though, which is just as well because this falling-weight system is certain to be a long way short of 100% efficient at turning the weight's gravitational potential energy into light.

Even if it's only 50% efficient, though, you've still got 36 milliwatts to play with, which is plenty to light one LED to a useful, though far from room-filling, brightness. Even if you pare off some of the above generous assumptions about weight and fall distance, you'll still be easily above 20 milliwatts, which is also usefully bright.

The GravityLight people say their invention is intended to replace kerosene lanterns, but it's definitely not going to have the room-filling brightness of a kerosene lamp with the wick turned well up. Given the numerous downsides of kerosene lighting, though, and the fact that a lot of poor people probably don't turn their lamps up to max very often, a GravityLight or two could well replace one.

The ridiculous Gravia concept thing had a heavier weight falling a similar distance, for a total of about 271 joules. But the designer idiotically claimed that the thing would have the light output of a forty-watt incandescent bulb for four hours. This was so impossible that it couldn't come anywhere near being done even in Physics Experiment Land, with a perfectly efficient dynamo and a perfectly efficient lamp.

With real-world hardware, in contrast, the GravityLight can work. I'm not totally convinced that for household lighting you wouldn't be better off with a couple of conductive objects a reasonable distance from each other in the galvanic series, some damp earth as an eletrolyte, and a Joule Thief to boost the output to run an LED. That sort of improvised battery can run for a very long time at the very low power a Joule-Thiefed LED requires, and its poor portability doesn't matter if you're using it as a hoursehold light (and is an advantage if you want to avoid your light being stolen...).

But the physics, at least, checks out for the GravityLight.

Weaponised victuals

The flameless ration heater that comes with every Meal, Ready-to-Eat contains a fine powder of iron, magnesium and salt, which gets hot when it gets wet.

You are supposed to put some water in the heater packet and then lean the MRE pack against, famously, a rock or something, while it warms up. The reaction shouldn't take more than a minute.

What, young soldiers with little to do have doubtless asked many, many times, would happen if you were instead to collect the powder from several ration heaters and put it in a sturdy sealed vessel, with some water?

Why, this would, thanks for asking!

Unlike relatives like the dry ice bomb, the heat of these "MRE bombs" will soften plastic, allowing the impressive inflation seen above.

If this just isn't dangerous enough for you, you could always use a metal canteen instead of a plastic bottle!

Note also the suggestion of putting the MREs' Tabasco sauce in the bomb, to add a chemical-warfare tang to the operation.

In Which I Try Not To Set A Reader's House On Fire

A reader writes:

I've been searching the internet (including your articles) for information on putting together a simple back-up power supply for my central heating system. One that when the power fails (which it will here in Greece when we get a good old storm), I just go down to the boiler room, disconnect the boiler system from the mains and hook it up to a back-up supply (for the two or three hours that it takes to get the utility repair man out of the taverna, up the pole and get my 230v back on line). Automatic systems are all very well, but isn't it nice to know what is actually going on, and also be in charge!

Most of the back-up power systems that I have found on the internet seem to be designed for computer systems (oh, and maybe a fridge). Well when it's wet and cold here and the power goes down, I am more interested in keeping warm than keeping my beer cold (although I do understand the importance of the later) and if my computer doesn't work, well, I still have my Ipod.

The reason I need back-up power is because my oil fired central heating system has a wood burning stove linked to it and must keep the system (in particular both the pump and the system controls) running. The boiler is rated at 140w, the pump 160w, 5 motorised valves 30w and the control box and lamp 70w - 400w in total.

Is there such a back-up supply suitable for my heating system or do I put up with shivering, writing unnecessary e-mails in the dark over a can of cold beer?


Anything involving monkeying with heating systems raises red flags with me, but I'm pretty sure I'm not about to give you advice that will lead to your death. I have, however, been repeatedly demonstrated to have very poor judgement in this regard. (Some friends of ours have officially notified their small children that not everything Daniel says should be accorded the same respect as things said by other adults. They were fine with this, though.)

There may also be some local law that makes this illegal, or requires a licensed electrician to install it, or something; I know nothing of Greek law.

