# Light bulb, diode, and capacitor step mains down to 12V DC

[Todd Harrison] needed a way to run a 12 volt PC fan from mains voltage. Well, we think he really just needed something to keep him occupied on a Sunday, but that’s beside the point. He shows us how he did this in a non-traditional way by using the resistive load of an incandescent light bulb, a diode, and a capacitor to convert voltage to what he needed. You can read his article, or settle in for the thirty-five minute video after the break where he explains his circuit.

The concept here is fairly simple. The diode acts as a half-wave rectifier by preventing the negative trough of the alternating current from passing into his circuit. The positive peaks of the electricity travel through the light bulb, which knocks down the voltage to a usable level. Finally, the capacitor fills the gaps where the negative current of the AC used to be, providing direct current to the fan. It’s easy to follow but the we needed some help with the math for calculating the correct lightbulb to use to get our desired output current.

1. Mike says:

He sure does talk a lot… even for a tutorial.

2. Mike says:

Your capacitor on the schematics is backwards.

3. Mikey Sklar says:

I use poly film caps and a bridge rectifier to make a cheap DC power supplies. These are great for charging batteries without a transformer (much more efficient). They also provide constant current which is sweet.

http://blog.holyscraphotsprings.com/2011/10/third-world-charger-tested.html

4. Stendall says:

Ohm’s law says that soon or later that capacitor will blow up.
You can get 12v, 8v, 100v, or whatever between capacitor leads, depending on the internal resistance of the fan.
It’s just a 2 resistor divider. The bulb and the fan resistances.
You can’t get the same voltage if you replace the fan by another fan with more or less internal resistance.
That’s a very, very bad way to “regulate” a voltage.

• meatman says:

Exactly. not a hack, but a “I dont know anything about electronics, but I know how to use youtube” event.

Tomorrow we show you how you can make money bu cutting out the ground wires in houses to sell the copper. The ground is not needed! Free money!

• Vonskippy says:

I am a Nigerian Prince, and am very interested in funding your research into ground removal free money. Please send me your banking info so that we can get started right away.

• raidscsi says:

I am a lonely nigerian princess whose father has recently passed away. His top scientist(my uncle) is very interested in starting work on this project. Maybe we can a lot of money and live together happily for long time.

5. jwk says:

Interesting. I’d have used a capacitor in series to knock the voltage down, then a diode. Might not have even needed a smoothing cap after, for a load like a motor.

6. Stendall says:

Some calculations on the fly:
If the fan consumes 0.150A at 12V that means a 80ohm of internal resistance while running.
It we take for good the bulb resistance, that give us 16.9V rms in the case of a full wave rectifier, and 8.4v rms in this half way rectifier case.

7. j t lien says:

You need a safety capacitor so that if the cap fails it fails as an open and not as a ahort. Very important in this kind of circuit.

8. rasz says:

fan is not a continuous resistance, it has pwm controller inside

• Will says:

No PWM on that fan buddy, it’s only a 3-pin connector – Ground, Voltage and Tach. The 4-pin fans (like most CPU cooler fans) have the 4th pin for PWM

• NewCommenter1283 says:

that kind of fan (brushless) uses a pwm controller REGARDLESS! the 4th wire is for when the controller is EXTERNAL.

normally, the PWM controller is INSIDE!
and is a mag.sens.+control+power-output I.C.

• rasz says:

exactly, every brushless motor has pwm controller inside “buddy” :)

9. Antonio says:

I didn’t watch the video, but read the article.

From what I know, a lightbulb will act as one resistor.

Don’t you need a voltage divider (2 resistors) to drop down the voltage?

• Daid says:

The fan “works” as the 2nd resistor.

This is a BAD idea. Lightbulbs have a very low resistance when they are switched from off to on. So during the first 10ms of switching it on there is some 200V on the fan. Only try this if you want to break your stuff. Also, don’t play with the mains unless you know what you are doing. (Hint: Above youtube poster doesn’t)

10. mahoney says:

Sure it works but it’s awfully inefficient!

11. MRC says:

I used a similar circuit to power a fan once. fan burned out after about 20 hours of runtime. YMMV, but some DC box fans don’t like being run like this.

12. Dax says:

There’s also the fact that the light bulb is not a constant resistance, but depends on its temperature.

So at start, the bulb is nearly a short, which is kinda bad for the capacitor and fan.

• Dax says:

Also, with a half-wave rectifier, the bulb’s temperature fluctuates more because half of the time there’s no current passing through it, so the output voltage varies quite considerably.

• Steve0 says:

You really think that the temperature is going to change that much between the rectified on and off of 60 Hz? When an LED is pulse width modulated to dim it, do you see it blinking?

• danman1453 says:

@ steve0, would the winding of the element act as a (very) poor inductor and peak higher than the intended voltage?

• boarder2k7 says:

@steve0 Your point is half right. PWM on an LED takes place at a very high frequency. The flicker imparted on this bulb is 60 cycles/second, and I know that I for one can see a half wave flicker. The range hood over my stove has a half brightness setting which is a half-wave, and it makes me nuts to use.

And if I can see the flicker, its definitely enough to produce voltage spikes. Of what value, I don’t know, but it isn’t stable for sure.

13. pin says:

i used to make a batery charger the same way loose the cap and put a car battery with a 1000 watt spot light bulb

14. pin says:
15. Hirudinea says:

He should have shown the finished product up front, its like searching through porn for the good stuff.

16. CJ says:

great video, nice t-shirt from the EEVblog! I would be just concerned about isolation from the mains, if something were to fail could be a shock hazard or fire hazard, Didn’t see a fuse either. Plus with no regulation the fan could burn out, here in MA our mains suck they vary from 105 to 130 V and 47hz to 62hz.

17. therian says:

This is bad, bad, bad. It dangerous, it inefficient, it not original.

• Glenn says:

Yes I agree, this is an example of what not to do. Its cheap and easy, but inefficient and bloody dangerous. Depending on how the house is wired, you could have mains voltage at the PC fan, in wires with insulation designed for 12v.

A better idea it to use a discared power adapter like a phone charger, plug pack, etc, we all have draws full of them.

Glenn

18. Jakeypoo says:

Add some Zeners to regulate the 12V

Transformer-less power supplies.
Common way to suck small current off the mains at a ‘regulated’ voltage.

19. Creepr says:

As my embedded systems design professor would probably say, “This circuit will catch on fire”.

20. jimbob says:

so now we’re posting “hacks” that are demonstrably wrong and dangerous ???

what the hell is anyone supposed to learn from this, or make use of ????

• dave says:

I agree. Less updates if this is what gets posted as filler.

I made a better power supply in class for a lab.

• FridgeFreezer says:

^^^ This

This is the sort of hack that you would only do out of absolute desperate necessity and never, ever tell anyone about – not parade it on the internet for all to see & emulate.

I know this blog is for hacks, but let’s have some quality/sanity control please!

Next week – now to test for voltage using only your testicles…

• Aussietech says:

+1

An outstanding example of what NOT to do. A bad idea, badly carried out.

21. This would be a good example of “This is what you do if you only have access to some parts”, but if this is supposed to be permanent… ugh, just ugh. Not likely to electrocute anyone, but very likely to end up with a dead fan and a dead cap.

That much said, I like “limited” hacks when I see good ones (no microcontrollers, only use 555s, etc.)

One thing I built recently (GPS aggregator and logger) for a dredging inspection company had the limitation that most, ideally all, parts should be available at a radio shack immediately. Guess someone still fixes stuff old style :)

22. Fabien T. says:

For safety and for math, microchip has this application note about transformerless psus:

23. hobz says:

Seems like the perfect setup for an indoor grow, bright light already there, i need to circulate air around plants but all i got is an old pc fan…

• Morbo says:

:) just what indoor grows need! Even more chance to get into the papers with a house fire!

24. Solenoid says:

This is very interesting since the circuit is easy to understand and he actually uses math (not banging rocks together like some…), also he explains what he’s doing and he’s using resistances that make sense for an uninitiated electrical engineer (even if non-linear).

As for the efficiency goes I’m sure there are other ways to do it. I myself am still a noob in this department and the only way to step down voltage I know is a voltage divider like his… but maybe with more linear resistances.

A dirty hack, but in a nice way.

25. Sean says:

Old valve(tube) tvs had a very similar arrangement with a ‘green concrete’ dropper resistor in place of the bulb.