OK, disclaimers over. If you're happy to have a setup that you have to go into the boiler-room to connect, then I think the best option would be an appropriately robust petrol-powered generator. You have to duct the exhaust outside, of course, or set the generator itself up outside. (It might be possible to plumb the generator exhaust into the boiler flue or something, but this could also be another piece of extremely dangerous advice.) Apart from the exhaust issue, though, it'd probably work nicely. Modern generators from the major manufacturers are reliable, quiet and not even all that expensive.

(Generators that serve the purpose of a UPS, cutting in automatically when power fails, are fancier and more expensive. Way more expensive, if you want one that won't give you even half a second of blackout.)

You could probably also use a suitably large off-the-shelf UPS, though, if the tromping into the basement and switching the cables and pulling the starting rope starts to pall. The wattage figures on your heating system's specification stickers are, like most such figures, likely to be over-estimates, so it's possible a quality UPS with as small a rating as 700 volt-amps (which are not quite the same as watts, as I discuss here) could do the job.

The power-rating issue is the same for generators as for UPSes, but I think generators are better at handling the initial "inrush" current when a motor starts. That can be high enough to cause a UPS to beep and shut down, even if the UPS is perfectly able to power the motor if it's already running. This is particularly the case for refrigerators, whose run power is quite low but whose compressors suck a lot of watts for a brief moment when they click on. A UPS trying to power such a motor will therefore work OK if you lose power when the motor's already running, but not if the motor needs to start from UPS power. In the same situation, a similarly marginal generator should just bog down and deliver lower voltage than it's meant to, which is in this case perfectly fine and should let the motor start up with no trouble.

The solution to this whole problem is, of course, to just get a UPS or generator with a higher volt-amp rating, or with a specific high surge capability that it may only be able to deliver for one second, but that'll do. A "home"-model UPS or small generator rated for a genuine 1500 VA (as opposed to the suspiciously high numbers on suspiciously cheap off-brand UPSes) might be adequate; if you turn out to need more than that, and decided to go with a UPS, then you'd have to pay extra for a commercial-market one.

Easy enough to find out what works, of course, if you can get a local dealer to let you borrow likely-looking generators and/or UPSes and try them out. Overload won't actually damage any half-decent UPS or generator; at worst, they'll just complain and shut down.

The real killer for a UPS solution would be run time. Generators can run for as long as you have fuel, of course, but three hours is a long time for a home-or-small-office UPS to be delivering a few hundred watts. Smaller UPSes may even overheat and die in such a situation.

If we presume the constant draw is, say, 400 watts, then that for three hours is 1200 watt-hours, and the battery-to-UPS-to-appliance chain is not 100% efficient, so you'd need more than 1300 watt-hours of batteries to get it done. The capacity of the standard little sealed-lead-acid brick batteries in small UPSes is maybe 90 watt-hours. Less, actually, if you don't want them to die young; standard lead-acid batteries don't like being run flat.

Commercial UPSes can usually be had with extended batteries, but regular readers will know that I recommend just hooking up some car batteries instead. The very cheapest car batteries are still good for 240 watt-hours or more, so it'd be inelegant but feasible to build an array out of them that could meet your needs.

Drop some extra dough on quality "deep-cycle" batteries that are actually meant to do this kind of job and you can easily get well over a thousand watt-hours from one 12V battery that two normal humans can probably move. Graduate to proper industrial batteries that only strongman contestants can move by themselves and you probably won't actually get a whole lot more capacity per kilogram, but probably will get a setup that'll work for many, many years with no more maintenance than an occasional distilled-water top-up. Industrial batteries and a commercial UPS should actually easily outlive a generator.

The smallest batteries in the current Trojan Industrial Line, for instance (PDF here), are six-volt with a 355-amp-hour rating even if you're running them flat over only five hours; two of those in series will give you 4000 watt-hours at 12 volts even into quite a large load.

But, again, none of this is necessary if a generator's acceptable to you. Since you specifically asked for a system that requires you to switch it over manually, a standard, quite inexpensive pull-start generator looks like just what you want.

(I invite commenters to point out the many ways in which I have, in the above, unknowingly endangered Bill's life.)