They were bloody lethal as well !

Friend of mine had a dropper burn out on his and couldn’t replace it so had a 60watt bulb hanging out the back on a piece of flex……also doubled as a power on indicator :-)

26. Marty Taylor says:

This is what is known as a transformerless power supply. Although this is a particularly bad example of an already dangerous design.

Microchip’s website has an application note about them see:

Transformerless power supplies may be resistive or capacative in design and are usually only used for very low power requirements.

I’ve seen such designs in electronic mosquito swatters. These devices look look like a tennis raquet with a wire mesh and are powered by NiCd batteries. You plug them into an electrical outlet to recharge the batteries. Since the charging circuit is in the handle there’s no room for a mains transformer.

Similar circuits can be found in cheap Chinese built night lights. They usually burn out the load resistor at the first spike in mains voltage.
And where I live that’s regular.

Such designs can prove lethal (as can any piece of equipment involving lethal voltages) but do have a place if used correctly. After all, that’s why some people come to ‘hack a day’ to read about the unconventional.

27. Richard says:

Potentially lethal this may be, but when I was a kid a lot of the older folk used a similar arrangement to charge the LT accumulators for their old valve radios, and the HT ran straight off the rectified but un-isolated mains on bigger sets, smaller ones used a green concrete resistance to drop it down to 100v or so but still with a live chassis.

And out mains voltage at the time was 240v.

Beats me how so many of them survived long enough to breed. :-)

28. Kiwisaft says:

next step is to use a piece of wet wood to replace the bulb

29. Per Jensen says:

Well that seems like a good idea – NOT! I would rather just remove the propeller from the fan and mount it on the motor shaft. The other way around is just waiting for someone to get electro uted.

30. truthspew says:

In my case I need 130VDC to trigger a coin relay on a 1D2 payphone.

So I can do an isolation transformer into a bridge rectifier and then filter caps to get a nice flat DC signal.

31. jeff says:

nice workbench and lab kit.

nice math and schematic.

but a shame all of it is wrong and dangerous.

32. jimmy says:

Kudos to commenter Sean–this is very close to valve amplifier power supplies.

Yeah, don’t use this with any modern circuitry, or anything sensitive to overvoltage. And yeah, the cap gets hit with higher voltage when the filament is cold, which is not good.

But I’ve got half-wave rectified amps with similar supplies that have worked for 40+ years. Just sub a fixed voltage-drop resistor for the bulb (and add a fuse, and an iso transformer if human hands can get near it).

It’s a matter of perspective whether to term the bulb “voltage-dropping” or “current limiting.” In this application the terms were often used interchangeably…

33. Stendall says:

… 470uF 80v dielectric capacitor …
It means dielectric for electrolytic.
Electronics it’s not about to have the biggest oscilloscope and don’t know who Kirchhoff was.

34. Martin says:

Wow.

I had to check it was not an April 1st joke.

The usual way of cooling a motor is to put a fan on its output shaft. If that would not fit then any other way to power the fan is preferable.

One way or another that electrolytic is going to blow. Whether the fan breaks or is disconnected sooner or later there is going to be 170V across the cap and a big mess.

The other point in the write up is the author saying he wants a little over voltage on the fan to cool the motor better. Small electric motors do not generally need that much cooling unless they are being overloaded or run stalled.

Usually I see great projects here, this is not one of them.

35. aEx155 says:

Seems like it would be safer to just get a 120v fan in the same form factor as the DC fan, because they have those…

36. Frillip says:

There’s a difference between ‘non-traditional’ and ‘completely shit’. Seriously, using a lightbulb for what is essentially half of a potential divider circuit? Aside from terrible output stability, let’s not even speculate about what would happen under locked rotor conditions. Also, presumably I’d want the fan to keep things cool, and adding a lightbulb that kicks out a silly amount of heat is just going to defeat the object?

Prepare for my submission of ‘How to step down mains to 12VDC PROPERLY’.

37. Chris says:

That kind of fan does have semiconductor electronics inside, which are sensitive to overvoltage. Fortunately, they draw more current at start-up; that and a bit of luck is the reason neither fan nor capacitor has blown.

While this is a terrible way to power a 12V fan, it is a good reminder that there are alternate and low-tech ways of doing things, that may come in handy in the right situation.

In fact, earlier this year I needed to make an X10 lamp module work properly, including phase-control dimming, with a high-impedance load. I knew I needed a lower-impedance parallel load, but had no idea what value, so I had to do it experimentally. My selection of power resistors is rather limited. So out came the light bulbs! Between those and the resistors, it helped me zero in on the right load, and I was able to order an appropriate power resistor for the permanent installation.

And back before I invested in a variac, I’ve used light bulbs as current limiters to track down shorts in various wall-powered equipment.

38. qwerty says:

This is so bad and dangerous. The correct way to reduce mains voltage without a transformer is by using a capacitor AND a resistor in series, then a diode (or a bridge), then a zener diode to regulate it and a second capacitor to level/filter the DC.
The capacitor in series would act as a resistor WRT alternate current, while the resistor in series limits the maximum current flowing through that capacitor when it’s initially in a discharged state to avoid being seen as a short.
The circuit also depends heavily on the load characteristics, or even its presence, which is even more bad. If the fan fails the electrolytic capacitor will explode in seconds.

39. Miroslav says:

Next up: you can use 2 nails instead of AC wall plug. Find 2 wires, connect to nails and submerge in pail filled with water. And presto: your own kettle … of death. Just in time for Halloween.

40. steve says:

Stupid. Besides that it doesn’t work, missing mains isolation is a total turn down.

41. tron9000 says:

A more verbose tutorial:

Diode halves AC voltage = 170V PK

170 / SQRT2 = 170 / 1.41 = 120.56 – AC RMS

120.56 -12 = 108.56V – voltage over resistor

R = V /I = 108.56 / 0.15 = 720 OHM -value of current limiting resistor

Pr = 108.56 * 0.15 = 16.2W – power rating of resistor > 16.2W

Smoothing cap: bigger the better (>470uF) and make sure its over 12VDC rated.

I know he’s trying to be through, but god man, its a simple circuit! the amount of math and science he’s putting into it! there’s not need!
If anything he’s scaring off people who want to learn electronics

42. RichR says:

Resistive droppers like this were the norm in radios, record players and TVs until the 1960s. Transformers were big, heavy and expensive. A ceramic wirewound resistor followed by a valve rectifier and a couple of electrolytic capacitors was perfectly capable of providing a good dc supply. Less common in the US due to the lower supply voltage, but very common in the rest of the world running on 220V+ ac supplies. Plus you can run from a dc electricity supply which many places had. Transformers obviously require an ac supply.

Capacitive droppers are common nowadays in things like clocks and appliances like ovens and fridges – transformers and switch mode supplies just aren’t necessary if you don’t have a high current requirement.

And yes – always use class X or class Y rated capacitors that fail safe across the mains…

43. Dan says:

Skip the lightbulb altogether, and just use a 4-diode bridge and a capacitor. This is how some of the new-fangled LED nightlights get mains down to a usable voltage. Go to your local home improvement and get a pack of nightlight led bulbs and take them apart, there’s almost nothing in there but diodes, a cap, and the LEDs. Trace the pcb to obtain the schematic. All that’s needed to calculate is the value of the cap, which must be something bipolar to handle ac.

44. Dan says:

I would add this, a capacitive dropper is theoretically 100% efficient, as it does not use the dissipation in heat that a resistive one does. The charge that builds up on the positive half of the AC cycle gets pushed to the load on the negative half, and vice versa. So all the charge on the cap finds its way to the load, nothing is wasted except for the parasitic inefficiencies of the caps and diodes themselves. A light bulb or anything else used as a resistor is diverting some of the energy away and wasting it on a non-load use, such as making light or heat.

45. ironring says:

Safety issues aside (and wow, there are some serious safety issues), this has got to be the most inefficient way to “regulate” voltage (and I use the term regulate very loosely). This is equivalent to motor speed control with a series resistor.

46. Nitesh gujar says:

How to make a battery of 12 volts with the usess of power regulator, diodes, capacitor transformer and resistor

47. Jim says:

Unconventional, yes. I built something similar to slow charge tool batteries. One diode, one 25w light bulb. It works. Housed in thermoplastic, nothing exposed, it doesn’t kill anyone. I left off the cap, who needs it. If something goes wrong, there’s a fuse.